Professional Documents
Culture Documents
Strøm, Kaare, Wolfgang C. Müller, and Torbjörn Bergman. 2006. Delegation and Accountability in Parliamentary Democracies. OUP Oxford PDF
Strøm, Kaare, Wolfgang C. Müller, and Torbjörn Bergman. 2006. Delegation and Accountability in Parliamentary Democracies. OUP Oxford PDF
Comparative politics is a series for students and teachers of political science that deals with
contemporary issues in comparative government and politics. As Comparative European
Politics it has produced a series of high quality books since its foundation in 1990, but now
takes on a new form and new title for the the millennium—Comparative Politics. As the
process of globalization proceeds, and as Europe becomes ever more enmeshed in world
trends and events, so it is necessary to broaden the scope of the series. The General Editors are
Max Kaase, Vice President and Dean of Humanities and Social Sciences, International
University, Bremen, and
Kenneth Newton, Professor of Government,
University of Southampton. The series is published in association with the
European Consortium for Political Research.
Edited by
KAARE STRØM
WO L F G A N G C . M Ü L L E R
TO R B J Ö R N B E RG M A N
1
3
Great Clarendon Street, Oxford OX2 6DP
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© the several contributors 2003
The moral rights of the authors have been asserted
Database right Oxford University Press (maker)
First published 2003
First published in paperback 2006
Publication of this book has been supported by a grant from the
Austrian Bundesministerium für Bildung, Wissenschaft und Kultur.
Publication of this book has been supported by a grant from the
project on Constitutional Change and Parliamentary Democracy
(Bank of Sweden Tercentenary Foundation 1996-0801)
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose the same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
Typeset by Newgen Imaging Systems (P) Ltd., Chennai, India
Printed in Great Britain
on acid-free paper by
Biddles Ltd., King’s Lynn, Norfolk
1 3 5 7 9 10 8 6 4 2
Preface to the Paperback Edition 2006
When Delegation and Accountability in Parliamentary Democracies was first published in
2003, it was a work of unusual ambition both theoretically and empirically. Empirically,
it was no small task to assemble data on the entire constitutional policy chain for the
entire post-World War II era for 17 West European democracies and to do so within a
coherent principal–agent framework. We would never have been able to do so without
the painstaking efforts of all our contributors, and we remain immensely grateful to
them for all their empirical and theoretical contributions.
Although we believe that the quality of the original publication was very high, we
have since discovered a few errors and misprints. This paperback edition contains a few
corrections, most of which had already been included in the second printing of the hard
cover version published in 2004. None of these corrections change any of our inter-
pretations or conclusions. When using our book, the reader should also be aware of a
few caveats. For all countries, we present data up to January 1, 2000. The one exception
is that for governments (cabinets) in power at that date, we present data for the forma-
tion but not for the full “life-cycle” of that cabinet. While we for most countries cover
the entire period since World War II, the tables and figures for Greece, Portugal and
Spain cover the period from the (re-)introduction of democracy in the 1970s. As for
France, our data begin with the inception of the Fifth Republic in 1958.
We have prepared an electronic file with the institutional data. These annual data
reflect the institutional arrangements of each country as of January 1. Those interested
in obtaining the institutional data in electronic form should visit the “Constitutional
Change & Parliamentary Democracies” project homepage (http://www.pol.umu.se/
ccpd/).
For help with the data file and for many other contributions to this edition, we are
indebted to Ivy Orr Hamerly. We are, once again, also grateful for the enduring support
and swift assistance of Claire M. Croft and Oxford University Press. It has been our
privilege and pleasure to work with them.
October 2005 Kaare Strøm, San Diego
Wolfgang C. Müller, Mannheim
Torbjörn Bergman, Umeå
This page intentionally left blank
Acknowledgements
Some academic projects are sustained by their intellectual rewards as well as by the
good companionship that goes into their creation. Such has been the case with this
volume. It has been long in the making. Each one of us has, in his own way and
throughout his career, been centrally concerned with governance in parliamentary
democracies. Indeed, we have worked jointly as well as individually in this field of
research. Yet, the labour that has resulted in this book began in earnest with a confer-
ence in Umeå in June 1997, at which the very first drafts of the chapters were pres-
ented. For each of us, this conference represented the largest scholarly enterprise we
had ever launched, both in terms of the sheer number of collaborators involved and in
terms of our scholarly ambitions. We were therefore suitably awed, and we have surely
wandered down our share of blind alleys, for which we apologize to our patient
contributors.
Recent years have witnessed a great and most gratifying revival in theoretically ambi-
tious and empirically comprehensive cross-national studies of democratic institutions.
We have nothing but the greatest respect for those scholars who are capable of pro-
ducing such works single-handedly. In our own study of European parliamentary
democracies, however, we were from the beginning convinced that teamwork was
essential and indispensable. At the same time, we have been acutely aware of the
administrative complications and potential inconsistencies that such forms of scholar-
ship imply. Therefore, our approach has been one that has combined a great deal of
collaboration with an elaborate and very detailed agenda.
Hence, the bulk of the data presented in this volume has been collected and com-
piled by a set of country specialists. In some cases, our specialists have been teams of
scholars; in other cases, they have been individuals. In either case, they have collected
and presented their data according to very specific guidelines. Thus, the country chap-
ters in this volume are based on a much more detailed set of questions than is usually
the case. We have surely occasionally driven our country experts to exasperation with
our long and numerous lists of open-ended as well as fixed-format questions about
institutional details in their respective national governance processes.
Generating this framework for our research was not a one-way process. The coun-
try authors helped define the questions that we asked in the first place and thus
contributed to our theoretical agenda as well as to the data that we present. The same
set of information was collected for each of our seventeen countries and for each year
in the 1945–2000 period (or for as much of that period as these countries have been con-
tinuously democratic). These questions were dutifully answered by our country spe-
cialists, who altogether prepared more than 250 draft tables. For the purposes of
economy and presentation, and again with the help of the contributors, most of this
information has now been compiled in the cross-national tables that constitute the
backbone of Chapter 4. Through these tremendous efforts, our collaborators have
viii Acknowledgements
helped us provide information about practically all postwar institutional changes that
impinge on the democratic delegation process in the parliamentary democracies of
Western Europe. We gratefully acknowledge their efforts.
The main sponsor of this scholarship has been the Bank of Sweden Tercentenary
Foundation, which through a grant to Torbjörn Bergman and the project on ‘Constitutional
Change and Parliamentary Democracy’ (Project No. 1996-0801) made the whole effort
possible. In addition, we have had more help and assistance than we can probably
remember to acknowledge. We want to thank especially all those who made our first
meeting in Umeå such a success. Also, let us mention Joseph B. Board, Gunnel
Gustafsson, Michael Metcalf, and Nelson W. Polsby—four individuals who in various
ways made lasting contributions at the formative stage of this project.
Since the original Umeå conference, we have enjoyed numerous opportunities to
meet with our contributors as our work progressed. The Political Academy (Vienna)
and its Director Günther Burkert Dottolo have supported this book by twice enabling
us to meet and develop our project. In 1998, we were also able to meet at the annual
meetings of the European Consortium for Political Research at the University of
Warwick. The same year, the University of Siena graciously hosted a group meeting at
its beautiful conference centre. We are indebted to Maurizio Cotta and to Luca
Verzichelli for this wonderful opportunity. The University of Kent similarly supported
a group meeting in 2000. We are grateful to the university and especially to Thomas
Saalfeld, who was hugely helpful in organizing this meeting. All of these meetings were
used at least in part to push forward our work on this volume. The Center of German
and European Studies at the University of California has supported the book through
several research assistantship grants, as has the Center for the Study of Democracy at
the University of California, Irvine. Last but certainly not least, we are grateful to the
Rockefeller Foundation, which, by granting a fellowship for collaborative residence to
Müller and Strøm, and graciously welcoming Bergman to join in, allowed us to meet
in the sunmer of 2001 at its beautiful and serene Villa Serbelloni in Bellagio, where we
completed many of the intellectual tasks of this volume.
Several individuals deserve special thanks for their contributions to the project.
Cindy Kite skillfully language edited the country chapters into more complete accounts
and more felicitous prose. Although such thanks must be kept to a minimum, we sim-
ply have to acknowledge the help of the two most prominent scholars on the Swedish
constitution, Erik Holmberg and the late Nils Stjernquist. And we gratefully acknowl-
edge the constructive suggestions of numerous colleagues who have commented on
earlier drafts of these papers, including James E. Alt, John D. Huber, and Mona Lyne.
Last, but certainly not least, we are indebted to Magnus Blomgren, Scott Kastner, and
Benjamin Nyblade for their painstaking and most conscientious research assistance as
well as for their creative and genuine contributions to our project.
All academics know about the perils of publishing. In this project, we have had noth-
ing but favourable experiences with the good offices of Oxford University Press. We
especially wish to thank our editor Dominic Byatt for his enthusiasm for our project
and his continuous support, and Alison Heard, Claire M. Croft, and Stuart Fowkes for
their warm and supportive efficiency in guiding our bulky volume through the various
stages of publication.
Acknowledgements ix
And to our fellow authors as well as our research assistants we would like to say that
while we—in what has been sometimes been referred to as the ‘troika’—are very happy
to see this book completed, we will always take pleasure in having worked with you. It
was great fun!
June 2003 Kaare Strøm
Wolfgang C. Müller
Torbjörn Bergman
This page intentionally left blank
Contents
List of Figures xiii
List of Tables xiv
List of Contributors xvi
I . I N T RO D U C T I O N A N D T H E O RY
I I . S U RV E Y
Index 751
List of Figures
2.1. A Delegation Model 38
2.2. Graphical Depictions of the Possible Situations 39
3.1. Delegation and Accountability Under Parliamentary and
Presidential Government 65
3.2. Political Parties in the Parliamentary Chain of Delegation:
An Empirical Assessment 68
3.3. Agency Loss (Policy Divergence) Under Parliamentary and
Presidential Government 78
4.1. The Causes of Political Party Cohesion 127
4.2. Types of No-Confidence Votes in Western Europe 156
4.3. The Policy Process Within Cabinet and (Potential)
Prime Ministerial Powers 180
4.4. Prime Ministerial Powers in Western Europe 191
22.1. Appointment Powers of Western European Presidents 679
22.2. Two-dimensional Map of Citizen Control 700
List of Tables
2.1. Common Forms of Delegation 34
2.2. How Moral Hazard Potential Affects Agency Loss 43
3.1. Policy Divergence Under Different Regime Types 81
3.2. Policy Divergence Under Different Regime Types: 82
Numerical Example
3.3. Regime Form and Agency Loss: A Summary 98
4.1. Decision-making Procedure: Ordinary Legislation 114
4.2. Strength of Upper Chambers 119
4.3. Decision-making Procedure: Constitutional Amendments 122
4.4. Mandatory and Minimum Constraints on Constitutional Amendments 126
4.5. The Legal Electoral Connection Between Citizens and MPs 131
4.6. Party Rules on Candidate Selection 137
4.7. Rules of Electoral Competition 140
4.8. Public Party Finance and Spending Caps 144
4.9. Rules of Cabinet Inauguration, Lower Chamber Rules Only 150
4.10. Rules of No Confidence (i.e. Initiative Rests with Parliament) 153
4.11. Rules of Confidence (i.e. Initiative Rests with Cabinet) 158
4.12. Rules of Parliamentary Dissolution 160
4.13. Confidence and Early Elections: The Behavioural Record 165
4.14. Parliamentary Accountability of Cabinet 169
4.15. National Scrutiny Process Before a Decision in the EU Council of 175
Ministers, Lower House Only (1993–2000)
4.16. Delegation Within the Cabinet: Prime Ministers and Cabinet Ministers 186
4.17. Conflict Management Mechanisms in Western European 195
Coalition Governments, 1945–99
4.18. Rules of Delegation from Cabinet Ministers to Civil Servants 199
4.19. Ministerial Duration and Mobility in Western Europe, 1945–84 204
4.20. Ideal-Typical and Actual West European Mechanisms of
Delegation and Accountability: Summary 208
5.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary 232
Dissolution: Austria 1945–2000
6.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary 262
Dissolution: Belgium 1946–2000
7.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary 288
Dissolution: Denmark 1945–2000
8.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary 310
Dissolution: Finland 1945–2000
9.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary 336
Dissolution: France 1959–2000
10.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary 358
Dissolution: Germany 1949–2000
11.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary 386
Dissolution: Greece 1977–2000
List of Tables xv
I N T RO D U C T I O N A N D T H E O RY
This page intentionally left blank
1
I N T RO D U C T I O N
Democracy is neither simple nor straightforward, but it is feasible. While the twentieth
century witnessed some of the greatest political atrocities in human history, it was also
the century in which democracy was gradually established and consolidated in large
parts of the world. At the end of the millennium, there were more stable democracies
than ever before, and there was no continent in which democracy was entirely foreign
(Lijphart 1999; Przeworski et al. 2000).
To fulfill its promise, democratic government must be benign as well as effective.
Many distinguished political thinkers have been fortunate enough to be able to pre-
sume government effectiveness—hence their major concern has been to be tame and
constrain government power. Yet, democratic government cannot exist without a gov-
ernment capable of enforcing law and order, as Weber (1972) reminded us, and
securing revenue, as Schumpeter (1954 [1918]) has pointed out. At the same time,
democracy must mean limited government—the exist-ence of a protected private
sphere in which the government is prohibited from regulating, and a set of human and
civil rights that it is obligated to respect. While the precise boundaries between the pub-
lic and the private sphere are constantly revised, there must be some form of private
sphere and some commitment to a set of citizen rights (Dahl 1971).
Popular government may take many forms, but contemporary exemplars are pre-
dominantly representative democracies. Thus, popular sovereignty is exercised through
delegation from citizens to individual politicians and collective actors, in particular polit-
ical parties. Just as citizens in democratic societies may delegate important tasks in their
private lives to specialists (e.g. their medical treatment to physicians and the education
of their children to teachers), they delegate the greater part of their public life to
politicians.1 Moreover, those who are directly elected in turn delegate to specialists
1
Or, to be more precise, such acts of delegation have happened in the past, or are assumed to have
taken place. Indeed, most citizens face a set of rules of delegation that was set up without their formal
participation.
4 Wolfgang C. Müller et al.
many of their more specific tasks. It is politicians who deliberate, make laws, and
take other political decisions. And, eventually, these decisions are implemented by
bureaucrats that politicians recruit and supervise.
You cannot always trust politicians, or, for that matter, anyone else to whom you
delegate. Hence, delegation is inherently risky. The danger of any delegation is that it
may turn into abdication, that the person who delegates may lose control rather than
get help. To avoid such disasters, delegation to politicians needs to be coupled with
some mechanism of accountability, some way in which these politicians can be checked
and controlled, and if necessary removed, if their behaviour or performance in office is
unsatisfactory. Democracy therefore means effective but limited government containing
mechanisms of delegation as well as accountability (Manin, Przeworski, and Stokes 1999: 29;
Powell 2000; Strøm 2000).
Parliamentary government, which we shall formally define later in this chapter, is
the most common way to organize delegation and accountability in contemporary
democracies. Of the thirty-six states with a minimum of 19 years of democratic
experience covered in Lijphart’s magisterial Patterns of Democracy (1999), as of 2003 only
five—the United States, Switzerland, Colombia, Costa Rica, and Venezuela—are not
parliamentary by our standards (see below).2 Woldendorp, Keman, and Budge (2000),
who apply less restrictive criteria and claim to have excluded only tiny states, count
forty-five parliamentary states by our criteria (though not necessarily by theirs). In fact,
about a third of the world’s population live under this regime form, a larger proportion
than for any other system of government. The states that feature parliamentary
constitutions span all continents and include some of the largest countries in the world
(India), as well as some of the most economically advanced and powerful ( Japan),
and some of those commonly considered to have the highest qualities of life
(Canada, New Zealand). Yet, Europe remains the heartland of parliamentarism. In
Western Europe (leaving aside microstates such as the Vatican), only Switzerland is not
parliamentary. And by and large, the great majority of European citizens accept their
parliamentary constitutions as legitimate vehicles for popular representation.
This volume examines political delegation and accountability in the parliamentary
democracies of Western Europe. Our aim is to add to our understanding of parliamentary
democracy both empirically and theoretically. Empirically, we survey the institutions
that govern political delegation and accountability in seventeen Western European
parliamentary democracies. Electoral and governmental institutions in Western Europe
display great variation, with consequences for democratic delegation and accountability.
While such observations are commonplace, our knowledge of the range that parlia-
mentary institutions take is still remarkably unsystematic and incomplete. Delegation
and Accountability in Parliamentary Democracies is certainly the most ambitious, as well
as the most comprehensive, account of the institutions of democratic delegation in
2
Lijphart (1999) also classifies Israel as non-parliamentary after the introduction of a directly elected
Prime Minister in 1996. A recent institutional reform has once again removed this anomaly, however, and
returned Israel to the set of parliamentary systems.
Parliamentary Democracy 5
3
This, of course, is not to say that the agency problems of representative democracy have previously gone
unnoticed. Dennis C. Mueller’s impressive Constitutional Democracy largely shares our normative concerns
and analytical perspective, seeing democracy as a giant agency problem (1996: 17). The empirical research
that comes closest to our concerns, although it does not share our analytical foundations, is the British
Democratic Audit project (Weir and Beetham 1999), which measures British democracy against a set of demo-
cratic standards and covers all steps in the chain of delegation. Finally, numerous individual authors, mainly
examining a single step in the chain of delegation in a single country, have identified problems of delegation
and accountability stemming from preference divergence and information asymmetries (see Chapters 2 and 3).
They have provided valuable information on delegation and accountability in their respective countries, as
can be seen from the references in the seventeen country chapters. We build on all this work.
6 Wolfgang C. Müller et al.
over much of national politics out of the hand of a hereditary monarchy. This histori-
cal meaning of the term parliamentary government (or parliamentarism) often creates
confusion when it is used in reference to modern representative democracies. This prob-
lem is addressed in the section titled Parliamentary Government. In the following
section, we compare the regime type of parliamentary democracy with its main rival,
presidentialism. Even though there has been a heated debate about the relative merits
of parliamentary versus presidential government, surprisingly little is known about
the strengths and weaknesses of parliamentary democracy as a mechanism of delega-
tion from voters via politicians to civil servants. We then begin to develop our own
principal–agent approach to the challenges of delegation and accountability in parlia-
mentary democracies. The final section outlines the rest of the book.
T H E H I S TO RY O F PA R L I A M E N TA RY G OV E R N M E N T
Unlike American federalism or presidentialism, parliamentary government was not the
product of deliberate institutional design. Rather, it gradually evolved in Britain over
several centuries.4 Norton (1981: 12) identifies 1688 as the beginning of parliamentary
dominance, when Parliament prevailed in its conflict with King James II, who was
forced to flee the country. Parliament was then in 1689 effectively able to select the new
King, William of Orange, and to impose significant constraints on his powers.
According to Lowell (1896: 3), the roots of parliamentary government can be traced
back to 1693, when the King first appeased the House of Commons by appointing a
government out of the party (the Whigs) that enjoyed a majority there. Important fur-
ther developments occurred in the late eighteenth century, when William Pitt the
Younger, based on solid support in the House of Commons, established the leading
role of the cabinet vis-à-vis parliament and, in conducting its daily business, vis-à-vis the
monarch. At that time the cabinet was still that of the Crown, though as a rule cabinets
were appointed that could count on the support of the House of Commons. Yet, the
Commons was still dominated by the King and the aristocracy, who controlled access
to the vast majority of seats.
In 1830, however, the House of Commons forced Wellington to resign over his
unwillingness, due to a deep split among the Tories, to engage in parliamentary reform.
Earl Grey’s subsequent Whig Cabinet persuaded the King to dissolve Parliament and
then won the general election on the issue of parliamentary reform. The reform passed
the House of Commons but was held up by the House of Lords. Grey then asked the
King to create enough new peers to force the reform through the Lords. When the
King refused, Grey, backed by the Whigs, tendered the cabinet’s resignation. The King
unsuccessfully tried to install a new Wellington cabinet but eventually had to invite the
Whigs back into office and accept their terms. In the end, the House of Lords
4
See, for example, Mackintosh (1968); Kluxen (1983); Norton (1981); Judge (1993) ch. 1; Cox (1987); see
also Goldsworthy (1999).
Parliamentary Democracy 7
pre-empted the appointment of many Whig peers and passed the reform bill—The
Representation of the People Act—which eliminated or reduced the weight of small
boroughs, where elections were often corrupt, and created new seats in towns and
cities. The electorate increased from 220,000 to more than 500,000. By 1866, increasing
wealth had swollen the ranks of voters to more than one million.
After the 1832 Reform Act, the Crown appears to have had scant influence over
cabinet appointments or deliberations—and thus became the ‘dignified part’ of the
British Constitution (Bagehot 1963 [1867]). The House of Commons, on the other
hand, criticized and could dismiss the cabinet and in most cases also determine cabinet
appointments. Over the subsequent decades, party government was established, and
the cabinet came to dominate the House of Commons. According to Gary Cox (1987),
this was an indirect response to heightened electoral competitiveness. Members of
Parliament (MPs) became more active because their re-election increasingly depended
on their visibility and constituency service. Yet, while parliamentary activism was indi-
vidually rational, it was collectively self-defeating. Too many active and ambitious MPs
seeking position-taking and credit-claiming opportunities threatened to overburden the
parliamentary agenda and induce institutional paralysis. Hence, those procedural rights
that had been abused by visibility-seeking MPs were abolished and the cabinet’s agenda
control strengthened.
Parliamentary behaviour also became more party-based as the larger constituencies
required both a different kind of campaigning and greater activity in Parliament. This
process accelerated with the electoral reform acts of 1867 and 1884, which enfranchised
a much larger electorate. Now the fate of cabinets was no longer decided on the floor
of the House of Commons, as it had been in the ‘golden age’ of parliamentary
government (1832–67), but by the general electorate. ‘Members of Parliament were
gradually relegated to being representatives of that opinion, and their freedom of par-
liamentary action was correspondingly diminished’ (Norton 1981: 16). Consequently,
cabinet members and MPs began to address the general electorate rather than
the House of Commons. Procedural reform continued and led to the evolution of a
government-managed parliament (Norton 1981: 19–20).
PA R L I A M E N TA RY G OV E R N M E N T
The terms parliamentarism, parliamentary government, and parliamentary democracy are
often used interchangeably, and so far we have not established any explicit distinction
between them. Before we engage in a more substantive discussion of parliamentary
government and democracy, however, we need to clarify our use of these key terms.
Let us first consider parliamentary government as a positive term denoting a particular
institutional arrangement.
The term parliamentary government was not used until 1832 in Britain and the late
1830s on the continent (von Beyme 1999: 29–33). Competing terms include represent-
ative government ( J. S. Mill 1984 [1861]), responsible government, and cabinet gov-
ernment (Bagehot 1963 [1867]). Note that the term parliamentary government
conventionally focuses on the relationship between parliament and government. There
is no necessary connection to citizens and full universal suffrage. Parliamentary
government is thus possible without democracy. Indeed, such early theorists of
parliamentary government (though they did not use this term) as John Stuart Mill and
Walter Bagehot favoured a limited franchise.
Parliamentary regimes are characterized by a number of specific and interrelated
institutional features, upon which its dynamics hinge. To illuminate these institutional
effects, a number of authors, most notably Bryce (1921), Loewenstein (1957, 1975),
Verney (1959), and von Beyme (1973, 1999), have sought to provide comprehensive def-
initions of parliamentary government. Their mutual differences notwithstanding, these
definitions collectively include the following criteria:
1. There is a dual executive (split between the head of government and the head of
state).
2. Parliament has formal or informal investiture powers.
10 Wolfgang C. Müller et al.
3. The cabinet is a collective decision-making body.
4. Cabinet Ministers are usually also members of parliament.
5. The cabinet is politically responsible to the parliamentary majority.
6. Parliament has means of control (interpellations, committees of inquiry, etc.)
over the cabinet.
7. The parliamentary majority can force the cabinet to resign.
8. This power is in most cases balanced by the Prime Minister’s power to dissolve
parliament.
The properties listed above certainly jointly describe a number of salient features of
parliamentary government. Yet, while such a comprehensive conception enhances our
configurative understanding, it does not cleanly differentiate between parliamentary
government and other regime forms. It is commonly acknowledged that not all parlia-
mentary systems exhibit this list of features in its entirety, and conversely that some of
these properties adhere to non-parliamentary systems as well. But a characterization
that specifies neither necessary nor sufficient conditions is not very helpful analytically.
Moreover, it is not clear whether some of these properties are simply the behavioural
consequences of others. Most comprehensive definitions do not make it clear what
are definitional properties of parliamentary government and what are contingent fea-
tures or mere behavioural regularities. Even if in principle these conceptual problems
could be solved, the problem remains that some of these criteria are difficult to opera-
tionalize. And the more (necessary) definitional criteria we add, the more likely we
exclude from the parliamentary category constitutions that otherwise have very much
in common.
Minimal Definitions
We believe in the virtues of simple and unambiguous classifications, and hence will not
attempt to provide a comprehensive definition. Rather, we prefer a minimal definition,
which should provide the simplest and sharpest analytical tool by which we can distin-
guish parliamentary systems from others. Several such minimal definitions have been
proposed. Lijphart (1984: 68, drawing on Epstein 1968) defines parliamentary govern-
ment as ‘a form of constitutional democracy in which executive authority emerges
from, and is responsible to, legislative authority’. The ‘emergence’ criterion denotes
that Prime Ministers are selected by parliament, though Lijphart includes in this cat-
egory cases in which the head of state appoints the Prime Ministers, who ‘emerge from
interparty bargaining’ (1984: 69).5 Similarly, Sartori (1997: 101) states that ‘parliament is
sovereign’ under parliamentarism and that this regime type requires ‘government to be
appointed, supported and, as the case may be, dismissed, by parliamentary vote’.
Stepan and Skach (1993: 3) define a parliamentary regime as a system of mutual
dependence: ‘(1) The chief executive power must be supported by a majority in the
5
In a later publication Lijphart (1999: 116–8) adds ‘collective or collegial executives’ as a third criterion,
and thus moves a step away from a minimal definition.
Parliamentary Democracy 11
legislature and can fall if it receives a vote of no confidence. (2) The executive power
(normally in conjunction with the head of state) has the capacity to dissolve the
legislature and call for elections’. And Shugart and Mainwaring (1997: 14–5) thus define
parliamentary in contrast to presidential democracy: ‘While in presidential systems the
head of government is popularly elected, this is not the case with the Prime Minister in
parliamentary systems. Furthermore, the electoral terms of the government and the
assembly are fixed in presidential systems but not in parliamentary systems’.
Useful as these contrasting definitions are, they are not quite satisfactory. There are
three problems. One is analytical: some of the terms of these definitions are ambiguous,
or do not effectively partition the world’s democratic regimes. The second problem is
empirical: Some definitions, if taken seriously, would lead us to categorize particular
political systems in ways that few analysts seem to find compelling. The third and most
important problem is theoretical: None of these definitions tell us what is democratic
about parliamentary democracies. We shall address the first two of these problems
below, and the third and most important one later in this chapter.
The first (analytical) problem is thus that many existing definitions of parliamentary
government are operationally ambiguous. Different scholars using the same criteria
may classify the same political systems differently, as Robert Elgie (1998) has amply
demonstrated. Also, there are political systems that by common agreement are neither
parliamentary nor presidential. Indeed, much of the literature on regime types has
concerned itself with such hybrids as the French Fifth Republic or Switzerland. In the
case of the French Fifth Republic, the executive does not emerge from the legislature,
whereas in Switzerland it does not depend on the legislature’s confidence. How, then, do
we categorize France? Is it parliamentary or presidential? And what about Switzerland?
Or Poland? Or Russia? In fact, a growing number of constitutions do not fit the standard
definitions. These constitutions, and the French Fifth Republic in particular, have spawned
a growing literature on ‘semi-presidential’ or ‘premier-presidential’ government
(Duverger 1980; Shugart and Carey 1992; Sartori 1997; Elgie 1998, 1999).
The second (empirical) problem is that if parliament’s role in the ‘emergence’
(Lijphart) or ‘appointment’ (Sartori) of the cabinet means some form of active selec-
tion, then this criterion simply does not very faithfully describe the process of govern-
ment formation in systems conventionally labelled as parliamentary. It is true that in
some such systems, for example Ireland, parliament actually elects a Prime Minister by
majority vote. In other systems, such as Italy, parliament must, after a new Prime
Minister has been appointed, approve the incoming cabinet in an investiture vote. In
most cases, however, there is simply no mechanism by which parliament directly selects
the Prime Minister or cabinet. Britain, the birthplace of modern parliamentary demo-
cracy, has no such institutional vehicle. Nor does Austria, Denmark, Finland, The
Netherlands, or Norway, just to name a few (for empirical surveys, see Laver and
Schofield 1990; Bergman 1993; De Winter 1995: ch. 4; and Chapter 4 of this volume).
In the real world, then, parliamentary government rarely means that the legislature
actually elects or appoints the executive (von Beyme 1973: 41–3). A similar problem
is that those definitions that include the executive’s dissolution power (e.g. Stepan
and Skach 1993; Shugart and Mainwaring 1997) by implication exclude cases that are
12 Wolfgang C. Müller et al.
normally understood as parliamentary. In declining order of severity, this problem
applies to Norway, whose constitution simply does not permit early parliamentary dis-
solution, Germany, in which it is limited to rare circumstances explicitly described in
the constitution, and Sweden, where dissolution is contained by the strong disincentive
that a new parliament would be elected only for the remainder of the dissolved
parliament’s term.
6
Riggs (1988: 252) also mentions presidential veto rights and the possibility of executive policies not
endorsed by the assembly as ‘natural consequences’ of presidentialist systems. However, they are ‘not the
essence of presidentialism’.
Parliamentary Democracy 13
parliamentary government is a system of government in which the Prime Minister and his or her
cabinet are accountable to any majority of the members of parliament and can be voted out
of office by the latter, through an ordinary or constructive vote of no confidence.7 For
simplicity, we shall also use parliamentarism as a shorthand for a parliamentary system
of government.8
PA R L I A M E N TA RY V E R S U S P R E S I D E N T I A L G OV E R N M E N T
Rules matter to the extent that they affect the game that is played. So also with polit-
ical regime types such as parliamentarism. Much of the current literature on parlia-
mentary government is concerned with its political effects compared to other regime
types. The most frequent alternative is presidentialism, and there is a sizeable literature
comparing their respective effects and merits.9 Bryce (1921: ii. 464–76) was probably the
first scholar systematically to assess the comparative merits of modern parliament-
arism and presidentialism (plus the Swiss executive council). Bryce compared these
regime types along three dimensions, captured in three questions:
1. Which of them best succeeds in giving prompt and full effect to the Will of the
People?
2. Which is best calculated to guard against errors into which the people may fall by
ignorance, haste, or passion?
3. Which secures the highest efficiency in administration?
Bryce did not provide a clear answer to his third question. Yet, in his words,
the Presidential system leaves more to chance than does the Parliamentarian. A Prime Minister
is only one out of a Cabinet, and his colleagues may keep him straight and supply qualities want-
ing in him, but everything depends on the character of the individual chosen to be President. He
may be strong or weak, wise or short-sighted. He may aim at standing above party and use his
authority and employ his patronage with a single eye to the nation’s welfare, or may think first
of his own power and his party’s gain, . . . (Bryce 1921: ii. 469)
7
For systems with bicameral legislatures, it suffices for the Prime Minister and cabinet to be accountable
to the majority in one chamber. Empirically, this is typically the lower chamber. Constitutions under which
the Prime Minister and cabinet are accountable to both chambers, such as Italy, are the exception.
8
We are aware that this common usage can be criticized because an ‘ism’ typically refers to a normative
perspective or movement. Indeed, in the nineteenth century parliamentarism was a democratic movement
against autocratic rule (Schütt-Wetschky 1984: 38–9).
9
Other democratic alternatives are assembly (or convention) government (Loewenstein 1957, 1975; von
Beyme 1973, 1999) and Swiss directoral government.
14 Wolfgang C. Müller et al.
Parliamentary system, because whatever the Executive needs it is sure to obtain from
its majority in the Assembly, subject of course, to any check which the existence of a
Second Chamber may provide’10 (Bryce 1921: ii. 475). With regard to Bryce’s second
question—which system guards better against errors caused by ignorance, haste, or
passion—he favoured presidentialism: ‘by dividing power between several distinct
authorities, it provides more carefully than does the Parliamentary [system] against
errors on the part either of Legislature or Executive, and retards the decision by the
people of conflicts arising between them’ (1921: ii. 474).
Before arriving at his verdict, Bryce discussed at some length the merits and demerits of
parliamentarism and presidentialism. In the present context it may suffice to summarize
his judgement on parliamentarism. Parliamentary government, according to Bryce,
concentrates the plenitude of popular power in one body, the Legislature, giving to its majority
that absolute control of the Executive which enables the latter, when supported by the
Legislature, to carry out the wishes of the majority with the maximum of vigour and prompt-
ness . . . The essence of the scheme is that the Executive and the majority in the Legislature work
together, each influencing the other; the Cabinet being in fact an Executive Committee of the
Legislature . . . The working of the scheme presupposes not only the existence of parties, but a
sentiment of party unity strong enough to induce the majority in the Legislature to entrust a
large discretion to the Cabinet, and to support it, except now and then in very grave matters, with
a trustful loyalty which assumes its action to have been right till proved to have been wrong.
(Bryce 1921: ii. 465)
This description led Bryce to identify the following merits of parliamentary govern-
ment: It is
calculated to secure swiftness in decision and vigour in action, and enables the Cabinet to press
through such legislation as it thinks needed, and to conduct both domestic administration
and foreign policy with the confidence that its majority will support it against the attacks of
the Opposition. To these merits there is to be added the concentration of Responsibility. For
any faults committed the Legislature can blame the Cabinet, and the people can blame both
the Cabinet and the majority. (Bryce 1921: ii. 465–6)
However, these merits of parliamentarism ‘are balanced by serious defects’ (Bryce
1921: ii. 466), among which Bryce counts excessive partisanship and erratic policy
making. In his view, parliamentary government nurtures an ‘unhealthy’ competition
between politicians and parties: ‘The system intensifies the spirit of party and keeps it
always on the boil.11 Even if there are no important issues of policy before the nation
there are always the Offices to be fought for. One party holds them, the other desires
them, and the conflict is unending, for immediately after a defeat the beaten party
begins its campaign to dislodge the victors’ (Bryce 1921: ii. 466).
10
Obviously, Bryce did not consider minority governments, which account for 37% of Western European
governments in the 1945–99 period (Müller and Strøm 2000: 561). Under this type of government, parliaments
are less likely to be the willing executor of government policies.
11
Bryce (1921: ii. 335) attributed the ‘decline of parliaments’, his famous thesis, to the rise of party
machines and the multiplication of parties, which, in turn, was said to cause cabinet instability and faulty
compromises in policy-making.
Parliamentary Democracy 15
T H E PA R L I A M E N TA RY C H A I N O F D E L E GAT I O N
The policy process in contemporary democracies, from voters all the way to the civil
servants that ultimately implement public policy, can be viewed as a process of
delegation, in which those authorized to make political decisions conditionally designate
others to make such decisions in their name and place. The basic normative assumption
is that policy-makers in a democracy should do what the citizens want them to do.12
In democratic societies, citizens delegate to politicians first and foremost through
elections. Democratic elections need to satisfy a number of criteria. They must be held
under conditions of political competition (freedom to form and join organizations
freedom of expression, alternative sources of information, and freedom of candidacy),
and there must be an independent and competent administration, as well as appropri-
ate judicial bodies, to implement the election law (for the ‘classic’ argument, see Dahl
1956: 84, 1971: 3, 1989: 221–2; for a recent application, see Choe 1997).
As the primary democratic institution, elections have a double nature, allowing
for delegation as well as accountability (Powell 2000). As mechanisms of delegation,
elections vest politicians and parties with a mandate, and as instruments of account-
ability they provide the opportunity to hold politicians and parties responsible for their
actions in office. In the former (mandate) perspective, parties and politicians present
their programmes and credentials to the voters and are elected on the basis of the
voters’ policy choices and their own perceived competence.
Elections can induce accountability to that extent that elective offices are attractive
and that incumbents therefore generally would like to remain incumbents or even
to rise in the political hierarchy, the motivation identified by Schlesinger (1966) as pro-
gressive ambition. The fear of electoral punishment is a strong incentive for incumbents
to remain in tune with their voters’ demands (see e.g. Sartori 1987: 156). Even if
representatives have no ambition to run again, or when term limits prohibit them from
doing so, the salutary effects of electoral accountability may operate indirectly. In such
cases, the electoral connection constrains politicians through their collective affiliation
with political parties, which seek to remain in business well beyond the terms of indi-
vidual politicians, and hence do their best to make incumbents with discrete ambition
(i.e. ambition that does not extend beyond their current term of office) behave as if
they would face the electorate again (Schlesinger 1991; Wittman 1995; Carey 1996;
Müller 2000).
Yet, democratic elections are only the first step in the delegation process. John Stuart
Mill’s prescription for the ‘proper office of a representative assembly’ thus stresses the
need for parliamentary delegation to the cabinet. Since ‘no body of men . . . is fit for
12
There are at least three arguments which can be made against this assumption. First, because of the
problems of preference aggregation, it may simply be impossible to know what citizens want (Riker 1982).
Second, if doing what the citizens want means doing what the majority (or plurality) of citizens want, there
are normative limits to majority rule (human rights). Third, elitists may claim that leaders know their fol-
lowers’ interests better than the latter do themselves (see Manin, Przeworski, and Stokes 1999). We return to
these issues in Chapter 3.
20 Wolfgang C. Müller et al.
action’, parliaments ‘ought not to administer’ (Mill 1984 [1861]: 249), ‘or to dictate in
detail to those who have the charge of administration’ (1984 [1861]: 250). A parliament
‘is as little fitted for the direct business of legislation as for that of administration’, Mill
(1984 [1861]: 254) maintained. Accordingly, it ‘has never been thought desirable that
Parliament should itself nominate even the members of a Cabinet’ (Mill 1984 [1861]:
253). Hence, parliament should not make administrative decisions, select ministers (or
even civil servants), or even legislate in a substantive sense.
Instead of the function of governing, for which it is radically unfit, the proper office of a repre-
sentative assembly is to watch and control the government: to throw the light of publicity on its
acts: to compel a full exposition and justification of all of them which any one considers
questionable; to censure them if found condemnable, and, if the men who compose the
government abuse their trust, or fulfil it in a sense which conflicts with the deliberate sense of
the nation, to expel them from office, and either expressly or virtually appoint their successors.
(Mill 1984 [1861]: 258).
Mill favoured a system in which parliament delegates extensively and limits itself to
the tasks of controlling the executive and to be an arena for public debate. There are,
however, many ways in which citizens can delegate to politicians, and many ways in
which politicians can delegate among each other. Let us therefore consider more sys-
tematically the nature of democratic delegation from voters to those who govern. This
is a chain in which we can identify at least four discrete steps:
P RO B L E M S O F D E M O C R AT I C D E L E GAT I O N
Delegation is often problematic, and delegation in national governments is no excep-
tion. Such problems have not gone unnoticed. Robert Michels, for example, was
both aware of the need to delegate to party leaders and highly critical of its effects:
‘democracy is inconceivable without organization’, he wrote, yet organization ‘gives
birth to the domination of the elected over the electors, of the mandataries
over the mandators, of the delegates over the delegators. Who says organization says
oligarchy’ (Michels 1962: 61, 365). In the same spirit Robert De Jouvenil ([1914] quoted
in Loewenberg and Patterson 1979: 21) coined the well-known dictum that two
13
Note that if we take the idea of singular agents to its logical extreme, the ideal-typical parliamentary
democracy is also a unitary state (such as Britain, New Zealand, or Sweden, but not Austria or Canada), with
a unicameral parliament (such as the Nordic countries or, again, New Zealand).
14
In our elaboration of this framework in Chapter 3, we recognize that even under an ideal-typical parlia-
mentary democracy, principals sometimes delegate to multiple agents. Yet, these agents do not generally
compete with one another.
22 Wolfgang C. Müller et al.
revolutionaries, only one of whom is a member of parliament, have less in common
than two members of parliament, only one of whom is a revolutionary. There are also
numerous references to ‘downstream’ delegation problems between ministers and civil
servants, from Max Weber (1972: 572–4), who in naming ministers ‘dilettantes’ and civil
servants ‘experts’ clearly indicated agency problems, all the way to the popular television
comedy series Yes, Minister (Lynn and Jay 1981).
The contemporary literature continues to express concern about delegation, though
perhaps in a more disaggregated fashion than in the classical literature. Yet, larger
claims about the possibility of democratic delegation have by no means been
abandoned. They range from those who hail the virtues of democratic delegation to
those who think that delegation equals abdication. Among those who celebrate the
successes of parliamentary delegation are Klingemann, Hofferbert, and Budge (1994),
who find that government expenditure in several policy areas responds to party manifestos.
Among the less sanguine analysts, there are those who claim that delegation works, but
only at election time. One important strand of modern democratic theory, the
competitive model, puts great store in accountability at election time. All voters can do
is to choose a set of politicians who appear more likely to do the ‘right things’ and to
punish politicians who have not done so (see Schumpeter 1942; Riker 1982). In a more
cynical formulation, then, the people rule directly only on election day; otherwise
they are ruled by the politicians to whom they have delegated the task of making
authoritative decisions (Schattschneider 1960: ch. VIII; Sartori 1987: 28–31, 86–9).
But even this view is too positive for some. Even if delegation works on election day,
it may have unintended and negative consequences. According to Mayhew (1974) the
‘electoral connection’ triggers legislative behaviour that is individually rational (by
enhancing re-election prospects) but that does not contribute to, and may even under-
mine, the collective goals of the United States Congress. Hence, individual congressmen
engage in ‘advertising’, ‘credit claiming’, and ‘position taking’. The ‘electoral connec-
tion’ thus induces members to shirk their legislative responsibilities, in which credit
claiming is more difficult. Similarly, Fiorina (1977: 39–47) argues that congressmen
devote their time and energy to casework and pork-barrel politics rather than legislation.
While legislation is likely to be controversial to at least some voters, pork-barrel politics
and casework are largely popular. In the most cynical interpretation, the lack of emphasis
that representatives put on legislation creates problems that are then resolved piecemeal by
the same politicians whose re-election is guaranteed by exactly these services.
Finally, the message of the most severe critics, is that delegation simply does not work.
Politicians and other public officials are autonomous, and citizens have no real influence
on what their representatives are doing. Consequently, politicians may misbehave in one
of two ways: policy drift or rent-seeking. In the former case, ‘Politicians may want to
pursue their own ideas even if these differ from those of citizens’ (Manin, Przeworski,
and Stokes 1999: 40). In the latter case, they may use political power to chase material
advantage. The ‘grabbing hand’ perspective in economics argues that politicians do not
maximize social welfare but instead pursue their own selfish objectives (Shleifer
and Vishny 1998: 4). The literature on rent-seeking politicians indeed claims that ‘gov-
ernments transfer wealth not just among subgroups of citizens but also directly to
Parliamentary Democracy 23
themselves’ (McChesney 1997: 35). Rent extraction by politicians becomes more likely
when institutions allow them to cloud responsibility for inefficient government action.
And according to influential theories of bureaucracy, non-elected public officials are
also likely to behave in ways that defy the voters’ preferences (Downs 1967; Niskanen
1971).15
Problems of delegation thus give rise to some of the most scathing criticisms of
modern democracies. Although much of this scholarship does not focus squarely on
parliamentary government, and although we may differ with many of its premises, it
holds major implications for our investigation of parliamentarism. Parliamentary systems
certainly face their own delegation problems, with critical implications for our assess-
ment of contemporary democracies. But although it is hugely consequential whether
parliamentary delegation does or does not work, we need sharper tools and a more
detailed plan of campaign before we can properly return to this debate. It is to these
needs that we next attend.
Agency Problems
Let us now return to the problems of delegation under parliamentary democracy. We
shall discuss these problems in the language of agency theory. The difference between
what the principal wants and what the agent delivers is known as agency loss (the tech-
nical definition will be made clear in the Chapter 2). Principals (in our case, citizens) may
be poorly served by their agents (politicians) for either of two reasons (or both). One
problem is that the agents may have preferences that differ from those of the principals.
The other problem is that the principals may not know enough about their potential
agents to get the best possible deal from them. The latter, informational, problem may
in turn come in two forms. First, principals may not be able to choose the right agents
in the first place (adverse selection). Second, principals may not be able to keep their agents
honest and diligent (moral hazard). In the Chapters 2 and 3, we shall expand on these
problems and their application to the chain of delegation in parliamentary democra-
cies. In this section we briefly review the extent to which the literature on parliamentary
democracy has identified adverse selection and moral hazard problems. While the initial
and final stages of the parliamentary chain of delegation have traditionally been seen as
beset with delegation problems, much less concern has been given to delegation from
the parliamentary majority to a Prime Minister or cabinet, or to delegation within the
cabinet. Yet, agency loss can occur in any link in the parliamentary chain of delegation,
and only if the whole chain works can the democratic promise be fulfilled.
Adverse Selection
After the first wave of European democratization crested immediately following the
First World War, problems of parliamentary democracy soon figured prominently in
15
For a theoretical critique of Niskanen (1971), see, for example, McLean (1987: 89–100) or Dunleavy
(1991). For an assessment of the empirical fit of Niskanen’s theory, see Blais and Dion (1991).
24 Wolfgang C. Müller et al.
the political debate. Among the principal protagonists in this debate were Gaetano
Mosca and Carl Schmitt, whose contributions had a more lasting impact than most oth-
ers. A core element of their critique was that parliament had been robbed of its power.
Although, of course, he never used this label, Gaetano Mosca was centrally concerned
with the problem of adverse selection, specifically with the dominance of parliament
by societal elites. In his view, elections were controlled by organized minorities, leaving
the great majority of people little effective choice of candidates (Mosca 1950: 135–6).
Consequently, parliaments would represent the interests of wealthy and powerful elites
rather than the popular majority. Mosca also saw this selection problem as a cause of
patronage politics. ‘Influential voters’, that is societal elites, would make MPs intervene
in administrative and judicial processes. Because the MPs would depend on the ‘influ-
ential voters’, and the government, in turn, on the MPs, ‘all moral and legal considera-
tions’ would be pushed to the background. The government and the representatives’
‘permanent and obvious conflict between duties and interests’ would lead to a situation
in which ‘the bureaucracy and the elected bodies, which are meant to check each other,
corrupt each other’ (Mosca 1950: 217).
As Mosca presented his concerns, they do not pertain specifically to parliamentary
regimes. The charge that powerful societal elites control elected politicians can be
levelled against any system of representation (see, for instance, Mills’ (1956) critique of
the United States). Mosca most likely associated these problems with parliamentary
government because it was the regime under which he lived. At the end of his career,
however, Mosca (1947, 1950: 215) conceded that parliamentarism had to be compared
with other real-world regimes, in particular those that preceded or succeeded it
(i.e. constitutional monarchy and Fascism, respectively). In this comparison, the delegation
problems of parliamentarism seem less damning. Mosca also suggested remedies for
the problems of parliamentary government (Mosca 1950: 219–22), namely
In the terms with which this volume will examine parliamentary democracy, most of
these suggested remedies constitute constraints on the agents in the democratic chain
of delegation. As we shall see in later chapters, it is intriguing to note the striking ways
in which much of Mosca’s reform agenda is being heeded, more than half a century
later. But that is a story to which we shall return in Chapters 4 and 22.
Moral Hazard
Mosca does not seem to have seriously considered the ways in which mass political
parties might help solve the delegation problems with which he was concerned by
elevating the representatives of more humble classes. Yet, although political parties
might have helped counter adverse selection, they brought their own problems.
Parliamentary Democracy 25
Writing in a time when parties had already become the dominant political players,
Ostrogorski (1907: ii. 712–17) lamented that they had practically destroyed political
accountability. Under party-based parliamentary government, he claimed, ‘the respons-
ibility which is supposed to govern parliamentary relations comes to nothing but general
irresponsibility’, because the MP’s responsibility ‘disappears in that of the party’ . . .
‘however incompetent or culpable’ ministers may be, ‘it is impossible to punish one of
them without punishing all’. Given party cohesion, under party government the fall of
a cabinet could only happen by ‘accident’. In between elections the government ‘can do
very much what it likes’. Parliamentary control ‘is almost non-existent’. Thus,
Ostrogorski’s main point is the lack of accountability under party-based parliamentary
government. In a word, his complaint is that parties foster moral hazard.
Likewise, in his famous English Government and Politics, Frederic Austin Ogg argued
that executive agenda control in practice had eroded the cabinet’s accountability to
parliament: ‘Armed with paramount rights of initiative, supported by procedural rules
drawn in their favour, and holding the power of life and death over Parliament itself,
the cabinet indicates what is to be done; and Parliament, on its part . . . dreading the
consequences of refusal, complies’ (Ogg 1936: 461).
Echoing the concerns of Ostrogorski, Carl Schmitt regretted the abandonment of a
separation of powers between parliament and government. In his view, ‘narrow and
narrowest committees of parties and party coalitions make decisions behind closed
doors’ (Schmitt 1969 [1923]: 52). Note his concern here with what the principal–agent
approach labels ‘hidden action’, which can be a source of moral hazard. Like ourselves,
Schmitt conceived of parliamentary democracy as a chain of delegation, with parlia-
ment as a committee of the people and the cabinet a committee of parliament. To him,
the value of parliamentary government stemmed from rational and public discussion
among the MPs. Once all the relevant information had been tabled and discussed, this
should lead to ‘true and right policies’. Clearly, under ‘proper parliamentarism’ some
learning and enlightenment would take place. This, however, would no longer be the
case if parliamentary debate were a mere ‘façade’. Thus, the fusion of executive and
legislative powers creates severe accountability problems. In Chapter 3 we shall return
to similar concerns about transparency.
Yet important early twentieth-century authors, in particular Max Weber and Hans
Kelsen (see also Herman Finer 1931), defended parliamentary government against real-
world alternatives that were all non-democratic. Weber’s writings on parliamentary
democracy were essentially an attempt to convince his more reluctant German fellow
citizens of its virtues.16 In contrast, Kelsen (1926, 1929) gave a more balanced
assessment. He considered parliamentary government the only possible way to realize
the idea of democracy in the world of his day (Kelsen 1926: 5). Yet, he also identified
delegation problems and suggested remedies. Specifically, Kelsen called parliamentary
government a ‘compromise between the democratic demand for freedom and the
principle of the division of labour’, hence delegation (Kelsen 1926: 7, 1929: 29). According
16
See ‘Wahlrecht und Demokratie in Deutschland’ (1917) and ‘Parlament und Regierung im neugeord-
neten Deutschland’ (1918), both reprinted in Weber (1988).
26 Wolfgang C. Müller et al.
to Kelsen, parliamentary government is characterized by a ‘fiction of representation’ that
consists of two components: (1) the idea that parliament is only the representative
(‘Stellvertreter’) of the people and (2) that the people can issue their will only in parlia-
ment and through parliament, despite the fact that they do not give the MPs binding
instructions. This leaves the parliament legally sovereign and hence the decisions of
parliament may deviate from the preferences of the voters (Kelsen 1926: 10, 1929: 30).
Thus the problem is the ‘irresponsibility of the MP vis-à-vis his voters’ (Kelsen 1926: 14).
Kelsen’s concerns resonate with many subsequent authors, for whom the fusion of
the parliamentary majority and the chief executive creates special accountability prob-
lems, especially on ‘minor’ issues. While the electoral connection may secure accounta-
bility on major issues, it is likely to be inadequate when issues fall short of catching the
voters’ attention or causing their alienation. In other words, accountability breaks down
when the ultimate principals, the voters, have a severe information problem. Their direct
representatives, the Members of Parliament, whose information is considerably better,
are bound by party loyalty (or discipline) and hence may not wish to ‘rock the govern-
ment’s boat’ (King 1976). In this respect, parliamentary government perverts the repre-
sentatives’ incentives: Were government MPs to take seriously their duty to hold the
executive accountable, they might raise the voters’ attention to the wrongdoings of their
co-partisans in the executive branch, with likely adverse electoral consequences. The
fusion of legislative and executive powers may also cause bad policies, if the parliamen-
tary majority invariably is willing to turn all government proposals into law. Thus, pol-
icy failures are likely to result from the abandonment of parliamentary scrutiny.
Phrased in the terms employed in this volume, Kelsen thought the main problem of
parliamentary government to be moral hazard and considered the electoral account-
ability mechanisms to be insufficient. Consequently, he suggested three remedies:
● the referendum,
● political parties that allow for participation between elections, and
● special corporatist parliaments, which assemble expertise not available in parliament.
These suggested remedies empower the ultimate principal, the voters, and aim at over-
coming the information problems vis-à-vis MPs. Note that, in language that Lupia
develops in the Chapter 2, Kelsen’s remedies focus in large part on ex ante mechanism
of agency control, whereas Mosca largely favoured ex post instruments (see Kiewiet and
McCubbins 1991). The debate between these competing vehicles of popular sover-
eignty continues to this day.
T H E ROA D A H E A D
Representative democracy means delegation, and delegation is problematic for several
reasons. In this chapter, we have briefly identified some delegation problems that will
be discussed at greater length in the following chapters. Preference divergence and
asymmetric information may make a mockery out of political representation and lead
to such problems as adverse selection and moral hazard. Because such agency problems
exist, democracy requires not only delegation, but also accountability. The aim of this
Parliamentary Democracy 27
volume is to explore the ways in which delegation and accountability works in the parlia-
mentary systems of Western Europe and thus to understand the conditions that favour
and constrain popular sovereignty. We focus on those political institutions that impinge
on delegation and accountability throughout the democratic chain of delegation. As
subsequent chapters will show, institutions can solve or at least ameliorate agency
problems in several ways.
The co-editors and contributors to Delegation and Accountability in Parliamentary
Democracy have had two main objectives. The first aim is to identify the problems of
delegation that parliamentary democracy implies and to identify and assess the institu-
tional solutions that might be brought to bear on them. Our second goal is to provide
a comprehensive and theoretically guided survey of the institutions of parliamentary
democracies, and to describe in greater detail how the democratic process of delega-
tion and accountability operates in the Western European heartland of parliamentary
democracy. With regard to both aims we try to identify trends over time. These are
interrelated objectives, but they are not modest ones.
Our empirical investigation includes all Western European countries that meet our
minimal definition of parliamentary government, that is, all those in which the chief
executive (generically referred to in this study as the Prime Minister) is accountable to,
and can be removed for purely political reasons by, any parliamentary majority. By this
criterion, the only West European country (apart from some mini-states) that during
2003 fails to enter our sample is Switzerland. Of course, by casting our net so widely,
we are including a set of political systems that are dissimilar in many ways. Thus, we
are including majoritarian as well as consensus democracies (Lijphart 1999), systems
with few or many veto players (Tsebelis 1995, 2002), systems with unrestricted
versus highly restrictive dissolution powers (Strøm and Swindle 2002), and systems with
single-member district as well as proportional electoral systems (Cox 1997).
Some may find that the most consequential and controversial of our decisions is that
we include countries oftentimes referred to as semi-presidential (Duverger 1980; Elgie
1998, 1999). It is customary to regard France, Finland (at least until the constitutional
reform of 2000), and Portugal (until the constitutional reform of 1982) as semi-
presidential, based on the Presidents’ direct election and the authority they have exercised
in legislation or policy-making and/or over the formation and termination of cabinets
as well as over the calling of elections. For the 1975–85 period Greece may be added to
the list of semi-presidential regimes, since the President—although not directly
elected—was vested with important powers. There is a broader category of semi-
presidentialism to which also Austria, Iceland, Ireland, and Portugal (since 1982)—even
though their respective effective presidential powers seem much more limited—belong
by virtue of the fact that their heads of state are popularly elected. Yet, we include all
of these countries in the present study because we believe that their common mechanisms
of delegation and accountability, and particularly the accountability of the chief executive
to parliament (our minimal definition of parliamentarism), are more important than
the features that set them apart. That is not to say that these distinctions are unimportant,
however, and in fact a central concern of ours will be to demonstrate how and to what
extent they actually do shape popular governance.
28 Wolfgang C. Müller et al.
This volume consists of three major sections. The first section contains two additional
theoretical chapters. Chapter 2, by Arthur Lupia, discusses the general problem of
delegation versus abdication and introduces the principal–agent framework within which
we seek to understand parliamentary democracies. Chapter 3, by Kaare Strøm, brings
together this analytical approach and our empirical project. It elaborates a conception of
the democratic chain of delegation, contrast the governance mechanisms that character-
ize parliamentary versus presidential systems, and discusses the implications of these
institutional differences for the problems of democratic delegation and accountability.
The second section, and the bulk of this book, consists of eighteen chapters. In the
first of these chapters (Chapter 4), we detail the most important mechanisms of delega-
tion and accountability that exist at each stage of the parliamentary chain. This chapter
is cross-national and presents a wealth of information in tabular form. It also discusses
the broad range of variation in institutional structures that we find across Western
Europe. Chapters 5–21 apply our perspective and questions to seventeen West European
countries, beginning with Austria and proceeding alphabetically through to the United
Kingdom. The country chapters all have a common structure and address the same
sets of questions. They cover the entire post-Second World War period (or the
period since the latest introduction of democratic government) and report relevant
changes in the rules and conventions governing the governance process. These eighteen
chapters present the most systematic and detailed cross-national survey that is available
on institutional arrangements (accountability mechanisms) in parliamentary democracies,
in 2003.
Each chapter focuses on four discrete stages in the delegation process: (1) delegation
from voters to parliamentary representatives, (2) delegation from parliament to the
Prime Minister and cabinet, (3) delegation within the cabinet, and (4) delegation from
individual cabinet members to civil servants. These detailed discussions are followed by
a section on external constraints, that is, constitutional and other mechanisms that
impinge on democratic delegation and limit the discretion of the relevant constitu-
tional actors. The country chapters give a full account of the mechanisms of delega-
tion and accountability. They also discuss how effective they are in solving the problems
of delegation.
In Chapters 22 and 23, as we return to the broader picture of parliamentary democracy
today, we assess our lessons. We are here particularly concerned with the prominence of
certain agency problems and with the growing importance of constraints. As we will
argue, the relationship between parliamentary democracy and institutional constraints is
becoming increasingly complex. While this may accentuate certain agency problems, cur-
rent trends also point toward certain promising developments. And while the historical les-
sons presented in this chapter remind us that parliamentarism has been and still is a very
robust system of government, the parliamentary democracies of Western Europe are fac-
ing a number of very important challenges.
REFERENCES
Ackerman, Bruce (2000). ‘The New Separation of Powers’. Harvard Law Review, 113: 633–729.
Bagehot, Walter (1963 [1867]). The English Constitution. London: Fontana/Collins.
Parliamentary Democracy 29
McCubbins, Mathew D., Noll, Roger D., and Weingast, Barry R. (1987). ‘Administrative Procedures
as Instruments of Political Control’. Journal of Law, Economics, and Organization, 3: 243–77.
McLean, Iain (1987). Public Choice. An Introduction. Oxford: Basil Blackwell.
Metcalf, Michael M. (1977). Russia, England and Swedish Party Politics 1762–1766. The interplay
between Great Power Diplomacy and Domestic Politics during Sweden’s Age of Liberty. Stockhom:
Almqvist & Wiksell International.
Michels, Robert (1962). Political Parties. New York: Free Press.
Mill, John Stuart (1984 [1861]). ‘Considerations on Representative Government’, in John Stuart
Mill. Utilitarianism. Liberty. Representative Government. London: Dent.
Mills, C. Wright (1956). The Power Elite. New York: Oxford University Press.
Moe, Terry M. (1984). ‘The New Economics of Organization’. American Journal of Political
Science, 28: 739–77.
—— (1997). ‘The Positive Theory of Public Bureaucracy’, in Dennis C. Mueller (ed.), Perspectives
on Public Choice. A Handbook. Cambridge: Cambridge University Press.
—— and Caldwell, Michael (1994). ‘The Institutional Foundations of Democratic Government:
A Comparison of Presidential and Parliamentary Systems’. Journal of Institutional and
Theoretical Economics, 150: 171–95.
Mosca, Gaetano (1947). Elementi di scienza politica, 4th edn. Bari: Gius Laterza & Figli.
—— (1950). Die herrschende Klasse. Munich: Leo Lehnen Verlag.
Mueller, Dennis C. (1996). Constitutional Democracy. Oxford: Oxford University Press.
Müller, Wolfgang C. (2000). ‘Political Parties in Parliamentary Democracies: Making Delegation
and Accountability Work’. European Journal of Political Research, 37: 309–33.
—— and Strøm, Kaare (2000). ‘Conclusions: Coalition Governance in Western Europe’, in
Wolfgang C. Müller and Kaare Strøm (eds.), Coalition Governments in Western Europe. Oxford:
Oxford University Press.
Niskanen, William A. (1971). Bureaucracy and Representative Government. Chicago: Aldine-Atherton.
Norton, Philip (1981). The Commons in Perspective. Oxford: Basil Blackwell.
Ogg, Frederic Austin (1936). English Government and Politics, 2nd edn. New York: Macmillan.
Ostrogorski, M. (1907). Democracy and the Organization of Political Parties. London: Macmillan.
Persson, Torsten and Tabellini, Guido (1999). ‘The Size and Scope of Government: Comparative
Politics with Rational Politicians’. European Economic Review, 43: 699–735.
Powell, G. Bingham, Jr. (1982). Contemporary Democracies. Participation, Stability, and Violence.
Cambridge, MA: Harvard University Press.
—— (2000). Elections as Instruments of Democracy. Majoritarian and Proportional Visions. New
Haven: Yale University Press.
Pratt, John W. and Zeckhauser, Richard J. (eds.) (1985). Principals and Agents: The Structure of
Business. Boston: Harvard Business School Press.
Przeworski, Adam, Alvarez, Michael E., Cheibub, José Antonio, and Limongi, Fernando (2000).
Democracy and Development. Cambridge: Cambridge University Press.
Riggs, Fred W. (1988). ‘The Survival of Presidentialism in America: Para-constitutional Practices’.
International Political Science Review, 9: 247–78.
—— (1997). ‘Presidentialism versus Parliamentarism: Implications for Representativeness and
Legitimacy’. International Political Science Review, 18: 253–78.
Riker, William H. (1982). Liberalalism Against Populism. San Francisco: W. H. Freeman.
Roberts, Michael (1986). The Age of Liberty. Sweden 1719–1772. Cambridge: Cambridge University
Press.
Sartori, Giovanni (1987). The Theory of Democracy Revisited: The Comtemporary Debate. Chatham:
Chatham House.
32 Wolfgang C. Müller et al.
Sartori, Giovanni (1997). Comparative Constitutional Engineering, 2nd edn. Houndmills: Macmillan.
Schattschneider, E. E. (1960). The Semisovereign People. New York: Holt, Rinehart, and Winston.
Schlesinger, Joseph A. (1966). Ambition and Politics. Chicago: Rand McNally.
—— (1991). Political Parties and the Winning of Office. Ann Arbor: University of Michigan Press.
Schmitt, Carl (1969 [1923]). Die geistesgeschichtliche Lage des heutigen Parlamentarismus. Berlin:
Duncker & Humblot.
Schumpeter, Joseph A. (1942). Capitalism, Socialism and Democracy. New York: Harper & Brothers.
—— (1954 [1918]). ‘The Crisis of the Tax State’, in Alan T. Peacock, Ralph Turvey, Wolfgang F.
Stopler, and Elizabeth Henderson (eds.), International Economic Papers. New York: Macmillan.
Schütt-Wetschky, Eberhard (1984). Grundtypen parlamentarischer Demokratie. Freiburg: Verlag Karl
Alber.
Shleifer, Andrei and Vishny, Robert W. (1998). The Grabbing Hand. Government Pathologies and
Their Cures. Cambridge, MA: Harvard University Press.
Shugart, Matthew S. and Carey, John (1992). Presidents and Assemblies. Cambridge: Cambridge
University Press.
—— and Mainwaring, Scott (1997) ‘Presidentialism and Democracy in Latin America: Rethinking
the Terms of Debate; in Scott Mainwaring and Matthew Soberg Shugart (eds.) Presidentialism
and Democracy in Latin America. Cambridge: Cambridge University Press.
Steffani, Winfried (1979). Parlamentarische und präsidielle Demokratie. Opladen: Westdeutscher
Verlag.
Stepan, Alfred and Skach, Cindy (1993). ‘Constitutional Frameworks and Democratic
Consolidation: Parliamentarism versus Presidentialism’. World Politics, 46: 1–22.
Strøm, Kaare (2000). ‘Delegation and Accountability in Parliamentary Democracies.’ European
Journal of Political Research, 37: 261–89.
–––– and Swindle, Stephen M. (2002). ‘Strategic Parliamentary Dissolution’. American Political
Science Review, 575–91.
Strong, C. F. (1963). A History of Modern Political Constitutions. New York: G. P. Putnam’s Sons.
Tsebelis, George (1995). ‘Decisionmaking in Political Systems: Veto Players in Presidentialism,
Parliamentarism, Multicameralism, and Multipartyism’. British Journal of Political Science,
25: 289–325.
—— (2002). Veto Players: An Introduction to Institutional Analysis. Princeton: Princeton University
Press and the Russell Sage Foundation.
Verney, Douglas V. (1959). The Analysis of Political Systems. London: Routledge & Kegan Paul.
von Beyme, Klaus (1973). Die parlamentarischen Regierungssysteme in Europa. Munich: Piper.
—— (1999). Die parlamentarischen Regierungssysteme in Europa, 3rd edn. Opladen: Westdeutscher
Verlag.
von Mettenheim, Kurt (1997). Presidential Institutions and Democratic Politics. Baltimore: Johns
Hopkins University Press.
Weaver, R. Kent and Rockman, Bert A. (eds.) (1993). Do Institutions Matter? Government
Capabilities in the United States and Abroad. Washington, DC: The Brookings Institution.
Weber, Max (1972). Wirtschaft und Gesellschaft. Tübingen: Mohr.
—— (1988). Gesammelte politische Schriften. Tübingen: Mohr.
Weir, Stuart and Beetham, David (1999). Political Power and Democratic Control in Britain. London:
Routledge.
Wittman, Donald A. (1995). The Myth of Democratic Failure. Chicago: University of Chicago Press.
Woldendorp, Jaap, Keman, Hans, and Budge, Ian (2000). Party Government in 48 Democracies
(1945–1998). Dordrecht: Kluwer.
2
I N T RO D U C T I O N
Delegation is a central concept in the conduct and the study of politics. Governments
great and small use delegation to increase the range of services that they can provide.
National governments, for example, delegate to defence ministries the task of main-
taining national security and delegate to finance ministries the task of managing the
economy. Indeed, the modern nation-state could not exist without delegation—for
without delegation, lawmakers would be forced personally to implement and enforce
every single law that they make. With delegation, national governments can address a
wide range of social issues simultaneously.
In most polities, the most prominent form of delegation is from lawmakers (e.g.
parliaments) to government agencies (e.g. economic, security, and public welfare agen-
cies). Such agencies house the persons responsible for enforcing laws and tend to
employ the nation’s leading experts in the agency’s policy area (e.g. economic agencies
employ leading economists). In many places, these agencies are known collectively as
the bureaucracy. In some places, people who work within these agencies are called
bureaucrats or civil servants.
We define delegation as an act where one person or group, called a principal, relies
on another person or group, called an agent, to act on the principal’s behalf. This
volume’s purpose is to examine delegation problems that exist in parliamentary
democracies. Table 2.1 displays prominent forms of delegation in these democracies.
Delegation allows political principals to accomplish desired ends with reduced
personal cost and effort. Implementing a single ministerial directive, for example, can
require more time and effort than any minister can muster. Only by delegating—getting
subordinates to do most of the work—can any high-level government official achieve
multiple objectives. Similarly, most citizens are not sufficiently qualified to provide the
social services that improve their lives (e.g. education, health care). Only by delegating—
getting specialists to do most of the work—can any citizen enjoy such services.
I thank Torbjörn Bergman, Sean E. Cain, Christopher DenHartog, James N. Druckman, Elisabeth R. Gerber,
Simon Hug, Mathew D. McCubbins, Wolfgang C. Müller, and Kaare Strøm for helpful comments. I worked
on this chapter while I was a fellow at the Center for Advanced Study in the Behavioral Sciences and
I thank them for their support.
34 Arthur Lupia
Table 2.1. Common Forms of Delegation
Principal Agent
While delegation allows principals to benefit from the expertise and abilities of
others, it can also be perilous. The perils arise from the fact that delegation entails a
transfer of power—every time lawmakers delegate to bureaucrats, they give away a
portion of their authority to govern. A great danger is that the people to whom power
is delegated will abuse the power they receive. Consider, for example, Weber’s famous
treatise about what happens when legislatures delegate to bureaucrats:
Under normal conditions, the power position of a fully developed bureaucracy is always over-
towering. The ‘political master’ finds himself in the position of the ‘dilettante’ who stands
opposite the ‘expert’, facing the trained official who stands within the management of administra-
tion. This holds whether the ‘master’ whom the bureaucracy serves is a ‘people’, equipped with
the weapons of ‘legislative initiative’, the ‘referendum’, and the right to remove officials, or a
parliament, elected on a more aristocratic or more ‘democratic’ basis and equipped with the right
to vote a lack of confidence . . . (Weber, quoted in Gerth and Mills 1946: 232)
If this danger is realized, then those entrusted with the responsibility of governing lose
control of what is done in the government’s name and anarchy can ensue.
While delegation is not a perfect means for political principals to achieve their object-
ives, it is the only feasible means. Anyone who wants to govern or provide services to
mass populations must find agents to carry out their plans. Therefore, understanding
the relations between principals and agents is critical to answering many questions
about parliamentary governance.
In this chapter, I present a theoretical framework that clarifies when principals can,
and cannot, use delegation to accomplish desired ends. I use the framework to clarify
how a number of factors, including political institutions, affect the success of delega-
tion. In subsequent chapters, other authors use the framework to examine the empir-
ical relationship between delegation and accountability in 17 West European
parliamentary democracies.
I continue as follows. In the section titled Measures of What Delegation
Accomplishes, I offer important definitions and terms. In the section that follows,
I introduce the theoretical framework. In the section on the Consequences of
Delegation, I use the framework to clarify when agents will act in their principals’ inter-
ests. In section titled How Institutions Affect Agency Loss, I explain how political insti-
tutions can alleviate the perils of delegation. In the conclusion, I link this chapter’s
insights to broader questions.
Delegation and its Perils 35
M E A S U R E S O F W H AT D E L E GAT I O N AC C O M P L I S H E S
In this section, I introduce needed definitions. I begin by defining delegation and
accountability. Delegation, as noted in this chapter’s opening sentences, is an act where
one person or group, called a principal, relies on another person or group, called an
agent, to act on the principal’s behalf. Defining accountability is not as simple.
People use the term accountability to characterize the effectiveness or efficiency of
governance, in general, and delegation acts, in particular. A problem with the term is
that it is used in different ways. For some, accountability is a process of control. In this
view, civil servants are accountable to ministers only if the ministers can influence a
civil servant’s actions. For others, accountability is a type of outcome. In this view, civil
servants are accountable to ministers only if the agent acts in the minister’s interests.
The ‘control’ and ‘outcome’ definitions of accountability mean different things.
A civil servant can provide outcomes that a minister likes without the minister exercising
any control. This happens when a civil servant and minister share the same policy
goals, the civil servant completely ignores ministerial directives, and pursuing her own
interests leads her to take actions that favour the minister nevertheless. Conversely, a
minister can exercise great control over a civil servant without achieving a desired
outcome. This happens when a minister can remove a civil servant if a certain level of
service is not achieved and the servant lacks the skills needed to achieve the goal.
This book adopts the ‘control’ definition of accountability. An agent is accountable
to a principal if the principal can exercise control over the agent and delegation is not
accountable if the principal is unable to exercise control. If a principal in situation A
exerts more control than a principal in situation B, then accountability is greater in
situation A than it is in situation B.
Having defined accountability in this way, I should now restate this chapter’s goal. In
what follows, I will clarify what delegation accomplishes by using a theoretical frame-
work to examine how common processes of control (i.e. the rules or institutions under
which delegation commonly occurs) affect the outcome of an act of delegation.
To describe outcomes, I employ two metrics—agency loss and success/failure.
Agency loss is the difference between the actual consequence of delegation and what
the consequence would have been had the agent been ‘perfect’. By perfect, I mean a
hypothetical agent who does what the principal would have done if the principal had
unlimited information and resources to do the job herself.
Agency loss describes the consequence of delegation from the principal’s perspect-
ive. To say that agency loss is high is to say that the outcome of delegation is different
from the principal’s ideal outcome. To say that agency loss decreases is to say that the
outcome of delegation moves closer to the principal’s ideal.
The agency loss metric, though simple, is easy to misinterpret. Agency loss is zero
when the agent takes actions that the principal would have taken given unlimited informa-
tion and resources. If people forget this fact when using the agency loss metric, then
delegation often seems to be a bad idea (e.g. ‘This act of delegation caused agency
loss. Therefore, it is bad.’). However, most political principals do not have unlimited
information and resources. Therefore, it is impossible for them to be perfect agents for
36 Arthur Lupia
themselves (e.g. doctors are agents that are employed to maintain good health; con-
sider what would happen if you tried to serve as your own doctor). Indeed, many prin-
cipals who try to do all that agents can do will experience delegation outcomes far
inferior to the ones that a hypothetical perfect agent would provide. Therefore, finding
that delegation causes agency loss does not imply that the outcome is bad or even
suboptimal.
A coarser metric for describing delegation outcomes is success and failure.
Delegation is successful if the outcome of delegation improves the principal’s welfare
relative to what would have happened if the principal had chosen not to delegate.
Following common parlance, I call the consequence of an agent’s inaction the status
quo. Delegation fails if it decreases the principal’s welfare relative to this status quo.
Throughout the chapter, I use both metrics to describe delegation outcomes. I use
agency loss when I want to make comparative statements about how differences
between the principal and agent or variations in political institutions affect the degree
to which the outcome of delegation reflects the principal’s desires. I use the
success/failure metric when I want to make simple statements about whether or not
an act of delegation makes the principal better or worse off than the status quo.
I should also note that my use of the term status quo is equivalent to the common
usage of the term reversion point. That is, if the agent fails to act, then the outcome of
delegation reverts back to the status quo. While I will stick with the term status quo,
authors of some later chapters prefer the term reversion point to mean the same thing.
T H E O R E T I C A L F R A M E WO R K
I now introduce a framework that clarifies the relationship between delegation, mech-
anisms of accountability, agency loss, delegation success, and delegation failure. The
framework builds on three simple premises.
1. Parliamentary democracy entails a chain of delegation
In such a chain, each link attaches a principal to an agent. The ideal–typical delegation
chain resembles a straight line and includes a link: that attaches voters to members
of Parliament; a link that attaches members of Parliament to the government; a link
that attaches the government to individual ministers; and a link that attaches ministers
to civil servants. In other words, citizens occupy one end of the ideal–typical chain
while civil servants who implement government edicts occupy the chain’s other end. In
this view, the effectiveness of institutional accountability mechanisms can be measured
by the extent to which agents at one end of the chain ‘feel a tug’ when principals at the
other end ‘pull the chain’.
Of course, most parliamentary democracies are not simple, singular delegation
chains. Sometimes multiple principals delegate to a single agent or prin-cipals grant
agents overlapping jurisdictions. Such practices change a chain of delegation’s shape.
Since most countries engage in such practices at least some of the time, we observe del-
egation chains varying widely in size and shape when we compare democracies empiri-
cally. Such variations are a product and a cause of each country’s unique experience
Delegation and its Perils 37
with democracy. It is, however, important to note that alongside such differences are
common properties that all delegation chains share, which leads to our framework’s
second premise.
While the first of these properties is obvious, the latter two require explanation. The
possibility of conflicting interests implies that the agent need not have the same policy
objectives or prefer the same outcomes as principals. An agent may want to help
the elderly while his principal detests them (i.e. the principal and agent desire different
policies). A principal may prefer that an agent work hard on a complex problem, while
the agent may prefer to spend afternoons fishing or at a bar (i.e. the principal and agent
may desire the same policy but differ in who should exert effort to achieve that policy
objective). Either difference is important because each can provide an agent with reasons
to act contrary to his principal’s desires.
The possibility of limited information implies that the principal, agent, or both are not
omniscient. When an agent lacks information, one possible result is incompetence (e.g. the
agent does not know how to do what a principal asks). Students of politics, however,
pay much more attention to what happens when principals lack information. Indeed, it
is commonly assumed that political principals, particularly voters and backbench MPs,
are ignorant of their agents’ activities. Our treatment of how limited information affects
delegation will share this focus—we shall concentrate primarily on what happens when
the principal lacks information about what her agent is doing.
While these three common properties are far from a complete description of any
delegation act, they are sufficient for us to derive some important results what delegation
accomplishes later in the chapter.
T H E C O N S E QU E N C E S O F D E L E GAT I O N
I now use insights from several formal models to clarify what delegation accomplishes.
I begin with a simple model that focuses on a single act of delegation. While this model
has origins in economics (e.g. Niskanen 1971), the version I present is due to Romer
and Rosenthal (1978). The sequence of events in their model is as follows and as
depicted in Fig. 2.1.1
Romer and Rosenthal model an act of delegation as a game between a principal and
an agent. To simplify my explanation of this and subsequent models, I henceforth refer
to the principal as a female and the agent as a male. In the Romer–Rosenthal model,
the agent moves first by choosing whether or not to change what was done in the past
(i.e. the agent chooses whether or not to act). A civil servant, for example, acts when
he interprets a government directive differently than it had been interpreted before. Put
another way, Romer and Rosenthal study an act of delegation after the principal has
already chosen to delegate.
Romer and Rosenthal conceive of the agent’s possible actions as points on a line.
They represent the agent’s choice of action as the choice of a point, X, on the line [0, 1].
This choice may or may not be the same as the pre-existing status quo—denoted as
SQ 僆 [0, 1]. SQ represents the outcome of delegation when the agent chooses not to act.
After the agent chooses whether or not to act and, if so, which action to take,
the principal reacts by accepting the action or rejecting it in favour of the status quo
(i.e. the principal chooses X or SQ). You can think of the principal’s decision to reject
the agent’s action as an outright veto of an agent’s decision or as a sanction against the
agent that is sufficient to induce him to undo what he has done.
Romer and Rosenthal assume that the principal and agent have preferences over the
possible outcomes of delegation. These preferences are represented by assuming that
each player maximizes a single-peaked utility function. The peak of each player’s utility
function is his or her ideal policy. So it is equivalent to say that each player’s objective is
to achieve a delegation outcome that is as close as possible to his or her ideal policy.
I denote the principal’s ideal policy as P 僆 [0,1] and the agent’s ideal policy as A 僆 [0,1].
Given our definition of agency loss and this notation, we can also say that P, the princi-
pal’s ideal policy, is the point at which agency loss is zero. Agency loss grows as does the
distance between P and the final outcome of the principal’s and agent’s interaction.
1
Throughout the chapter, I present theoretical insights with a minimum amount of mathematical nota-
tion. For those interested in the mechanics of these theories, I provide citations throughout the text. Some
of the material in this section follows directly from an argument first published in Lupia (2000).
Delegation and its Perils 39
Situation 1
∗
SQ P=A
Situation 2
∗
SQ A P
or
∗
SQ P A
Situation 3
∗
SQ P A
Situation 4
∗
A∗ SQ P
Figure 2.2. Graphical Depictions of the Possible Situations. The star indicates the
predicted outcome of delegation in the (complete information) Romer and
Rosenthal model.
A key assumption in the Romer–Rosenthal model is that the principal and agent have
complete information. This assumption simplifies the model considerably and allows
important insights to be derived efficiently. After stating these results, I will describe
other models where limited information plays a key role.
The Romer–Rosenthal model reveals fundamental causes of agency loss. It does
so by separating four mutually exclusive and collectively exhaustive situations. Each
situation provides a distinct insight about the optimal actions of principal and agent
given their beliefs about the other’s response. I depict each situation in Fig. 2.2.
In situation 1, the principal and agent have identical ideal policies. As a result, they
are in perfect agreement about what the agent should do. Thus, the model predicts that
agent proposes the principal’s (and his own) ideal policy and the principal accepts it. In
this situation, there is no agency loss.
In all other situations, the outcome of delegation is different. That is, in any case
where the principal and agent do not have identical ideal points, the outcome of
delegation in the Romer–Rosenthal model entails some agency loss. In situations 2 and 3,
for example, the principal and agent have different ideal policies, but both players’ ideal
policies are on the same side of the status quo. In these situations, the principal and
agent agree on the desired direction of policy change but not on the magnitude of
change (e.g. both prefer higher taxes but disagree on the rate of increase).
40 Arthur Lupia
In situation 2, the principal’s ideal policy is farther from the status quo than it is from
the agent’s ideal policy. In other words, this is the case where the principal prefers the
agent’s ideal policy, A, to the status quo, SQ. In this situation, the Romer–Rosenthal
model predicts that the agent chooses his ideal policy and the principal accepts the
agent’s action. This occurs because the agent knows that the principal prefers some
degree of change (i.e. A) to no change at all (i.e. SQ). The agency loss in situation 2 is
the distance between the agent’s ideal policy and the principal’s ideal policy. To see
why, note that if the agent proposes X ⫽ A and the principal accepts the agent’s
proposal, then agency loss is the negative of the difference between A and P. If the
agent rejects this action, then agency loss is the negative of the difference between
P and SQ. Since A is closer to P than SQ, by the definition of situation 2, then the
principal is better off accepting the agent’s action.
In situation 3, the principal’s ideal policy is closer to the status quo than it is to the
agent’s ideal policy. This means that the principal would rather maintain the status quo
than accept the agent’s ideal policy. Since the agent knows this, he will not act to achieve
his ideal policy. He will, however, choose another action. In equilibrium, the agent
chooses the policy closest to his own ideal policy that the principal will accept.2 The
agency loss in situation 3 is at least as great or greater than the agency loss in any of the
previous cases and is just less than the distance between SQ and P.
In situation 4, the principal and agent ideal policies are on opposite sides of the
status quo (e.g. the principal wants higher taxes, the agent wants lower taxes). Here, no
agent action is mutually agreeable to the principal and the agent. Therefore, the
outcome is the status quo and the agency loss is the distance between SQ and P.
Comparing these situations to each other offers two insights. First, as we move from
a lower numbered situation to a higher numbered situation, agency loss increases if it
changes. Second, moving from lower numbered situations to higher numbered situ-
ations implies either an increase in the distance between A and P or an increase in the
distance between P and SQ relative to the distance between A and P. Put another way,
if the agent’s ideal policy or the status quo moves away from the principal’s ideal policy
then agency loss can grow. That agency loss grows with the distance between P and A
is straightforward—as the agent finds himself sharing fewer interests with the principal,
he gains fewer rewards from serving the principal’s interests. That agency loss can grow
with the distance between P and SQ follows from the fact that a bad status quo makes
a wider range of proposals attractive to the principal (i.e. beggars can’t be choosers).
It is worth emphasizing that the Romer–Rosenthal model’s predictions depend heavily
on two simplifying assumptions. The first assumption is that the principal can only
accept or reject the agent’s action. If the principal is given greater powers relative to the
agent (i.e. she can force the agent to take actions closer to P), then agency
loss will decrease. Similarly, if the principal is given lesser power relative to the agent
(i.e. she cannot reject agent actions), agency loss will increase.
2
If SQ is greater than A and P, then the agent will propose P ⫺ |SQ ⫺ P| ⫹, where ⬎ 0 and small. If
SQ is less than A and P, then the agent will propose P ⫹ |SQ ⫺ P| ⫺ .
Delegation and its Perils 41
The second simplifying assumption is that the principal and agent have complete
information. For many acts of delegation, this assumption is wholly unrealistic. I now
focus on what delegation accomplishes when information is incomplete. I define informa-
tion as incomplete when a person is unable to predict the consequence of their actions
with certainty. Within a delegation relationship, a principal, agent, or both can have
incomplete information.
Principals and agents make use of two kinds of information.
● Information about actions: They want to know what the other has done.
● Information about people and their context. If people lack information about others’
actions, they can often use information about a person (i.e. their skills or prefer-
ences) or their circumstances to deduce that person’s actions.
If a principal or agent lacks either kind of information, the consequences of delegation
can change dramatically. The extent to which the agent will attempt to act in accordance
with the principal’s preferences is particularly sensitive to whether or not principal and
agent have asymmetric information. Information is asymmetric when one person
knows more about the consequences of their actions than does the other.
As is customary in the literature on delegation, I will focus on the case where the
principal knows less than the agent. While agents can lack information about their prin-
cipals, students of democracy are far more concerned about the extent to which lim-
ited information hinders political principals’ abilities to delegate successfully. Indeed,
when principals do not know enough to track their agents’ actions, the transfer of
power inherent in an act of delegation can cause it to be equivalent to abdication—
a concern that motivates much of the cited work on delegation. I adopt the same
emphasis here.
When principals know less than agents, two kinds of problems arise: moral hazard
and adverse selection. Rasmusen (1989: 133) describes these terms as follows:
● Moral hazard with hidden actions: Smith and Brown begin with symmetric information and
agree to a contract, but then Smith takes an action unobserved by Brown . . .
● Adverse selection: [Smith knows things about himself that Brown does not.] Smith and
Brown then agree to a contract. Information is incomplete.
The classic example of moral hazard comes from the realm of insurance. When a
person buys a life insurance policy, he knows that his insurance company will pay for
possible outcomes of risky behaviours. Since the insurer cannot prevent its clients from
all risks, it faces a moral hazard problem—its clients may take riskier actions after a pol-
icy is issued. Moral hazard causes delegation problems when it gives the agent an
opportunity to take actions of which the principal would not otherwise approve.
Adverse selection problems arise when an agent has attributes that a principal cannot
observe. A classic example of adverse selection arises in the context of health insurance.
As Kreps (1990: 626–7) notes: ‘If premiums are set at actuarially fair rates for the popu-
lation as a whole, insurance may be a bad deal for healthy people, who will then refuse
to buy. Only the sick and dying will sign up. And premium rates must then be set to
reflect this.’ Also see Akerlof (1970) about similar properties in used car markets.
42 Arthur Lupia
Adverse selection is a problem in many aspects of delegation—most notably
elections. To see this problem, suppose that only two types of people choose to run for
elective office—those who are truly public-spirited and those who line their own pockets
with public resources. If candidates know their type and voters do not, the problem of
adverse selection exists. If voters want only to elect public-spirited individuals, then all
candidates have an incentive to portray themselves as public-spirited. The extent of
agency loss for voters in this delegation will depend on the extent to which they can
separate the dutiful from the knaves.
In what follows, I will first use a model to show how moral hazard affects delegation.
Then, I will use similar logic to clarify the impact of adverse selection.
Notes: A* denotes the agent’s ideal point in the situation where he and the principal’s ideal point are on oppo-
site sides of SQ. In this situation, A* is worse for the principal than SQ. ⑀ denotes a number that is greater
than zero, but very small. Note that ⫺|X ⫺ P| refers to the utility level of the principal given ideal policy, P,
and delegation outcome, X. So, for example, when the outcome of delegation is the principal’s ideal policy,
as it is in the complete information version of situation 1, then the principal’s utility is ⫺|P ⫺ P| or 0. When
the outcome of delegation is the agent’s ideal policy, as it is in several instances in the table, then the principal’s
utility is ⫺|A ⫺ P|.
The main lesson, therefore, is that a principal lacking information about actions is
not sufficient to cause agency loss. What matters more is whether or not the principal
can learn enough to wield whatever power she may have over the agent effectively. In
the following section, I explain when accountability mechanisms can, and cannot, pro-
vide such information.
H OW I N S T I T U T I O N S A F F E C T AG E N C Y LO S S
Moral hazard and adverse selection give the principal an incentive to gather informa-
tion about her agent. One means by which such information can be generated is
institutions. Institutions affect incentives and can be constructed in ways that affect
agent’s incentives to reveal information about their motivations or actions. In the 1980s
and 1990s, formal theorists spent considerable effort attempting to clarify what kinds of
institutional designs can help delegation succeed. In this section, I highlight a few of these
insights.
Delegation and its Perils 45
Formal theorists classify institutional designs that affect the provision of information
to political principals as ex ante or ex post mechanisms.
● Principals use ex ante mechanisms to learn about their agents before they act, typically
as they select particular individuals to serve as their agents. Such mechanisms are
appropriate if a principal anticipates adverse selection problems.
● Principals use ex post mechanisms to learn about an agent’s actions after the fact.
Such mechanisms are a way to deal with moral hazard problems.
The most common forms of ex ante control are screening and selection rules. Screening
and selection rules are devices that principals can use to sort good agents from bad
before delegating to them. Such rules are an important part of the political institutions
that govern many delegation relationships. Consider, for example, the extensive review
process endured by candidates for cabinet ministries and, supreme courts and or Prime
Ministerial candidates.
Contracts—task or time specific agreements between principals and agents—are also
a common means for ex ante control. The point of such agreements is to establish
shared interests between principals and agent, for example, by giving the agent a cut of
any gain that the principal experiences as a result of the agent’s efforts. It is common
for contracts to specify an ex post mechanism of control (e.g. a clause such as ‘If at date x,
you have performed task y, I will reward you. Otherwise, I will sanction or sue you.’).
In what follows, I survey models that reveal the conditions under which common
ex ante control mechanisms such as those just described can help political principals
reduce agency loss. I then discuss the effects of some ex post mechanisms.3
3
Readers who are interested in a more comprehensive theoretical treatment of how information and
institutions affect choices such as those inherent in delegation can consult introductory texts such as
McMillan (1992: parts II and III), intermediate texts such as Kreps (1990: chs 16 and 17) and Salanié (1998),
advanced texts such as Laffont and Tirole (1993), or seminal articles such as Crawford and Sobel (1982),
Milgrom and Roberts (1986), Farrell and Gibbons (1989), Banks (1991), Farrell and Rabin (1996).
46 Arthur Lupia
the alternative that the first agent proposed when lacking competition. If A2 is not
between P and A, then adding A2 has no effect. Therefore, in the complete information
setting of the Romer–Rosenthal model, setting up a competition between agents can
reduce agency loss.
Case 2: Competition Sometimes Fails Competition need not have the same effect
when the principal faces information problems. To see why, consider the analogy of the
government putting up for auction the right to be an agent. Such a selection procedure
occurs when a government hires a private firm to perform a public service (e.g. muni-
cipal construction projects). One could also use an auction analogy to characterize the
battles for cabinet positions that occur during coalition formation negotiations.
When a principal forces potential agents to compete for the right to be her agent, she
can induce the potential agents to reveal information about themselves—in particular
their ability and willingness to perform the tasks she asks of them. The conditions
under which such a competition best reduces agency loss are as follows:
● it is easy to determine whether or not the agent can perform the task in question,
● potential agents know whether or not they are willing and able to perform the task
in question,
● potential agents know whether or not other potential agents are willing and able
to perform the task in question,
● it is relatively easy and costless for the principal to punish the agent for failing to
accomplish the task ex post.
If all of these conditions are met, then potential agents have an incentive to truthfully
report their willingness and ability to serve the principal. In equilibrium, the most effi-
cient potential agent becomes the actual agent. This outcome occurs because it solves
the adverse selection problem. In this case, no agent has an incentive to make a false
claim about his attributes (in the form of his bid) just to win the auction. In this situa-
tion, auctioning the right to be an agent minimizes agency loss.
If, however, any of these conditions are not met, then competition need not have
such a pleasant effect. For example, if the principal cannot easily determine whether
the agent has done what he promised to do, then the moral hazard problem reappears.
Alternatively, if potential agents are uncertain about their own ability or their com-
petitors’ abilities, then they may gain an incentive to understate or overstate their own
desire for the job. Such an outcome can lead to increased agency loss.
Case 3: The Hold-Up Problem Moreover, if the principal lacks information about the
agent’s abilities and cannot punish the agent, then the principal is exposed to an
extreme form of moral hazard known as the ‘hold up’ problem (e.g. see Williamson
(1989) and Salanié (1998: ch. 7)). In a ‘hold up’ problem, a question arises about who
will obtain the benefits from resources already committed to a particular activity.
To see how a hold up problem works, suppose that when a principal selects a par-
ticular person to be her agent, that agent then acquires valuable expertise that those
who were not selected to be the agent cannot acquire. This agent can then use his
expertise as leverage to renegotiate the terms of his relationship with the principal and
Delegation and its Perils 47
extract additional benefits for himself. Consider, for example, a person who becomes
identified with excellent service in a particular cabinet post. Upon achieving such a status,
the ‘excellent’ minister may seek increased power in a wider range of government
decisions. As the government’s effectiveness or public standing becomes increasingly
dependent on the ‘excellent’ minister’s performance, they may have a greater incentive
to let this agent have his way—even if it increases agency loss.
At this point, any reduction in agency loss that may have resulted from earlier
competition may be lost. In sum, while competition among agents may seem like a natural
solution to adverse selection and moral hazard problems, the benefits of competition
are realized only if certain informational and institutional conditions are met.
Additional Caveats
Several additional factors work against competition among agents as a means of
increasing accountability and decreasing agency loss. First, competition can preclude
the most qualified potential agents from seeking the job. For if the skills required to be
an effective agent differ from the skills required to win the competition (as is sometimes
alleged in mass elections) or if the most skilled candidates for a position lack the
resources needed to prevail in competition for a job (as may be the case with qualified
experts in a field who lack political connections or the field of candidates who choose
to run for the US presidency), then competition may produce a suboptimal outcome
(also see Aghion and Tirole 1997 for a more general discussion of participation con-
straints). Second, competition among agents can entail negative externalities. In the
rush to provide a particular service, for example, competition may cause efforts to be
duplicated and lead to inefficiency. Such caveats are not meant to imply that competi-
tion is the wrong way to seek accountability or reduce agency loss. However, focusing
on these caveats reminds us that means such as competition do not automatically
produce desired ends (see Salanié 1998 for a recent and more technical review).
Ex Post Mechanisms
In many democratic delegations, a principal may lack the information she needs to dis-
tinguish agent actions that help her from agent actions that hurt her (i.e. she may lack
knowledge sufficient to reduce agency loss).4 Such principals have three ways to gain
information about her agent’s actions: direct monitoring, attending to the what the
agent says about his activities, or attending to third party testimony about agent actions.
Relying on others for information can be advantageous because it requires less of the
principal’s time and effort than does conducting her own investigation. Such depend-
ence, however, can also entail substantial peril. Not all people from whom principals
can seek advice are trustworthy or knowledgeable. Therefore, a principal who wants to
exert some degree of control over her agent’s actions must be very selective about
which advice she follows. She has an incentive to seek information providers who
provide credible reports of agent activities and to avoid information providers who
provide vague or misleading reports.
Some procedural requirements, for example, attempt to induce interested third
parties to share information with the principal that the principal might not otherwise
receive, as is the case when government officials hold public hearings on agency actions
and invite affected interests to provide reports or give testimony. McCubbins and
Schwartz (1984) call this process ‘fire alarm oversight’ because of its similarity to
concerned citizens pulling fire alarms to alert firefighters about dangerous activities.
Such processes allow interested individuals to ‘signal’ political principals about what an
agent is doing. It is useful to note McCubbins and Schwartz contrast fire alarm oversight
with ‘police patrol oversight’ in which principals do much of the oversight themselves.
If a principal can achieve the same quality of oversight using fire alarms and police
patrols, fire alarm oversight would be preferred because others pay the cost of learning
about the agent’s activities. The institutional context in which fire alarm oversight is
attempted can have a great effect on a principal’s ability to learn about her agent’s
activities. Fire alarm oversight is most effective when it occurs amongst institutions that
feature a means for assessing the information provider’s credibility. Institutions can
affect what people choose to say and what people choose to believe. When they have this
effect on people who signal principals, they affect the principals’ ability to judge the
credibility of others.
To clarify how institutions affect delegation, I describe the results of a model (Lupia
and McCubbins 1998) that includes the potential for moral hazard described in Lupia
(1992), and adds a third party who can send signals to the principal about what the
agent is doing.
In the model, principals may be uncertain about the quality of the third party’s
advice. Specifically, Lupia and McCubbins assume that principals may be uncertain
about how much these advice givers know about the agent’s actions and about whether
or not the third party is attempting to mislead the principal. They further assume that
4
The following argument draws extensively from a parallel one in Lupia and McCubbins (2000).
50 Arthur Lupia
signals are sent in the context of institutional forces including the threat of verification
(i.e. a device that helps the principal distinguish true claims from false claims). Such
threats can affect what people choose to say and what people choose to believe. When
institutions have this effect on people who signal principals, they affect the principals’
ability to judge the credibility of others.
Institutions increase the likelihood of verification when they provide new actors with
incentives to provide information about an agent’s actions. An institution such as the
European Union (EU), for example, can now provide information about domestic level
activities (see Lupia 2000). A domestic agent who, absent the threat of verification, will
pursue a personal agenda, rather than the agenda of his principal (in his own country),
must now consider the potentially verifying activities of actors in other EU nations. When
EU membership increases the population of potential verifiers, domestic agents are faced
with a higher likelihood that their principals will detect and reject his actions. In such a
case, the emergence of the European Union can reduce agency loss domestically.
Using the model described earlier, Lupia and McCubbins (1998, paraphrased) gener-
ate the following result about when the principal can use signals to overcome moral
hazard problems with her agent:
Proposition 1. The principal can distinguish agent actions that are better for her than the
SQ from those that are worse only if:
● the principal’s prior knowledge is sufficient for this task or
● she can correctly perceive a third party to have common interests and the knowledge she
desires or
● institutional factors, such as verification, provide effective substitutes for her lack of knowl-
edge about the third party’s knowledge and interests.
This proposition implies that the principal need not know very much about what her
agent is doing in order to delegate successfully. Indeed, if institutional forces are suffi-
cient to induce the principal to follow the third party’s advice, then delegation can suc-
ceed even if the principal initially lacks knowledge about her agent, and even if she
believes that any third party from which she could learn has conflicting interests.
Put another way, if a principal has access to reliable advisers then she need not be
very knowledgeable about her agent to act as if she was so knowledgeable. Moreover,
in cases where a principal may lack information about those who attempt to advise her,
institutions can substitute for the type of information she lacks. Consider, for example,
the position of someone—a juror or a judge—who must render a verdict in a trial.
There will be cases when the juror is presented with a witness whose motives or know-
ledge she cannot verify on her own. However, if the principal is in a setting where fac-
tors such as a high likelihood of verification are present (e.g. when a valuable
reputation is at stake and cross-examination is likely), then she need know little about
the witness’s interests to make a reliable credibility assessment. The incentives imposed
on the witness by the institutional context in which she testifies may be sufficient for
the principal to render an accurate judgement about the witness’s credibility.
Another common form of ex post accountability mechanism is what Kiewiet and
McCubbins (1991: 34) call ‘institutional checks’. Such checks are means by which principals
Delegation and its Perils 51
can reduce agency loss by increasing the likelihood of verification for those who
provide information about agent actions as Kiewiet and McCubbins attest: ‘Rather than
striving for an unbiased source of information, a principal may do better obtaining
biased reports from different agents who have conflicting incentives. The view that
legal proceedings should be adversarial rather than administrative is based on the same
logic. Conversely, checks are disabled when agents’ incentives cease to be in conflict.’
When it is possible to construct such checks, many problems caused by the principal’s
lack of direct information about an agent’s activities can be alleviated. However, like
contracts, institutional checks are not always feasible or effective. As was the case with
contract design, if such checks prevent agents or information providers from under-
taking risky but potentially beneficial activities, then agents and third parties will avoid
such actions, which may lead to increased agency loss ex post.
C O N C LU S I O N A N D I M P L I C AT I O N S
It is impossible to provide a simple statement that describes how common account-
ability mechanisms affect delegation and its perils. However, reviewing the insights of
this chapter can clarify when delegation is capable of accomplishing particular things.
● When a principal and agent desire precisely the same policy outcome and the
agent is willing and able to put forth the work needed to achieve the objective,
then delegation succeeds and there is no agency loss.
● As the agent’s willingness or ability to serve the principal decreases, then the pos-
sibility for agency loss arises.
● When an agent is not fully willing or able to serve the principal, the principal can
minimize agency loss if she has complete information about the agent’s actions.
● If the principal lacks such information, then her ability to reduce agency loss
depends on her ability to acquire other kinds of information.
● Through ex ante mechanisms, principals can attempt to learn about important agent
attributes such as their policy motivations, the extent to which they are willing to
exert effort on the principal’s behalf, and their ability. With such information in hand
principals can reduce agency loss. In many cases, however, such mechanisms are
impossible to construct (i.e. sometimes it is impossible to learn what you need to
know about an agent in advance) and ex ante strategies will be less effective.
● Through ex post mechanisms, principals can attempt to learn about agent actions
from others. With such information to hand, principals can effectively distinguish
agent actions that benefit them from actions that do not—a skill that allows them
to reduce agency loss. Institutions can help by clarifying for principals whose
reports of agent actions are credible.
Political principals delegate with the hope that their agents will work towards ends
they desire. While delegation can have this effect, it need not always. In cases where
agents are willing and able to serve their principals, delegation increases those prin-
cipals’ abilities to accomplish their goals. Otherwise, the perils of delegation arise. Given
the many demands on voters, members of parliament, and members of government,
52 Arthur Lupia
it is unrealistic to expect them to have detailed information about their agents’ prefer-
ences or actions. The perils commonly associated with such information asymmetries,
however, are not inevitable. Ex ante and ex post accountability mechanisms can be made
part of a country’s political institutions—and those entrusted with the responsibility of
governing can retain control of what is done in the government’s name.
I conclude this chapter by describing what such insights imply about delegation
chains. As we look across the parliamentary democracies of Western Europe, we see
within it delegation chains of many shapes. While these chains differ in important
ways, it is important to recognize that the links in these chains are all constructed from
the same basic material—principal–agent relationships that are affected by information
and institutions in systematic ways.
To see the effect of this insight on delegation chains, I begin with a simple example.
Suppose, for a moment, that a country’s chain of delegation is a straight line—one link
connects voters to MPs, one link connects MPs to governments, and so on. In such a
chain, a broken link implies a broken chain. So, for citizens at one end of the chain to
hold accountable civil servants at the other end, every link of the chain must be strong.
Reducing agency loss in such cases requires that the necessary conditions for successful
delegation are satisfied at every link in the chain.
If this condition is not satisfied, the chain is broken. As a result, principals on one side
of the break have no direct control over agent actions on the other side of the break.
The agents on the other side may, if such principals are lucky, take actions that are good
for the principal with the result being a minimum of agency loss. This happy outcome
is not accountability as defined above because the principal has no direct control.
What would it take to satisfy this condition? Not as much as a follower of Weber
might think. Citizens at one end of a straight chain holding accountable civil servants
at the other end does not, for example, require that every principal have complete
information about the actions of their agents. Instead, if at each link in such a chain of
delegation, any of the conditions in Proposition 1 hold, principals have control over
their agents—and agency loss can be no worse than if the principals had never deleg-
ated. If, moreover, a policy exists that both the principal and agent prefer to the status
quo (as is the case in situations 1–3 in Fig. 2.2 and Table 2.2) at every stage in the chain,
then delegation will make all principals better off than if they had not delegated.
But what happens if a country’s delegations are not linked in a serial fashion? The
answer depends on whether failed delegation at one link in the chain can be remedied
by actions at a later link. If so, then it is easier to reduce agency loss—the conditions
for successful delegation must be satisfied at only one of the two substitutable links.
Consider, for example, cases where a policy-maker has the option of working with any
of several bureaucratic agencies to perform a particular task. If the policy-maker
cannot delegate successfully to agency A, delegation may nevertheless succeed if he
can delegate the task to an agency B. If the strength of the delegation chain depends on
this policy-maker’s ability to delegate successfully, he need only find one reliable agent
to keep the chain strong.
As important as the width of the chain is its length. In many cases, the longer the
chain, the greater the likelihood of agency loss. Consider, for example, a government’s
Delegation and its Perils 53
attempt to affect what teachers say to children. In this case, the chain of delegation
extends from the government to cabinet ministers to high-level civil servants in the
nation’s capital to possibly many additional stages all the way until a directive arrives in
a classroom itself. If a teacher simply ignores the mandate, then delegation fails. Indeed,
for certain policies, it may be necessary that all agents in the chain approve of the change.
When such a requirement holds, the length of the chain becomes increasingly relevant as
more links entails more people who have a veto, and greater difficulty for voters or MPs
who want agents to ‘feel a tug’ when they pull the chain.
The chapters that follow give empirical flesh to this theoretical backbone. When
these kinds of knowledge are combined, thy should increase the extent to which
students and real-world observers understand the most effective and efficient ways to
delegate. If they do, then volumes such as this will continue to give governments large
and small ability to delegate successfully—a product of great value in an increasingly
complex world.
REFERENCES
Aghion, Philippe and Tirole, Jean (1997). ‘Formal and Real Authority in Organizations’. Journal of
Political Economy, 105: 1–29.
Akerlof, George (1970). ‘The Market for Lemons: Quality Uncertainty and the Market
Mechanism’. Quarterly Journal of Economics, 89: 488–500.
Banks, Jeffrey S. (1991). Signaling Games in Political Science. Chur, Switzerland: Harwood
Academic Publishers.
Brehm, John and Gates, Scott (1997). Working, Shirking, and Sabotage: Bureaucratic Response to a
Democratic Public. Ann Arbor: University of Michigan Press.
Crawford, Vincent P. and Sobel, Joel (1982). ‘Strategic Information Transmission’. Econometrica,
50: 1431–51.
Dahl, Robert A. (1967). Pluralist Democracy in the United States: Conflict and Consent. Chicago: Rand
McNally.
Farrell, Joseph and Gibbons, Robert (1989). ‘Cheap Talk with Two Audiences’. American Economic
Review, 79: 1214–23.
Farrell, Joseph and Rabin, Matthew (1996). ‘Cheap Talk’. Journal of Economic Perspectives,
10: 103–18.
Gerth, H. H. and Mills, C. Wright (eds.) (1946). From Max Weber: Essays in Sociology. New York:
Oxford University Press.
Harsanyi, John (1967). ‘Games with Incomplete Information Played by “Bayesian” Players, I: The
Basic Model’. Management Science, 14: 159–82.
Kiewiet, D. Roderick and McCubbins, Mathew D. (1991). The Logic of Delegation. Chicago: The
University of Chicago Press.
Kreps, David M. (1990). A Course in Microeconomic Theory. Princeton: Princeton University Press.
Laffont, Jean-Jacques and Tirole, Jean (1993). A Theory of Incentives in Procurement and Regulation.
Cambridge, MA: MIT Press.
Lupia, Arthur (1992). ‘Busy Voters, Agenda Control, and the Power of Information’. American
Political Science Review, 86: 390–404.
—— (2000). ‘The EU, the EEA, and Domestic Accountability: How Outside Forces Affect
Delegation within Member States’. Journal of Legislative Studies, 6: 15–32.
54 Arthur Lupia
Lupia, Arthur and McCubbins, Mathew D. (1998). The Democratic Dilemma: Can Citizens Learn
What they Need to Know? Cambridge: Cambridge University Press.
—— and McCubbins, Mathew D. (2000). ‘When is Delegation Abdication: How Citizens Use
Institutions to Help Delegation Succeed’. European Journal of Political Research, 37: 291–307.
McCubbins, Mathew D. and Schwartz, Thomas (1984). ‘Congressional Oversight Overlooked:
Police Patrols versus Fire Alarms’. American Journal of Political Science, 28: 165–79.
McMillan, John (1992). Games, Strategies, and Managers: How Managers can use Game Theory to Make
Better Business Decisions. New York: Oxford University Press.
Milgrom, Paul and Roberts, John (1986). ‘Relying on the Information of Interested Parties’. Rand
Journal of Economics, 17: 18–31.
Niskanen, William A. (1971). Bureaucracy and Representative Government. Chicago: Aldine-
Atherton.
North, Douglass C. (1981). Structure and Change in Economic History. New York: Norton.
Rasmusen, Eric (1989). Games and Information: An Introduction to Game Theory. Oxford:
Blackwell.
Romer, Thomas and Rosenthal, Howard (1978). ‘Political Resource Allocation, Controlled
Agendas, and the Status Quo’. Public Choice, 33: 27–44.
Salanié, Bernard (1998). The Economics of Contracts: A Primer. Cambridge, MA: MIT Press.
Shepsle, Kenneth A. (1979). ‘Institutional Arrangements and Equilibrium in Multidimensional
Voting Models’. American Journal of Political Science, 23: 27–60.
Spence, A. Michael (1974). Market Signaling: Informational Transfer in Hiring and Related Screening
Processes. Cambridge, MA: Harvard University Press.
Williamson, Oliver E. (1975). Markets and Hierarchies: Analysis and Antitrust Implications.
New York: The Free Press.
—— (1989). ‘Transaction Cost Economics’, in Richard Schmalensee and Robert D. Willig (eds.),
Handbook of Industrial Organization, Volume I. Amsterdam: North-Holland.
3
Parliamentary Democracy
and Delegation
I N T RO D U C T I O N
Madison’s dilemma has been the greatest challenge of modern politics, and it is the
central concern of this volume. Democracy, as we argue in Chapter 1, seeks to solve it
in various ways making the governors responsible to the citizens. Among such designs
for popular sovereignty parliamentary government is most common. But even though
parliamentary systems have progressively empowered their citizens, they rarely feature
much direct popular governance. Not only do citizens typically not make most policy
decisions themselves, they often do not even select or instruct the actual decision-
makers, but delegate in ways that can be disconcertingly indirect.
Parliamentary democracy, then, implies delegation and a corresponding need for
accountability. As Chapter 2 has demonstrated, all delegation, before and after we grant
authority to others to act in our name and place, entails potential agency problems.
Indeed, since antiquity political observers have commented on the dangers of repres-
entation and delegation. The choice and design of a parliamentary system has particu-
lar implications for the forms and severity of political agency problems. And what
makes parliamentary and other regimes democratic is precisely the mechanisms by
which citizens, the ultimate principals, can select and control their representatives. This
book examines the generic features of parliamentary delegation and accountability, as
well as its myriad gradations and variations.
I am grateful to Torbjörn Bergman, Gary W. Cox, Robert L. Brown, James N. Druckman, Simon Hix, Carl
LeVan, Arthur W. Lupia, Mona Lyne, Wolfgang C. Müller, Benjamin Nyblade, Bjørn E. Rasch, and Asbjørn
Skjæveland for helpful comments and suggestions.
56 Kaare Strøm
In this chapter, I first identify three motivations for political delegation and then discuss
the agency problems that may arise and the need for mechanisms of accountability,
such as constraint and competition. In its ideal–typical form, parliamentary democracy
is a particular regime of delegation and accountability, with two salient properties:
singularity and indirect delegation. These institutional features, and the prevalence
of cohesive political parties, help explain the commonly observed phenomenon that
under parliamentary government, the executive branch dominates policy-making
through its control of the political agenda. Yet, by no means are all parliamentary
systems the same and the chapter identifies various subtypes of parliamentary democracy.
Subsequent chapters will indeed demonstrate a great deal of institutional variation
among European parliamentary systems, as well as a contemporary trend towards
weaker political parties and stronger alternative accountability mechanisms.
Building on the ideal–typical features of parliamentary democracy identified above,
I develop a delegation model in which I show that agency loss, in the form of policy
slippage, is greater than under either of two versions of presidentialism. The focus then
shifts to two other causes of democratic agency problems, namely non-policy motivations,
such as leisure-shirking and rent seeking, and incomplete information. When information
is scarce, the perils of delegation include adverse selection and moral hazard. I discuss
various institutional remedies and show that parliamentary democracy prioritizes one
of these agency problems, adverse selection, and facilitates accountability primarily
through ex ante screening by cohesive political parties. On the other hand, parliamentary
democracy is not equally well protected against moral hazard. All told, its distinctive
properties render parliamentary democracy efficient but often lacking in transparency.
W H Y D E L E GAT E ?
Delegation is a feature of all modern governments. Since delegation means voluntarily
giving over authority to others, we must ask why anybody would do so, and indeed
why everybody (or at least every large-scale political system) does. I assume that cit-
izens in democratic societies delegate for reasons that satisfy their criteria of good gov-
ernance. As Robert A. Dahl puts it, ‘I have three main criteria for judging whether I
shall accept as valid and rightful, and therefore binding on me a process for making
decisions on matters that affect me. First, a process may insure that decisions corre-
spond with my own personal choice. Second, a process may insure decisions informed
by a special competence that would be less likely under alternative procedures. Third,
a process may be less perfect than other alternatives according to the first two criteria
but, on balance, more satisfactory simply because it economizes on the amount of
time, attention, and energy I must give to it’ (Dahl 1970: 8).
While the first consideration, which Dahl calls the Criterion of Personal Choice,
militates against delegation, the two others (the Criterion of Competence and the
Criterion of Economy, respectively) constitute key parts of its rationale. In addition,
the social choice literature has demonstrated some general pitfalls of decision-making
in large groups, such as the problems of preference aggregation, collective preference
intransitiv-ities, and collective action. Citizens may thus decide to delegate power
Parliamentary Democracy and Delegation 57
because they face problems of (1) capacity (including Dahl’s Criterion of Economy),
(2) competence, or (3) social choice or collective action (for a similar argument, see Rasch 2000).
Capacity
Although popular sovereignty means that in a general sense the citizenry rules, these
citizens are not always capable of making all (or most) necessary decisions. Their
capacity to do so is limited by a variety of resource constraints, one of which is simply
time: ‘The amount of time devoted to making decisions is limited both by the unwill-
ingness of the people involved to spend an inordinate amount of time on decisions and
by deadlines set by events that will not wait. Because time has value, scheduled meet-
ings often have a more or less fixed amount of time within which to do their business;
the cost of not completing the agenda during that meeting is another meeting or not
getting it done at all’ (Dahl 1970: 44–5).
Hence, citizens may often be persuaded to delegate their authority simply to save
time, and to ensure that necessary decisions are made. One specific version of such
concerns is what economists call transaction costs. Applied to politics, transaction costs
refer to any costs that collective decision-making bodies, such as voters or parliament-
arians, confront in reaching, implementing, and enforcing policy decisions. These costs
may be considerable, particularly when it is difficult to anticipate all the contingencies
that might arise, and they may suffice to persuade these political actors to delegate their
decisions (Epstein and O’Halloran 1999).
Competence
A second motivation for delegation is competence, or rather the principal’s lack of
such. This is the classic motivation for delegation between private citizens and various
professionals. Competence is, for example, the main reason that most people take their
medical concerns to a physician and their legal matters to a lawyer (and not, for example,
vice versa). The knowledge and skills advantages that such professionals possess
motivate ordinary citizens to demand their services at substantial cost. Increasingly,
competence motivates political delegation as well. Most ordinary citizens recognize
that they do not have the requisite insight to determine production and safety standards
for the energy industry or to select the optimal interest rate for government lending. As
Dahl puts it, ‘I should not like to be a surgical patient in an operating room governed by
the principle that one man’s opinion is as much entitled to a hearing as another’s. I do not
want to be tried in a court where every question of procedure might be appealed to a
vote of judge, jury, attorneys, news-papermen, spectators—and defendant. I do not
want a referendum to decide what drugs and food additives are safe enough to be made
available to the public’ (Dahl 1970: 30–1). Contrary to other motivations for delegation,
however, the concern for competence does not drive delegation from voters to elected
representatives as much as it does delegation further ‘downstream’, especially between
cabinet ministers and civil servants.
58 Kaare Strøm
T H E P R I N C I PA L – AG E N T F R A M E WO R K
Delegation is an extremely general, and consequential, way in which citizens in all sorts
of societies try to respond to such challenges. Representative democracy implies a chain
of delegation from voters to those who govern. In this chain, those authorized to make
political decisions (principals) conditionally designate others (agents) to act in their name
and place. The agency-theoretic research tradition, also known as the principal–agent
approach, has generated a growing set of models to illuminate delegation and its
effects. As Lupia has shown in Chapter 2, principal–agent models are a powerful tool
for political analysis.
Yet, no model does full justice to reality, and our view of representative democracy
implies a number of assumptions which to some readers may be controversial, and
which I therefore spell out:
1. The political community is given and bounded. Of course, in the real world this is
never strictly true. As the ‘garbage can’ model of decision-making suggests, many
political organizations have fluid boundaries and decision-makers that come and
go, grow in importance, or fade into the background (Cohen, March, and Olsen
1972; for a critique see Bendor, Moe, and Shotts 2001). In this book, the political
communities are the respective nation-states of Western Europe. But although national
political institutions in Western Europe remain the most important arenas of decision-
making, many decisions are reached through a tug-of-war between national, subna-
tional, and supranational agents. We discuss these issues below and in subsequent
chapters.
2. The preferences of principals as well as agents are exogenously given (i.e. not
explained within our models). Delegation models thus pay no attention to the formation
of preferences.
3. All principals and agents act rationally on the information available to them.
Rational behaviour implies that actors have complete and transitive preferences over
the potential outcomes and that they consistently act so as to obtain more-preferred
outcomes over less-preferred ones. The rationality assumption is less a hard empirical
fact than the most fruitful benchmark for thinking about institutional design and per-
formance. Generally, the verisimilitude of the rationality assumption increases with the
professionalism of the political actors. Thus, political rationality may, for example,
more reasonably be attributed to leading cabinet members than to ordinary voters. This
simply reflects the relative costs of behaving irrationally. A voter may suffer no serious
repercussions if she acts erratically or irrationally in the voting booth. A Prime Minister
behaving similarly in front of the parliamentary opposition may have much more rea-
son to regret such acts.
4. Principals face information scarcities, and information is critical. Informational
assumptions are key to delegation models, which yield particularly acute insights when
information is scarce and consequential.
60 Kaare Strøm
5. Politics is hierarchical, and the principal’s preferences are privileged. The role of the
agent is to propose means among which the principal can choose, not to help the prin-
cipal determine her political goals.1 Moreover, it is the principal’s preferences alone that
are the benchmark of such concepts as agency loss or slack. Thus, the principal’s most
preferred policy position (ideal point) is not only analytically critical but also norma-
tively privileged. When the ultimate principals are the voters, then there is indeed
strong normative reason to be concerned with whether or not their preferences find
expression. But, as we shall discuss below, even in a democracy it is not obvious that the
voters’ desires should always be decisive.
In this book, we apply the delegation framework to parliamentary democracies.
Note, however, that delegation may be neither sufficient nor necessary to solve the
collective decision-making problems that I have discussed. Collective agents may face
the same problems of preference aggregation, collective action, and coordination as
their principals.
Delegation can also be jeopardized by constraints on the agent, when for reasons
beyond his own abilities or preferences, the agent cannot satisfy the principal’s demands.
Here we shall ignore constraints that arise from, for example, the mortality of human
beings, the scarcities of life on earth, or the laws of physics. The more relevant political
constraints may be rules that prohibit certain forms of agency (for example, no demo-
cratic agent can be forced to commit crimes against humanity), or that force agents into
behaviour that neither they nor their principals would have freely chosen.
For example, even if parliamentarians or cabinet members may indeed consider
themselves to be agents of their voters, they also have other constitutionally prescribed
roles to play. Many constitutions, for example the German (Article 38) and Italian
(Article 67) ones, explicitly designate parliamentarians as representatives of all citizens,
and not simply those that elected them. Such broad constitutional accountability is typ-
ically unenforceable, but it can at least give elected officials a normative justification for
ignoring their voters’ preferences. And politicians do indeed tend to embrace this con-
ception of their agency (Wessels 1999).2
Moreover, liberal democracies may have constitutional norms that disallow the
representation of certain popular preferences (e.g. racial discrimination, cruel punish-
ment, or confiscatory government takings). Often, such norms are entrenched in
a constitutional document or a bill of rights. The constitution may also insist on certain
principles, such as the sovereignty or indivisibility of the nation, that were particularly
1
For simplicity, when I henceforth refer generically to principals or agents, I shall, unless otherwise noted,
assume that both principal and agent are individuals and that the principal is female and the agent male.
2
In their study of role perceptions among Swedish members of parliament, for example, Esaiasson and
Holmberg note that ‘one alternative from the history of representational doctrine is missing, however—
Burke’s famous credo of “the nation as a whole”. The reason for this is quite simply that Burke has been too
successful; the norm that national interests come first is so well-established in most European parliaments
that scholars consider it meaningless to probe parliamentary representatives on this matter’ (Esaiasson and
Holmberg 1996: 62).
Parliamentary Democracy and Delegation 61
dear to its framers. In such cases, government officials represent not only their voters
and other superiors, but also the constitutional founders or some set of rights and liber-
ties to which the constitution is committed.
Thus, office-holders in parliamentary democracies are constrained and frequently
common agents, with manifold responsibilities and accountabilities. When we nevertheless
consider the agency framework a powerful tool with which to analyse parliamentary
government, it is because political institutions and the normative ideas of popular
sovereignty can help keep politicians from ignoring their obligations to their principals.
Institutions are often constructed expressly for the purpose of aggregating prefer-
ences in a manageable and reasonably time-efficient manner. Electoral systems do so
for citizen preferences, and voting rules and procedures for legislators. Although all
institutions can be manipulated and none is universally fair, many do give defensible
and predictable expression to the more inchoate desires and beliefs of large numbers of
individual principals. Institutions are also often deliberately designed to allow political
principals to overcome their coordination and collective action problems. This is, for
example, the role that the literature on legislatures assigns to committees (Shepsle and
Weingast 1987; Mattson and Strøm 1995). Arguably, the more consistently an institu-
tion (such as an electoral system) both gives defensible expression to popular prefer-
ences and promotes decisional efficiency, the more likely it is to survive. Thus, the
demand for popular control and governmental effectiveness helps select suitable
institutions.
Delegation models all simplify and distort reality. Yet that may be a price worth
paying if such simplifications help us shed light on modern democratic constitutions.
As in the case of the rationality assumption, I shall in the following suspend disbelief,
as the proof of the framework is in the ‘eating’.
AG E N C Y P RO B L E M S A N D AC C O U N TA B I L I T Y
A good politician is quite as unthinkable as an honest burglar.
H. L. Mencken (1880–1956)
Any delegation of authority entails the risk that the agent may not faithfully pursue the
principal’s interests. If the agent has preferences and incentives that are not perfectly
compatible with those of the principal, delegation may generate agency problems, as
discussed by Lupia in Chapter 2. Agency loss occurs when agents take action that is
different from what the principal would have done, had she been in the agents’ place.
As Lupia tells us, for agency loss to occur there must be some minimum of preference
divergence between principals and agents as well as asymmetric information. If principals
are fully informed about their agents’ behaviour, or if they have identical preferences,
there is no reason to worry about agency loss. Alas, in politics principals often face both
preference heterogeneity and asymmetric information at the same time.
Agency problems under incomplete information may take the form of hidden
information (principals do not fully know the competencies or preferences of their agents
62 Kaare Strøm
or the exact demands of the task at hand) and/or hidden action (principals cannot fully
observe the actions of their agents). The former condition can give rise to adverse
selection, the latter to moral hazard. The former problem may lead principals to select
the ‘wrong’ agents, who do not have the most appropriate skills or preferences. The
problem of moral hazard, on the other hand, arises when agents, once selected, have
incentives and opportunity to take unobservable action contrary to the principal’s interests.
Agents may ‘misbehave’ in several different ways. Brehm and Gates (1997) differen-
tiate between dissent-shirking (failing to serve the principal due to differences in policy
preferences), leisure-shirking (failing to serve the principal due to a preference for
leisure over policy goals), and sabotage (actively undermining the principal’s policy
goals). Here, I shall distinguish between (1) policy divergence (or policy slippage),
(2) leisure-shirking, and (3) rent seeking. I shall discuss these forms of agency loss in
greater detail below. Next, however, let us consider the mechanisms by which principals
may seek to counteract such problems.
Accountability
As Lupia points out in Chapter 2, we can think of accountability as a property of polit-
ical procedures or as an outcome. In this chapter, our focus will largely be on the for-
mer meaning. Political accountability, then, refers to mechanisms by which agency loss
may be contained. As Fearon puts it, ‘relations involving accountability are agency rela-
tions, in which one party is understood to be an “agent” ’ who makes some choices on
behalf of a ‘principal’ who has powers to sanction or reward the agent (Fearon 1999:
55. An agent is accountable to his principal if (1) he is obliged to act on her behalf, and
(2) she is empowered to reward or punish him for his performance in this capacity
(Fearon 1999: 55; Lupia in Chapter 2). And in the words of Manin, Przeworski, and
Stokes, ‘Governments are accountable if citizens can discern representative from
unrepresentative governments and can sanction them appropriately . . . An “accounta-
bility mechanism” is thus a map from the outcomes of actions (including messages that
explain these actions) of public officials to sanctions by citizens’ (Manin, Przeworski,
and Stokes 1999: 10).
Accountability implies that principals have two kinds of rights vis-à-vis agents:
a right to demand information, and a capacity to impose sanctions. Three salient forms
of sanction are the ability to (1) block or amend decisions made by the agent (veto
power), (2) deauthorize the agent (remove him from office or curtail his authority), and
(3) impose specific (monetary or other) penalties. Principals may have one of these
rights (e.g. some form of veto power, or the right to obtain information) without the
others. Accountability varies in the comprehensiveness of the principal’s right to informa-
tion and sanctions, as well as in its exclusiveness. Governance structures in which the
principal’s rights are comprehensive and exclusive, that is to say, where for each agent
there is only one principal, and where that principal has extensive information rights
and sanctions, conform to the classic model of a hierarchy. Where, on the other hand,
Parliamentary Democracy and Delegation 63
there are multiple principals with less comprehensive rights toward the agent, we can
speak instead of a plurarchy.
Strictly speaking, accountability focuses on the rights and sanctions (oversight) that
the principal retains after she has contracted with the agent. Yet, in a broader sense
principals can accomplish some of their control objectives even before delegating. In
Chapter 2, Lupia identifies several means by which principals can contain agency costs,
such as (1) contract design, (2) screening and selection mechanisms, (3) monitoring and
reporting requirements, and (4) institutional checks (Lupia in Chapter 2; also Kiewiet
and McCubbins 1991; Aghion and Tirole 1997).
The former two are mechanisms by which principals seek to contain agency losses
ex ante, that is, before entering any agreement. Principals may then subject potential
agents to careful scrutiny and/or insist on retaining as much residual power as possible.
Contract design typically seeks to establish shared interests, or incentive compatibility,
between principals and agents, so that their preferences are aligned. Screening and selection
represent efforts to sort out good agents from bad ones before a contract is made. In
the case of screening, the principal does the sorting, whereas selection refers to costly
action (for example, the acquisition of recognized credentials) by which the agent
demonstrates his suitability. In politics, parties help voters screen candidates for public
office and parliament screens potential cabinet members. Selection occurs when aspiring
politicians seek minor offices, or run for hopeless seats, to prove their worthiness for more
attractive offices. Also, particular educational institutions, such as the ENA (in France) or
Christ Church college, Oxford (in Britain), may offer valuable political credentials.
The remaining control mechanisms operate ex post facto. That is to say, they are ways
to contain agency losses after the contract has been made, by providing the principal with
information that she might not otherwise receive. Such ex post accountability may rely on
information produced by the principal (known as monitoring or police patrols), by the agent
(reporting), or by some third party (institutional checks or fire alarms). In politics, legislators
may seek to control executive agencies through committee hearings in which ministers
or civil servants have to appear and testify (see Mattson and Strøm 1995). Alternatively (or
additionally), executive agencies may regularly have to report to parliament. Finally,
parliaments employ institutional checks when they subject executive agencies to legal
scrutiny or external audits, or in some other way submit them to the veto powers of
a third party. As the concept implies, such practices are particularly common in checks-
and-balances (collegial or veto player) systems, but by no means confined to such.
Many important control mechanisms serve as vehicles of ex ante as well as ex post
control. They can be used both to select agents in the first place and to subject them
to sanctions and possible ‘deselection’ after the fact. The most basic mechanism of
representative democracy, elections, is clearly of this kind. Voters use elections both
prospectively (to select office-holders) and retrospectively (to sanction incumbents).
Principals face two serious constraints on oversight. First, even though the principal
may find some forms of oversight (such as selection, reporting, fire alarms) less
burdensome than others, for example because they rely on third-party efforts, all oversight
is costly. And reliance on third-party oversight provides no free lunch, since it introduces
64 Kaare Strøm
additional agency problems (when can you trust the reports you get?) (Lupia and
McCubbins 1994). Since oversight is costly, the principal obviously wants to maximize its
effectiveness relative to its cost. Second, principals may face collective action and coordi-
nation problems with respect to their oversight activities. For example, in most large-scale
societies, voters cannot actively deliberate over the selection and supervision of all their
agents. And while it is typically in the interest of all principals that someone bears the cost
of monitoring their agents, no individual principal may have an incentive to do so.
These problems notwithstanding, principals possess a substantial menu of oversight
mechanisms that facilitate accountability. In the larger picture, political regime types
are designed to facilitate accountability (as an outcome) through either or both of two
major mechanisms: competition and constraint. Roughly speaking, in a competition
regime the dominant sanction is deauthorization, whereas a constraints regime is more
beholden to veto powers and penalties (see above). Parliamentary democracy differen-
tiates itself from presidentialism in its affinity for competition relative to constraint. But
as we shall see, we can also distinguish subtypes of parliamentary as well as presidential
constitutions on the basis of their reliance on these respective mechanisms.
PA R L I A M E N TA RY D E M O C R AC Y A N D D E L E GAT I O N
Let us now consider more explicitly the institutions of parliamentary democracy as a
framework for delegation and accountability. There are many ways in which the dis-
tinction between parliamentarism and other regime forms, such as presidentialism,
could be drawn (see Chapter 1). Our approach in this volume is two-pronged. First, we
seek a minimal definition that differentiates as cleanly and parsimoniously as possible
between parliamentarism and other constitutional designs. Chapter 1 thus defines par-
liamentary government as a constitution in which any majority of the members of par-
liament can vote the Prime Minister and his or her cabinet out of office at virtually any
time and for whatever reason they deem sufficient.
While this minimal definition allows us to distinguish as cleanly as possible between
parliamentary and other constitutions, it does not tell us much about their policy
processes. For this purpose, we need a second, ideal–typical conception. We thus iden-
tify the policy process of parliamentary democracy as a chain of delegation that includes
four discrete steps or links:
1. From voters to elected representatives (legislators and others),
2. From legislators to the chief executive (the Prime Minister) and his or her cabinet,
3. From the cabinet and the chief executive to the ‘line ministers’ (typically individual
cabinet members) that head the different executive departments, and
4. From cabinet members, in their capacity as heads of different executive departments
(ministries), to civil servants within their agencies.
This chain of delegation consists of a series of agency relationships. For example, the
Prime Minister is both the agent of the parliamentary majority and the principal of the
line ministers in his cabinet. Democracy as popular sovereignty means that ordinary
citizens are the ultimate principal. There is a reverse chain of accountability that
Parliamentary Democracy and Delegation 65
mirrors the chain of delegation, from civil servants and other office-holders back to the
voters. In Chapter 1 we defined parliamentary democracy in its ideal–typical form as a
chain of delegation and accountability, from the voters to the ultimate policy-makers, in which
at each link a single principal delegates to one and only one agent, or to several non-competing
ones, and in which each agent is accountable to one and only one principal. Figure 3.1 gives a
stylized representation of the chain of delegation under parliamentary democracy,
compared with presidentialism of the US variety.
Under either constitution, the chain of delegation approximates the shape of an
hourglass (turned horizontal). Representation begins with a multitude of principals
(the citizens) and ends with a large number of agents (civil servants). In between, how-
ever, the parliamentary chain narrows down more than does the presidential one. Simply
put, parliamentary delegation relationships take the form of a long and singular chain,
whereas presidential ones look more like a grid. Specifically, the parliamentary chain
narrows down to a single link in which the cabinet connects the elected representatives
of the people and the administrators of the state. It is thus in the relationship between
the legislative majority and the executive that the two regime forms most clearly
diverge. This relationship, of course, is also the focal point of our minimal definition of
parliamentary government. Note also the singularity of principals and agents in the
parliamentary system, as compared with the more complex presidential model (for a
similar conception, see Palmer 1995). In short, classical parliamentarism tends to be
hierarchical, whereas presidentialism typically means plurarchy.
There are two critical institutional features implicit in our very definition of
parliamentary democracy, namely indirect delegation and singularity. On both dimensions,
66 Kaare Strøm
presidentialism is an interesting counterpoint. It is, then, also possible and useful to
place real-world cases on a continuum that spans from one of these ideal types to the
other. Figure 3.1 captures these prominent and important contrasts in constitutional
design.
Indirect Delegation
Parliamentary democracy implies indirect delegation and accountability, as few agents
are directly elected by, or accountable to, the citizens. In the classical Westminster
model, the voters directly elect members of the House of Commons only. All other
agents of the people are indirectly appointed and accountable, often in several stages.
Presidential systems such as the United States fall towards the other end of this con-
tinuum in that the voters directly elect a large number of political offices. In a pres-
idential and federal system such as the United States, for example, this is true not only
of the chief executive at different levels of government (Presidents and Governors), but
also of a variety of other executive officers at the state level, such as attorneys general,
secretaries of state, education commissioners, insurance commissioners, and others.
There are thus in parliamentary democracies several stages of delegation from
voters to ultimate policy-makers. Hence, the challenge is to ensure effective representation
across several links in an indirect chain of delegation and accountability. In reality, of
course, parliamentary systems vary in the range of elective offices and thus in the
directness of delegation. Some parliamentary systems may resemble presidential systems
in having more directly elective political offices, although nowhere in our sample does
this include the Prime Minister or any other cabinet member.
Singularity
A second notable and distinctive feature of parliamentary democracy is the singularity
principle. It implies that each principal employs a single agent (who may, however, con-
sist of several individuals), or a set of non-competing agents, and that each agent
reports to a single principal. Parliamentary democracy in its pure form thus implies an
absence of mutual constraint among agents. Westminster democracies approximate
this conception, and may indeed have served as the model for Laver and Shepsle’s
(1996) conception of parliamentary government: an executive branch organized into a
set of mutually exclusive jurisdictions, each headed by a single and supreme minister
exercising autonomous jurisdictional control. Again, reality exhibits variation among
parliamentary as well as presidential systems, as the singularity principle is nowhere
perfectly legislated or observed. Two or more agents, such as executive agencies, may
compete for the attention of the same principal, for example a particular cabinet
member. Or an agent may see himself as accountable to two or more principals at
different stages of the delegation process. For example, the Prime Minister may at the
same time seek to satisfy the parliamentary and the popular majority. I shall discuss
variations on the parliamentary theme below. First, however, let us consider some of
the specific challenges that follow from a typically parliamentary constitution.
Parliamentary Democracy and Delegation 67
AC C O U N TA B I L I T Y U N D E R PA R L I A M E N TA RY D E M O C R AC Y
Let us now identify some characteristic features of accountability (oversight) under-
parliamentary democracy. A long and singular chain of delegation offers ample oppor-
tunities for agency loss. As Lupia points out in Chapter 2, with some caveats, the longer
the chain of delegation, the greater the potential for agency loss. Also, the singularity
of agents means that principals cannot rely on agents to check one another, or to
make competing bids from which she can choose. To counteract these dangers, the
Westminster model of parliamentary democracy (although no longer found in its pure
form in Britain) relies heavily on a mechanism that ex ante sorts agents by preferences
and subjects them to centralized authority. This mechanism is party government.
Voters Voters
Party Party
MPs MPs
Party Party
Government Government
Party Party
Ministers Ministers
Party Party
Delegation Inter-party
weak to strong competition/cooperation
government’. As Müller puts it, ‘Parties are least able to control the final stage in the
delegation chain—the behaviour of civil servants. The fact that the role of political parties
deteriorates the more the delegation chain develops, reflects increasing informational
asymmetries and the relevance of normative constraints’ (Müller 2000: 330).
Political parties are complex collaborative devices for mutual gain, formed because
candidates for public office and voters find them useful for their respective purposes.
But to serve these purposes, political parties have to satisfy two incentive conditions.
First, they have to provide sufficient inducements for political office-holders to submit
to the cohesion (discipline) that they impose. Second, the policy cohesion that parties
induce among office-holders must be sufficient that voters find the party label inform-
ative and useful. In addition, it is helpful (though not strictly necessary) for parties to
attract activists that can help them in the recruitment, training, and selection of potential
candidates for office, as well as provide inexpensive campaign labour.
The first of these conditions means that party leaders must control the recruitment
(election and appointment) of government personnel, as well as government policy-
making. For party government to be efficient, parties must be the only, or at least the
preferred, way in which legislators realize their main political goals: re-election, political
Parliamentary Democracy and Delegation 69
power, and policy objectives. John Aldrich (1995) identifies three sets of incentives
among politicians for party formation: collective action problems, social choice problems,
and political ambition (see also Cox and McCubbins 1993). I have already discussed
collective action and social choice problems as motivations for delegation, and such
incentives exist among legislators as well as among citizens at large. To provide effective
answers to these challenges, parties must be able to structure the legislative and executive
arenas so as to facilitate effective decision-making, and they must control access to the
offices to which politicians aspire.
At the same time, the second condition implies that parties must be sufficiently effective
‘informational economizing devices’ that voters and activists will sustain them. Voters
can only rely on party labels, or in other words delegate policy aggregation to party
leaders, if these are capable of enforcing policy agreement in government. If legislative
politics is anarchic, then party labels can be of no use to the voters. Furthermore, as
Michael F. Thies points out, the policies underpinning the party label must resonate
with voter preferences: ‘Mixing apples and oranges is fine if you can sell yourselves as
the fruit party, but if voters still like apples or oranges, but not both, they will no longer
be able to learn much from the “Fruit Party” label that will help them with their vot-
ing decisions’ (Thies 2000: 249). The evolution of meaningful party labels in Europe
was facilitated by the profound social cleavages generated by the industrial and national
revolutions (Lipset and Rokkan 1967). In Britain, due to the prior resolution of religious
conflict and the temporary suppression of ethno-national ones, the predominance of
the class cleavage helped foster two successive two-party systems with strong partisan
attachments.
Finally, for voters to attach themselves to political parties, there must also be electoral
institutions that translate popular votes into partisan representation in a reasonably
meaningful and transparent way. In the classical Westminster model, this is indeed the
case. Under a competitive two-party system and plurality single-member districts, the
electoral menu is simple and the process transparent. Moreover, the process is non-
perverse in that voters can never hurt themselves (would never experience rational
ex post regret) if they vote for their most-preferred party.
Two institutional features of parliamentary democracy reinforce centralized party
authority: the confidence vote and the dissolution power. We have already noted that
it is the parliamentary majority’s dismissal power, through a vote of no confidence, that
fundamentally sets parliamentary government apart from other regime types. But
paradoxically, the flip side of this procedure enables the cabinet and the party leadership
to dominate the legislative branch. The confidence vote enables the Prime Minister to
attach the cabinet’s fate to some bill before parliament. This is indeed a ‘doomsday
device’ (Lupia and Strøm 1995), by which the entire cabinet can be removed in one fell
swoop. Yet, this ability to raise the stakes and redefine the parliamentary agenda
often enables Prime Ministers to quell policy dissent within their respective parties or
coalitions. It also allows them to manipulate the legislative policy process (Huber 1996).
The dissolution power is another doomsday device that similarly allows the Prime
Minister or cabinet to threaten the parliament with an immediate electoral verdict.
Most parliamentary democracies feature some provision for early dissolution. In the
70 Kaare Strøm
Westminster model, within the limits of the maximum constitutional term, the election
date is fully controlled by the Prime Minister and his party. Consequently, this dissolution
power can be used strategically for partisan purposes (Strøm and Swindle 2002).
Consequences of Partisanship
The prominence of partisanship in parliamentary democracies has important rami-
fications. One has to do with the pool from which political agents are selected. In
parliamentary democracies, ex ante screening is reinforced through internal selection of
agents such as parliamentary candidates and cabinet members, particularly in socially
cohesive (often externally created) parties. The stronger the ‘partyness’ of society
(Sjöblom 1987), and the more parties represent distinct social groups, the more they
tend to rely on prior screening devices, such as strong extra-parliamentary membership
organizations, to make sure that candidates represent ‘their kind’ and the best of their
kind. Thus, candidates for office are recruited by and from the party organization.
The second arena in which internal selection is used is parliament. In many parlia-
mentary democracies, such as the United Kingdom, law or convention requires that
cabinet members simultaneously hold parliamentary office. Thus, the cabinet is a true
subset of parliament. This is a common form of contract design in Westminster
systems, and some observers see it as the defining feature of parliamentary government
(e.g. Hernes and Nergaard 1989; Lane and Narud 1992).3 Even in countries without such
requirements, cabinet members often have substantial parliamentary experience (for
cross-national surveys, see Andeweg and Nijzink 1995; Saalfeld 2000).
Governance through cohesive political parties helps align preferences, and to some
extent information, along the chain of delegation. The purpose of the concomitant
practice of selecting agents internally is mainly informational. Internal, rather than
external, selection of agents has obvious informational advantages, which, as Juan Linz
laments, may not be available in presidential systems: ‘The presidential candidates do
not need and often do not have any prior record as political leaders. They may not be
identified with a party with an ideology or program and record, and there may be little
information about the persons likely to serve in the cabinet’ (Linz 1994: 11).
Another consequence of Westminster parliamentarism is that the cabinet enjoys
virtually monopolistic agenda control. Agenda control, or the ability to control the set of
policy options under deliberation, can be a source of great power. McKelvey’s (1976)
famous ‘chaos’ theorem shows that an agenda setter can be a virtual policy dictator.
Yet, the value of agenda control depends on the context. Thus, Tsebelis (2002) shows
that the power of agenda setters generally declines (or at least does not increase) with
the number of veto players.
In agency models, as Lupia points out in Chapter 2, generally the agent proposes and
the principal disposes. That is to say, when we select an agent, we simultaneously vest
3
Applied to the relationship between voters and parliamentarians, internal delegation implies a residence
requirement for elected representatives. Although such rules, or at least norms, exist in some countries, they
are not uniquely associated with parliamentarism. Since the introduction of civil service reforms, internal
delegation is not generally a feature of the relationship between cabinet members and civil servants.
Parliamentary Democracy and Delegation 71
that person (or group or organization) with agenda powers. Thus, for example, when
a cabinet member (she) delegates a policy issue to a civil servant (he), she specifically
asks him to prepare a policy proposal that she may later approve or veto. This recognition
helps us reconcile two apparently contradictory facets of the democratic policy
process. On the one hand, constitutions prescribe a process of delegation that moves
from voters and ‘downstream’. On the other hand, students of decision-making often
point out that reality looks more like our ‘upstream’ process of accountability: proposals
are initiated by agents and work their way back to the ultimate principals for final
approval. Delegation in fact means that authorization moves in a forward direction
(from principals to agents) and policy proposals backwards. To put it baldly, agents have
agenda control and principals veto power.
The simple delegation scheme of parliamentary democracy implies a concentration
of agenda control in the cabinet. That is to say, since voters delegate along a singular
chain of delegation, agenda control over the policy process is placed squarely, and
indeed almost monopolistically, in the cabinet. The cabinet’s agenda control can be
derived directly from the parliamentary chain of delegation. Policy initiative rests with
the executive because the cabinet is the sole agent of the parliamentary majority. Co-
partisanship supports such delegation by aligning political preferences and incentives.
The fact that civil servants generally are not partisan means that cabinet members will
be more reluctant to delegate broadly to agents further ‘downstream’.
Under presidentialism, the legislature and the executive are instead coequal agents of
the citizens. Since the President is not accountable to the legislators, they are under-
standably reluctant to grant him exclusive policy initiative. Yet, since there is simply no
hierarchy between legislature and executive, it is not obvious that agenda control
instead rests with the former. Rather, legislators may still look to the President and the
executive branch for expertise and policy initiatives, but will delegate to the executive
only when such delegation can be sustained by common partisan bonds, that is, when
a single, cohesive party controls both the legislature and the executive branch (as, for
example, in the United States during the New Deal and after the 2002 elections). In such
situations, the preferences of President and legislature are likely to be well aligned.
Otherwise, we might instead see a tug-of-war between legislature and executive over
agenda control.4
Ineffective Oversight
Whereas Westminster parliamentary democracy thus relies heavily on partisanship and
internal selection to contain agency problems, it makes far less use of another class of
accountability mechanisms, namely oversight after the fact of delegation. We can think
about a continuum of vehicles of ex post oversight, from those that are integral to the
chain of parliamentary delegation (consisting of the voters’ representatives or their
4
In his veto player analysis, George Tsebelis (2000, 2002) similarly argues that in parliamentary systems
agenda control rests with the executive branch, whereas under presidentialism it is lodged with the legis-
lature. As suggested above, I agree with the first part of this argument, but not necessarily with the latter.
72 Kaare Strøm
direct agents), to those that are most clearly external (not directly or indirectly account-
able to the voters). Internal mechanisms include such procedures as parliamentary
committee hearings. The usefulness of such mechanisms will depend heavily on the
degree to which agent preferences are aligned through effective partisanship. The more
effectively partisanship works, the less incentive for politicians to use these mechanisms.
Ex post constraint comes in two forms: (1) multiple (or partitioned) democratic principals,
and (2) external checks (such as constitutional courts) that are outside the direct popular
chain of delegation. Presidential systems, such as the United States, typically feature a
policy process with both multiple principals and external checks or constraints.
Parliamentary systems are less likely to have such a plethora of external constraints.
Recall that in its pure form, parliamentary democracy implies a singularity of principals (as
well as agents) and more generally a lack of external constraints or checks on policy mak-
ers. The Westminster model in fact comes close to this ideal type. In Dicey’s (1959: 39–40)
formulation, for example, parliamentary sovereignty means that Parliament ‘has the right
to make and unmake any law whatever; and, further, that no person or body is recognized
by the law of England as having a right to override or set aside the legislation of
Parliament’.5 Any democratic constraint in such a system would thus have to be internal.
Yet, in a Westminster parliamentary system, even internal, ex post oversight tends to
be weak and ineffectual. This claim may at first encounter sound odd, since it is the very
ex post accountability of the cabinet to the parliamentary majority that defines
parliamentary government. But this ex post accountability depends almost entirely on
electoral competition. And the indeterminacy of election dates under most parliamentary
constitutions also complicates ex post electoral accountability. In the pure Westminster
model, within the limits of the maximum constitutional term, the Prime Minister and
his or her party fully control the election date. Consequently, this dissolution power can
be used strategically for partisan purposes (Strøm and Swindle 2002). The manipulability
of election dates reduces the effectiveness of ex post electoral accountability and adds
to the informational asymmetry between ins and outs.
Between elections, accountability is tenuous for three reasons. First, parliamentary
democracy lacks institutional mechanisms for credible ex post oversight. Second, par-
liamentary systems (at least in their pure form) lack the capacity to determine when
sanctions are appropriate. Third, and most critically, party government implies that the
parliamentary majority lacks motivation to sanction its agents. Let us consider these
issues in turn.
First, while the hallmark of parliamentarism, the no confidence vote is blunt and
unwieldy and there are few alternative sanctions. The impeachment procedure (which
elsewhere has been employed not just in the United States, but also in Brazil and
Venezuela) has in most parliamentary systems fallen into disuse. Nor are other and less
severe forms of censure or reprimand common. And recall provisions, which feature in
some presidential constitutions, are essentially unknown (see Chapter 4).
A second problem is that parliamentary systems do not have the monitoring capacity
necessary to determine when sanctions might be appropriate. Presidential systems tend
5
As quoted in Norton (1994: 63).
Parliamentary Democracy and Delegation 73
VA R I E T I E S O F PA R L I A M E N TA RY G OV E R N M E N T
Parliamentary democracy, we have noted, typically features strong ex ante partisan
control of democratic agents, and few and weak ex post checks and constraints. Yet, not
all parliamentary democracies are typical. Indeed, real-world parliamentary systems
exhibit substantial variation in accountability regimes along two critical dimensions.
One ranges from the cohesive and competitive party government depicted above to a
weaker and more fragmented pattern. The other dimension runs from the lack of
external accountability mechanisms to greater structural constraint. And while these
two dimensions are analytically distinct, there is reason to believe that they may be
empirically and historically correlated, in the sense that a weakening of partisan bonds
has contributed to an evolution towards more effective constraint (see Chapter 22).
Political scientists have, of course, duly noted this variation. Weaver and Rockman’s
(1993) insightful work divides parliamentary systems into three different subtypes. The
key to this differentiation is the partisan composition of governments, which in turn is
74 Kaare Strøm
driven by differences in electoral systems and accounts for substantial differences in
policy performance. Similarly, Arend Lijphart’s (1984, 1999) influential work differentiates
between Westminster and consensus democracy along two dimensions. The first
dimension, which Lijphart seems to privilege, has much to do with the electoral system
and the number of political parties. Finally, in a seminal series of publications, George
Tsebelis (1995, 2000, 2002) argues that with respect to the capacity for policy change,
‘the logic of decision-making in presidential systems is quite similar to the logic of
decision-making in multiparty parliamentary systems’ (Tsebelis 1995: 292). The critical
variable is not regime type but the number of veto players. All else equal, the larger the
number of veto players, the lower the capacity for policy change. Like Lijphart and
Weaver and Rockman, Tsebelis distinguishes between two-party and coalitional
parliamentary systems. Contrary to the latter, however, he goes on to claim that the
multiple veto players of coalitional multiparty parliamentary democracies generate
policy-making dynamics akin to presidential systems. A further implication of Tsebelis’
argument is that deviation from our ideal–typical model along either dimension
(party cohesion and external constraint, respectively) should have similar consequences.6
These issues are certainly worthy of our attention. To the extent that parliamentary
systems deviate from the Westminster model, by having lesser party cohesion or
greater ex post constraint, how is democratic delegation affected? Below, I therefore
discuss party cohesion first and institutional constraint subsequently.
Party Cohesion
Real-world parliamentary systems have been drifting away from the simplicities of the
Westminster model, towards less cohesive parties, for about a hundred years. The most
important causes of this evolution have been the introduction of proportional repre-
sentation (PR) and the consequent fragmentation of parliamentary party systems
(Rokkan 1970; Boix 1999). Since about First World War, most parliamentary systems,
and particularly those in continental Europe, have not been competitive two-party
systems. Concomitantly, multiparty government has become the norm (see Müller and
Strøm 2000).
As noted, political parties are especially critical to the operation of parliamentary
systems, as they function as devices of ex ante screening and preference alignment.
Hence, the question is whether regimes that involve many (perhaps four or five)
6
Tsebelis acknowledges that the vetoes of institutional veto players are enforceable in ways that those of
partisan veto players are not. If one party in a coalition government, for example, refuses to go along with a
particular policy proposal, it cannot prevent the rest of the governing parties from passing the bill anyway,
provided they can gather enough support on the floor of parliament. Tsebelis argues that the dissenting party
can effectively prevent such outcomes by threatening to withdraw from the coalition and thus forcing
a resignation or renegotiation. While there are surely cases in which a coalition party can have its way by
threatening to withdraw, it is far from obvious that all governing parties have such bargaining power. In fact,
a dissenting coalition party should be able to veto policy a proposal only if its threat to bring down the
government is credible, that is to say, if its walk-away value is sufficiently high (see Lupia and Strøm 2006).
Some parties have high walk-away values because they are pivotal and have other potential coalitions
that they could join. But others are not so privileged, for example, if they are surplus parties in an ‘oversized’
coalition, of if their policy preferences are relative extreme.
Parliamentary Democracy and Delegation 75
parties simultaneously in executive policy-making differ from those in which one party
controls the cabinet. The answer is not entirely straightforward. The value of parties is
determined not by their official status, but by their ability to condition the behaviour of
voters and their agents. The fact that for a long time Christian Democrats in Italy were
organized as a single party (the DC), whereas in Germany there were two such
parties (the CDU and the CSU), does not mean that the former served as a more effective
vehicle of agency control than the latter. The contrary is probably closer to the truth.
Thus, the ‘partyness’ of a given polity is determined less by the number and size of polit-
ical parties than by the policy cohesion and electoral competition they generate.
The value of partisanship hinges on the incentives that it offers politicians on the one
hand and voters on the other. These factors are in turn a function of rules concerning
ballot access, electoral seat allocation, legislative agenda control, confidence relation-
ships, parliamentary dissolution, and other choices that matter to politicians. They also
depend on the organization and strategies of the political parties themselves. Parties
have developed a range of devices by which they induce cohesion by controlling access
to a range of perquisites that politicians desire, such as ballot access, coveted committee
assignments, campaign funds, and the like.
The PR system directly affects the relationship between voters and parliamentarians
less than the relationship between parliament and the cabinet. At least closed-list PR
systems do nothing to weaken the internal cohesion of political parties. Such electoral
systems may, however, jeopardize the cabinet cohesion that is traditionally found in the
Westminster system. Interparty coalitions must strive hard to generate and control
similar inducements and then to distribute them in such a way as to foster coalitional
cohesion. Such coalition governance institutions are feasible, but not simple or universally
effective. Coalition parties often sign extensive policy agreements and set up institu-
tional arrangements, such as coalition committees, inner cabinets, or party summits, to
enforce them. Yet, such agreements are notoriously difficult to enforce, and there are
numerous cases in which they have failed (Strøm and Müller 1999).
The generally lower level of cohesion in multiparty systems is due not just to enforce-
ment problems, but also to a lack of suitable incentives for politicians. As Palmer (1995)
notes, interparty coalitions are unlikely to align electoral incentives as effectively as a sin-
gle political party. Whereas the cohesion of copartisans is brought about in large part by
the expectation of a common electoral fate, this tie is much less likely to bind in multi-
party coalitions. Some electoral systems, such as the German mixed-member system or
those with apparentement provisions, permit politicians to make coalitions more ‘party-
like’ by forging pre-electoral agreements that effectively allow voters to treat the coali-
tion as if it were a party. But even though such institutions may help, electoral coalitions
are typically too fragile to be genuine substitutes for parties. Under the most common
electoral systems, members of interparty coalitions are much more likely than coparti-
sans to be in mutual competition for votes. Although the electoral fortunes of coalition
parties are likely to be positively correlated, there are many examples of parties that ben-
efit at the expense of their coalition partners (Rose and Mackie 1983).
Party system fragmentation also affects voter incentives. To the extent that interparty
coalitions are unstable, of course, such commitments become less valuable to the voters.
76 Kaare Strøm
The more parties break their own commitments, the less reason voters have to count on
them. Party cohesion helps voters control access to the executive branch through the
ballot box. Multiparty politics clouds and complicates the relationship between electoral
contests and executive power. Voters can no longer assume that voting for their most
preferred party will necessarily be their best strategy. Moreover, the entire linkage
between electoral results and control of the executive branch may become tenuous. In
extreme cases (as in Italy 1945–87), the responsiveness of government formation to elec-
toral results may become negative, meaning that parties that lose votes in elections sub-
sequently are more likely to enter government than those that gain (Strøm 1990).
These problems of transparency and responsiveness are more severe in some multi-
party systems than in others. In terms of their competitive characteristics, multiparty
systems may usefully be subdivided into alternational versus pivotal types. In alterna-
tional systems, two blocs of parties (internally cemented, perhaps, by effective agree-
ments) compete for the voters’ favour and alternate in power much in the way that
individual parties do in classical two-party systems. In pivotal systems, on the other hand,
there is a single party, or occasionally a bloc of smaller parties, that forms of the core of
the governing coalition, often because it controls the median legislator on the dominant
policy dimension. This pivotal party may occasionally change coalition partners, but
there is no wholesale alternation in power as a result of electoral contests. It is in pivotal
systems that the bond of partisanship is most severely tested. Institutions such as the
German mixed-member system or the French double ballot may drive multiparty
systems towards the alternational subtype. Yet, particularly in pivotal party systems,
coalition partners cannot assume that they are doomed to sink or swim together and
often have electoral incentives to defect. Hence, the glue that keeps politicians in line and
subjects them to popular control is particularly suspect in pivotal party systems.
Institutional Constraints
The Westminster model of parliamentary democracy concentrates power in the hands
of a virtually unconstrained democratic executive. Yet, in the real world many parlia-
mentary systems have a larger number of constitutional constraints on the power of
the Prime Minister and his or her cabinet. Germany is an obvious example, where
the power of the chief executive is constrained by an indirectly elected Bundesrat, by
a federal constitution, by a constitutional court, by the institutions of the European
Union, including an enterprising judiciary and an autonomous central bank, and by
a practice of social partnership. As noted above, constraints come in two basic forms:
partitions and checks. Partitions, such as federalism, consociationalism, or corporatism,
subdivide democratic principals into different subgroups. Checks, such as judicial review,
presidential vetoes, or abrogative referendums, subject the decisions of democratic agents
to various veto gates. In principal-agent terms, the former category corresponds to a sit-
uation of multiple principals, whereas the latter (at least in some cases) is akin to one of
multiple, checking agents.
Such constraints may play a role similar to that of political parties in containing
agency costs in the democratic chain of delegation. Yet, they do so in different ways and
at different stages. For reasons I have already discussed, party cohesion subverts the
Parliamentary Democracy and Delegation 77
7
A President in a parliamentary state, for example, can have each of these powers in different domains.
He can be a dictator as commander of the armed forces, for example, or in some of his appointment powers.
He can be a veto player with respect to ordinary legislation. Of course, we could further distinguish between
presidents whose vetoes are only suspensory, those whose vetoes could relatively easily be overriden, and
those whose vetoes are nearly impossible to override (see Shugart and Carey 1992). Furthermore, the
President can be a decisive player with respect to parliamentary dissolution if he can freely dissolve parlia-
ment but there are other circumstances in which parliament may also be dissolved without his involvement
(see Strøm and Swindle 2002). Finally, in areas in which he has no constitutional powers, a President can be a
powerful player, if by expressing his endorsement or displeasure, giving exhortatory speeches, or dragging
his feet, he can affect that benefits that other players derive from particular initiatives.
78 Kaare Strøm
I N S T I T U T I O N S A N D P O L I C Y D I V E RG E N C E
Having laid out the institutional features of parliamentary versus presidential constitu-
tions, it is now time to consider their implications for specific democratic agency problems.
In discussing specific agency problems in contemporary democracies, I proceed in
three steps. First, I develop a delegation model in which I assume complete information
and thus isolate agency problems that stem strictly from policy preference divergence
between principals and agents. The model will highlight specific agency problems that
arise from the institutional features of parliamentary and other constitutions. Next
I consider, in a less formal way, agency problems associated with various non-policy
motivations. Finally, I discuss two critical agency problems, adverse selection and moral
hazard, that may arise under asymmetric information, when principals lack certain
critical information about their agents.
Let us, then, first examine parliamentary agency problems that may arise even when
the democratic principal is fully informed about the preferences of her agent. Consider
Fig. 3.3, which illustrates a delegation game, loosely in the tradition of Romer and
Rosenthal (1978) and Lupia in Chapter 2, but with one principal and two agents. This
Parliamentary democracy
t
accep X
accept
A1 choose x ∈ℜ1 A2 P reject
reje
ct RP
RP
x X
choose x ∈ℜ1 ose
A1 cho
choose y
P Y
reje
ct
A2 choose y ∈ℜ1
RP
Figure 3.3. Agency Loss (Policy Divergence) Under Parliamentary and Presidential
Government.
Parliamentary Democracy and Delegation 79
game may take three different forms, depending on the relationship between the two
agents: a hierarchical model, in which one agent is subordinated to the other, a competitive
model, in which the agents contest for the favour of the principal, and a collegial model,
in which both agents have to approve any proposal that is made to the principal, but in
which no agent is systematically subordinated to the other.
These models roughly correspond to major contemporary regime types. We can
identify the principal as the median voter, and the agents as the median member of the
national legislature and cabinet, respectively. (For simplicity, I here ignore the civil
service.) The hierarchical model then corresponds to a parliamentary regime, and the
two latter models to what I shall call competitive and collegial presidentialism, respectively.
The common features of these models are as follows: One or both agents are
allowed to make policy proposals. The principal can either accept one proposal, or
reject any or all proposals. If the principal accepts some proposal, that proposal is
implemented and the game ends. If no proposal is accepted, a reversion point (which
could represent the status quo) prevails. I assume that any proposal can be represented
by a point on a policy continuum and that all players have Euclidean preferences and a
most-preferred position (an ideal point) on this policy dimension. The solution concept
is subgame perfect Nash equilibrium.
For simplicity, I set the value of the reversion point (status quo) to zero. For the same
reason, I assume that no player is indifferent between any proposals, and that no player
has an ideal point identical to that of any other player, or to the reversion point.
Relaxing any or all of these restrictions should add complexity but not change any of
the general results. All players are assumed to have full information concerning all
aspects of the game. I further assume that it is costless to make or accept proposals, but
that if indifferent between a proposal and the reversion point, any player will opt for
the latter. Under these assumptions, no agent will make or endorse a proposal that will
not be accepted, or that will not represent an improvement for him, relative to the
reversion point. Similarly, if the principal is indifferent between a proposal and the
reversion point, she will reject the proposal.
The three games differ in their structure. The parliamentary game proceeds as
follows: Agent 1 decides whether to propose. If he does, agent 2 either rejects or endorses
the proposal. If agent 1 fails to make a proposal, or if agent 2 rejects it, the game ends
and the reversion point prevails. If, however, agent 1 proposes and agent 2 endorses,
then the principal either accepts or rejects that proposal as it is. Under competitive presid-
entialism, on the other hand, agents 1 and 2 decide separately and independently
whether to propose. The principal can then choose between any proposals and the
reversion point. Finally, under collegial presidentialism, the principal initially vests agenda
control in one of her two agents. Along either branch of the game tree, the subgame
that follows is identical to the parliamentary game. In other words, the agent vested
with agenda control may propose, the other agent endorses or vetoes this proposal, and
if the second agent endorses, the principal finally accepts or rejects that proposal.
The two versions of presidentialism differ in their reliance on competition and
constraint, respectively. The collegial model may be the more familiar representation
of real-world checks-and-balances systems. Yet, even the competitive model captures
80 Kaare Strøm
important elements of procedural choice inherent in presidential regimes. Under US
presidentialism, for example, legislators may establish a range of different executive
agencies with similar or overlapping jurisdictions. They sometimes design these insti-
tutions so as to allow some choice over which agents to utilize. And particularly in
federal systems, policy advocates may have some choice whether to press their concerns
in state or federal institutions. In the United States, for example, anti-tobacco activists,
gun control advocates, and right-to-life groups are among those that have used different
channels of access in such a competitive fashion.
The model allows us to examine the magnitude of policy divergence under different
institutions and preference configurations. I measure this loss as the distance between
the outcome (whether an accepted proposal or the reversion point) and the principal’s
ideal point.8 Table 3.1 presents the outcomes of all three games under the ten different
preference configurations under which they vary.9 For each preference configuration
and regime type (institutional rule), Table 3.1 reports the magnitude of agency loss.
Note that in some cases the rank ordering of the ideal points and the reversion point
does not suffice to precisely determine the equilibrium. Rather, the latter depends on the
relative distance between these points. Because of these complications, I report two sets
of values for each preference configuration: one representing a best-case scenario (min-
imal policy divergence) and one a worst-case scenario (maximum policy divergence).
The general lessons of these games can be fairly easily summarized: the parliamentary
regime is susceptible to greater agency loss than either presidential model. Out of the
latter, competitive presidentialism generally minimizes policy divergence, with the
collegial model between parliamentarism and competitive presidentialism. There is no
arrangement of preferences under which agency loss is greater under either presidential
model than under parliamentary institutions. Under most preference configurations,
agency loss under parliamentarism varies within the same bounds as under collegial
presidentialism. Yet, there are two configurations (numbers VII and X) under which the
latter presents a more favourable best-case scenario. These are both cases in which all
players favour a change in the same direction relative to the reversion point, but in
which the parliamentary agenda setter (agent 1) has the most status quo-oriented
preferences. The contrast between parliamentarism and competitive presidentialism is
more clear-cut: this presidential variety is superior (has smaller policy divergence in
either the best- or the worst-case scenario, or both) in seven out of ten configurations
8
As Lupia points out in Chapter 2, agency loss could be measured either against the reversion point or
against the principal’s ideal point. Since in this model I am assuming complete information as well as ration-
ality, the principal will never accept a proposal that is inferior to the reversion point. Hence, only the second
measure of agency loss is meaningful in a model of this kind.
9
In all, there are twenty-four different ordinal preference configurations—different ways in which the
reversion point and the ideal points of the various players can be arrayed. Note that we shall be concerned
only with the magnitude, but not the direction, of policy divergence. Therefore, we can disregard half of
these preference configurations, since each of them is simply a ‘mirror image’ of one other configuration, in
which the order of the ideal points is reversed. For simplicity, I also ignore any preference ordering in which
both agents are at the opposite side of the reversion point from the principal. In no such situation would
rational and fully informed players adopt any alternative to the reversion point, regardless of institutions.
That leaves us with the ten preference configurations represented in Table 3.1.
Table 3.1. Policy Divergence Under Different Regime Types
Note: The entries in each cell represent the magnitude of policy divergence, measured as the absolute distance between the equilibrium and the
principal’s ideal point. Vertical bars denote absolute values. See the text and Fig. 3.2 for descriptions of the models.
82 Kaare Strøm
Table 3.2. Policy Divergence Under Different Regime Types: Numerical Example
Note: R represents the reversion point. The entries in each cell represent the magnitude of
policy divergence, measured as the absolute distance between the equilibrium and the principal’s
ideal point. In each case, the numerical values of the various positions are 1, 0.67, 0.33, and 0,
respectively. See the text and Fig. 3.2 for descriptions of the models.
and is otherwise equal. In a majority of best-case scenarios (six out of ten), competitive
presidentialism generates no agency loss at all. The corresponding numbers are four
for collegial presidentialism and three for parliamentarism. Moreover, there are four
configurations in which competitive presidentialism never exhibits any agency loss,
whereas this is never true for parliamentarism or for collegial presidentialism.
To make the results more accessible, Table 3.2 presents a set of numerical examples.
Here I have placed the reversion point and all three player ideal points, in every prefer-
ence configuration represented in Table 3.1, on a policy scale from zero to one, and
roughly equidistant from one another. Specifically, the numerical values of the differ-
ent positions are 0, 0.33, 0.67, and 1, respectively. Note that the reversion point in this
example is allowed to vary, rather than being set to zero. Agents are allowed to make
policy proposals with precision down to the second decimal point (e.g. 0.51).
Otherwise, all features of this example are identical to those represented in Table 3.1.
The numerical results are no great surprise, though the advantages of collegial presi-
dentialism over parliamentarism are a bit more obvious here.
The greater policy divergence under parliamentary institutions is largely due to
the agenda power of the agenda-setting (first) agent (the cabinet). Parliamentarism is
particularly susceptible to agency loss when this agent has relatively extreme preferences,
while at the same time the reversion point is unattractive to the principal. While these
results may at first blush seem abstract and arcane, it is not difficult to find empirical
examples that conform. Unattractive reversion points are particularly likely to obtain
immediately after such events as regime transitions and major international crises
(such as war). Then, parliamentary cabinets may indeed use their agenda control to
push policies that are controversial to fairly large numbers of voters, as seems to have
been the case with the Vaclav Klaus post-transition cabinet in the Czech Republic, or
Parliamentary Democracy and Delegation 83
with some social democratic post-Second World War cabinets, such as Clement Attlee’s
in Britain or Einar Gerhardsen’s in Norway, or even Margaret Thatcher’s administration
in the wake of the British 1978–9 ‘winter of discontent’. This is not to say that such
radical policy departures may not occasionally be desirable for economic or political
reasons, as many transition theorists have indeed maintained. Yet, the monopolistic
agenda control, which characterizes parliamentary democracy, leads to a substantial
potential for policy choice that deviates from the preferences of the median voter.
In the discussion above, I have considered only the magnitude, but not the direction,
of policy divergence from the principal’s ideal point. If in a less abstract setting we con-
sider democratic delegation as an ongoing process, however, we may want to examine
also the direction in which adopted policies tend to deviate. If in repeated delegation we
think of the mean absolute value of policy divergence as the efficiency of representation,
then we can think of a measure that also captures the direction of this divergence as
capturing bias. In a competitive two-party system in which the parties do not converge
to the position of the median voter, there may over time be considerable inefficiency,
but potentially no significant bias. Pivotal parliamentary systems may have quite different
characteristics. Since the pivotal party (bloc) is typically fairly centrist, such systems
may be less susceptible than Westminster polities to the large inefficiencies that may
result from vesting agenda control in a cabinet with relatively extreme preferences. Yet,
the pivotal system may exhibit more bias, particularly if the pivotal party is non-centrist
on secondary dimensions of political competition.10 Constrained systems, whether
parliamentary or presidential, may also tend to exhibit more bias than those based on
accountability through competition, since constraining institutions (e.g. courts) may
have systematically different preferences from those of the median voter.
N O N - P O L I C Y M OT I VAT I O N S
So far I have implicitly assumed that politicians care only about policy outcomes.
Although there is good reason to begin our exploration here, this assumption cannot
give us a fully satisfactory account of democratic agency problems. We need to con-
sider other motivations that political agents may have. In doing so, I shall proceed more
informally than above, so as to discuss a variety of concerns in delegation, rather than
rigorously modelling the effects of each in isolation. Readers who are differently
inclined may take the discussion that follows as theoretical conjecture.
Non-policy motivations in politics may encompass a great swath of the human
condition, including such powerful forces as love, jealousy, compassion, vanity, spite,
and greed, as much of the classical literature has indeed recognized. Here, however,
I shall focus on a few motivations that are particularly common, generalizable, and
liable to cause agency problems, namely (1) lack of effort and (2) the pursuit of, and use
of public office for, private advantage.
10
In the Italian First Republic (1948–92), for example, the DC was consistently able to dominate the cab-
inet through its median position on the economic left–right axis. Because of DC dominance, however,
Italians also had to accept very traditional social policies (e.g. on divorce) and a much cosier relationship with
the Vatican than the median voter might have preferred.
84 Kaare Strøm
Leisure-shirking is certainly not unknown in politics. There are plenty of folk tales
about lazy bureaucrats, or about elected politicians that use their offices to enjoy the
‘good life’ ( junkets, lavish receptions, and possibly worse) rather than to serve their
constituents. Many delegation models seek to capture such distractions by recognizing
that political service is costly and that agents thus pay a cost (in leisure opportunities,
for example) for developing policy proposals that can be presented to their principals.
For delegation to work, it is thus necessary not only that the agent is able to make
proposals that serve the principal’s interests, but also that doing so is worth the agent’s
effort. In more technical terms, the agent’s incentive conditions have to be satisfied.
This is clearly a non-trivial consideration in real-world representational politics. In
many societies, there is widespread concern that the rewards from public service
simply do not suffice to motivate the ‘best and brightest’ to pay the large personal and
sometimes financial costs of political involvement. In this respect, however, the
distinctive features of parliamentary democracy would seem to count in its favour.
While its singularity of representation and the concentration of agenda power may
exacerbate agency losses, they simultaneously enhance the agents’ incentives to get
involved. Moreover, vigorous electoral competition may effectively sanction lazy politi-
cians, whereas judicial and other constraints are less likely to work (inactivity is rarely
illegal). Thus, Westminster parliamentarism should be particularly favoured.
Yet, the trade-off between leisure and service to the principal still does not
fully capture the range of concerns facing individuals in public office. Clearly, they
frequently have options other than either serving the public interest or slacking off.
One complication in democratic delegation is the presence of political rents. A rent,
according to James M. Buchanan, ‘is that part of the payment to an owner of resources
over and above that which those resources could command in any alternative use’
(Buchanan 1980: 4). Political rents, then, are whatever a public official can gain by virtue
of holding that office, over and above what he or she would otherwise receive from vol-
untary transactions. Political office-holders may, for example, be able to gain speaker’s
fees or honoraria, memberships in exclusive associations, campaign contributions, busi-
ness contracts, or even cold, hard cash, that they would not otherwise have received.
Such rents may be offered quasi-voluntarily by favour-seekers, or politicians may
extract them through various forms of extortion, such as threats of regulation or taxa-
tion (McChesney 1997). Any office that allows its occupant to make discretionary polit-
ical decisions with significant distributional consequences is likely to carry some value
in rents. The closer to the apex of executive decision-making, the greater the rents that
tend to attach to political offices. And the greater the rents, the greater
the temptations towards self-serving behaviour. Hence, the risk of agency problems.
If political agents are self-interested and respond to political and economic incentives,
then rent-seeking behaviour should covary positively with the magnitude of available
rents and negatively with the expected payoffs from alternative uses of their resources.
Hence, societies may be particularly susceptible to rent-seeking if they feature (1) a gen-
erous public sector (Goel and Nelson 1998), (2) few alternative opportunities for social
advancement, and (3) ineffective constraints on the ability of politicians to use their offices
for private gain. Sadly, this description fits certain resource-rich developing societies, such
as Nigeria, to a tee, and rent-seeking is by all accounts correspondingly rampant.
Parliamentary Democracy and Delegation 85
This may also explain why, as Susan Rose-Ackerman (1999: 1) notes, political corruption
is such a depressingly persistent feature of the world’s poorest societies, regardless of the
promises that may surround new office-holders (see also Montinola and Jackman 2002).
For those who wish to ‘better themselves’ economically, many poor countries simply
offer few alternatives to rent-seeking political activities. Corruption is a special form of
rent extraction, in which public offices are not only exploited for private advantage, but
in which public favours are also traded or conferred in ways that are contrary to the law.
Asymmetric information may foster corruption, and vice versa. Corrupt officials
have strong incentives to withhold information about their actions. Once an agent has
violated the law, or in other ways the trust of the principal, transparency can be fatal.
This tendency prevails not only because incumbents develop large asymmetries of
information, but also because becoming corrupt is a bit like losing one’s virginity: there
is no going back. Once a politician has succumbed to such a temptation, he is unlikely
to find the straight-and-narrow path enticing, as his career will continue to be at risk.
Since penalties for corruption are often insensitive to the number of violations the
agent has committed, each additional act of betrayal often carries a minimal marginal
cost. Hence, sadly, honest politicians may become corrupt over time, but corrupt ones
rarely become honest. This is probably the best argument in favour of term limits,
which otherwise detract from electoral accountability (see Carey 1996).
Do parliamentary systems foster rent-seeking political behaviour that may undermine
representational ideals? Clearly, although some parliamentary systems (e.g. Italy and
Belgium) have decidedly mixed reputations, most are not among the rent-seeking
basket cases of this world. One reason may be comparatively effective judicial checks. A
second reason may lie in the strength of civil society and independent mass media. An
even more critical factor may be the fact that most parliamentary countries have wealthy
and diverse economies that offer many opportunities for social betterment outside of pol-
itics. On the other hand, many parliamentary systems, particularly in Europe, have large
and well-financed public sectors that may attract rent seekers (see Boix 2001). The overall
picture for our countries is thus mixed, with one generally unfavourable and two
favourable conditions. Hence, the challenge that rent-seeking poses may vary significantly
from one parliamentary system to another, but should not be extreme in Western Europe.
Yet, it is ex ante typically more difficult to gauge a candidate’s rent-seeking propen-
sities than his or her policy preferences. This is both because candidates are more likely
to lie about the former, and because agency loss in the form of rent-seeking or cor-
ruption, rather than policy divergence, is more likely to result from temptation rather
than design. Hence, it is difficult to contain rent-seeking behaviour strictly through ex
ante controls, and parliamentary systems that are weak both in external constraints and
electoral competitiveness, such as pivotal ones, should be particularly vulnerable.
I N F O R M AT I O NA L P RO B L E M S
So far, I have considered only such challenges to parliamentary governance as stem
from the preferences and incentives of politicians. There is, however, a whole other
source of agency problems, namely information. Democratic representation can be
jeopardized not only when agents have interests and desires that go against those of
86 Kaare Strøm
their principals, but also when principals cannot easily figure out what their agents are
up to. Preference divergence between principal and agent is frequently accompanied
and exacerbated by asymmetric information favouring the agent, and it is typically the
interaction of divergent preferences and poor information that causes the really signific-
ant agency problems.
Agency problems under incomplete information may take the form of hidden
information (principals do not fully know the competencies or preferences of their
agents or the exact demands of the task at hand) and/or hidden action (principals cannot
fully observe the actions of their agents). The former condition can give rise to adverse
selection, the latter to moral hazard. The former problem may lead principals to select
the ‘wrong’ agents, who do not have the most appropriate skills or preferences. The
problem of moral hazard, on the other hand, arises when agents, once selected, have
incentives and opportunity to take unobservable action contrary to the principal’s interests.
Adverse Selection
The city where those who rule are least eager to do so will be the best governed.
(Plato)
Adverse selection is one generic problem facing democratic principals. In selecting
agents, principals look for candidates that have the necessary information and skills.
Separating qualified candidates from unqualified ones is no great problem if political
qualifications and predispositions are easily discernible. But the problem of many
delegation relationships is precisely that the principal does not know very much about
the task at hand or the aptitude that it requires. At the same time, and perhaps for the
same reasons, the qualifications of potential agents may be very difficult to judge.
Furthermore, potential agents must also have the requisite motivation. Simply put,
agents must be willing as well as able, and in order to attract such agents, principals
must offer adequate incentives. The trouble is that the incentives that principals offer
may sometimes systematically attract the least desirable agents.
As Lupia notes, the classical illustration of adverse selection derives from the insur-
ance business. An insurance company, in this case the principal, offers insurance to
motorists (the agents). To succeed, the insurance company has to set its rates suf-
ficiently high to recover the claims that will be made, but not so high that it fails to
attract any customers. Yet, some motorists are inherently less desirable customers than
others. Even though there is considerable randomness in traffic accidents, certain drivers
are much more likely to be involved than others, because they have poor skills, little
experience, bad driving habits, or risk-acceptant personalities. The problem is that the
insurance company is much less likely to be able to identify them than the bad drivers
themselves. Even though motorists may flatter themselves about their driving skills,
many problem drivers recognize their liabilities, but conceal them from the insurance
company as best as they can. At the same time, these customers will use their private
information to shop for insurance. If the insurance company is unable to discriminate
between good and bad drivers, the latter customers will crowd out the former.
Whatever level the insurance company sets its rates, it will attract motorists for whom
Parliamentary Democracy and Delegation 87
that is favourable. And the higher these premiums are set, the more systematically it
will adversely select only the very worst motorists.
Adverse selection has been recognized as a political problem at least since Plato, whose
observation is quoted above. Delegating power to politicians is risky because the indi-
viduals most attracted to politics may not be the most desirable rulers. Indeed, it may be
precisely the most power-hungry or greedy, and perhaps those most likely to abuse
authority, that are drawn to politics. Although most people would prefer to be ruled by
the likes of Cincinnatus rather than Macbeth, sadly the latter may be in greater supply.
Hence, many students of political representation see the selection of ‘good’ agents as
the primary democratic challenge. Brehm and Gates (1997: 202) find that in federal, state,
and local bureaucracies in the United States, ‘the problem of adverse selection trumps the
problem of moral hazard’. James Fearon (1999) similarly maintains that the most impor-
tant task for voters is to select ‘good types’ for office: ‘a candidate with similar policy pref-
erences, who is relatively honest and principled (hard to buy off ), and who is skilled’
(Fearon 1999: 68). It is not difficult to agree with Fearon’s desiderata, and deviation from
any of these ideals is indeed a serious real-world concern. Democracy is jeopardized if
aspiring politicians have atypical preferences, are disproportionately willing to sacrifice
honesty and policy principles, or are lacking in skills. The latter problem, one sometimes
hears, is that those people become politicians who lack the ability to succeed elsewhere.
Glenn R. Parker (1996) paints a disturbing picture of how adverse selection has
contributed to the moral and political decline of the US Congress. Over time, he
argues, the intrinsic rewards of congressional service, such as the value of producing
public goods, have declined relative to the opportunities for rent extraction.
As a result, those legislators who gain the most from intrinsic rewards (i.e. those with long tenure
in Congress, since intrinsic returns are a function of longevity in office) exit. As Congress
becomes populated by rent-seeking legislators, through both recruitment and attrition, institu-
tional rules and arrangements are manipulated to facilitate wealth-earning opportunities . . . If
adverse selection is afflicting our political institutions, the market for politicians may come to
resemble George Akerlof ’s characterization of the market for used cars in the United States—a
market for ‘lemons’. (Parker 1996: 48)11
The highly respected Nigerian writer Chinua Achebe likewise holds adverse selection
responsible for many of his country’s woes: ‘Nigeria . . . is a country where it would be
difficult to point to one important job being held by the most competent person we
have . . . We have displayed a consistent inclination since we assumed management of
our own affairs to opt for mediocrity and compromise, to pick a third and fourth eleven
to play for us. And the result: we have always failed and will always fail to make it to
the world league’. (Achebe 1983: 19–20. Emphasis in the original.)
Moral Hazard
Power tends to corrupt, and absolute power corrupts absolutely.
(Lord Acton (1834–1902))
11
In this context, a ‘lemon’ is a low-quality used car.
88 Kaare Strøm
Moral hazard is an equally basic problem in delegation. Whereas adverse selection
stems from the agent’s private information, moral hazard has to do with hidden action
on his part. That is to say, moral hazard arises when the principal cannot, after entering
a contract, fully observe the agent’s actions that may affect the outcome. Such informa-
tion asymmetry may jeopardize successful delegation as much as adverse selection
does. To return to the automobile insurance example, moral hazard refers to the
behaviour of drivers after they have obtained insurance. The problem is that although
the insurance company wants its policy holders to take all reasonable precautions while
driving, it cannot actually observe whether they do so. This problem is compounded by
the drivers’ incentives. Once insured, they no longer face the same financial risk. In the
event of an accident, the insurance company will pay (except, perhaps, for a minor
deductible). Even previously good drivers may be tempted to relax their precautions
and take greater risks, with obvious adverse consequences for the insurer.
Again, it is not difficult to see the political analogue. Once politicians have been
elected, and especially after they have been elevated to executive office, they face a myriad
of temptations. The most innocuous ones may be to use their authority to advance
their own political agenda, rather than the platform on which they were elected or their
constituents’ preferences. A more sinister possibility is that politicians use their offices
and influence covertly for personal gain or to benefit their families, friends, business
associates, fellow tribesmen, or other acquaintances.
The magnitude of moral hazard is likely to be systematically related to certain sig-
nificant parameters of delegation. First, the larger the potential political rents, the more
serious the problem of moral hazard. That is to say, the more attractive the spoils of
office are relative to alternative social rewards, the more we should worry about moral
hazard. Second, the greater discretion politicians are granted, and the more
generous the spoils under their control, the more severely their character will be tested.
Recall Lord Acton’s pithy formulation above. Third, the weaker the oversight mech-
anisms, the larger the threat of moral hazard. Finally, the longer incumbents stay in
office, the more serious the moral hazard problem may become, especially with respect
to corruption and other abuses of office.
Representative democracy clearly entails problems of moral hazard, which is
indeed the motivation for John Ferejohn’s (1986) model of electoral competition (see
also Barro 1973). Dominant models of electoral competition, Ferejohn notes, have
a ‘disturbing feature’, namely the possibility that ‘once in office, the politician’s preferences
may diverge from those of his constituents and that he may therefore choose policies at
variance from his platform’ (Ferejohn 1986: 5). Ferejohn’s concern is with the extent to
which electoral accountability can constrain such behaviour. He assumes the politician to
be ‘an agent of the electorate whose behaviour is imperfectly monitored...In other
words, the voter’s problem is to police moral hazard rather than to find and select the
more capable of benevolent officeholders’ (Ferejohn 1986: 11–2).
Agency problems stemming from incomplete information can thus prevent citizens
in contemporary democracies from getting fully satisfactory service from their political
agents. Let us now more specifically consider informational challenges in the parlia-
mentary chain of delegation. Recall that this policy process can be divided into at
Parliamentary Democracy and Delegation 89
least four distinct stages of delegation, namely those from voters to parliamentarians,
from parliamentarians to the cabinet and its chief executive (Prime Minister), from
the cabinet to individual ministers, and finally from individual cabinet members to civil
servants. Each stage has its informational perils.
Electing Parliamentarians
The simple act of voting enables a relatively small number of representatives to be chosen
from among the many that may feel the call. Yet, picking good agents is far from easy.
Candidates may well have preferences that differ from those of the voters, especially as
regards their leisure and political rents. As Manin, Przeworski, and Stokes (1999: 40) put it:
Politicians may want to pursue their own ideas even if these differ from those of citizens. Some
may care most about advancing their careers against fellow politicians, within the government or
the same party. Some may seek perks. Some may want to get rich at the expense of citizens, while
in office or after leaving it. Some may be most concerned about recognition by foreigners. In all
these cases politicians will want something whose pursuit is injurious to citizens.
Such candidate preferences are rarely transparent ex ante and candidates are under-
standably reluctant to reveal them. It is also often difficult to ascertain ex post whether
elected officials have acted upon them. Elections thus clearly feature the risk of adverse
selection as well as moral hazard, and theories of representation are centrally con-
cerned both with selecting the ‘most suitable’ representatives and with preventing
those elected from abusing their powers for ideological purposes or personal advan-
tage. As Mitchell (2000: 336) points out, ‘of all the links in the delegation chain from
individual voter to Prime Minister, the link from voter to MP may be fraught with the
greatest danger.’ Similarly, in commenting on agency problems between politicians and
bureaucrats, William A. Niskanen (1991: 15) observes that ‘[m]y judgment is that there
is a much more substantial agency problem between politicians and voters’.
The severity of agency problems between voters and representatives is due in part to
the institutions that regulate this stage of delegation. Electoral institutions are complex
and manipulable: the rules concerning ballot access, the rules that allocate seats among
the competing political parties or candidates, and in multimember districts the rules that
determine the selection of individual representatives within contending lists. Politicians
and political parties often engineer these rules for personal benefit or to render the
choices of the voters, for example their preferences over individual candidates, ineffectual.
Consider, for example, the practice of gerrymandering legislative district boundaries.
The informational challenges facing voters are also gigantic, since it is very difficult
to gather systematic data on all potential representatives. Finally, but perhaps most par-
ticularly, agency problems at this stage are due to the great asymmetries of information
between ‘amateur’ voters and full-time ‘professional’ politicians. Delegation is thus
driven by competence concerns as much as by the voters’ lack of capacity, and the
threat of agency loss is exacerbated by the interaction of these factors. In trying to
reduce their informational asymmetries, voters must rely primarily on the screening
efforts of political parties, independent interest groups, and mass media. The former
mechanism (parties) is likely to be more effective in parliamentary systems, the latter
90 Kaare Strøm
ones (interest groups and media) in presidential ones. The problem, of course, as in all
reliance on third parties, is that parties, interest groups, and media may have their own
political agendas.
Forming Cabinets
The next step in the chain of delegation is when coalitions of parliamentarians form to
select a Prime Minister and cabinet. The rules vary, so that in some countries parliament
delegates specifically to a Prime Minister who in turn selects a cabinet, whereas
elsewhere parliament selects the cabinet as a whole. Yet, it is very rare for the Prime
Minister not to be selected internally among the members of parliament, and indeed
among the leaders of the largest parliamentary parties. The same rules and expectations
apply, though often less strictly, to other cabinet members. In the Westminster model,
these are quite rigid requirements; in other systems they are at least conventions or
strong regularities.
The information problems at this stage of delegation are generally less formidable
than between voters and parliamentarians. Cabinet members are linked to parliamen-
tarians by copartisanship and are often personally familiar to them. Yet, although this
delegation relationship involves professional politicians with common backgrounds
and experiences, it may entail adverse selection as well as moral hazard. Regardless of
how familiar potential cabinet members are, they will have private interests not known
to their backbenchers. Once in office, they will generally also have opportunities to act
on such interests, even to the detriment of their respective parties. For example, some
office holders use the perquisites of office to enrich themselves or at least to support
a lifestyle that they would not otherwise have been able to enjoy. In less aggravated cases
(as when Mona Sahlin of the Swedish Social Democrats charged personal expenses to
her ministerial account), such behaviour may cause embarrassment and electoral
liability to copartisans. In more serious cases (as in the cases of Italian politicians such
as Bettino Craxi in the 1980s), it may drain the public coffers and/or undermine the
popular legitimacy of the regime. Hence, parliamentarians must look for institutional
and other solutions to such agency problems at the very core of government. While
some accountability mechanisms (confidence procedures) are constitutionally
entrenched, for the most part parliamentary democracies have to rely on the more
informal policing efforts of parliamentarians and political parties. As we have seen,
such accountability is often rendered ineffective by an incentive system that induces
parties to close ranks behind their leaders.
imply.12 In multiparty coalitional systems such as Italy, the Prime Minister’s control over
cabinet appointments and dismissals is thus often sharply curtailed. Cabinet decision
rules vary between hierarchical ones, in which the Prime Minister is entitled to define
the ‘sense of the meeting’, and collegial ones, in which issues are ultimately decided by
majority vote.
Andeweg (2000) points out that delegating authority to a team of cabinet members
involves its own peculiar challenges. In his words,
the process of delegation from government to ministers appears to differ fundamentally from dele-
gation processes elsewhere in this democratic chain. The complication is that, as we commonly
understand the term, the government, which delegates to heads of departments, also consists of
heads of departments (with exceptions for the Prime Minister and ministers without portfolio). In
other words, the principal is made up of its own agents (Andeweg 2000: 377).
As Andeweg suggests, the delegation relationship between Prime Minister and other
ministers is often not entirely clearly defined. Besides, the Prime Minister may have
good reason to worry about agency problems on the part of his line ministers. Even if
these ministers share the bond of copartisanship, their personal career incentives may
diverge sharply. Indeed, the Prime Minister may literally be the only thing that stands
between some of them and the realization of their ultimate career objectives. Hence,
a norm that disposes Prime Ministers to appoint to the cabinet their most prominent
and ambitious copartisans may entail particularly acute forms of adverse selection.
When cabinet members do not even share the common interest that membership in
the same political party provides, preference divergence within the cabinet may be even
greater. And the anecdotal literature on cabinet politics is full of stories about moral
hazard and treachery.
responsibility or role, distinct from that of the Minister. Subject to the conventions limiting the access of
Ministers to papers of previous Administrations, it is the duty of the civil servant to make available to the
Minister all the information and experience at his or her disposal which may have a bearing on the policy deci-
sions to which the Minister is committed or which he is preparing to make, and to give to the Minister hon-
est and impartial advice, without fear or favour, and whether the advice accords with the Minister’s view or
not’ (Armstrong 1989: 141–2).
Parliamentary Democracy and Delegation 93
impose penalties for lying or misbehaving, the more successfully she can contain
moral hazard.
Except for the last stage of delegation, between cabinet ministers and civil servants,
political parties are the main vehicles of accountability in parliamentary systems.
Parties are instruments by which preferences are aligned before authority is delegated
and they serve especially to protect against adverse selection. Institutional checks
(constraints or veto gates) are mechanisms by which public policy is constrained later in
the policy process, and they guard most effectively against moral hazard. If the dominant
agency problem is adverse selection, then, as Lupia points out in Chapter 2, ex ante
mechanisms of screening and selection have clear advantages. Consequently, the institutions
of Westminster (and other) parliamentary democracy have much to recommend them-
selves. If, on the other hand, moral hazard is the greater concern, there is a stronger case
for ex post control. If so, in turn, presidential constitutions are at a comparative advantage.
T H E B E AU T Y O F PA R L I A M E N TA R I S M : E F F I C I E N C Y
Let us now summarize the effects of parliamentarism on democratic agency problems.
The merits of any constitutional design must be judged partly on the basis of proce-
dural qualities, such as efficiency, coordination, transparency, and credibility.
Westminster parliamentarism scores particularly well on the first two of these dimen-
sions. It permits policy coordination and fosters efficiency. The distinctive features of
parliamentary delegation—simplicity, indirectness, and cabinet agenda control—gener-
ate a set of discernible advantages. One is efficiency, which refers to three salient prop-
erties of parliamentarism. First, parliamentarism favours policy efficiency in the sense that
agents face fewer institutional (external) checks and constraints (veto players). Thus, in
the vocabulary of Cox and McCubbins (2001), parliamentarism promotes decisiveness.
Second, since agents serve only one principal and are thus unlikely to face mutually con-
tradictory or ambiguous demands, parliamentarism fosters administrative efficiency.
Third, because it gives agents greater inducements to exert themselves on behalf of
their principals, parliamentarism implies incentive efficiency. The first of these three
sources of efficiency follows directly from our definition of parliamentary democracy.
Let us, however, briefly consider the other two advantages of parliamentary rule.
First, parliamentary democracy implies advantages in administrative efficiency. The
fact that each agent faces a single principal means that agents are less likely to face
conflicting demands than under presidentialist or other multiple-principal systems.
Agents are less likely to have their hands tied by contradictory requests from two or
several principals. Hence, parliamentary systems exhibit greater administrative
efficiency than presidential systems, an effect that may be most notable at the stage of
policy administration. These effects are consistent with the argument of Terry Moe
and Michael Caldwell (1994), who note that for reasons of greater administrative
simplicity, environmental and other legislation has been much more successful, and
environmental administration much more cost-efficient, under British parliamentarism
than under US presidentialism.
94 Kaare Strøm
Besides reducing constraint and gridlock, the simplicity of parliamentary demo-
cracy may induce greater effort among principals and agents alike. In addition to their
policy concerns, both principals and agents face trade-offs between work and leisure
(Brehm and Gates 1997). Differences in representative institutions affect the effort that
agents put into their tasks and principals into monitoring the agents (Aghion and Tirole
1997). Agents exert greater effort when they are given more latitude to choose the
means by which to realize their goals. As long as the agents’ preferences do not deviate
greatly from the public interest, this may be a good thing. For example, to the extent
that the task of civil servants is to find technically efficient solutions to coordination
problems, extensive oversight and constraint by multiple principals is probably
misguided. The relative scarcity of constraints and monitoring thus means that parlia-
mentary democracy is comparatively resistant to leisure-shirking.
Similarly, the institutional simplicity of parliamentary democracy may enhance the
principals’ incentive to exert effort in monitoring agents. If such oversight allows
the principal (e.g. the members of a particular parliamentary committee overseeing an
executive agency) to claim exclusive credit, and if their efforts to control their agents
are not thwarted by other ‘overseers’, then these principals should put more effort into
their constitutional responsibilities. Hence, parliamentary democracy is at an advantage,
particularly compared to collegial presidentialism, in providing incentives for principals
to monitor and for agents not to shirk.
These expectations are borne out by the case of Nigeria. When Nigeria returned to
civilian rule in 1979, after 13 years of military government, it changed from a parlia-
mentary to a presidential constitution that included numerous checks and balances.
Hence, the country offers a relative rare natural experiment in regime change. Whereas
the parliamentary First Republic (1960–6) allowed the Northern-dominated majority to
run roughshod over the opposition, the presidential Second Republic (1979–83) legisla-
ture quickly became embroiled in gridlock and ineffectiveness. What ensued, according
to Osaghae (1998: 114), was ‘rivalry between the two houses, leading to unnecessary
delay and controversies in the passing of bills’. To the chagrin of Nigerian citizens, their
politicians increasingly turned to ‘opportunistic and self-interested behaviour’ (Diamond
1988: 52), in other words, leisure-shirking (or worse).
In addition to these salutary effects of simplicity, parliamentarism benefits from the
role that political parties play as devices of preference alignment and screening (see
Müller 2000). Party cohesion under a Westminster system allows voters to make
reasonably well-informed choices and ensures a certain amount of responsiveness and
accountability in government. When governing parties are cohesive and reasonably
large (as under single-party majority government), they may also more easily be able to
pursue policies favouring encompassing rather than distributional interests (Olson
1982). Finally, as we have seen, parliamentary regimes may be better equipped to deal
with problems of adverse selection. To the extent that the main problem in politics is
to select the right ‘type’ of representative, the advantage should lie with regimes
that devote more resources to the prior screening of candidates, as is the nature of
parliamentarism.
Parliamentary Democracy and Delegation 95
T H E AC H I L L E S H E E L : T R A N S PA R E N C Y
Just as parliamentary democracy has identifiable advantages over its competition, so
too does it have its weak points. Simplicity has its costs as well as its benefits. One such
disadvantage stems from the singular and indirect chain of parliamentary accountability,
which increases susceptibility to policy divergence, as the model above demonstrates.
Recall that in this model, agency loss is decisively greater under parliamentarism than
under either form of presidentialism, and that such agency loss could be particularly
severe when the agenda setter is relatively extreme and the reversion point unattractive
to the principal.
Indeed, to the extent that the chain of delegation is longer and more indirect that this
simple model assumes, the agency loss may be even greater. As Barbara Geddes (1994)
argues, a singular chain of delegation is only as strong as its weakest link (see also
Lupia, Chapter 2). If legislators, for example, are poor agents of their voters, then it
may not at all be desirable to enhance civil servant accountability to them. This is
true at least as long as agents do not compete for the favour of their principals. In
other words, if principals can bypass a particular link in the chain of delegation,
then agency losses in that link are less critical. But what characterizes parliamentary
democracy is precisely that such competition does not occur and that principals therefore
cannot bypass their agents. Therefore, severe agency loss at any stage of delegation is
a particularly serious concern.
Broadly speaking, more complex systems, such as presidentialism or parliamentary
systems with more external constraints, favour credible and stable outcomes. In Cox
and McCubbins’ (2001) terms, they tend to be more resolute, which means for one thing
that agents are less likely to take action that will make their principals worse off than
they previously were. For the same reasons, presidential systems are likely to produce
greater policy stability, and they are more able credibly to commit themselves to
particular policy goals (such as fiscal rectitude).
Moreover, presidential systems in particular are more likely to generate transparency,
because they contain mechanisms by which agents are forced to share information and
principals can learn. The more principals can learn, of course, the more they can ulti-
mately trust their agents, and the less the likelihood of agency loss. Learning is partic-
ularly important in dynamic settings in which poorly informed principals facing
repetitive challenges attempt to acquire the requisite information from their agents.
The informational perspective on legislative organization identifies some of the precon-
ditions for effective political learning. For example, Gilligan and Krehbiel model the rela-
tionship between legislatures and their committee members as a signalling game, in
which committee members invest resources in specialization and information acquisi-
tion and then use their private information to advise the non-specialist members (the
floor) on the quality of pending proposals. The more heterogeneous the preferences of
the committee members, the more confidence floor members can have in committee
signals. Thus, learning improves with the diversity of the sources of information available
to the legislators (Gilligan and Krehbiel 1989; Krehbiel 1991).
96 Kaare Strøm
These lessons can be generalized to the agency relationships we are considering
here. Generally speaking, because they feature multiple agents with potentially diver-
gent preferences, presidential constitutions are more likely to feature the conditions
under which principals can rely on the signals they receive from their respective agents,
at least insofar as these signals converge. Under parliamentary systems, all else being
equal, principals (e.g. parliamentarians) are less likely to be able to rely on the infor-
mation fed to them by their agents, in part because these agents have homogeneous
preferences induced by copartisanship.
Moreover, the informational advantages of presidentialism are reinforced by the fact
that more of the policy bargaining between different constitutional actors takes place
in the public domain, rather than behind closed doors. In a presidential system, policy
bargaining typically takes the form of proposals and counter-proposals that are shuttled
back and forth between different chambers or branches of government. In contrast,
parliamentary bargaining tends to take place behind close doors in cabinet or in coalition
committees or party summits. Thus, political bargaining is displaced from a public to
a private arena, where it will be less informative to the citizens. The more bargaining is
confined to this sphere of ‘invisible politics’ (Sartori 1976), to private negotiations
within political parties, or within government coalitions, the less transparent the policy
process, and the more uncertainty voters may have about their representatives. Thus,
transparency is a general liability of Westminster systems, and the lack of transparency
may in itself exacerbate the problems of transparency mentioned above.
Thus, comparing the foreign policy of Japan to that of the United States, Peter
F. Cowhey notes that
there are . . . no built-in checks on reversals of policy promises. This poses problems when trying
to limit foreign or domestic demands for politically risky policies. So, the LDP invented a mixture
of institutional constraints and self-binding political pledges to manage the political and diplo-
matic bargaining risks of foreign policy. At the same time, prolonged LDP rule permitted exten-
sive delegation to the bureaucracy. This delegation created a significant veil over the Japanese
policy process that reduced the transparency and credibility of foreign policy promises (Cowhey
1995: 209).
T H E P E R I L S O F C O N S T I T U T I O NA L D E S I G N
Policy-making in parliamentary democracies can be understood as a chain of delegation
that begins with the voters and ends with the civil servants (or other agents) that
implement government decisions. In Chapter 2, Lupia introduced a class of analytical
models, principal–agent models, which can illuminate such delegation relationships
and generate precise understandings of their effects. In this chapter, I have sought to
integrate these insights in a way that helps us understand contemporary parliamentary
constitutions.
All political systems imply particular delegation regimes, and any delegation runs the
risk of agency loss. But democracies are not necessarily created equal in this sense.
Thus, I have shown that parliamentary democracies are better protected against some
such problems, such as adverse selection, than against others, such as moral hazard.
Yet, the contrast between parliamentary and presidential constitutions is not simply
one between a regime designed to combat adverse selection and one focused on moral
hazard. There are other values that are tied up in these institutional ‘packages’. Thus,
the ideal–typical parliamentary constitution is one that promotes efficiency by
eliminating redundancies and taking incentives seriously. On the other hand, presidential
constitutions promote credibility and transparency by protecting against hasty and
potentially ill-considered policy change and by forcing disclosure of the pull and tug of
policy-making in the public arena.
The broader lessons are that agency problems are stubborn and that they come in
different forms. Democratic delegation is necessary, consequential, and at the same
time fraught with danger. Political parties and institutional checks exist in large part for
these reasons. In turn, and for better or worse, they affect the policy process in their
own ways and sometimes in complex interaction. There is no simple institutional fix for
all agency problems. Thus, neither parliamentary nor presidential constitutions can
effectively safeguard against dishonest or incompetent public servants. As long as
principals and agents differ in their preferences and information, some agency losses
must be expected, but such agency losses may come in different forms under different
constitutions. And the choices between them represent real trade-offs, as summarized
in Table 3.3, which is based partly on the models presented above, partly on the less
formalized arguments that I have presented.
Yet, parliamentary systems are by no means all alike, and any serious assessment of
their merits and demerits clearly requires more detailed and configurative description
and analysis. In order to understand the effects of institutions such as parliamentarism
on the policy process, our account thus has to begin and end with that process. We
need to map out and understand the democratic policy process all the way from voters
to civil servants. If we, for example, think of deviations from the Westminster model
in terms of veto players, it is important to realize that a proliferation of institutional veto
players (external constraints) has very different consequences from a fragmentation of
98 Kaare Strøm
Table 3.3. Regime Form and Agency Loss: A Summary
the party system (partisan veto players). An increase in the former should under many
circumstances enhance ex post control and thus counteract problems of moral hazard
especially. A fragmentation of the party system, on the other hand, may very well lead
to weaker ex ante controls and consequently to an increased susceptibility to adverse
selection problems and other agency problems. Needless to say, if the goal is to contain
agency loss, the former effect is much more desirable than the latter.
To understand parliamentarism, we must thus read its fine print, which is the
purpose of the next eighteen chapters of this book, which collectively constitute its
bulk. These chapters will take us through the four main links in the chain of delega-
tion, as these steps have been outlined here and elsewhere. The data will be presented
in summary cross-national form and further analysed in Chapter 4. The following sev-
enteen chapters provide more configurative portraits of the respective national chains of
demo-cratic delegation. Our emphasis will be on identifying mechanisms of delegation
and accountability more than agency loss. Thus, these chapters will lay out the national
policy process and describe the constraints on policy-making and the boundaries within
which it is appropriate to characterize the constitutional actors as agents (directly or
indirectly) of the citizens. Each chapter focuses specifically on the role of parties
as bonding instruments used to align incentives and permit citizen control of the
policy process. In sum, these data provide an unprecedented guide to contemporary
European parliamentary democracies, as well as to the larger lessons and broader
prospects of democratic representation, at the dawn of the twenty-first century.
APPENDIX
This appendix describes, for each of the three games depicted in Fig. 3.2 and each of the ten pref-
erence configurations, the largest (worst-case) and smallest (best-case) agency loss, relative to the
Parliamentary Democracy and Delegation 99
principal’s ideal point. In each scenario, I assume for simplicity that the status quo and the reversion
point are identical and equal to zero. While there may in many cases be a range of conditions
associated with either the best-case or worst-case outcome, only one example of each is offered.
All references below to distance refer to the absolute policy distance between the ideal points of
the respective players. Recall that all models assume complete information, that it is costly to
make any proposal, and that if otherwise indifferent, the principal will opt for the reversion point.
Therefore, the agency loss will never be greater than P (in case no proposal is made), or P ⫺
(in case a proposal is made and accepted).
Competitive Presidentialism
The best-case scenario, with no agency loss, occurs when P is minimal (equal to ). Then agent
2 will propose the principal’s ideal point, and she will accept. The worst-case scenario is when the
principal is closer to agent 2 than to the reversion point. Agent 2 will then propose his ideal point
(A2), and the principal will accept. The agency loss will be P ⫺ .
Collegial Presidentialism
The best-case scenario is when the distance between the principal and the reversion point (P) is
minimal (equal to ). Then either agent will have to propose the principal’s ideal point, which
will be accepted. The worst-case scenario occurs when the distance between the principal and
agent 2 is just smaller than the distance between the principal and the reversion point (or equal to
P ⫺ ). Then agent 2 will propose his own ideal point (A1), which will be accepted. The agency
loss will be P ⫺ .
Competitive Presidentialism
The agents’ proposals will converge to the ideal point of the principal, and there will be no
agency loss.
Collegial Presidentialism
The best-case scenario, with no agency loss, is when the distance between the principal and agent
2 is smaller than that between agent 2 and the reversion point. Then both agents will make
100 Kaare Strøm
acceptable proposals that will converge to the principal’s ideal point. The worst-case scenario is
when the distance between agent 2 and the reversion point is minimal (equal to ). Then the
agents’ proposals will converge to the ideal point of agent 2, which will be accepted. The agency
loss will be P ⫺ .
Competitive Presidentialism
The agents’ proposals will converge to the ideal point of the principal, and there will be no
agency loss.
Collegial Presidentialism
Since agent 2 will not make or accept any proposal greater than 0, and the principal
will reject any proposal smaller than 0, no proposal will be made. Hence, the best-case (and
worst-case) scenario is the reversion point, under which agency loss is P.
Competitive Presidentialism
Since agent 1 prefers the reversion point to any proposal that would be acceptable to the princi-
pal, any acceptable proposal will come from agent 2, who will indeed propose his ideal point. The
best-case scenario is when agent 2 is at a minimal distance, , from the principal. The worst-case
scenario is when agent 2 is similarly close (distance equal to ) to the reversion point. The agency
loss will then be P ⫺ .
Collegial Presidentialism
Since agent 1 will make or accept no proposal smaller than 0, and the principal will reject any
proposal greater than 0, no proposal will be made. Hence, the best-case (and worst-case) scenario
is the reversion point, under which agency loss is P.
Competitive Presidentialism
The agents’ proposals will converge to the ideal point of the principal, and there will be no
agency loss.
Parliamentary Democracy and Delegation 101
Collegial Presidentialism
Since agent 1 will make or accept no proposal smaller than 0, and the principal will reject any
proposal greater than 0, no proposal will be made. Hence, the best-case (and worst-case) scenario
is the reversion point, under which agency loss is P.
Competitive Presidentialism
As under parliamentarism, the precise ideal point of agent 2 is immaterial. The best-case scenario
is when the distance between the principal and the reversion point (P) is minimal (equal to ).
Then either agent will propose the principal’s ideal point, which will be accepted. The worst-case
scenario occurs when the distance between the principal and agent 1 is just smaller than the dis-
tance between the principal and the reversion point (or equal to P ⫺ ). Then agent 1 will pro-
pose his own ideal point (A1), which will be accepted. The agency loss will be P ⫺ .
Collegial Presidentialism
Agent 2 will approve any proposal acceptable to agent 1 and the principal, but will never make a
more favourable proposal than agent 1 does. The best-case scenario is when the distance between
the principal and the reversion point (P) is minimal (equal to ). Then agent 1 will have to pro-
pose the principal’s ideal point, and agent 2 and the principal will accept. The worst-case scenario
occurs when the distance between the principal and agent 1 is just smaller than the distance
between the principal and the reversion point (or equal to P – ). Then agent 1 will propose his
own ideal point (A1), which will be accepted. The agency loss will be P ⫺ .
Competitive Presidentialism
The agents’ proposals will converge to the ideal point of the principal, and there will
be no agency loss.
Collegial Presidentialism
The best-case scenario, with no agency loss, is when the distance between the principal and agent
1 is just smaller (by ) than that between agent 1 and the reversion point. Then the principal will
delegate to agent 2, who will propose P, and agent 1 and the principal will accept. The worst-case
102 Kaare Strøm
scenario is when the distance between agent 1 and the reversion point is minimal (equal to ).
Then either agent will propose agent 1’s ideal point (A1), which will be accepted. The agency loss
will be P ⫺ .
Competitive Presidentialism
Both agents will propose, and the principal will accept the larger proposal, which will be agent
2’s proposal of his own ideal point. In the best-case scenario, the distance between the principal
Parliamentary Democracy and Delegation 103
and agent 2 is minimal and the agency loss equal to . In the worst-case scenario, the agency loss
will be P ⫺ 2.
Collegial Presidentialism
Agent 2 will always make a proposal that agent 1 and the principal prefer to the reversion point.
Agent 1 will never make a more favouable proposal than agent 2. The best-case scenario is when
the distance between the principal and agent 2 is minimal and when agent 1 is closer to agent 2
than to the reversion point. Then agent 2 will propose his ideal point and agent 1 and the princi-
pal accept. The agency loss is . In the worst-case scenario, agent 1 is at a minimum distance, ,
from the reversion point. Agent 2 will then have to propose agent 1’s ideal point, and the agency
loss will be P ⫺ .
REFERENCES
Achebe, Chinua (1983). The Trouble with Nigeria. London: Heinemann.
Aghion, Philippe and Tirole, Jean (1997). ‘Formal and Real Authority in Organizations’. Journal of
Political Economy, 105: 1–29.
Aldrich, John (1995). Why Parties?: the Origin and Transformation of Political Parties in America.
Chicago: Chicago University Press.
Andeweg, Rudy B. (2000). ‘Ministers as Double Agents? The Delegation Process Between
Cabinet and Ministers’. European Journal of Political Research, 37: 377–95.
—— and Nijzink, Lia (1995). ‘Beyond the Two-Body Image: Relations Between Ministers and
MPs’, in Herbert Döring (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt am
Main: Campus and New York: St. Martin’s Press.
Armstrong, Sir Robert (1989). ‘The Duties and Responsibilities of Civil Servants in Relation to
Ministers’, in Geoffrey Marshall (ed.), Ministerial Responsibility. Oxford: Oxford University Press.
Arrow, Kenneth J. (1951). Social Choice and Individual Values. New York: John Wiley.
Bagehot, Walter (1963[1867]). The English Constitution. London: Fontana/Collins.
Barro, Robert J. (1973). ‘The Control of Politicians: An Economic Model’. Public Choice, 14: 19–42.
Bendor, Jonathan, Moe, Terry M., and Kenneth W. Shotts (2001). ‘Recycling the Garbage Can: An
Assessment of the Research Program’. American Political Science Review, 95(1): 169–90.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
Boix, Carles (1999). ‘Setting the Rules of the Game: The Choice of Electoral Systems in
Advanced Democracies’. American Political Science Review, 93: 609–24.
—— (2001). ‘Democracy, Development, and the Public Sector’. American Journal of Political
Science, 45: 1–17.
Brehm, John and Gates, Scott (1997). Working, Shirking and Sabotage. Ann Arbor: University of
Michigan Press.
Buchanan, James M. (1980). ‘Rent Seeking and Profit Seeking’, in James M. Buchanan, Robert
D. Tollison, and Gordon Tullock (eds.), Toward a Theory of the Rent-Seeking Society. College
Station: Texas A&M Press.
Carey, John M. (1996). Term Limits and Legislative Representation. Cambridge: Cambridge
University Press.
Cohen, Michael, March, James G., and Olsen, Johan P. (1972). ‘A Garbage Can Model of
Organizational Choice’. Administrative Science Quarterly, 17: 1–25.
Cowhey, Peter F. (1995). ‘The Politics of Foreign Policy in Japan and the United States’, in Peter
F. Cowhey and Mathew D. McCubbins (eds.), Structure and Policy in Japan and the United States.
Cambridge: Cambridge University Press.
104 Kaare Strøm
Cox, Gary W. (1987). The Efficient Secret. Cambridge: Cambridge University Press.
—— and McCubbins, Mathew D. (1993). Legislative Leviathan. Cambridge: Cambridge University
Press.
—— —— (2001). ‘The Institutional Determinants of Economic Policy Outcomes’, in Stephan
Haggard and Mathew D. McCubbins (eds.), Presidents, Parliaments, and Policy. Cambridge:
Cambridge University Press.
Dahl, Robert A. (1970). After the Revolution? New Haven: Yale University Press.
Diamond, Larry (1988). ‘Nigeria: Pluralism, Statism, and the Struggle for Democracy’, in Larry
Diamond, Juan J. Linz, and Seymour Martin Lipset (eds.), Democracy in Developing Countries.
Boulder: Lynne Rienner.
Dicey, A. C. (1959). An Introduction to the Study of the Law of the Constitution, 10th edn. London:
Macmillan. (First edition 1885.)
Epstein, David, and O’Halloran, Sharyn (1999). Delegating Powers. Cambridge: Cambridge
University Press.
Esaiasson, Peter and Holmberg, Sören (1996). Representation from Above: Members of Parliament and
Representative Democracy in Sweden. Aldershot: Dartmouth.
Fearon, James D. (1999). ‘Electoral Accountability and the Control of Politicians: Selecting Good
Types versus Sanctioning Poor Performance’, in Adam Przeworski, Susan C. Stokes, and
Bernard Manin (eds.), Democracy, Accountability, and Representation. Cambridge: Cambridge
University Press.
Ferejohn, John (1986). ‘Incumbent Performance and Electoral Control’. Public Choice,
50: 5–25.
Geddes, Barbara (1994). Politician’s Dilemma: Building State Capacity in Latin America. Berkeley:
University of California Press.
Gilligan, Thomas and Krehbiel, Keith (1989). ‘Asymmetric Information and Legislative Rules
with a Heterogeneous Committee’. American Journal of Political Science, 33: 459–90.
Goel, Rajeev, and Nelson, Michael A. (1998). ‘Corruption and Government Size: A Disaggregated
Analysis’. Public Choice, 97: 107–20.
Hardin, Garrett (1968). ‘The Tragedy of the Commons’. Science, 162: 1243–8.
Hennessy, Peter (1986). Cabinet. Oxford: Basil Blackwell.
Hernes, Gudmund and Nergaard, Kristine (1989). Oss imellom. Oslo: FAFO.
Huber, John D. (1996). Rationalizing Parliament. Cambridge: Cambridge University Press.
Kiewiet, D. Roderick and McCubbins, Mathew D. (1991). The Logic of Delegation. Chicago:
University of Chicago Press.
Krehbiel, Keith (1991). Information and Legislative Organization. Ann Arbor: University of
Michigan Press.
Lane, Jan-Erik, and Narud, Hanne M. (1992). ‘Regjering og nasjonalforsamling: Bemanning av de
to statsorganene’. Norsk Statsvitenskapelig Tidsskrift, 8: 303–16.
Laver, Michael J. and Shepsle, Kenneth A. (1996). Making and Breaking Governments. Cambridge:
Cambridge University Press.
Lijphart, Arend (1977). Democracy in Plural Societies. New Haven: Yale University Press.
—— (1984). Democracies. New Haven: Yale University Press.
—— (1999). Patterns of Democracy. New Haven: Yale University Press.
Linz, Juan J. (1994). ‘Presidential or Parliamentary Democracy: Does It Make a Difference?’ in
Linz, Juan J. and Valenzuela, Arturo (eds.), The Failure of Presidential Democracy. Baltimore: The
Johns Hopkins University Press.
Lipset, Seymour M. and Rokkan, Stein (eds.) (1967). Party Systems and Voter Alignments. New York:
Free Press.
Parliamentary Democracy and Delegation 105
Lupia, Arthur and McCubbins, Mathew D. (1994). ‘Learning from Oversight: Fire Alarms and
Police Patrols Reconstructed’. Journal of Law, Economics, and Organization, 10: 96–125.
—— and Strøm, Kaare (1995). ‘Coalition Termination and the Strategic Timing of Parliamentary
Elections’. American Political Science Review, 89: 648–65.
———— (2006). ‘Coalition Governance Theory: Bargaining, Electoral Connections and the
Shadow of the Future’, in Kaare Strøm, Wolfgang C. Müller, and Torbjörn Bergman (eds.), Cabinets
and Coalition Bargaining: the Democratic Life Cycle in Western Europe. Oxford: Oxford University Press.
McChesney, Fred S. (1997). Money for Nothing: Politicians, Rent Extraction, and Political Extortion.
Cambridge, MA: Harvard University Press.
McCubbins, Mathew D. and Schwartz, Thomas (1984). ‘Congressional Oversight Overlooked:
Police Patrols versus Fire Alarms’. American Journal of Political Science, 28: 165–79.
McKelvey, Richard D. (1976). ‘Intransitivities in Multidimensional Voting Models and Some
Implications for Agenda Control’. Journal of Economic Theory, 12: 472–82.
Manin, Bernard, Przeworski, Adam, and Stokes, Susan C. (1999). ‘Elections and Representation’,
in Adam Przeworski, Susan C. Stokes, and Bernard Manin (eds.), Democracy, Accountability, and
Representation. Cambridge: Cambridge University Press.
Mattson, Ingvar and Strøm, Kaare (1995). ‘Parliamentary Committees’, in Herbert Döring (ed.),
Parliaments and Majority Rule in Western Europe. New York: St. Martin’s Press.
Mitchell, Paul L. (2000). ‘Voters and their Representatives: Electoral Institutions and Delegation
in Parliamentary Democracies’. European Journal of Political Research, 37: 335–51.
Moe, Terry M. and Caldwell, Michael (1994). ‘The Institutional Foundations of Democratic
Government: A Comparison of Presidential and Parliamentary Systems’. Journal of Institutional
and Theoretical Economics, 150: 171–95.
Montinola, Gabriella R. and Jackman, Robert W. (2002). ‘Sources of Corruption: A Cross-Country
Study’. British Journal of Political Science, 32: 147–70.
Müller, Wolfgang C. (2000). ‘Political Parties in Parliamentary Democracies: Making Delegation
and Accountability Work’. European Journal of Political Research, 37: 309–33.
—— and Strøm, Kaare (eds.) (2000). Coalition Governments in Western Europe. Oxford: Oxford
University Press.
Niemi, Richard G. and Weisberg, Herbert F. (1968). ‘A Mathematical Solution for the Probability
of the Paradox of Voting’. Behavioral Science, 13: 317–23.
Niskanen, William A. (1991). ‘A Reflection on Bureaucracy and Representative Government’, in
André Blais and Stéphane Dion (eds.), The Budget-Maximizing Bureaucrat: Appraisals and
Evidence. Pittsburgh: University of Pittsburgh Press.
Norton, Philip (1994). The British Polity, 3rd edn. New York: Longman.
Olson, Mancur, Jr. (1965). The Logic of Collective Action. Cambridge, MA: Harvard University
Press.
—— (1982). The Rise and Decline of Nations. New Haven: Yale University Press.
Osaghae, Eghosa E. (1998). Crippled Giant: Nigeria since Independence. Bloomington: Indiana
University Press.
Palmer, Matthew S. R. (1995). ‘Toward an Economics of Comparative Political Organization:
Examining Ministerial Responsibility’. Journal of Law, Economics and Organization, 11: 164–88.
Parker, Glenn R. (1996). Congress and the Rent-Seeking Society. Ann Arbor: University of Michigan Press.
Rasch, Bjørn Erik (2000). Demokrati: Ideer og Organisering. Bergen: Fagbokforlaget.
Riker, William H. (1982). Liberalism Against Populism: A Confrontation Between the Theory of
Democracy and the Theory of Social Choice. San Francisco: W. H. Freeman.
Riker, William H. and Ordeshook, Peter C. (1973). Introduction to Positive Political Theory.
Englewood Cliffs: Prentice Hall.
106 Kaare Strøm
Rokkan, Stein. (1970). Citizens, Elections, Parties. Oslo: Universitetsforlaget.
Romer, Thomas and Rosenthal, Howard (1978). ‘Political Resource Allocation, Controlled
Agendas, and the Status Quo’. Public Choice, 33: 27–44.
Rose, Richard and Mackie, Thomas T. (1983). ‘Incumbency in Government: Asset or Liability?’,
in Hans Daalder and Peter Mair (eds.), Western European Party Systems: Continuity & Change.
London: Sage.
Rose-Ackerman, Susan (1999). Corruption and Government: Causes, Consequences, and Reform.
Cambridge: Cambridge University Press.
Saalfeld, Thomas (2000). ‘Members of Parliament and Governments in Western Europe: Agency
Relations and Problems of Oversight’. European Journal of Political Research, 37: 353–76.
Sartori, Giovanni (1976). Parties and Party Systems: A Framework for Analysis. Cambridge: Cambridge
University Press.
—— (1994). ‘Neither Presidentialism nor Parliamentarism’, in Juan J. Linz and Arturo Valenzuela
(eds.), The Failure of Presidential Democracy: Comparative Perspectives. Baltimore: The Johns
Hopkins University Press.
Shepsle, Kenneth A. and Bonchek, Mark S. (1997). Analyzing Politics. New York: W. W. Norton.
—— and Weingast, Barry R. (1987). ‘The Institutional Foundations of Committee Power’.
American Political Science Review, 81: 85–104.
Shugart, Matthew S. and Carey, John M. (1992). Presidents and Assemblies. Cambridge: Cambridge
University Press.
Sjöblom, Gunnar (1987). ‘The Role of Political Parties in Denmark and Sweden, 1970–1984’, in
Richard S. Katz (ed.), Party Governments: European and American Experiences. Berlin: de Gruyter.
Strøm, Kaare (1990). Minority Government and Majority Rule. Cambridge: Cambridge University Press.
—— and Müller, Wolfgang C. (1999). ‘The Keys to Togetherness: Coalition Agreements in
Parliamentary Democracies’. Journal of Legislative Studies, 5: 255–82.
—— and Swindle, Stephen M. (2002). ‘Strategic Parliamentary Dissolution’. American Political
Science Review, 96: 575–91.
Thies, Michael F. (2000). ‘On the Primacy of Party in Government: Why Legislative Parties Can
Survive Party Decline in the Electorate’, in Russell J. Dalton and Martin P. Wattenberg (eds.),
Parties without Partisans. Oxford: Oxford University Press.
Tsebelis, George (1990). Nested Games: Rational Choice in Comparative Politics. Berkeley: University
of California Press.
——(1995). ‘Decisionmaking in Political Systems: Veto Players in Presidentialism, Parliamentarism,
Multicameralism, and Multipartyism’. British Journal of Political Science, 25: 289–325.
—— (2000). ‘Veto Players and Institutional Analysis’. Governance, 13: 441–74.
—— (2002). Veto Players: An Introduction to Institutional Analysis. Princeton: Princeton University
Press and the Russell Sage Foundation.
Weaver, R. Kent and Rockman, Bert A. (eds.) (1993). Do Institutions Matter? Government
Capabilities in the United States and Abroad. Washington, DC: The Brookings Institution.
Wessels, Bernhard (1999). ‘Whom to respresent? Role orientations of legislators in Europe’ in
Hermann Schmitt and Jacques Thomassen (eds.), Political Representation and Legitimacy in the
European Union. Oxford: Oxford University Press.
PART II
S u r vey
This page intentionally left blank
4
I N T RO D U C T I O N
Contemporary democracies necessitate political delegation from the ultimate principal,
the citizens, to all sorts of agents. Chapter 3 identifies three particularly important
motivations for delegation, namely problems of (a) capacity (since not all citizens have
the time or resources to make all important decisions for themselves), (b) competence
(which is to say that they also do not have the skills to do so), and (c) social choice and
collective action (which is to say that the members of any large group have trouble
reaching and enforcing mutually beneficial agreements among themselves). While
the last category may appear somewhat academically arcane, in truth it applies to
decision-making in any community. So do, of course, the former two problems,
although these seem to have become particularly acute in contemporary advanced
democracies, with their technical advancement, expansive governments, and exceedingly
complex policy challenges.
As we have also learned in the preceding chapters, delegation poses a myriad of
potential problems. The first three chapters of this book have explained how preference
divergence and asymmetric information between citizens and politicians can cause
agency problems. Preference divergence can stem from conflicting visions about policy
but also from leisure- or rent-seeking motivations on the part of politicians. Information
problems materialize in two critical forms: adverse selection and moral hazard, which
have been discussed in Chapters 2 and 3.
To deal with such problems, democratic constitution makers have devised various
mechanisms of political accountability. As Arthur Lupia points out in Chapter 2, there
are at least two ways to think about accountability: as policy outcomes or processes of
control. In the ‘outcome’ perspective, accountability means that the policy outcomes
meet the principal’s preferences (whether or not the agent’s intentions or hard work
were in any way responsible). Viewed in terms of control, on the other hand, the relevant
Most of the credit for this chapter goes to our country experts who have painstakingly collected the inform-
ation provided in the tables. In addition to those that appear as authors in this volume, we want to thank
Indridi H. Indridason and Josep M. Reniu for their invaluable help.
110 Torbjörn Bergman et al.
question for accountability is whether principals can effectively direct the behaviour of
their agents (whether or not such direction leads to desired outcomes).
Our concern in this and the following chapters, with the exception of the final one
(Chapter 23), will largely be with accountability as a procedure, rather than an outcome.
As discussed in Chapter 3, accountability implies two kinds of rights that principals
have vis-à-vis agents: a right to demand information, and a capacity to impose any or
all of three types of sanctions: (a) blocking or amending decisions made by the agent
(veto power), (b) deauthorizing the agent (remove him from office or curtail his author-
ity), and (c) imposing specific (monetary or other) penalties on the agent.
To determine when and how to impose sanctions, principals rely on various mechan-
isms of accountability. An accountability mechanism is any device by which a principal
(a) can get information about an agent’s intentions, skills, and behaviour or (b) can
sanction or reward the agent. The information can come in oral, written, digital, pictorial,
or any other form imaginable. Sanctions can also take many forms, and may include
positive rewards such as new tasks, perks, or inducements. Of course, accountability
mechanisms may be more or less effective. As information devices, at best they tell the
principal what she needs to know. At worst, they provide faulty information about the
agent’s preferences (the adverse selection problem) or behaviour (the moral hazard
problem). As sanctions, at best they ensure that agents behave according to their principals’
preferences. At worst, badly designed accountability mechanisms can reward or sanction
ineffectively, or, for that matter, so severely that they constrain agents from taking
action that the principal would have desired.
Accountability mechanisms vary in the comprehensiveness of the principal’s sanctions
and right to information, as well as in their exclusiveness. Accountability mechanisms in
which the principal’s rights are comprehensive and exclusive are classically hierarchical.
Under multiple principals with less comprehensive rights towards the agent, we speak
instead of plurarchy. As mentioned in Chapter 3, classical parliamentarism tends to be
hierarchical, whereas presidentialism is typically more plurarchical.
Delegation and accountability are two complementary and interdependent aspects
of the democratic policy process. We shall refer to this larger policy process as demo-
cratic governance. We distinguish between government and governance. Government is
a set of institutions for public policy-making. Governance is a process that takes place
in part in those institutions but also requires a linkage between principals (citizens) and
political agents. This chapter examines national governance in European parliamentary
democracies. These decisions typically involve a great number of actors interacting
formally and informally. Yet, in this chapter and indeed in this volume we are
concerned primarily with top-level national governance.
Despite a number of excellent national or cross-national studies of individual
mechanisms of delegation and accountability, and even though Western Europe
includes some of the best-researched countries in the world, we lack a study that
applies a unified perspective to the entire national process of democratic delegation and
accountability. Our point of departure is the ideal–typical parliamentary democracy
defined in Chapter 3. With the theoretical framework from previous chapters to
guide us, we can identify cross-national differences as well as discrepancies between
Democratic Delegation and Accountability 111
1
In the text and tables we sometimes use such common acronyms as PM (Prime Minister) and MP
(Member of Parliament).
112 Torbjörn Bergman et al.
RU L E S O F G OV E R NA N C E
The rules of parliamentary governance are not particularly secret. Yet, identifying the
relevant national rules is not entirely straightforward. Certainly, any serious account
must begin with the formal constitution, but it cannot stop there. Some constitutional
rules are in practice simply obsolete. Many Western European democracies used to be
monarchies in which the heads of state were much more than figureheads, from
which constitutional provisions have survived that currently have little meaning and rel-
evance. Hence, we largely ignore the formal rules concerning hereditary monarchs
unless these rules shape the expectations of other important political players (such
as party leaders or PMs). We also realize that actors not recognized by constitutions
may be of greater relevance to the policy process than are some constitutional actors.
The prime example is political parties, which are scarcely mentioned in many older
constitutions, yet dominate much of the public policy process in Western Europe
(Müller 2000a).
Constitutional provisions do not always serve the purpose for which they were
designed. For one thing, it may be notoriously difficult to pinpoint the ‘real’ intention
behind a certain institutional design (Müller 2002). Furthermore, gifted politicians—
masters in the ‘art of manipulation’ (Riker 1986)—can use institutions in ways not
intended or foreseen by their creators. For example, the secret vote in parliament (as
used in Italy from the nineteenth century until 1988) originally was invented to protect
MPs from external pressure and to allow them to vote according to their conscience.
Yet, in the age of party politics it had two kinds of unintended consequences: (a) moral
hazard for MPs (who could vote against their constituency and/or party without
repercussions) and (b) less predictable parliamentary outcomes. And the unsuccessful
employment of the German constructive vote of no-confidence in 1972 (also by secret
ballot) showed that under certain conditions the secret vote can be an instrument
by which party leaders control their MPs—exactly the opposite of the original intent
(see below).
Strong conventions, which generate expectations and can entail costs for those who
violate them, can matter more than written rules. Despite great similarities in formal
rules, the Presidents in France and Austria (or for that matter Iceland) perform quite
different roles, with the former being the country’s leader and the latter essentially a
figurehead (see Duverger 1980; Elgie 1999 and below). Hence, we agree with Bogdanor
(1988: 5) that it is more important to look at the ‘working constitution’ than at the
written one.
For these reasons, our empirical investigation here goes much beyond the simple
collection of seventeen (or, not counting the United Kingdom, sixteen) national
constitutions to encompass written rules of different sorts, including ordinary legislation
and, for example, parliamentary standing orders. We sometimes also include
constitutional practices (strong conventions) that the actors involved understand to be
de facto rules that it would be costly to disregard. But since our interest is in identify-
ing rules that have some capacity to induce compliance, there is a real and significant prob-
lem in adopting too broad a notion of what constitute rules of governance. We do not
Democratic Delegation and Accountability 113
M A K I N G LAW S A N D C O N S T I T U T I O N S
Some institutions are more fundamental than others. The rule of law, which most
Western European democracies generally respect, implies that most important political
decisions require legislative authorization. And within the body of law, constitutional
rules are typically privileged. Constitutions (and constitutional practice) provide the
basic ‘contract design’ for delegation between the main democratic principal (‘the citi-
zens’) and its agents. Among other things, they specify (a) the extent of the power del-
egated to these agents and (b) how the people can hold them accountable.
Constitutions also regulate (c) the relationships between different types of agents, for
instance, between parliament and cabinet. Below we describe for each country the pro-
cedures for ordinary legislation as well as constitutional amendments, and we account
for the organization of the legislature.
Making Laws
Consider first the making of ordinary law. Table 4.1 presents the actors that together
with the lower (or only) chamber are constitutionally designated participants in ordinary
Table 4.1. Decision-making Procedure: Actors and Rules in Ordinary National Legislation
Country Head Cabinet Lower chamber Decision-making procedure in lower Upper chamber Other constitutional Which actors
and year of state (Actor 2) (Actor 3) chamber (Actor 4) actor (Actor 5) co-decide? (Refers
(Actor 1) to actors in
Quorum Decision Voting previous columns)
requirement rule procedure
Austria
1945–2000 Formal Can propose Discretionary power 1/3 of all MPs Simple Anonymous Suspensive vetoa n.a. n.a.
majority
Belgium
1945–2000 Formal Can propose Co-decides Absolute Simple Recorded Co-decides n.a. 3, 4
majority majority
Denmark
1945–52 Formal Co-decides Co-decides Absolute Simple Recorded Co-decides n.a. 2–4
majority majority
1953–2000 Formal Co-decides Co-decides Absolute Simple Recorded n.a. n.a. 2, 3
majority majority
Finland
1945–2000 Suspensive Can propose Discretionary power No Simple Recorded n.a. n.a. 1–3
vetob majority
France
1959–2000 Can refer to Can propose Discretionary power No Simple Recorded Can propose n.a. n.a.
constitutional majority
courtc
Germany
1949–2000 Formal Can propose Discretionary power Absolute Simple Recorded Co-decidesd n.a. n.a.
majority majority
Greece
1975–2000 Suspensive Can propose Discretionary power No Simple Recorded n.a. n.a. n.a.
vetoe majority
Iceland
1945–2000 Can refer to Can propose Discretionary power Absolute Simple Recorded n.a. n.a. n.a.
referendumf majority majority
Ireland
1945–2000 Can refer to Can propose Discretionary power 20 MPs Simple Recorded Can propose n.a. n.a.
Supreme majority
Courtg
Italy
1948–87 Suspensive Can propose Co-decides Absolute Simple Secret Co-decides n.a. 3, 4
vetoh majority majority
1988–2000 Suspensive Can propose Co-decides Absolute Simple Anonymous Co-decides n.a. 3, 4
vetoh majority majority
Luxembourg
1945–2000 Formal Can propose Discretionary power Absolute Simple Recorded n.a. Council of Statei n.a.
majority majority
The Netherlands
1945–2000 Formal Co-decides Co-decides Absolute Simple Recorded j Veto n.a. 2–4
majority majority
Norway
1945–2000 Formal Can propose Discretionary power Absolute Simple Recorded j n.a. n.a. n.a.
majority majority
Portugal
1976–2000 Vetok Can propose Discretionary power Absolute Simple Recorded n.a. Constitutional 1, 3, 5
majority majority Tribunall
Spain
1978–2000 Formal Can propose Co-decides Absolute Simple Anonymous Co-decides Autonomous 3, 4
majority majority Communities and
popular initiativesm
Sweden
1945–70 Formal Co-decides Co-decides No Simple Recorded Co-decides n.a. 2–4
majority
1971–4 No role Can propose Discretionary power No Simple Recorded n.a. n.a. 2, 3
majority
Table 4.1. (Contd.)
Country Head Cabinet Lower chamber Decision-making procedure in lower Upper chamber Other constitutional Which actors
and year of state (Actor 2) (Actor 3) chamber (Actor 4) actor (Actor 5) co-decide? (Refers
(Actor 1) to actors in
Quorum Decision Voting previous columns)
requirement rule procedure
1975–2000 No role Can propose Discretionary power No Simple Recorded n.a. n.a. n.a.
majority
United Kingdom
1945–2000 Formal Can propose Discretionary power 40 MPs Simple Recorded Suspensive n.a. n.a.
majority veton
Notes: The table reports the default voting method. In cases where there is a choice between two or more options and no clear default, we report the most restrictive one. The entry ‘Can pro-
pose’ means that the actor has proposal rights but that this proposal is subject to discretionary approval by other actor(s). The entry ‘Co-decides’ means that two or more actors have to agree
on a proposal or it fails. The entry ‘n.a.’ means not applicable in this case.
An absolute majority is defined as 50% ⫹ 1 of all MPs. Simple majority is defined as 50% ⫹ 1 of the MPs present and voting. Most countries count a simple majority of ‘votes cast’, but in Belgium,
Denmark, Italy, Luxembourg, Portugal, and Sweden, the counting rule is based on a ‘simple majority of valid votes’. The latter counting rule excludes blank votes.
a
The Bundesrat has suspensive veto power over legislation, that is, it can issue a veto, but the Nationalrat can decide the issue by holding a second vote. In this case the decision rule is simple
majority. The quorum requirement in the Nationalrat is 50% of all MPs.
b
The Finnish President has a suspensive veto. Until 1987, the President could delay legislation until a new general election had been held and the new parliamentary majority voted against his
veto. From 1987, a President can delay a law for 3 months, after which point the parliament can override the veto in a simple majority vote.
c
The Prime Minister, the President of the National Assembly, the President of the Senate, and, since 1974, any 60 deputies or senators, hold the same right.
d
Slightly more than half of the laws, including virtually all important ones, require the consent of both ‘chambers’. With regard to the remaining laws, the Bundesrat can delay a bill’s passage
by a suspensive veto. Such an objection can be overturned by an absolute majority of the Members of the Bundestag. If the Bundesrat rejects a regular bill by a two-thirds majority, its veto can
only be overturned in a vote in which at least two-thirds of the Bundestag MPs participate and at least 50% ⫹ 1 of the Bundestag MPs reject the veto. The budget bill can only be introduced by
the cabinet. The Bundesrat may express an opinion, but cannot veto the federal budget. The Bundesrat may express an opinion, but cannot veto the federal budget.
e
The Parliament can override the veto with absolute majority of all MPs.
f
If the President in Iceland vetoes a new law, the law is subjected to a referendum in which the people can reject it in a simple majority vote.
g
The Irish President can either sign a new law or refer it to the Supreme Court to test its constitutionality.
h
The Parliament can override the veto.
i
The advice of the Council of State is compulsory (but non-binding) for every bill, proposal, and amendment. The Council of State can decide whether or not legislation must be put to a sec-
ond vote (Article 59). The Constitution stipulates, in effect, that all bills must pass a second vote in the Parliament no earlier than three months after the first vote. This second vote can only be
avoided if plenary sessions of the Parliament and Council of State decide that it is unwarranted. The Council of State thus has a suspensive veto, which in practice is rarely used. The Council
of State can also initiate (non-binding) proposals for the amendment of legislation.
j
In the large majority of cases Dutch MPs vote by the show of hands and voting is recorded by party rather than individual MP. Yet, any individual MP can demand a roll-call vote.
k
The Parliament can override the veto with absolute majority of all MPs with regard to some policy areas but only with two-thirds majority with regard to others. The number of policy areas
requiring a two-thirds majority was increased in the 1982 constitutional reform.
l
From 1976 to 1982, the Council of Revolution had the power to veto laws it considered unconstitutional. The 1992 constitutional revision abolished the council, and replaced it with the
Constitutional Tribunal, which was granted the power to veto laws on account of their unconstitutionality.
m
Bills can be proposed by the Autonomous Communities or through popular initiatives (500,000 signatures are required).
n
The House of Lords has the power to delay enactment of a bill (suspensive veto) for a year (de facto for 8 months). Until 1949 this was 2 years.
Democratic Delegation and Accountability 117
law-making, and the role of such constitutional actors as the cabinet and the head of
state. The respective roles of these actors can range from none to full discretionary
power.2
2
Note, however, that Table 4.1 does not include advisory bodies that provide non-mandatory or non-binding
reviews of legislative proposals before they enter into law. Nor do we include constitutional courts or other
forms of judicial review that take place mainly after laws have gone into effect.
3
Austria, Italy, and Spain feature the people’s initiative, that is, the right for a predetermined number of
citizens to propose legislation.
4
The Swedish upper chamber, which over time came to be a bastion of the Left, became a notable exception.
118 Torbjörn Bergman et al.
legislation. These three countries instead use anonymous votes, which means that the
votes of individual MPs may be observed but that they cannot be found in the official
parliamentary record.
Upper chambers make the chain of delegation more complex (Table 4.2). The rela-
tionship between lower and upper chambers is often complex, not least in Germany
where the Bundesrat represents the governments of the federal Länder.5 In bicameral
systems, the cabinet is generally accountable to the lower (or first) chamber (such as the
House of Commons). In the symmetrical legislatures of Belgium (until 1995), Italy, and
Sweden (until 1971), the cabinet is accountable to both chambers, but even here the
lower chamber is considered pre-eminent in national law-making (see also Tsebelis and
Money 1997; Patterson and Mughan 1999; Russell 2000; Baldwin and Shell 2001).
The main law-making influence of upper chambers stems from their authority to
block or substantially delay legislation. Such veto power comes in several shades. The
strongest veto power is where legislation cannot be enacted unless positively approved
by the upper chamber (the consent requirement). A somewhat lesser version is the
mere veto power, in which case a bill is enacted unless vetoed by the upper chamber.
This distinction mirrors the one between positive and negative parliamentarism
(cf. Bergman 1993). Some constitutions feature both the consent requirement and the
veto power, but apply them to different policy areas. To keep things simple, we have in
Table 4.1 coded ‘veto power’ whenever the upper chamber actively or passively can
block legislation. In some cases, this power is considerably restricted, as in Austria,
where only laws altering the respective jurisdictions of the Länder and the federal
government require the consent of the upper chamber (the Bundesrat). Even if they
lack veto powers, upper chambers may be able to delay legislation, through a suspensive
veto power or through procedural rules that guarantee the upper house a substantial
period for consideration. Since in politics time is often scarce, even the power to delay
decisions can be valuable. The one upper chamber that has only delaying power, the
British House of Lords, can de facto stall legislation for up to 8 months. Hence, a
government in the last 8 months of a parliamentary term depends on the cooperation
of the Lords. In a few cases, even the 8-week delay that the Austrian Bundesrat can
inflict has proven effective in preventing last-minute government legislation.
Whenever the upper chamber’s consent is necessary or it has permanent (as opposed
to suspensive) veto power, it is de facto a critical actor. The Italian Senate has exactly
the same legislative authority as the Chamber of Deputies. The other upper chamber
that is involved in a broad range of legislation, the Dutch Eerste Kamer, cannot initiate
5
It is debatable whether the German Bundesrat is appropriately labelled an ‘Upper Chamber’. The
Bundesrat represents the governments of the German federal states (Länder). Most laws are made jointly by
the Bundestag and the Bundesrat. In addition, the Bundesrat can freely veto national legislation (subject to
Bundestag override), and any national constitutional amendment requires a two-thirds Bundesrat majority
(see below). It also has the power to initiate bills at the national level and has done so with increasing fre-
quency. Moreover, Bundesrat members have the right to speak in the Bundestag and often do so on national
matters. For example, before the 1998 election SPD leaders Gerhard Schröder and Oskar Lafontaine used
their privileges as Bundesrat Members to challenge Chancellor Helmut Kohl in the Bundestag.
Democratic Delegation and Accountability 119
Austria
1945–84 No No No No No No Yes
1984–2000 No No No Yesb No Yesb Yesc
Belgium
1945–95 No Yes Yes n.a. Yes n.a. Yes
1995–2000 No No No Yesd No Yesd Yes
Denmark
1945–53 No No Yese Yes Yese Yes Yes
1953–2000 n.a. — — — — — —
Finland n.a. — — — — — —
Francef No No No Yesg No Yesg Yes
Germany No No No Yesh No Yesh Yes
Greece n.a. — — — — — —
Iceland n.a. — — — — — —
Ireland No No Noi Noj Noi Noj Yes
Italy Yes Yes Yes n.a. Yes n.a. Yes
Luxembourg n.a. — — — — — —
The Netherlands No No Yes n.a. Yes n.a. Yes
Norway n.a. — — — — — —
Portugal n.a. — — — — — —
Spain No No No Yes No No Yese
Sweden
1945–70 No Yesk Yes n.a. Yes n.a. Yes
1971–2000 n.a. — — — — — —
United Kingdom No No No No No No Yesl
a
This includes constitutional amendments, but excludes money bills.
b
Constitutional amendments that introduce shifts of competences between the federal level and the Land level require the
consent of the second chamber (the Bundesrat).
c
About 8 weeks.
d
The Senate has equal rights with the House for initiating and deciding on constitutional amendments and other institu-
tional reforms, regionalist matters, ratification of treaties, and the organization of the judiciary. The Senate is exclusively
competent to resolve conflicts between the national and subnational legislatures (Article 77).
e
Consent not required for the budget which is not considered in this table.
f
The French Senate has the same legislative powers as the Asemblée nationale until the Cabinet intervenes that can demand
a final decision from the Assemblée nationale. The Assemblée nationale can stick to its original decision, accept a compromise
drafted by the conference committee, or take on board amendments by the Senate (Article 45).
g
Laws interpreting the Constitution (lois auxquelles la Constitution, institutional acts of Parliament) require the consent of
both chambers if they are concerned with the Senate. If the Senate is not affected such laws can be accepted by the absolute
majority of all members in the Assemblée nationale (Article 46).
h
Slightly more than half of the laws require the consent of both chambers.
i
With the support of one-third of the Dàil members the Irish Seanad is entitled to appeal to the President of the Republic
and demand a referendum if the Dàil passes a bill of ‘national importance’ which had been rejected by the Seanad (Art. 27).
The President then decides on the holding of such a referendum. Then the bill can not become law until either it passes the
referendum or the Dàil passes it again following a general election. In practice Article 27 has never been invoked.
j
Even constitutional amendments can be passed without consent of the Seanad (Art. 46).
k
The Lower Chamber was considered more important, but both counted for cabinet formation, stability, and termination.
l
Two years until 1949, one year since then. Due to the counting rules the delay is de facto 8 months since the time is cal-
culated from the introduction of bills in the House of Commons.
Sources: Country chapters; Constitutions; Arter (1991); Tsebelis and Rasch (1995: 369–70, 372–3); Patterson and Mughan
(1999); Russell (2000).
120 Torbjörn Bergman et al.
or amend legislation, and is thus much more of a veto player, strictly defined (see the
country chapters for details). Although the Dutch upper chamber has comprehensive
jurisdiction, the German Bundesrat, whose consent is required for slightly more than
half of all legislative acts, is in practice much more influential. Its role in law-making
when its partisan majority is different from that of the Bundestag has given rise to the
thesis that Germany is a ‘grand coalition state’ (Schmidt 1996). In the other consent
requirement cases, relating to Austria, Belgium, France, Ireland, and Spain, the upper
chamber must consent on a few select issues only (see the footnotes to Table 4.2 for
details). As Table 4.2 demonstrates, the overall trend is that upper houses have become
less important or even abolished (see Longley and Olson 1991). For instance, until its
1995 constitutional reform, Belgium had a very influential upper chamber.6 Denmark
and Sweden abolished their upper chambers in 1953 and 1971, respectively.
Constitutions
Constitutions are political meta-rules that distinguish themselves not only by their
genesis, but also by their political entrenchment. Nonetheless, they vary a great deal in
their comprehensiveness and specificity. In accordance with the literature (e.g. Huber
et al. 2001: 336–8), we take the number of words to indicate the density of regulation
and hence as a proxy for the volume of specially protected rules and rights. If we focus
6
The Belgian constitutional reform of 1995 removed important powers (investiture and no-confidence
motions, budget matters, and the right to conduct interpellations of ministers) from the upper chamber.
Democratic Delegation and Accountability 121
only on the main constitutional documents, which may not be the only document with
constitutional status, their length (in the most recent English translation) varies
between 3,600 (Iceland) and 32,600 words (Sweden). While some constitutions
(Austria, Germany, and Greece) contain more than 20,000 words, others have fewer
than 10,000 (Luxembourg, Denmark, France, and Norway). The mean length of con-
stitutions is around 15,800 words and the median 13,300 words (see also Table 4.4).
Compared to ordinary law-making, more restrictive rules are used for constitutional
amendments. In practice, constitutional amendments can be difficult to pass even when
the formal requirements are few. Even in the United Kingdom, where passing the func-
tional equivalent of a constitutional amendment closely follows the procedure for ordinary
legislation, an important institutional change (such as a reform of the upper chamber) is
normally preceded by years of deliberation and debate. Most countries, however, protect
the constitution from amendment through one or more special requirements.
Table 4.3 describes the rules for constitutional amendments. In most countries the
procedure is much the same as for ordinary legislation. There are two major excep-
tions. One is that decision rules differ, and the second is that in many countries there are
intervening constraints. The most common decision rule is a two-thirds majority, used in
seven countries (Austria, Belgium, Germany, Luxembourg, The Netherlands, Norway,
and Portugal). Italy employs an absolute majority rule (i.e. more than half of all MPs
must vote in favour). Three Nordic countries (Denmark, Iceland, Sweden), Ireland, and
the United Kingdom use the simple majority rule (with some variation in counting
invalid or blank votes). In the remaining four countries, the decision rule is contingent
upon such matters as the size of the majority in the upper chamber or the authorship
of the original proposal (see the footnotes for Table 4.3 for details).
The most common constraint, used in seven countries, is the requirement of an
intervening election between two parliamentary votes. In France and Ireland, constitu-
tional amendments are instead put to a referendum. Denmark probably has the most
restrictive amendment rules, which require both an election and a referendum. The
remaining countries impose no mandatory intervening constraint, but many allow a
minority of MPs to subject the amendment to a national referendum. Countries that
allow a change of the constitution after only one vote, Austria, Finland, Portugal, and
Spain, use a qualified majority threshold to restrict amendments (these rules are detailed
in Table 4.3). In the Finnish case, though, an even larger majority (five-sixths) must first
agree to use such a procedure (otherwise there must be an intervening election).
Note that in Finland and France, the President can play a significant role in amend-
ing the Constitution. In Finland, it is the President who decides, in a plenary sitting of
the Council of State and at the suggestion of the Cabinet, whether the government will
present any legislative proposal to the Parliament.7 This has given strong Presidents
considerable agenda control over all types of legislation. In France, the President can
initiate a constitutional amendment by submitting a proposal to the Parliament. After
7
Note that a major Finnish constitutional reform, which went into effect in March 2000, significantly
altered the President’s powers. See Chapters 9 and 22 for more details.
Table 4.3. Decision-making Procedure: Constitutional Amendments
Country and Head of Cabinet Lower Decision-making procedure in lower chamber Upper chamber Other Which actors Mandatory
year State (Actor 2) chamber (Actor 4) constitutional co-decide intervening
(Actor 1) (Actor 3) Quorum Voting actor (refers to actors constraints
requirement Decision rule procedure (Actor 5) in previous
columns)
Austria
1945–2000 Formal Can propose Discretionary Absolute 2/3 majority Anonymous Can propose Popular n.a. Nob
power majority initiativea
Belgium
1945–2000 Formal Can propose Co-decides Absolute 2/3 majority Recorded Co-decides n.a. 3, 4 Electionc
majority
Denmark
1945–52 Formal Co-decides Co-decides Absolute Simple Recorded Co-decides n.a. 2–4 Election ⫹
majority majority referendumd
1953–2000 Formal Co-decides Co-decides Absolute Simple Recorded n.a. n.a. 2, 3 Election ⫹
majority majority referendumd
Finland
1945–2000 Suspensive Can propose Discretionary No 2/3 majorityf Recorded n.a. n.a. 1–3 No
vetoe power
France
1959–2000 Can Formal Co-decides No Simple Recorded Co-decides n.a. 3, 4 Referendumh
proposeg majority
Germany
1949–2000 Formal Formal Co-decides Absolute 2/3 majority Recorded Co-decides n.a. n.a. No
majority
Greece
1975–2000 None None Discretionary No 3/5 majorityi Recorded n.a. n.a. n.a. Election
power
Iceland
1945–2000 Formal None Discretionary Absolute Simple Recorded n.a. n.a. n.a. Election j
power majority majority
Ireland
1945–2000 Formal Can propose Discretionary 20 MPs Simple Recorded Formal n.a. n.a. Referendumk
power majority
Italy
1948–2000 Formal Can propose Co-decides Absolute Absolute Anonymous Co-decides n.a. 3, 4 Nol
majority majority
Luxembourg
1945–2000 Formal Co-decides Co-decides 3/4 of all MPs 2/3 majority Recorded n.a. Council n.a. Electionn
of Statem
The Netherlands
1945–2000 Formal Co-decides Co-decides Absolute 2/3 majority Recordedd Veto p n.a. 2–4 Electionq
majority
Norway
1945–2000 Formal Can propose Discretionary Absolute 2/3 majority Recorded n.a. n.a. n.a. Electionr
power majority
Portugal
1976–2000 None None Discretionary No 2/3 majority Recorded n.a. n.a. n.a. No
power
Spain
1978–2000 Formal Can propose Co-decides Absolute 2/3 majoritys Recorded Co-decides n.a. 3, 4 No
majority
Sweden
1945–70 Formal Co-decides Co-decides No Simple Recorded Co-decides n.a. 2–4 Electiont
majority
1971–4 Formal Co-decides Co-decides No Simple Recorded n.a. n.a. 2–4 Electiont
majority
1975–2000 None Can propose Discretionary No Simple Recorded Co-decides n.a. 2–4 Electiont
power majority
Table 4.3. (Contd.)
Country and Head of Cabinet Lower Decision-making procedure in lower chamber Upper chamber Other Which actors Mandatory
year State (Actor 2) chamber (Actor 4) constitutional co-decide intervening
(Actor 1) (Actor 3) Quorum Voting actor (refers to actors constraints
requirement Decision rule procedure (Actor 5) in previous
columns)
United Kingdom
1945–2000 Formal Can propose Discretionary 40 MPs Simple Recorded Suspensive n.a. n.a. No
power majority vetou
Notes: The table reports the default voting method. In cases there is a choice between two or more options and no clear default, we report the most restrictive one. The entry ‘Can propose’ means
that the actor has proposal rights but that this proposal is subject to discretionary approval by other actor(s). The entry ‘Co-decides’ entails that two or more actors have to agree on a proposal or it fails.
An absolute majority is defined as 50%⫹1 of all MPs. Simple majority is defined as 50%⫹1 of the MPs present and voting. Most countries count a simple majority of ‘votes cast’, but in Denmark and Sweden
the counting rule is based on a ‘simple majority of valid votes’. The latter counting rule excludes blank votes.
a
100,000 voters can initiate a constitutional amendment.
b
A ‘total’ revision of the Constitution requires acceptance by referendum. In other amendments a one-third minority of the MPs can demand such a referendum.
c
After a constitutional amendment, Parliament is dissolved, elections are held, and the new Parliament makes the final decision.
d
The Parliament passes the bill and a new election is called. The new Parliament must accept the bill without amendments and then the bill is submitted to a referendum within 6 months.
e
The Finnish President has a suspensive veto. Until 1987, the President could delay legislation until a new general election had been held and the new parliamentary majority voted against his veto. From 1987,
a President can delay a law for 3 months, after which point the Parliament can override the veto in a simple majority vote.
f
The Constitution can be changed by two identical decisions with an intervening election. The first decision (before the election) is a simple majority decision. The decision after election must be carried by a
two-thirds majority. Alternatively, a two-thirds majority can amend the Constitution in one decision. However, this is only possible if the Parliament first decides to use this procedure by a five-sixths majority.
g
The French President has the authority to initiate amendments to the Constitution by submitting a proposal to the Parliament. After an amendment has been accepted in both chambers of the National
Assembly by two-thirds majority of the MPs present, he can submit it to a joint session of Parliament (National Assembly and Senate) at which it is finally accepted if it is supported by a three-fifths majority of
all voting MPs. A referendum is not needed for the amendment to become law under this procedure.
h
After being proposed and adopted in the legislature, a constitutional amendment that does not emanate from the President is decided upon in a referendum.
i
A constitutional amendment must be passed by two successive Parliaments with an intervening election. An absolute majority (151 MPs) is required for one of the decisions and three-fifth majority (180 MPs)
in the other, irrespectively of order. Before the election, Parliament decides the issue either by a majority of three-fifths (180 MPs out of 300) or by absolute majority of all MPs (151). After the election, the
majority required for adoption depends on the size of the majority in favour in the proposing Parliament.
j
Constitutional amendments must be passed twice in identical form, with an intervening parliamentary election.
k
Lower chamber passes a Bill—an Act to Amend the Constitution, which decides the wording of the referendum question to be put to the people, who decide the issue. Referendums are decided by a simple
majority of valid votes cast.
l
Constitutional amendments must be passed twice with at least 3 months intervening. When an amendment has been approved by a majority below two-thirds of all members, it can be submitted to popular
referendum, if one-fifth of the members of either house or 500,000 voters or 5 regional councils so demand.
m
The advice of the Council of State is compulsory (but non-binding) for every bill, proposal, and amendment. The Council of State can decide whether or not legislation must be put to a second vote (Article
59). All bills must pass a second vote in the Parliament no earlier than three months after the first vote. This second vote can only be avoided if plenary sessions of the Parliament and Council of State decide
that it is unwarranted. The Council of State thus has a suspensive veto, which in practice is rarely used. The Council of State can also initiate and call for the modification of legislation.
n
After a constitutional amendment, a declaration is issued by the Parliament; that Parliament is dissolved and the new Parliament makes the final decision after the election.
o
In constitutional amendments the decision rule is two-thirds majority in the upper chamber.
p
In the large majority of cases Dutch MPs vote by the show of hands and voting is recorded by party rather than individual MP. Yet, any individual MP can demand a roll-call vote.
q
Constitutional amendments are proposed by law; these need simple majority. Then, Parliament must be dissolved and the newly elected Parliament must accept the amendments with a two-thirds majority.
r
Proposals for constitutional amendments are first adopted by simple majority. After the next election, the newly elected Parliament must accept the amendment with a two-thirds majority (of all votes cast).
s
Constitutional amendments must be approved by a three-fifths majority of each Chamber or, alternatively, by an absolute majority of the Senate and a two-thirds vote in Congress, with a referendum being
called upon request of one-tenth of the members of either chamber. A complete revision of the Constitution or amendments affecting basic principles, fundamental rights, or the Crown, require a two-thirds
majority of each chamber and ratification by referendum.
t
The Riksdag has to make two identical decisions, one before a general election to the lower chamber and one after such an election. From 1980, in addition, a minority of one-third can decide that
the constitutional amendment should be submitted to a referendum at the next general election.
u
The House of Lords has the power to delay enactment of a bill (suspensive veto) for a year (de facto for 8 months). Until 1949 this was 2 years.
Democratic Delegation and Accountability 125
8
Lijphart (1999: 218–23) generates four categories of supermajority requirements: (a) greater than two-
thirds majority, (b) two-thirds, (c) between two-thirds and ordinary, and (d) ordinary majority. To this catego-
rization, Lijphart makes two adjustments. To control for plurality electoral systems, which facilitate
supermajorities, Lijphart deducts one rank from their majority requirement score. And to adjust for the
effects of strong bicameralism, specifically the lack of intercameral congruence, Lijphart moves such coun-
tries up one category (in our sample, this is relevant only for Germany).
9
Lijphart’s classification leads to the following placements of our seventeen countries, from high to low
rigidity: (a) Germany; (b) Austria, Belgium, Finland, Luxembourg, The Netherlands, Norway, Portugal, and
Spain; (c) Denmark, France (since 1974), Greece, Ireland, Italy, and Sweden (since 1980); (d) France (before
1974), Iceland, Sweden (before 1980), and the United Kingdom.
10
Some of the values that Lutz’s rules assign to Western European countries are contestable. For instance,
Denmark has a score of 2.75 whereas Ireland’s is 3. Yet, we would consider constitutional amendments more
demanding in Denmark than in Ireland (see below).
126 Torbjörn Bergman et al.
Table 4.4. Mandatory and Minimum Constraints on Constitutional Amendments, 2000a
Country Mandatory and minimum constraints on Index of Number of
constitutional amendments relative words
difficulty of in main
Referendum Intervening Qualified Two decisions amendment constitutional
election parliamentary in Parliament process text
majority
and Ireland) change the rules of the game without express popular consent on final deci-
sion. In contrast, both of the latter countries’ constitutions by implication require a ref-
erendum on any changes to the EU treaties. And in both countries such amendments
have in fact been rejected when they were first put before the people: in Denmark in 1992
(the Maastricht Treaty) and in 2000 (the euro currency) and in Ireland in 2001 (the Nice
Treaty), resulting in considerable embarrassment to the national governments and the
European Union.
Democratic Delegation and Accountability 127
Lutz notes an interesting and strong relationship between the rules and the actual
frequency of constitutional amendments—the more demanding the referendum
process, the fewer the amendments. As it turns out, there is also a clear negative correla-
tion between the scope of the constitutionally protected area (as measured by the
number of words in the Constitutions) and the relative difficulty of constitutional
amendment. In other words, in Western Europe either the constitutionally protected
area is large but its protection relatively weak (Portugal or Sweden), or the scope is
small but the protection strong (Italy or Norway). (Yet, although we lack quantitative
data for Britain, it would clearly score low on both dimensions.) Indeed, the alternatives
would not be very appealing. Combining a narrow protected area with permissive
amendment rules invites majority tyranny while combining a large protected area with
restrictive amendment rules may cause political sclerosis. Thus, political systems may
achieve similar entrenchments of fundamental rights and procedures through different
means. Focusing narrowly on a single indicator of constitutional rigidity could there-
fore be misleading.
PA RT Y C O H E S I O N
Before examining in greater depth the chain of delegation in European parliamentary
democracies, let us consider the organizations that keep the different links in that chain
together. In parliamentary democracies, political parties traditionally perform that
function. Chapter 3 has argued that their strength (cohesiveness) is determined by their
value to politicians as mechanisms of bonding and to voters as informational eco-
nomizing devices. Party cohesion is the degree to which the public office-holders of any
one party act mutually consistently. As Fig. 4.1 shows, party cohesion results from the
individual cost–benefit calculations of the relevant intra-party actors. If many of them
find it beneficial to go against the party line, cohesion will be low. These indi-
vidual calculations, in turn, are based on the politicians’ preferences, which presumably
include both policy- and career-related goals. If these individual preferences within the
Preferences Institutions
Party State
rules rules
Policy-related
Intra-party actors’ Level of
Career-related cost–benefit calculations party cohesion
Party system
conditions
party by and large do not conflict, party cohesion will be high even if no discipline is
imposed. Yet, often intra-party preferences are not so well aligned. Under such condi-
tions parties need to be ‘strong’ in order to be cohesive, as Krehbiel (1993) has pointed
out. A ‘strong’ party can cause its members to follow the party line even if that con-
flicts with their preferences. And given the negative electoral effects of a lack of cohe-
sion, parties have a collective interest in being strong.
But how can parties achieve such conformity? It is party rules and their interplay
with government rules (particularly those that govern the first two steps of the delega-
tion chain) that, under appropriate institutional design, provide the proper incentives.
Quite simply, such rules make it more likely that politicians achieve their career and
policy goals if they act in line with their party comrades (Müller 2000a). But institutions
can also provide the ‘wrong’ kinds of incentives (as did the secret vote in the Italian
Parliament until 1988).
Finally, party system conditions affect the cost–benefit calculations of intra-party
actors. Maintaining cohesion is relatively easy for parties in single-party majority
cabinets facing strong competition and most difficult for parties in coalition govern-
ments characterized by great preference diversity. Single-party cabinets are most likely
to feature preference identity within the cabinet (and between the cabinet and the relev-
ant parliamentary party), particularly in majority situations. In contrast, parties in
coalition governments are forced to make concessions to the other government parties.
So do weak minority governments that have to accommodate other parties in
parliament. The greater the concessions, the more strain on party cohesion. Figure 4.1
summarizes our argument.
Western European parties generally display very high levels of parliamentary cohe-
sion, indeed so high that they are hardly ever systematically measured (for some excep-
tions, see Saalfeld 1995b; Bowler et al. 1999; Jensen 2000; Müller et al. 2001). And, as
Laver and Schofield (1990: 15) have noted, ‘parties do in practice tend to go into and
come out of government as single actor’. Yet, 51 of 331 Western European coalition
cabinets of the 1945–99 period (15.4 per cent) terminated due solely or in part to
intra-party conflict (Müller and Strøm 2000a: 586). Thus, in these cases the members of
at least one coalition party were not agreed that government participation was worth
the price.
In parliamentary voting, the major British parties are generally referred to as models
of strength and cohesion (e.g. Sartori 1997). British cabinets rarely fall over intra-party
conflict. At most, the same party continues to govern under a new Prime Minister
(such as John Major after the fall of Margaret Thatcher). This is due to the often com-
fortable majorities provided by plurality single-member elections (which then allow for
some parliamentary dissent) and to the power, generated by electoral insecurity, of
preference alignment between MPs and cabinet members. Given that a parliamentary
seat is a prerequisite for executive office, and there are a large number of such execu-
tive offices, and since many parliamentary seats are electorally marginal, dissolution is
a powerful threat. If elections were called and the incumbents lost, a large number of
politicians—members of the government and MPs in marginal seats—would lose
political office.
Democratic Delegation and Accountability 129
Yet, if anything, many continental European parties seem to outrank the British ones
in parliamentary cohesion. Major backbench ‘rebellions’ (Cowley 2000) are extremely
rare events in much of continental Europe, as are parliamentary defeats resulting from
a lack of party cohesion. We can, however, identify two countries where the parties are
clearly not very cohesive and hence constitute a challenge to governance: France and
Italy. In Fifth Republic France, all three parties that have held the presidency and the
prime ministership—the Gaullists, the UDF, and the Socialists—are weak. The non-
socialist parties are first and foremost vehicles for the personal ambitions of individual
politicians. They are characterized by ‘the delights and poisons of factionalism, patron-
age and personal rivalry’ (Fysh 1996: 90), as evidenced in the frequent breakdown of
party cohesion in presidential elections since the 1970s. More recently they have fis-
sured over European integration. The UDF, which always was a federation of different
parties, has split over the issue of cooperation with the Front National (Knapp 2004).
The Socialist Party traditionally has harboured several factions (Cole 1989), one of
which broke away in the 1990s (creating the Mouvement des Citoyens) (Szarka 1999).
More than anything else, the 2002 presidential election witnessed the dramatic inabil-
ity of the Socialist Party to keep the Left together. Yet, French cabinets manage to sur-
vive and govern. Government institutions, in particular the electoral system (which
forces political parties into stable alliances) and, as we will see below, government-
friendly parliamentary institutions, allow cabinets that achievement.
Italy has similarly weak parties. David Hine (1993: 109) concisely summarizes the
impact of factionalism in the so-called first Italian republic: ‘The Italian Christian
Democratic Party, and to a lesser extent its coalition partners, have deep internal divi-
sions . . . [These divisions] ensure that authority to make policy is rarely delegated to a
cohesive leadership group in government’. Political parties in the so-called second
republic seem no stronger. As Verzichelli (2003) has shown, about 22 per cent of the
lower-chamber MPs and 20 per cent of the Senators changed their party affiliation in
its first two legislative terms. Indeed, since 1945 more coalition cabinets have been ter-
minated due to intra-party conflict than in any other country (seventeen of the first fifty
postwar cabinets) (Verzichelli and Cotta 2000), and absenteeism and frequent break-
downs of voting cohesion in Parliament have been characteristic of Italian politics. As
yet, it remains unclear whether the parties of the second Italian republic will eventually
consolidate and attain greater cohesion than those of the first.
D E L E GAT I O N F RO M VOT E R S TO M P s
Above we have surveyed the procedures by which the parliamentary democracies of
Western Europe make fundamental political decisions. In this and the following sec-
tions, we shall examine how, and to what extent, citizens can select and control the
politicians that make such decisions. We trace this selection and control through the
four stages of democratic delegation, beginning with the transfer of authority from
voters to members of parliament.
The simple act of voting is the first and fundamental stage of democratic delega-
tion, and elections are the fundamental source of democratic legitimacy and
130 Torbjörn Bergman et al.
accountability. Though the incumbents of other political offices may be variously
selected, in all democracies citizens elect the members of at least one legislative
chamber. Under parliamentary democracy, that may be about the extent of their
direct delegation. The defining feature of parliamentarism is precisely that the core
executive (the cabinet) is not popularly elected. Typically, parliamentary systems also
feature fewer other directly elected officials (executive offices, judges, etc.) than presi-
dential ones. Thus, parliamentary elections are the linchpin of popular selection and
control.
Voters have a tough time delegating effectively. Even under the best of circumstances
they have little opportunity to coordinate their behaviour. Moreover, many voters also
have scant information about the individual agent (MP) to whom they delegate (see
below). After all, voters are the ‘amateurs’ of representative democracy and politicians
the ‘professionals’. Thus, great information asymmetries exist, which in turn make
agency loss more likely than in situations in which the distribution of information is
more symmetrical.
Agency control in this first stage of delegation can thus be exercised either directly
by the voters, or indirectly through political parties. Political parties may help voters by
aligning the preferences of their representatives. Yet, there is a clear tension here in con-
stitutional design. Mechanisms that help parties control their respective representatives
may leave ordinary voters with scant influence, whereas institutions that empower the
voters may weaken cohesion and control through political parties. Thus, intra-party
candidate choice, for example, may enhance voter control over their elected represen-
tatives, but often at the cost of weakening party cohesion. In the classical parliamentary
model, this tension has been resolved in favour of the parties, but our survey below will
show whether this model still fits the contemporary democracies of Western Europe.
Austria
1945–8 List PR 3 12 n.a. No No No Yes 4 No 1, 3, 6, 7, 8 6, 7, 8
1949–70 List PR 3 12 n.a. No Yes No Yes 4 No 1, 3, 6, 8 6, 8
1971–91 List PR 4 12 n.a. No Yes No Yes 4 No 1, 3, 6, 8 6, 8
1992–2000 List PR 1 7 n.a. Yes Yes No Yes 4 No 1, 3, 6, 8 6, 8
Belgium
1945–8 List PR 2 31 n.a. No Yes No Yes 4 Yes 1, 2, 3, 4, 5, 6, 8 1, 2, 3, 4, 5, 6, 8
1949–65 List PR 2 32 n.a. No Yes No Yes 4 Yes 1, 2, 3, 4, 5, 6, 8 1, 2, 3, 4, 5, 6, 8
1966–94 List PR 2 33 n.a. No Yes No Yes 4 Yes 1, 2, 3, 4, 5, 6, 8 1, 2, 3, 4, 5, 6, 8
1995–2000 List PR 2 22 n.a. No Yes No Yes 4 Yes 1, 2, 3, 4, 5, 6, 8 1, 2, 3, 4, 5, 6, 8
Denmark
1945–9 List PR 2 8 31 No Yes No Yes 4 Yes 10 10
1950–2 List PR 2 8 44 No Yes No Yes 4 Yes 10 10
1953–63 List PR 2 11 40 No Yes No Yes 4 Yes 10 10
1964–2000 List PR 2 15 40 No Yes No Yes 4 Yes 10 10
Finland
1945–50 List PR 1 31 n.a. No Yes No Yes 3 Yes 4, 8 4, 6, 7
1951–4 List PR 1 33 n.a. No Yes No Yes 3 Yes 4, 8 4, 6, 7
1955–61 List PR 1 33 n.a. No Yes No Yes 4 Yes 4, 8 4, 6, 7
1962–8 List PR 1 20 n.a. No Yes No Yes 4 Yes 4, 8 4, 6, 7
1969–71 List PR 1 20 n.a. Legally required Yes No Yes 4 Yes 4, 8 4, 6, 7
. but not binding
1972–4 List PR 1 20 n.a. Legally required Yes No Yes 4 Yes 4, 8 4, 7
but not binding
1975–86 List PR 1 24 n.a. Legally required Yes No Yes 4 Yes 4, 8 4, 7
but not binding
1987–94 List PR 1 29 n.a. Legally required Yes No Yes 4 Yes 4, 8 4, 7
but not binding
1995–2000 List PR 1 31 n.a. Legally required Yes No Yes 4 Yes 4, 8 4, 7
but not binding
Table 4.5. (contd.)
Country and year Electoral Smallest Largest No. of Primariesa Intra Inter Does an Length Can Conditions Rules of
formula constituency constituency suppl. party party extra of term Parliament under which ineligibility for
seats pref pref election expel its an MP loses parl officec
voting voting end the members? his/her seatb
electoral
period?
France
1959–84 Majority- 1 1 n.a. No No No Yes 5 Yes 1, 4, 5, 9 4, 5, 6
plurality
1985–6 List PR 2 24 n.a. No No No Yes 5 Yes 1, 4, 5, 9 4, 5, 6
1987–9 Majority- 1 1 n.a. No No No Yes 5 Yes 1, 4, 5, 9 4, 5, 6
plurality
1990–8 Majority- 1 1 n.a. No No No Yes 5 Yes 1, 2, 4, 5, 9 2, 4, 5, 6
plurality
1999–2000 Majority- 1 1 n.a. No No No Yes 5 Yes 1, 2, 4, 5, 9 2, 4, 5, 6
plurality
Germany
1949–52 List PR 1 43d n.a. No No No Yes 4 No 1, 2, 3, 7, 10, 11 8
1953–64 List PR 1 89 n.a. No No No Yes 4 No 1, 2, 3, 7, 10, 11 8
1965–78 List PR 1 80 n.a. No No No Yes 4 No 1, 2, 3, 7, 10, 11 8
1979–2000 List PR 1 77 n.a. No No No Yes 4 No 1, 2, 3, 7, 10, 11 8
Greece
1975–80 List PR 1 25 n.a. No Yes No Yes 4 Yes 1, 3, 8, 9 4
1981–4 List PR 1 28 n.a. No Yes No Yes 4 Yes 1, 3, 8, 9 4
1985–8 List PR 1 32 n.a. No No No Yes 4 Yes 1, 3, 8, 9 4
1989–95 List PR 1 32 n.a. No Yes No Yes 4 Yes 1, 3, 8, 9 4
1996–2000 List PR 1 38 n.a. No Yes No Yes 4 Yes 1, 3, 8, 9 4
Iceland
1945–58 Plurality ⫹ 1 8 11 No No No Yes 4 No 9 3, 8
PR comb.
1959–70 List PR 5 12 11 No No No Yes 4 No 9 3, 8
1971–86 List PR 5 12 11 No No No Yes 4 No 9 3, 8
1987–2000 List PR 5 18 9 No No No Yes 4 No 9 3, 8
Ireland
1945–6 STV 3 7 n.a. No Yes Yes Yes 5 Yes 4, 5, 6, 10 4, 5
1947–2000 STV 3 5 n.a. No Yes Yes Yes 5 Yes 4, 5, 6, 10 4, 5
Italy
1946–94 List PR 1 54 n.a. No Yes No Yes 5 No 1, 3 1, 3, 6
1994–2000 Plurality ⫹ 1 10 n.a No No No Yes 5 No 1, 3 1, 3, 6
PR comb.
Luxembourg
1945–53 List PR 6 20 n.a. No Yes Yes Yes 6e Yes 3, 4, 8, 11 5, 6, 8
1954–63 List PR 6 20 n.a. No Yes Yes Yes 5 Yes 3, 4, 8, 11 5, 6, 8
1964–83 List PR 6 23 n.a. No Yes Yes Yes 5 Yes 3, 4, 8, 11 5, 6, 8
1984–8 List PR 7 25 n.a. No Yes Yes Yes 5 Yes 3, 4, 8, 11 6, 8
1989–2000 List PR 7 23 n.a. No Yes Yes Yes 5 Yes 3, 4, 8, 11 6, 8
The Netherlands
1945–2000 List PR n.a.f n.a. n.a. No Yes No Yes 4 No 3 3, 4, 6, 9
Norway
1945–52 List PR 3 8 n.a. No No No n.a. 4 No 3, 4, 12 4
1953–84 List PR 4 13 n.a. No No No n.a. 4 No 3, 4, 12 4
1985–8 List PR 4 13 n.a. No No No n.a. 4 No 3, 4, 12 4
1989–2000 List PR 4 15 8 No No No n.a. 4 No 3, 4, 12 4
Portugal
1976–2000 List PR 5 56 n.a. No No No Yes 4 Yes 7, 8 3, 4, 6, 8
Spain
1977–2000 List PR 1 33 0 No No No Yes 4 No 1, 3, 4, 6 3, 4, 6, 8, 9
Sweden
1945–69 List PR 3 24 n.a. No No No No 4 No 1, 9 1, 6
1970–4 List PR 2 33 40 No No No No 3 No 1, 9 1, 6
1975–88 List PR 2 33 39 No No No No 3 No 1, 3, 9 1, 6
1989–94 List PR 2 33 39 No No No No 3 No 1, 3, 9 1
1995–6 List PR 2 33 39 No No No No 4 No 1, 3, 9 1
1997–2000 List PR 2 34 39 No Yes No No 4 No 1, 3, 9 1
United Kingdom
1945–2000 Plurality 1 1 n.a. No No No Yes 5 Yes 2, 10 3, 4, 5, 6, 8
Note: None of the countries in this study have legally required rules regarding (a) term limits, (b) mandatory retirement age, or (c) legally enforced recall mechanisms.
a
Response to question whether parties represented in the parliament use primary elections and whether they are legally required.
b
0 ⫽ None (other than voluntary resignations), 1 ⫽ Election annulled, 2 ⫽ Campaign or election violation, 3 ⫽ Election or appointment to incompatible office. The automatic resignation can
be temporal (as, e.g., in Norway and Sweden) or for the rest of the term, 4 ⫽ Incompatible occupation, 5 ⫽ Bankruptcy or incompatible economic interest, 6 ⫽ Insanity, 7 ⫽ Membership of ille-
gal organization, 8 ⫽ Automatically upon conviction of certain crimes, 9 ⫽ By discretionary court decision upon conviction of certain crimes, 10 ⫽ By parliamentary decision upon conviction
of certain crimes, 11 ⫽ By parliamentary decision in the absence of conviction of a crime, 12 ⫽ By successful impeachment through special procedure.
c
0 ⫽ None (other than lack of citizenship), 1 ⫽ Election annulled, 2 ⫽ Campaign or election violation, 3 ⫽ Election appointment to incompatible office, 4 ⫽ Incompatible occupation,
5 ⫽ Bankruptcy or incompatible economic interest, 6 ⫽ Insanity, 7 ⫽ Membership of illegal organization, 8 ⫽ Automatically upon conviction of certain crimes, 9 ⫽ By discretionary court
decision upon conviction of certain crimes, 10 ⫽ By parliamentary decision upon conviction of certain crimes, 11 ⫽ By parliamentary decision in the absence of conviction of a crime, 12 ⫽ By
successful impeachment through special procedure.
d
The number of list seats allotted for Northrhine-Westfalia, the largest list constituency in Germany, changes frequently. In the table we report the largest number of constituency seats during the
four time periods. The full list is as follows: 1949–52: 43; 1953–6: 72; 1957–60: 88; 1961–4: 89; 1965–8: 80; 1969–71: 78, 1972–9: 75; 1980–2: 76; 1983–6: 75; 1987–90: 72; 1991–3: 75; 1994–2000: 77.
e
In Luxembourg 1945–53, half of the seats were up for re-election every third year.
f
In The Netherlands the whole country is one constituency.
134 Torbjörn Bergman et al.
Electoral rules more narrowly defined are more critical to the voters’ ability to
contain agency costs.11 Such rules define the terms of parliamentarians and their most
important ex post constraint: re-election. They also regulate the competition between
candidates, and between and to some extent within political parties. The classical elect-
oral system, under parliamentary as well as presidential constitutions, was the ‘single
member district plurality’ (SMDP) system. Today, only the United Kingdom retains this
system for all seats in the national parliament. The typical Western European electoral
system is instead List Proportional Representation (PR). The finer details vary, but a
large majority of our countries (thirteen out of seventeen countries) use such a system.
France is unusual in that although it has experimented with PR, it currently employs a
two-stage single member district (SMD) system, which combines a majority principle
(SMDM) at the first stage with a plurality rule at the second. Italy has recently moved
to a mix of PR and SMDP. Germany formally also features such a combination but
since the parties typically place the same candidates at the top of both types of lists, in
practice the system functions as PR. Its mixed-member system has nonetheless been a
model for many recent electoral reforms (Shugart and Wattenberg 2000). The Irish
single-transferable vote (STV) system combines PR with inter-party candidate choice.
Legal rules concerning ballot access are one important way in which the people’s rep-
resentatives are controlled both before and after taking office. Simon Hug (2001: 178–81)
has compiled information on the conditions new political parties (i.e. parties not yet rep-
resented in parliament) must meet to get on the ballot. His empirical analyses, however,
demonstrate that these rules do little to explain the success or failure of new parties. We
focus here on the rules that govern the ballot access of individual candidates. Except in
Finland, intra-party candidate selection is not strictly regulated by the state. Finnish elect-
oral law, however, (since 1969) requires political parties to hold ‘primary elections’. Groups
as small as fifteen party members can nominate parliamentary candidates, and when the
number of nominated candidates is higher than the district magnitude (the number of
representatives elected per district), parties are legally required to let their members vote.
In practice, medium-sized and large parties must hold such intra-party elections. Formally,
membership influence is somewhat restricted in that the party’s district executive body
may replace up to a quarter of the candidates put forward by local branches. In practice,
however, the district party organizations tend not to avail themselves of this right.12
In our PR systems, district magnitude varies substantially both within and across
countries. We thus find districts as small as one (1) representative (which obviously does
not permit any proportionality in seat distribution), whereas the largest of the German
Länder, Northrhine-Westphalia, in 1998 elected as many as seventy-seven Bundestag
11
The electoral process is governed by an often complex set of institutions, but the most important rules
are those that (a) determine which candidates get access to the electoral ballot, and (b) translate the popular
vote into legislative (or other) representation (Katz 1997). It is common to refer to the former as rules con-
cerning ballot access and the latter as electoral rules (or system) proper. However, like Gary Cox (1997: 38)
we wish to avoid that somewhat artificial distinction.
12
As the country chapters and Table 4.4 show, other European parties have similar candidate selection
mechanisms that include ‘membership voting’. But since these are not legally mandated, they are not
reported here.
Democratic Delegation and Accountability 135
members via the Land list (70 since 2002) in addition to the MPs elected in single-
member constituencies (64 in 2002). In The Netherlands, the whole country is one con-
stituency with 150 members.
A substantial district magnitude is necessary for PR to have many of its putative effects
(such as the enhancement of proportionality). In closed-list PR systems, which in our
sample includes Germany, Norway, Portugal, and Spain, the leading list candidates of
major parties are often very likely to be elected and are thus somewhat cushioned against
adverse election results. And the greater the district magnitude, the more effectively par-
ties can use ballot placement to reward or punish incumbent MPs. Thus, district magni-
tude is an important part of electoral system design, and a large magnitude facilitates
centralized party control (Taagepera and Shugart 1989; Cox 1997).
On the other hand, a low district magnitude, and particularly a magnitude of one, is
traditionally seen as a key to a strong constituency link, which helps voters identify
their representatives and hold them accountable for government policy (e.g. see,
Loewenberg and Patterson 1979: 45–8). Do SMD systems therefore ensure greater MP
accountability more generally? Such accountability requires both voter information
and the ability to coordinate in accordance with popular preferences. SMD electoral
systems should have an advantage in informational simplicity compared to electoral
systems that link large numbers of voters to large pools of MPs.13 Based on data from
fifteen to nineteen countries from three continents, Norris (2001) shows that voters in
SMD systems indeed have a better knowledge of candidates and more contacts with
MPs than those in multimember systems, although mixed-member systems outper-
form both on these criteria.14 The case also seems straightforward with respect to
the ease of coordination. Generally, under SMD it takes comparatively little voter
coordination to unseat an unpopular MP. Since there is only one incumbent per district,
all that dissatisfied voters need to do is to coordinate on a serious competitor.
Some students of electoral systems (e.g. Mitchell 2000) argue that the high number of
‘safe seats’ in such SMD systems as the United Kingdom means that in reality only a rel-
atively small subset of seats are contestable. Thus, the Conservative Party is extremely
unlikely ever to lose Canterbury, while the Labour Party is equally unlikely to be defeated
in several constituencies in South Wales. According to Linton and Southcott (1998: 52, 79)
only 114 of the 432 seats (26.3 per cent) in which Labour and the Conservatives were the
two leading parties in the 1997 elections were ‘marginal’ (i.e. a Conservative or Labour
lead of less than 10 per cent of their combined vote). In other words, under first-past-the-
post (SMD) systems the competitiveness of individual districts matters, and in Britain the
geographical distribution of voter preferences diminishes accountability.15
13
However, in a case study of the now-defunct Italian Democrazia Cristiana, Golden and Chang (2001) show
that intense intra-party competition for preference votes had the unintended consequence of corrupting MPs.
14
Thus, 76% of German and 69% of Norwegian voters know at least one candidate, compared to 61% of
British voters. The picture is more mixed with regard to contacts with MPs. Such contacts are claimed by 15%
of the voters in Norway, 13% in the United Kingdom, 11% in Germany, 5% in The Netherlands, and 3% in
Spain (Norris 2001: 890–1).
15
A good summary of the British electoral system debate is in The Report of the Independent Commission on
the Voting System. London: Stationary Office, 1998 (Cm 4090-I).
136 Torbjörn Bergman et al.
The SMD systems present the voters with only one candidate per party. As of 2000,
voters in seven countries (Austria, Belgium, Denmark, Finland, Greece, The Netherlands,
and Sweden) can select specific candidates from the list of their chosen party. Intra-party
preference voting has become somewhat more widespread over the postwar period. It
was established in Austria in 1949 and Sweden in 1997, whilst Greece has gone back and
forth, but currently also allows for preference votes. Italy, however, has gone in the oppo-
site direction. A 1991 referendum abolished the intra-party preference vote, and in 1994
Italy introduced the single-member plurality system for three-quarters of the MPs while
the remainder is elected by closed-list PR. At least in general elections, voters in Ireland
and Luxembourg enjoy the privilege of being able to choose both among candidates and
parties. The Irish electorate votes for individual candidates in multimember constituen-
cies and can rank candidates from different parties. This often proves highly effective. In
2002, for instance, Fine Gael, the main opposition party, lost its deputy leader and
nine out of sixteen of its frontbench team due to such preference voting. Voters in
Luxembourg can also rank specific candidates, regardless of party.
Although PR electoral systems increasingly allow for intra-party preference voting,
this trend is by no means overwhelmingly strong. Moreover, in practice, relatively few
MPs owe their seats in parliament to this instrument. This is in part because intra-party
preference voting allows voters only to reward, but not punish, candidates
(by striking them out or moving them to the bottom of the list). Thus, leaving aside the
STV system, the effects of preference voting range from insignificant to modest. In some
countries, where candidates are ranked on party lists, preference voting favours a small
subset of candidates that would not otherwise have been elected (as Austria, Belgium,
The Netherlands, and Sweden). But as a check on the party leadership, such preference
voting is less effective since it is not the people at the top of the list that lose their seat
but rather the candidates at the end of the list of the ‘safe seats’. And even when only
preference votes are decisive and there is no bias in favour of high-ranking candidates (as
in Denmark, Finland, Luxembourg, and Italy before 1994), unpopular incumbents can
be deselected only if a sufficient number of voters converge on the same replacement
candidate. In practical terms this requires intensive coordination and low-ranking list
candidates with significant electoral appeal. Yet, the potential effects of preference vot-
ing are surely much greater than the actual effects that we have observed thus far.
Intraparty Rules
Since political parties play such a crucial role in the screening of parliamentarians, particu-
larly in closed-list PR systems, Table 4.6 reports on intraparty selection rules for all parties
that have been represented in the cabinet. Only Icelandic parties hold ‘open’ primary elec-
tions (contests in which the right to vote is not restricted to party members) among can-
didates from the same party. Note that Icelandic electoral law does not require primaries,
which instead reflect the political parties’ response to citizens’ calls for more openness
(see Chapter 12). But while the Austrian People’s Party at one time (in 1994) also sub-
scribed to primary elections, this party statute soon became a dead letter, which the party
chose not to observe in the early election of 1995 or in the regular one in 1999.
Primary elections defined more permissively as involving a vote among the party
members over a list of nominated candidates have become increasingly common. This
Democratic Delegation and Accountability 137
Austria
1945–66 No None No No
1967–90 No None No Yes
1991–2000 Yesc Conditionally or partially bindingc No No
Belgium
1945–68 No Fully binding No No
1969–80 No None No Yes
1981–2000 No None Yes Yes
Denmark
1945–2000 No None No No
Finland
1959–2000 No Conditionally or partially binding No No
France
1959–2000 No None No No
Germany
1949–2000 No None No No
Greece
1974–2000 No None No No
Iceland
1945–69 No None No No
1970–2000 Yes Conditionally or partially binding No No
Ireland
1945–95 No None No No
1996–2000 No Conditionally or partially binding No No
Italy
1946–2000 No None Yes No
Luxembourg
1945–2000 No None No No
The Netherlands
1946–2000 No None No No
Norway
1945–2000 No None No Yes
Portugal
1976–2000 No None No No
Spain
1977–2000 No None No No
Sweden
1945–2000 No Non-binding No No
United Kingdom
1945–2000 No None No No
a
Only parties that have been represented in the cabinet are included.
b
The categories ‘Non-binding’, ‘Conditionally or partially binding’, and ‘Fully binding’, imply an increasing degree of influ-
ence of membership voting. The categories are reported only when half or more of all parties in the respective country use
this selection procedure.
c
Still valid party statue but not observed after 1994.
138 Torbjörn Bergman et al.
is the minimal definition of a ‘primary election’ in Tables 4.5 and 4.6. The country chap-
ters provide further information on membership involvement and the competition for
ballot access. Yet, only in a minority of countries does a majority of all relevant parties
(here defined as parties that have participated in the cabinet) allow members to vote on
candidate selection. Currently at least some of the major parties in Austria, Finland,
Iceland, Ireland, The Netherlands, and Sweden permit such membership participation,
though again practice does not in all cases live up to their claims.
In the vast majority of parties, MPs formally owe their nomination to subnational
party units (constituency organization, district, or Land organization). This is true even
if we take into account the informal influence of the national party organization,
which, of course, is hard to measure. Generally, national party organizations can exer-
cise some influence over subnational candidate nomination, either by special quotas (as
in Austria) or by offering the right ‘type’ of candidate (Gallagher and Marsh 1988; Katz
and Mair 1992).16 Anecdotal evidence suggests that it is much harder for the national
party to veto the nomination of specific candidates unless they resort to such dramatic
means as expelling MPs from the parliamentary party.
Most studies of parties and candidate selection suggest that there is a significant
electoral advantage to incumbency. Not only do incumbents often control the con-
stituency organizations that nominate candidates, but they also benefit from greater
name recognition and from past opportunities to claim credit for policy initiatives and
constituency service, and generally to advertise themselves to their voters (Mayhew
1974). Except for some green parties, only in Belgium and Italy have established parties
imposed term limits on their MPs. Enforcement must rely exclusively on party-internal
mechanisms (Müller 2000a). Except for some Austrian, Belgian, and Norwegian par-
ties, Western European political parties have also refrained from introducing a manda-
tory retirement age for politicians. Moreover, incumbents are in a position to take many
other steps to enhance their standing with their electorate in ways that are difficult to
monitor. These advantages may translate into a significant moral hazard problem inter-
nal to political parties and their candidates.
and predictable, or subject to political manipulation. In fact, in all our states but
Norway, the constitution permits early (or extra) parliamentary elections. Under the
German Basic Law, such premature termination is severely restricted (although it has
occurred twice). Sweden discourages dissolution by requiring that elections be held at
predetermined and regular intervals, regardless of any intervening (extra) election.
Fixed election dates reduce the information advantages of the incumbents. Even with-
out such institutional means to determine the election date, they can plan the political
agenda so that the parliamentary term ends with the enactment or, preferably, imple-
mentation, of popular reforms, for example, tax cuts, or at the peak of the business
cycle (Nordhaus 1975; Tufte 1978; Willett 1988; Mueller 1989: 277–86). Yet, exogenous
political shocks, for example, a sudden economic downturn, can easily bring such plans
to nought. Hence being able to schedule parliamentary elections at will is a real boon
for incumbents, and the record suggests that governing parties do exploit this advan-
tage (Strøm and Swindle 2002).
Citizen control over parliamentarians may also depend on the rules concerning cam-
paigning and electoral competition. Since political candidates seek to gain support from
an electorate that often has little prior knowledge of them, the length of campaigns
matters. The longer the campaign, the more opportunities voters have to scrutinize the
candidates on offer. In some countries, campaign length is directly or indirectly regu-
lated. Where opportunities for early parliamentary dissolution are either non-existent
or tightly circumscribed (see above and Table 4.12), parties and candidates can con-
fidently anticipate election dates and thus schedule their own campaign activities. Where
early dissolution is more likely (such as in the United Kingdom, Denmark, or Ireland),
direct regulation of campaign duration is more important. The shorter the campaign,
the greater the potential incumbency advantage. Table 4.7 shows how the length of
political campaigns is regulated. In half of the twelve countries that we can classify, the
minimum campaign is shorter than a month, in three (Austria, Germany, and Italy) it
is roughly one month, and in three countries (Finland, Luxembourg, The Netherlands)
it is considerably longer.
Table 4.7 also includes other legal rules that pertain to the fairness of interparty com-
petition.18 If incumbent politicians are able to manipulate the agenda and stack the deck
in their favour, fairness clearly suffers. But while there is agreement on some minimal
conditions that make elections free and fair (e.g. one person–one vote, no election-
related violence or coercion, no electoral fraud), other normative standards, or at least
their operational implications, are less clear. Thus, the rules of competition may
systematically disadvantage new competitors and hence sustain a cartel of traditional
parties (Müller 1993; Katz and Mair 1995). Three common forms of discrimination are
against (a) the opposition in general, (b) smaller parties, and (c) new parties.
18
See Elklit and Svensson (2001) and Choe (1997) for a more comprehensive discussion of free elections.
Table 4.7. Rules of Electoral Competition
Country and year Minimum Public broadcasting time Commercial advertising Restrictions on
length of permitted?c the publications
a b
campaign Free time for Allocation Free time for parties Allocation of opinion polls
(days) Parliament- not represented in (in comparison
ary parties Parliament to parl. parties)
Austria
1945–68 21 Yes Proportional No n.a. Not permitted No
1969–91 32 Yes Proportional No n.a. Not permitted No
1992–2000 31 Yes Proportional No n.a. Not permitted No
Belgium
1945–84 n.a. Yes Proportional Yes Minimum quota Not permitted No
1985–91 n.a. Yes Proportional Yes Minimum quota Not permitted Yes
1992–2000 n.a. Yes Proportional Yes Minimum quota Not permitted No
Denmark
1945–2000 21 Yes Equal Yes Equal Not permitted No
Finland
1945–2000 50 No n.a. No n.a. Unrestricted No
France
1959–76 20 Yes Proportional Yes Minimum quota Not permitted No
1977–2000 20 Yes Proportional Yes Minimum quota Not permitted Yes
Germany
1949–2000 30 Yes Proportional Yes Minimum quota Not permitted No
Greece
1975–88 30 Yes Proportional Yes Minimum quota Not permitted No
1989–95 30 Yes Proportional Yes Minimum quota Unrestricted No
1996–2000 30 Yes Proportional Yes Minimum quota Unrestricted Yes
Iceland
1945–2000 n.a. Yes Equal Yes Less time Permitted No
(over minimum
quota)
Ireland
1945–64 21 No n.a. No n.a. Not permitted No
1965–76 21 Yes Proportional No n.a. Not permitted No
1977–2000 21 Yes Proportional Yes Minimum quota Not permitted No
Italy
1948–75 30 Yes Equal No n.a. Not permitted No
1976–92 30 Yes Equal Yes Minimum quota Not permitted No
1993–2000 30 Yes Proportional Yes Minimum quota Not permitted Yes
Luxembourg
1945–58 60 n.a. n.a. n.a. n.a. n.a. n.a.
1959–67 60 No n.a. No n.a. Not permitted No
1968–83 60 Yes Proportional Yes Minimum quota Not permitted No
1984–2000 60 Yes Proportional Yes Minimum quota Not permitted Yes
The Netherlands
1945–89 43 Yes Equal No n.a. Not permitted No
1990–2000 43 Yes Equal Yes n.a. Unrestricted No
Norway
1945–93 n.a. No n.a. No n.a. Not permitted No
1994–8 n.a. No n.a. No n.a. Not permitted Yes
1999–2000 n.a. No n.a. No n.a. Restricted Yes
Portugal
1976–94 20 Yes Proportional Yes Minimum quota Not permitted Yes
1995–2000 13 Yes Proportional Yes Minimum quota Not permitted Yes
Spain
1977–84 21 Yes Proportional Yes Equal Not permitted No
1985–2000 15 Yes Proportional Yes Minimum quota Not permitted Yes
Sweden
1945–59 80 No n.a. No n.a. Not permitted No
1960–4 80 Yes Equal No n.a. Not permitted No
1965–82 80 No n.a. No n.a. Not permitted No
1983–8 50 No n.a. No n.a. Not permitted No
1989–2000 35 No n.a. No n.a. Not permitted No
United Kingdom
1945–83 20 Yes Proportional Yes Equal Not permitted No
1984–2000 16 Yes Proportional Yes Equal Not permitted No
reformers in the United States, for instance, argue that such caps strengthen the power
of ideas vis-à-vis the power of money. On the other hand, such caps may help govern-
ment parties or parties with great media access, which need fewer resources of their
own to develop and advertise their policies. The country chapters do give some indica-
tion of the problems that new parties face in contesting established ones. However,
judging from the country chapters, there is no obvious correspondence between the
legal caps on campaign spending in Belgium, France, Greece, Ireland, Italy, Portugal,
and Spain, and more favourable conditions for new parties.
Table 4.8 reports how political parties and electoral campaigns are financed, which
again has much to do with electoral contestability (allowing entry to new challengers)
and fairness. Note that these are data on direct and public party subventions only, that is,
money that goes directly to the extra-parliamentary party and/or the parliamentary
party group.19 Existing studies of direct public party finance freely acknowledge the
problems of cross-national comparison (e.g. see Katz and Mair 1992).20 These problems
are truly staggering with regard to private sources of party finance. In this context, we
therefore satisfy ourselves with an inventory of when public party financing was intro-
duced and whether parties outside parliament benefit.
As Table 4.8 shows, since the 1970s public party finance has been a universal feature
of Western European democracies. One danger of public party finance is that it may
help sustain a cartel of established parties and petrify the existing party system (Paltiel
1981: 170). Thus, in some countries (Belgium, Finland, Iceland, Ireland, Italy,
Luxembourg, The Netherlands, and the United Kingdom) a quarter of a century or
more has passed without any extension of public funding to non-parliamentary parties.
Cartelization is less likely if public money is extended to parties not represented in
parliament, which is currently the case in seven countries (Austria, Denmark, France,
Germany, Greece, Norway, and Sweden). Here, parties with and without parliamentary
representation were in most cases granted public subsidies simultaneously or within a
few years. However, if anything, the party systems of the former countries have
been less stable than those of the latter, which suggests that the exclusion of non-
parliamentary parties from state subsidies is not sufficient to freeze party systems.21
19
Thus, we do not cover indirect forms of public party or campaign finance, such as public staff for the
party in public office, free telephone access, free mail, or free broadcasting time in ‘public service’ media.
20
Drawing on his massive edited volume on comparative party finance (Nassmacher 2001), Karl-Heinz
Nassmacher (2002) ranks eight Western European countries according to per capita annual public party
finance. Austria (USD 19.00) and Sweden (USD 11.00) stand out, followed by Germany and France (USD 4.90
and 4.20, respectively), whilst Spain, Italy, The Netherlands, and the United Kingdom spend only between
USD 1.50 and USD 0.07 per capita.
21
This is confirmed by more intensive empirical research on the effects of public party finance in a smaller
number of countries (Pierre et al. 2000; Nassmacher 2001: 191–2).
Table 4.8. Public Party Finance and Spending Caps
Country and year Does public party Public money for parties not in Legal cap on spending?
financing exist? Parliament?
Austria
1945–62 No No No
1963–74 Yes No No
1975–2000 Yes Yes No
Belgium
1945–70 No No No
1971–88 Yes No No
1989–2000 Yes No Yes
Denmark
1945–64 No No No
1965–85 Yes No No
1986–2000 Yes Yes No
Finland
1945–66 No No No
1967–2000 Yes No No
France
1959–88 No No No
1989–2000 Yes Yes Yes
Germany
1949–53 No No No
1954–8 Yes No No
1959–66 Yes No No
1967–2000 Yes Yes No
Greece
1975–83 No No No
1984–95 Yes Yes No
1996–2000 Yes Yes Yes
Iceland
1945–70 No No No
1971–2000 Yes No No
Ireland
1945–62 Yes No Yes
1963–96 Yes No No
1997–2000 Yes No Yes
Italy
1948–74 No No No
1975–92 Yes No No
1993–2000 Yes No Yes
Luxembourg
1945–64 No No No
1965–2000 Yes No No
The Netherlands
1945–63 No No No
1964–2000 Yes No No
Norway
1945–69 No No No
1970–4 Yes No No
1975–2000 Yes Yes No
Portugal
1976–87 No No Yes
1988–2000 Yes No Yes
Democratic Delegation and Accountability 145
Spain
1977–84 Yes No No
1985–2000 Yes No Yes
Sweden
1945–65 No No No
1966–71 Yes No No
1972–2000 Yes
United Kingdom
1945–74 No No No
1975–2000 Yes No No
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
The next stage in the parliamentary chain of delegation is one step removed from the
voters: the selection and control of the core executive (Prime Minister and cabinet).
Constitutionally, this is a crucial and pivotal phase in the democratic chain of delegation
(Müller 2000a: 323–7). In ideal-typical parliamentary systems, MPs are the only directly
elected national agents of the people. Parliamentarians in turn directly or indirectly del-
egate to the Prime Minister and other members of the core executive. These are indeed
the only agents directly accountable to parliament, and this very linkage is the crux of
parliamentarism.
In the link between parliament and executive, the potential for agency loss is large
and obvious. Parliamentarians face problems of adverse selection as well as moral haz-
ard. The many perquisites available to cabinet members may well draw the ‘wrong
kind’ of aspirants. Moreover, even well-intentioned politicians face serious temptations,
since parliamentary oversight is often ineffective. Given the myriad of opportunities
that cabinet members have to make policy and other decisions, moral hazard may
indeed be the most likely agency problem.
Agency loss among cabinet members may take several forms. It is widely known
from the scandal sheets and anecdotal evidence (though not extensively documented in
the academic literature) that cabinet members often gratify themselves with fancy cars
and living quarters, use government airplanes for private trips, host unnecessarily lavish
receptions, or make other inappropriate uses of government money. While ministers
tend to see such profligacy as ‘fringe benefits’ for their hard word, if disclosed it usually
inflicts damage on their party and may endanger the re-election prospects of their co-
partisans (who therefore have an incentive to prevent ministers from such behaviour).
And if ministers use their office as a ticket to high society and its indulgences, the bit-
ter fruits of leisure shirking and rent-seeking may follow. In serious cases, cabinet
members become ‘kleptocrats’ (Rose-Ackerman 1999), draining the public coffers for
their own benefit (or those of their families and friends). In international comparison,
the democracies of Western Europe are probably not the most seriously afflicted, but
the incentive for agency loss is nonetheless there.
Even politicians who have previously been loyal partisans may not always be reliable
agents as cabinet members. Government ministers have typically reached the peak of
their political careers and hence will not self-interestedly act as ‘robotic party delegate(s)’
but ‘opt for a quiet life or establishing a personal reputation as an efficacious decision-
maker, both of which might well maximize their post-ministerial career prospects in
business or elsewhere in public life’ (Dunleavy and Bastow 2001: 3, 6).23 Under such
23
Blondel shows that in Scandinavia, The Netherlands, Austria, Britain, and France, the ministerial career
is ‘truly a springboard for a third career’ (1991: 170).
Democratic Delegation and Accountability 147
circumstances ministers may follow the cues of civil servants, their department’s clientele,
or the media, whenever doing so seems more rewarding than following the party line or
popular preferences. Thus, cabinet members may make policy decisions contrary to the
preferences of the parliamentarians that stand behind them.
As this brief discussion makes clear, the link between MPs and ministers, even if they
belong to one and the same party, is problematic. This is even more the case in coali-
tion governments where typically no party link exists between the ministers and the
majority of MPs. Moreover, recruitment patterns may be such that cabinet members
have little in common with MPs. When cabinet ministers are drawn from the pool of
MPs they are more likely to share preferences with those members that remain in
parliament. In any case, they have gone through the same recruitment processes and
share the parliamentary experience. But as De Winter (1991: 48) has shown, in the
1945–84 period 12 per cent of cabinet members in Western Europe were outsiders who
had had practically no prior political experience.24
Yet, even cabinet members recruited from the ranks of MPs may have interests and
incentives not known, and often not acceptable, to their colleagues in parliament. And
once in office, they have ample opportunity to act on them. Besides, there is a long-
standing argument concerning the skills of MPs as cabinet ministers. According to
Richard Rose (1976: 364), the chief skills of MPs ‘concern personal relations and the
effective oral presentation of ideas’. In the words of Bruce Headey (1974), although
former MPs are well qualified to act as ‘policy selectors’ and ‘ambassador ministers,’
cabinet members also need to be ‘policy initiators’ and ‘executive minister’ (i.e. they
must be able to direct a large organization—their department). As Headey (1974)
concludes, MPs are not specially trained to serve as department managers and, as
Rose (1976: 368) puts it, ‘the role for which prior parliamentary work offers least prepa-
ration is that of a policy initiator’. To the discomfort of advocates of strong parliamentary
government, in which the cabinet members are recruited from parliament exclusively,
Rose (1976: 369) concludes that ‘there is nothing in parliamentary work that gives this
experience to MPs before they become ministers’. MPs may be able to acquire the
requisite qualifications prior or parallel to their parliamentary career, for instance, in
business. Otherwise, however, limiting the choice of cabinet members to the pool of
government MPs may help the willingness of the spirit, but not the strength of the flesh.
Thus, MPs cannot take it for granted that candidates for cabinet positions have the
capacity and incentives to act in the parliamentary majority’s best interest. Hence,
parliamentarians must look for institutional and other ways to screen candidates and
ensure accountability at the very core of government. Four mechanisms are particu-
larly spectacular and make for national and often international headlines: cabinet
inauguration, no-confidence votes, confidence votes, and parliamentary dissolution.
The former two are mechanisms by which the parliamentary majority can seek to con-
trol the cabinet through scrutiny and the threat of replacement. The latter two are
24
According to De Winter (1991: 48), ‘outsiders’ were particularly common in The Netherlands (37.5%),
Luxembourg (24.1%), Sweden (21%), and Austria (20.2%), and practically unknown in Ireland (0%), the United
Kingdom (2.6%), and Italy (2.4%). Greece, Iceland, Portugal, and Spain were not included in this study.
148 Torbjörn Bergman et al.
countervailing powers held by the cabinet (sometimes specifically by the PM or Head
of State). They similarly contain threats that can be used to diffuse parliamentary dis-
sent or more generally to align preferences between the cabinet and the parliamentary
majority. These four mechanisms can be combined in various ways and we believe that
together with the design of the electoral systems, these combinations are the most
important variations in delegation design and accountability mechanisms among par-
liamentary democracies.
Cabinet Inauguration
One of the most celebrated and public events in modern democracies is government
inauguration. In parliamentary regimes parliamentarians select, or at least tacitly accept
or reject, a PM and his or her cabinet. This is largely a party-dominated process
(Andeweg and Nijzink 1995). Thus, parliamentarians—as we stress throughout this
book—are constrained by the political parties that select them as candidates and that by
virtue of their ‘brand name’ often guarantee their election. Thus, the distribution of
seats and bargaining power among the political parties in parliament usually determines
the outcome of the formation process.25 Yet, the tools (mechanisms) provided by the
constitution shape the behaviour of both MPs and political parties in this process.
Contemporary constitutions typically designate the constitutional actor that appoints
the PM (a power most commonly vested in the Head of State) and determines whether
or not an incoming head of government needs to submit to a vote of investiture. In the
latter case, to enhance the ability of parliament to contain adverse selection problems, the
constitution may even prescribe a policy statement (government declaration) to be deliv-
ered by the presumptive PM. Historically, cabinet appointment was the prerogative of the
Monarch. As parliamentary government developed, Parliaments wrested this prerogative
out of his control. In some countries Monarchs were replaced by Presidents elected either
directly (e.g. Austria, Finland, France, Iceland, Ireland, and Portugal), or by an electoral
college composed largely or entirely of MPs (Germany, Greece, and Italy). In the other
countries, the Monarch is still the Head of State. With the exception of Sweden (since
1975), where the Speaker of Parliament nominates the PM, their respective constitutions
formally leave it to the Monarch to appoint the PM. Although the actual role of the
Monarch in government formation is largely formal and symbolic, even the most formal
and symbolic involvement can influence coalition bargaining (see Bogdanor 1984; De
Winter 1995). Among most Western European republics, the President under normal cir-
cumstances similarly plays only a formal role. However, several country chapters identify
episodes in which the President has been more consequential in the making of cabinets
(and other major political decisions). Clearly, the Presidents of Italy, Portugal, and Austria
have a greater role in cabinet appointment than the German President. And the
Presidents of France and Finland (until March 2000) have or have had a discretionary role
in government formation that clearly exceeds that of their counterparts elsewhere.
25
To date, the most systematic efforts to chart cabinet inauguration processes cross-nationally are Laver
and Schofield (1990: ch. 4), Bergman (1993, 1995: ch. 3), De Winter (1995), and Rasch (2001).
Democratic Delegation and Accountability 149
Table 4.9 presents the formal rules of cabinet eligibility, appointment, and inauguration.
Three eligibility regimes can be distinguished: (a) those that require membership in
parliament (Ireland, United Kingdom), (b) those that permit it, and (c) those that treat
cabinet office as incompatible with holding a seat in parliament (Belgium since 1995,
France, Luxembourg, The Netherlands, Norway, Portugal, and—since 1974—Sweden).
Those countries that allow the combination of parliamentary and executive office can be
further differentiated according to the behavioural patterns that have evolved (Andeweg
and Nijzink 1995: 160). Thus, simultaneous membership (the so-called ‘dual mandate’) is
less frequent in Austria and Finland than in the other countries that permit it.
The bulk of Table 4.9 describes the involvement of Parliament in cabinet inauguration.
In Belgium, Germany, Greece, Ireland, Italy, Luxembourg, and Spain, cabinets are essen-
tially inaugurated after they have won an investiture vote in Parliament. The seven coun-
tries that require a cabinet to win a majority vote (simple or absolute) in the Parliament
before or immediately after (Belgium, Italy, and Luxembourg) it takes office have a positive
investiture rule (positive parliamentarism). Elsewhere, the test of a new cabinet’s viability
is negative: it is up to the Parliament to unseat any cabinet it does not tolerate. Portugal
and Sweden (since 1975) have an investiture vote, but the requirement is only that the
Prime Minister candidate not lose by an absolute majority. Bergman (1993) refers to this as
a negative formation rule (or negative parliamentarism). In the remaining eight countries,
there is no inauguration (investiture) vote. Instead it is up to the Parliament, at any time,
to unseat the PM (and cabinet).26
Among countries that require an investiture vote (either positive or negative), in six
cases the Parliament votes on the PM candidate. In Belgium, Italy, and Luxembourg the
vote is instead on the full cabinet. In all cases except Ireland and Sweden, the Head of
State nominates or appoints the PM (to be confirmed by a vote in Parliament). In
Ireland any MP is entitled to propose a Prime Ministerial candidate to the Dáil. The
successful candidate for PM then proposes his cabinet members to the Dáil for its
approval. The President only gets involved after the investiture vote by formally
appointing the PM and the other cabinet members. In Sweden the Speaker of
Parliament nominates the PM candidate. Six of the nine investiture countries feature a
mandatory screening procedure, in which the designated PM lays out the government’s
policy programme, presents the proposed cabinet to Parliament, and subjects both to
plenary debate and an investiture vote. The investiture countries also differ in their vot-
ing rules. Except for Germany, which subscribes to the secret vote, in all investiture
countries cabinets are confirmed through a recorded vote.27
26
These countries (including Portugal and Sweden) all have negative formation rules. This decision rule
contrasts starkly with much of the rational choice literature on coalition formation, which assumes that
forming a cabinet (i.e. winning) requires the explicit support of an absolute majority of MPs (Bergman 1995).
Yet, as early as the 1920s, Axel Brusewitz (1929) noted the negative formation rule in his account of Swedish
parliamentarism.
27
Only Ireland, Portugal, and Sweden do not require a quorum of at least 50% of their MPs to invest a
new cabinet. Decision rules for cabinet inauguration vary considerably. They are most demanding in
Germany, Italy, and Spain, where a majority of all MPs (50% ⫹ 1 MP) is required. Greece and Ireland satisfy
themselves with majorities of MPs present and voting.
Table 4.9. Rules of Cabinet Inauguration, Lower Chamber Rules Only
Country and year Cabinet position Appointment by Inauguration by Parliament Parliamentary voting rules
and membership Head of State
in Parliament Inauguration Formal proposal What is Mandatory ex ante Minimum Decision Voting
by parl. by? proposed? screening of quorum rule procedure
program? requirement
Austria
1945–2000 Compatible Formal No n.a. n.a. n.a. n.a. n.a. n.a.
Belgium
1945–1994 Compatible Formal Yes Monarch Cabinet Yes Absolute Simple Recorded
majority or majority
higher
1995–2000 Incompatible Formal Yes Monarch Cabinet Yes Absolute Simple Recorded
majority or majority
higher
Denmark
1945–2000 Compatible Formal No n.a. n.a. n.a. n.a. n.a. n.a.
Finlanda
1945–2000 Compatible Discretionary role No n.a. n.a. n.a. n.a. n.a. n.a.
France
1959–2000 Incompatible Discretionary role No n.a. n.a. n.a. n.a. n.a. n.a.
Germany
1949–2000 Compatible Formal Yes President PM No Absolute Absolute Secret
majority or majority
higher
Greece
1975–2000 Compatible Formal Yes President PM Yes Equal to or less Simple Recorded
than half of all majority
MPs
Iceland
1945–2000 Compatible Formal No n.a. n.a. n.a. n.a. n.a. n.a.
Ireland
1945–2000 Required Formal Yes Individual MPsb PM No 20 MPs Simple Recorded
majority
Italy
1948–2000 Compatible Formal Yes President Cabinet No Absolute Simple Recorded
majority or majority
higher
Luxembourg
1945–2000 Incompatible Formal Yes Monarch Cabinet Yes Absolute Simple Recorded
majority or majority
higher
The Netherlands
1945–2000 Incompatible Formal No n.a. n.a. n.a. n.a. n.a. n.a.
Norway
1945–2000 Incompatible Formal No n.a. n.a. n.a. n.a. n.a. n.a.
Portugal
1976–2000 Incompatible Formal Yes President PM Yes No Less than an Recorded
absolute
majority
voting
against
Spain
1978–2000 Compatible Formal Yes Monarch PM Yes Absolute Absolute Recorded
majority or majority
higher
Sweden
1945–73 Compatible Formal No n.a. n.a. No n.a. n.a. n.a.
1974 Incompatible Formal No n.a. n.a. No n.a. n.a. n.a.
1975–2000 Incompatible No Yes Speaker of PM No No Less than an Recorded
Parliament absolute
majority
voting
against
United Kingdom
1945–2000 Required Formal No n.a. n.a. n.a. n.a. n.a. n.a.
28
Echoing the distinction between positive and negative formation rules, De Winter (1995: 137) distin-
guishes between positive and negative resignation rules. Positive resignation rules require a motion of no
confidence to carry an absolute majority in parliament (i.e. a majority of all MPs) while negative resignation
rules only require a simple majority (i.e. a majority of those MPs present and voting).
Table 4.10. Rules of No-Confidence Votes (i.e. Initiative Rests with Parliament)
Country and year Directed against Proposal power a Constraints Voting rules
PM/cabinet Individual Time Time in days No. of Quorum Decision rules Voting
collectively ministers proposals requirement procedures
Austria
1945–2000 Yes Yes Min. no. of MPs Yes 2 No Equal to or less Simple Recorded
than half of all majority
MPs
Belgium
1945–94 Yes Yes Any member Yes 7 No Absolute Simple Recorded
majority or majority
higher
1995–2000 Yes Yes Any member Yes 7 No Absolute Absolute Recorded
majority or majority
higher
Denmark
1945–2000 Yes Yes Any member No n.a. No Absolute Simple Recorded
majority or majority
higher
Finland
1945–2000 Yes Yes Min. no. of MPs No n.a. No No Simple Recorded
majority
France
1959–2000 Yes No Min. no. of MPs Yes 2 Yes No Absolute Recorded
majority
Germany
1949–2000 Yes No Min. no. of MPs Yes 2 No Absolute Absolute Secret
majority or majority
higher
Greece
1975–2000 Yes Yes Min. no. of MPs Yes 2 Yes Absolute Absolute Recorded
majority or majority
higher
Table 4.10. (contd.)
a
Country and year Directed against Proposal power Constraints Voting rules
PM/cabinet Individual Time Time in days No. of Quorum Decision rules Voting
collectively ministers proposals requirement procedures
Iceland
1945–2000 Yes Yes Any member Yes 2 No Absolute Absolute Recorded
majority or majority
higher
Ireland
1945–2000 Yes No Any member Yes 3 Yes Equal to or less Simple Recorded
than half of all majority
MPs
Italy
1948–2000 Yes Yes Min. no. of MPs Yes 3 No Absolute Simple Recorded
majority or majority
higher
Luxembourg
1945–2000 Yes Yes Min. no. of MPs No n.a. n.a. Absolute Simple Recorded
majority or majority
higher
The Netherlands
1945–2000 Yes Yes Min. no. of MPs No n.a. n.a. Absolute Simple Recorded
majority or majority
higher
Norway
1945–2000 Yes Yes Any member No n.a. No Absolute Simple Recorded
majority or majority
higher
Portugal
1976–2000 Yes No Recognized Yes 3 Yes No Absolute Recorded
party group majority
or min. no. of MPs
Spain
1978–2000 Yes No Min. no. of MPs Yes 5 No Absolute Absolute Recorded
majority or majority
higher
Sweden
1945–70 Yes Yes Any member No n.a. No No Simple Recorded
majority
1971–2000 Yes Yes Min. no. of MPs Yes 3 No No Absolute Recorded
majority
United Kingdom
1945–2000 Yes No Any member Yes 1 No Equal to or less Simple Recorded
than half of all majority
MPs
Required majority
Simple Absolute
Ordinary Austria, Belgium (until France, Greece, Iceland,
1995), Denmark, Portugal, Sweden (since
Finland, Italy, Ireland, 1971)
Luxembourg, The
Netherlands, Norway,
Sweden (until 1971),
United Kingdom
Constructive (No cases) Belgium (since 1995),
Germany, Spain
election of his or her successor; it forces the opposition parties to agree on a joint PM can-
didate. Logically, we can thus distinguish four types—simple majority, absolute majority,
constructive simple majority, and constructive absolute majority—in increasing order of
restrictiveness. Figure 4.2 places our seventeen countries accordingly.
More than half of the Western European countries require a vote of no confidence
to be carried by the usual (simple) majority (leaving aside some variation in counting
blank and invalid votes). Government stability is more strongly protected in the coun-
tries in the two quadrants of the right side of Fig. 4.2. This is particularly true when the
vote of no confidence needs to be constructive. Since the German constructive vote of
no confidence is generally seen to have contributed to the postwar stability of German
politics, it has become an export product, which was adopted by Spain in the 1970s, by
Belgium in 1995, as well as by several countries not included in this volume. Yet, the
German no-confidence vote is by secret ballot, which may jeopardize its restrictive-
ness.29 If the most severe threat to government—and indeed to democracy—is that
extremists of different stripes may join forces in order to remove a centrist cabinet (as
happened most prominently in Weimar Germany; Finer 1946: 1104–6), the construct-
ive vote of no confidence may be a powerful protection.
29
The secret vote did become an issue in the first application of the German constructive vote of no con-
fidence in 1972, though there are different interpretations what actually happened. After several government
MPs had resigned from their respective parties and the government seemed to have lost its majority,
CDU/CSU leader Rainer Barzel tried to topple Chancellor Willy Brandt through a constructive no-confi-
dence vote. To compromise the secrecy of the vote, however, the coalition parties demanded that the bulk
of their MPs abstain and allowed only a few ‘secure’ supporters of Chancellor Brandt to participate.
According to one interpretation it was this tactic that worked: Barzel received 247 votes (while ten ‘no’ votes
and two invalid votes were cast), whereas to be elected he would have needed an absolute majority of 249.
Hence, Brandt survived, although in a direct confrontation he might have won fewer votes than Barzel.
According to another interpretation the balance was tipped by one allegedly bribed CDU member who,
shielded by the secret vote, did not vote for Barzel. In this case the secret vote worked in two ways for Brandt:
To the extent that it was undermined it reduced the risk of defection from the Brandt camp, to the extent it
worked it allowed defection from the Barzel camp.
Democratic Delegation and Accountability 157
Regardless of its specific form, the no-confidence vote is a very powerful accountability
mechanism, but it is also a blunt one, and there are at least two circumstances that seri-
ously restrict its use. First, it takes a majority (in some countries an absolute majority) of
MPs to succeed. This is hard to achieve for members of the parliamentary opposition,
mainly because the MPs of government parties in the vast majority of cases do not want
to rock the boat, regardless of how much they may regret the actions of the cabinet or indi-
vidual ministers. The second problem is that a cabinet is a ‘package deal’ in policy terms.
Removing the PM and cabinet from power also means removing those (if any) parts of the
policy package that are favourably viewed by the MPs and substituting those of an alter-
native cabinet. This consideration, of course, particularly applies to government MPs and
explains why they do not want to rock the boat. Yet, even opposition MPs may fare better
in policy terms under the incumbent government than under its likely replacement, and
they may also face considerable uncertainty about the policy consequences of government
removal. Hence, there are few circumstances in which a majority of MPs can be motivated
to support a no-confidence motion.
Austria
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Belgium
1945–94 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
1995–2000 Specific regulation and Yes Yes Any member of Yes 168 Absolute Absolute Recorded
convention the Cabinet majority or majority
higher
Denmark
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Finland
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
France
1959–2000 Specific regulation only Yes No PM only No n.a. No Simple Recorded
majority
Germany
1949–2000 Specific regulation only Yes No PM only Yes 48 Absolute Absolute Secret
majority or majority
higher
Greece
1975–2000 Specific regulation and Yes No Cabinet collectively Yes 96 Equal to or less Simple Recorded
convention only than half of all majority
MPs
Iceland
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Ireland
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Italy
1948–2000 Specific regulation and Yes No Cabinet collectively Yes 24 Absolute Simple Recorded
convention only majority or majority
higher
Luxembourg
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
The Netherlands
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Norway
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Portugal
1976–2000 Specific regulation only Yes No Cabinet collectively Yes 72 Absolute Simple Recorded
only majority or majority
higher
Spain
1978–2000 Specific regulation only Yes No PM only Yes 24 Absolute Simple Recorded
majority or majority
higher
Sweden
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
United Kingdom
1945–2000 Convention only n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Austria
1945–2000 Yes Dissolution No No Dissolution power In conjunction with attempts to unseat No
power (not (not exclusive)a the Head of State
exclusive)a
Belgium
1945–4 Yes Vetob Can proposec Can proposec No Pending or following constitutional No
amendment
b c c
1995–2000 Yes Veto Can propose Can propose No Pending or following constitutional Time restrictions; dissolution
amendment requires majority consent of the
House (unless the House
introduced but failed to pass a
constructive motion of no
confidence)
Denmark
1945–52 Yes No Dissolution No No Pending or following constitutional No
power amendment
(exclusive)d
1953–2000 Yes No Dissolution No No Pending or following constitutional A new Cabinet cannot dissolve
power amendment Parliament before it and
(exclusive)d Parliament have met
Finland
1945–1990 Yes Vetod No No No No Only when Parliament is in session
1991–2000 Yes Vetod Can proposec No No No Only when Parliament is in session
France
1959–2000 Yes Vetod Can proposec No No No Time restrictions; not during
emergency or similar
Germany
1949–2000 Yes Vetob Can proposec No No No No
Greece
1975–85 Yes Dissolution No Can proposec No If Parliament is unable to elect a PM No
(exclusive)d
1986–2000 Yes Vetob No Can proposec No If Parliament is unable to elect No
a PM; if the President resigns
Iceland
1945–2000 Yes Dissolution No No No Pending or following constitutional No
power amendment; in conjunction with
(exclusive)d attempts to unseat the Head of State
Ireland
1945–2000 Yes Vetod Can proposec No No No No
Italy
1948–2000 Yes Dissolution No No No No Time restrictions; only when
power Parliament is in session
(exclusive)d
Luxembourg
1945–2000 Yes Vetob Can proposec Can proposec No Pending or following constitutional No
amendment
The Netherlands
1945–2000 Yes Vetob No Can proposec No Pending or following constitutional No
amendment
Norway
1945–2000 No n.a. n.a. n.a. n.a. n.a. n.a.
Portugal
1976–81 Yes Dissolution No No No If Parliament defeated the third Not during emergency or similar
powerf cabinet formed in a single not during an extension of the
(exclusive)d legislature President’s term of office
1982–2000 Yes Dissolution No No No No Time restrictions; not during
powerg emergency or similar
(exclusive)c
Spain
1978–2000 Yes Formalh Dissolution No No Pending or following constitutional Not during emergency or similar
power amendment; if Parliament is time restrictions
(exclusive)d unable to elect a PM
Sweden
1945–70 Yes Vetob No Can proposec No No Time restrictions
1971–4 Formalh Dissolution No
power
(exclusive)d
Table 4.12. (contd.)
Country and year Constitution Right to dissolve Automatic constitutional provision Restrictions on these powers to
permits early dissolve
dissolution? Head of PM Cabinet Parliamentary
State majority
Notes: For the column on ‘Automatic constitutional provision’ we have used the following cross-national coding categories: ‘Pending or following constitutional amendment’; ‘In conjunction with
attempts to unseat the Head of State’; or ‘If Parliament is unable to elect a PM’. For the column on ‘Restrictions on these powers to dissolve’ we have used the following cross-national coding
categories: ‘Time restrictions’; ‘ Not during states of emergency or similar; or ‘Only when Parliament is in session’. When these have not adequately reflected the real variation, the country experts
have provided a brief description of the relevant conditions. ‘n.a.’ means not applicable in this case.
a
May dissolve on own initiative, but does not have the power to prevent other constitutional agents from dissolving.
b
Has discretionary power to deny proposal by other constitutional agent, but does not have the power to initiate proposal.
c
May propose only, subject to discretionary approval of other constitutional agent.
d
May dissolve on own initiative and has exclusive discretionary power to dissolve.
f
Formally, dissolution requires support by the Council of the Revolution.
g
Note, however, that the President can only dismiss the cabinet after ‘consulting’ with the Council of State.
h
Formal role only.
Democratic Delegation and Accountability 163
Everywhere but in Norway (for reasons explained below), a lost confidence vote can
result in the dissolution of parliament. Alternatively, dissolving parliament can be a
deliberate choice made by an otherwise secure PM or cabinet (Strøm and Swindle
2002). The conditions under which such dissolution can take place are therefore
important. Table 4.12 identifies the rules of parliamentary dissolution. Below, we dis-
cuss the roles of the most important players.
Of all parliamentary accountability mechanisms, dissolution allows for the greatest
role for the Head of State. In all but two of the sixteen countries that in principle allow
parliamentary dissolution, the Head of State can or must be involved (the exceptions
are, by convention, Denmark30 and, since 1975, Sweden). Empirically, we can identify
five different types:
1. The Head of State’s role is purely formal in Spain, and was so in Sweden between
1971 and 1974.
2. In Portugal (since 1982) the Head of State has proposal power with respect to dis-
solution but does not make the final decision.
3. In eight countries the Head of State has or has had veto power: he or she can refuse
to call early parliamentary dissolution even when it is proposed by other players:
Belgium, Germany, Greece (since 1986), Ireland (though only if the PM, who pro-
poses dissolution, has lost the confidence of Parliament), Luxembourg, The
Netherlands, the United Kingdom, and Sweden (between 1945 and 1970). There
is considerable variation in the role the Head of State actually plays. For example,
the Basic Law tightly constrains the German President.31
4. Three Presidents, those of Austria, France, and Iceland can dissolve Parliament
on their own initiative. Hence, formally, they are decisive players, as defined in
Chapter 3. Strictly speaking, in Austria dissolution requires a cabinet proposal and
Iceland requires the counter-signature of the Prime Minister (although the
President is nominally free to appoint and dismiss the cabinet). In Austria, the
President’s power is not exclusive, as the Austrian Parliament is the only assem-
bly that has the discretionary right to dissolve itself. In France, the President needs
to consult with the PM and the Presidents of both chambers of Parliament, but
that is merely a formal requirement and the President is the only actor who can
dissolve Parliament. All three Presidents are restricted by constitutional clauses
that limit their dissolution powers, though the French President faces more severe
restrictions. Yet, the Icelandic President has never dissolved Parliament and
the Austrian President has not dissolved Parliament since 1930, while presidential
30
Section 32 of the Danish Constitution still states ‘The King may at any time issue writs for a new elec-
tion with the effect that the existing seats be vacated upon a new election’.
31
Although the German Constitution permits the early dissolution of the Bundestag, the strategic use of
this instrument is cumbersome. The two early dissolutions to date have occurred under extraordinary cir-
cumstances. The first dissolution took place after the failed no-confidence vote against Willy Brandt in 1972,
when there was no government majority but also no alternative majority. The second case happened in 1983
after the change from the SPD–FDP to the CDU/CSU–FDP coalition. The government ‘manufactured’ a
defeat on a vote of confidence in order to legitimize an early election (Poguntke 1999).
164 Torbjörn Bergman et al.
dissolutions are a regular feature of French politics (having occurred in 1962,
1968, 1981, 1988, and 1997).
5. According to the letter of their respective constitutions, the Presidents of Finland
(from 1991 on the proposal of the PM), Greece (1975–85), Italy (after consultation
with the Speakers of both houses of Parliament), and Portugal (1976–81) have the
most extensive dissolution powers. These Presidents have (had) the exclusive right
to dissolve on their own initiative and are thus dissolution dictators.
As we have already seen, Heads of State are not the only actors in the dissolution
process. In most countries, the PM has an important role in the dissolution process.
Empirically we can distinguish three types.
1. In Denmark, Spain, and Sweden (1971–74), the PM can or could dissolve parliament
on his or her own initiative. Moreover, except for automatic provisions, in these
countries only the respective PMs can effectively decide to dissolve. According to
the terminology developed in Chapter 3, they are dissolution dictators.
2. In Finland (since 1991), France, Germany, and Ireland, the PM has the power to
propose early dissolution. Formally, this is also true for Belgium, Luxembourg,
the United Kingdom and other monarchies. It is then within the discretion of the
Head of State to accept or decline such a proposal, though in the monarchies a
refusal might precipitate a constitutional crisis.
3. In contrast, the constitutions of Austria, Finland (before 1991), Greece, Iceland,
Italy, The Netherlands, Portugal, and Sweden (except 1971–74) do not give a spe-
cial role to the PM. Here, dissolution power is either in the hands of the Head of
State or the full cabinet.
One might naively expect parliamentary government to imply that parliaments them-
selves could decide to dissolve early. Yet, this is the case only in Austria, where the
Parliament is one of two actors that can dissolve in their own right, and to a lesser
degree in Belgium, where early dissolution requires either conflict between the Cabinet
and Parliament (a failed motion of confidence or a successful no- confidence motion),
or a Cabinet resignation, which in turn may be a way to preempt a vote of no confidence
by the parliamentary majority.
Constitutional mechanisms do not necessarily tell us the whole story about the prac-
tical importance of the four critical powers we have surveyed: investiture mechanisms,
no-confidence votes, confidence votes, and parliamentary dissolution procedures.
There are cross-national variations in the use of these powers that do not conform to
the variation in formal rules. Table 4.13 reports on the actual employment of these four
mechanisms for every cabinet formed in the relevant period.32 We count a new cabinet
with every (a) general election, (b) change of party membership in the cabinet, or
(c) new PM.33 Table 4.13 details the actual voting figures for investiture, no-confidence,
32
Note that the term ‘cabinet’ includes only the PM and those ministers (with or without portfolio) that
have voting rights in the government’s collective body. In this book, the term ‘government’ is used in a
broader sense to include the junior ministers and other political appointees that resign with the cabinet.
33
For reasons detailed in that book, we use the same counting rule as Müller and Strøm (2000b).
Table 4.13. Confidence Votes and Early Elections: The Behavioural Recorda
Country and year No. of cabinets No. of No. of no No. of confidence votes No. of early
formed investiture confidence votes (under specific elections
votes institutional rule only)
Austria
1945–2000 22 n.a. 22 n.a. 9
Belgium
1945–2000 34 34 n.d.b n.a/0 c 11
Denmark
1945–2000 31 n.a. 2 n.a. 20
Finland
1945–2000 45 n.a. 127d n.a. 4
France
1959–2000 23 n.a. 32 71 5
Germany
1949–2000 26 18 2 3 2
Greece
1975–2000 11 10 3 2 6
Iceland
1945–2000 26 n.a. 1 n.a. 7
Ireland
1945–2000 22 39 n.d.e n.a. 16
Italy
1945–2000 52 55 24 232 7
Luxembourg
1945–2000 17 17 1f n.a. 2
The Netherlands
1945–2000 23 n.a. 36 n.a. 5
Norway
1945–2000 26 n.a. 25 n.a. n.a.
Portugal
1976–2000 15 11 1 5 4
Spain
1977–2000 8 8 2 2 6
Sweden
1945–2000 26 9 2 n.a. 1
United Kingdom
1945–2000 20 n.a. 26 n.a. 10
Total 427 201 306 315 115
a
Except for the number of cabinets, the table does not include any data on the cabinets that were still in power on
1 January 2000. ‘n.a.’ means not applicable in this case, ‘n.d.’ means no data.
b
In Belgium, there is no study or inventory of the number of no-confidence votes introduced by the Chambers. It is common
practice that the opposition parties, after a government declaration or interpellation, introduce no-confidence motions. However,
since the Second World War, not a single government has been forced to resign because of a motion of no confidence.
c
Not applicable before 1995, 0 since 1995.
d
The Eduskunta minutes do not list or index no-confidence votes on individual ministers separately from the no-confidence
votes that have been directed at the PM/full Cabinet. Nor is this systematically available in the literature. Consequently, the
number of no-confidence votes listed here also includes a smaller number of votes directed at individual ministers.
e
There is no official source or index that lists the number of no-confidence votes.
f
In Luxembourg, there is no study or inventory of the number of no-confidence votes introduced by the Parliament. Here,
we record only the single no-confidence vote on which we have systematic information, precisely because it was successful.
166 Torbjörn Bergman et al.
and confidence votes.34 Except as otherwise noted, Table 4.13 reports only those no-
confidence votes that were directed against the PM or the whole cabinet and hence, if
successful, would have caused the cabinet’s fall.35 No-confidence data are missing for
Belgium, Luxembourg, and Ireland, where no official account is accessible.
Table 4.13 demonstrates considerable variation with regard to all the four account-
ability mechanisms presented here: investiture votes, no-confidence votes, confidence
votes, and early elections. Ireland and Italy alone have had more investiture votes than
cabinets. In Ireland the high number of investiture votes results from the separation of
the vote for PM from the en bloc vote, which typically occurs only a few hours later,
for the rest of the cabinet (Chubb 1982: 183–4). Otherwise, a higher number of investi-
ture votes than cabinets indicates that prospective cabinets occasionally lose investiture
votes. This has happened in Italy where five cabinets have not survived their vote of
investiture (Strøm 1990: 143), and in Ireland in 1989, when several parties presented
unsuccessful Prime Ministerial candidates (O’Leary 1991: 143).
While an investiture vote may be constitutionally required for a cabinet to assume
office, other confidence and no-confidence votes are not. Party leaders deliberately
schedule them, because they expect to win or to send a signal to the electorate or other
political players. Among incumbents, Finnish cabinets are most frequently challenged
through no-confidence motions, though most of these are not serious attempts to
unseat the cabinet but rather demonstrations of opposition discontent with govern-
ment policies.36 In France and Italy, where it is used most often, cabinets frequently put
their existence on the line, through the confidence vote, to back up their proposals in
Parliament. This mechanism has been most successful in France. If the government
invokes Article 49-3, it either provokes the opposition into introducing a no-confidence
vote or the relevant government proposal is automatically accepted (without any for-
mal vote in Parliament) (Chapter 9; Huber 1996b). In Italy, the confidence vote was
until 1988 less suited to shielding controversial legislation from dissident MPs, because
after a successful vote of confidence the government proposal still required a final sep-
arate vote (which in a good number of cases was lost, due to the secret vote and fac-
tionalism in the governing parties) (Müller-Wirth 1992; Hine 1993: 191–3).
Finally, Table 4.13 identifies whether the Parliament was dissolved early. We define
‘early dissolution’ to include cases (a) in which the cabinet resigned after it was defeated
in Parliament, (b) in which the cabinet resigned by convention for other constitutional
reasons (e.g. the assumption of office by a newly elected Head of State), and in which
these events were immediately followed by elections, or (c) in which election occurred
in the first nine-tenths of the constitutional interelection period (CIEP) (see Strøm and
34
Note that, for obvious reasons, for cabinets in power on 1 January 2000—the end date of our empirical
survey—we only record investiture data.
35
Unfortunately, the Finnish parliamentary records do not allow us to distinguish these no-confidence
votes from those directed against individual ministers. To the best of our knowledge, the large majority of
the Finnish no-confidence votes recorded in Table 4.13 were directed against the PM or the whole Cabinet,
though Table 4.13 also contains a few votes against individual ministers.
36
As noted above, some Finnish no-confidence motions were directed against individual ministers.
Democratic Delegation and Accountability 167
Swindle 2002).37 Thus defined, early parliamentary dissolution happens most fre-
quently in Denmark, Greece, and Ireland, where this fate befalls about two-thirds of all
cabinets or more. These are countries in which cabinets cannot expect—or do not
want—to serve the maximum term.
The Parliament–Cabinet connection is the defining link of parliamentary govern-
ment, and the accountability mechanisms that we have just detailed are the most visible
ones. Lupia and Strøm (1995) refer to no-confidence votes and dissolution powers as the
‘doomsday devices’ of parliamentary democracy. Elections and the different types of
confidence votes are the most important parts of the constitutional structure that shapes
and helps constrain the behaviour of politicians and parties. The country chapters will
place the large variations in these rules and their consequences in historical context.
Parliamentary Questions
Parliamentary democracies also feature a variety of less dramatic ways for parliamen-
tarians to control their agents in the cabinet. In contrast to mechanisms through which
ministers may be directly sanctioned and removed from power, Mény and Knapp (1998:
208–11) refer to these as ‘unsanctioned’ control mechanisms. Although less dramatic,
these can be important mechanisms in the ongoing parliamentary oversight of the
executive branch. Within the arsenal of such weapons, parliamentary questions occupy
a central position (Wiberg 1995). Parliamentary questions are among the most visible
mechanism by which individual MPs, especially those from opposition parties, can hold
ministers publicly accountable even when there is no immediate prospect of an elec-
tion or a change of government. Most parliaments feature a variety of procedures
whereby individual members can question the PM, other members of the cabinet,
or—in committees—civil servants.
Questions may take oral or written form, they may or may not effectively force the
relevant cabinet member to respond, and they may or may not allow the questioner to
proceed with more serious parliamentary challenges, such as motions of no confid-
ence. MPs may question ministers for a variety of reasons, involving the desire to get
personal credit for an issue or simply name recognition (Mayhew 1974). Individual MP
motives thus may have little to do with any serious ambition to oversee the cabinet or
enforce the policy preferences of the parliamentary majority. That, however, does not
take away from the importance which MPs and cabinet ministers sometimes attach to
parliamentary questions. Ministerial accountability in this form may indeed reduce the
potential for moral hazard.
Parliamentary questions have three functions: (a) to provide information, (b) to criticize
government action (or non-action), and (c) to test the honesty or ability of cabinet
members. The first function of parliamentary questions is straightforward: MPs,
particularly from the opposition parties, are at an informational disadvantage vis-à-vis
the government. By asking questions they can force cabinet members to share at least
some information with them and indeed the interested public. In many cases, of course,
37
As the country chapter notes, this counting is quite different from the British ‘convention’, in which
dissolutions in the last year of a government’s term are not considered ‘early’.
168 Torbjörn Bergman et al.
the questioner is not totally ignorant of the plans and actions of the government, often
because of the answers to previous (‘innocent’) questions or because of ‘fire alarms’,
that is, hints from interested parties (co-partisans ‘on the ground’, interest groups,
individual citizens, or even partisan or frustrated civil servants).
The remaining functions of parliamentary questions are less public-spirited.
Questions that implicitly criticize government action (or the lack of such) are designed
to force the minister to publicly admit inconvenient or embarrassing facts. If—rather
than admitting failure or answering evasively—ministers lie to parliament, opposition
MPs may be able to convict them. At least in some countries it is a long-standing con-
vention that ministers who have been proven guilty in this respect have to resign or be
dismissed by the PM (e.g. see Finer 1946: 975; Woodhouse 1994: 49; Brazier 1997: 273).
Since ministers often ‘are apt to think that a questioner possesses a larger fund of
knowledge than he in fact does’ (Finer 1946: 866), they will mostly avoid blatantly lying
to parliament. Yet, their answers may not be particularly comprehensive or informa-
tive. As Finer (1946: 865) has noted, the British House of Commons is willing to accept
‘the substitution of wit for confession’. After all, parliamentary politics is a party game
and a minister ‘scoring’ for his team (party) will be able to rely on the support of his or
her teammates as long as he or she does not grossly violate the rules of the game.
Nonetheless, there are limits to such evasive tactics and often a series of questions leads
the executive to reveal more of its private information than it had intended.
Table 4.14 presents the relevant instruments in broad categories. Written parliamen-
tary questions have existed in most Western European Parliaments throughout the
postwar period. In Denmark, however, they were introduced as late as 1947. In
Luxembourg, questions have in parliamentary practice existed during the entire post-
war period, but were codified in the standing orders only in 1965. In Denmark, France,
Ireland, The Netherlands, Portugal, and Sweden, any individual MP may at any time
ask a written question about anything that falls within ministerial competence. Except
for Ireland, the same countries also permit MPs to ask oral questions. Elsewhere, there
are restrictions, though it is difficult to judge their practical importance. There is a
minor trend towards allowing MPs to interrupt the Parliament’s regular schedule to ask
‘urgent’ questions. This has happened in Denmark and Greece, and—with the tempo-
rary exception of Luxembourg—no country that has allowed such questions has since
abandoned them. Apparently, once MPs have been allowed to ask questions to scruti-
nize cabinet members, it is not easy to constrain their ability to do so. This may be due
to the fact that parliamentary rules of procedure enjoy special protection (as in Austria,
which requires a two-thirds majority to change them) or to political considerations,
such as the public liability of restricting openness.
At the same time, we see trends to rationalize and streamline parliamentary rules of
procedure to limit the opposition’s capacity for dilatory action. Parliamentarism in the
Fifth French Republic and, to a lesser extent, Greece, has always been designed to ensure
the government’s predominance. Other (partly) restrictive reforms of the parliamentary
rules of procedure were adopted in Italy in the 1980s and in Austria in the 1990s.
Table 4.14 also indicates whether the Parliament requires the cabinet to report on
major issues that otherwise fall under the latter’s authority. Such issues may include
investigations that the cabinet is conducting, the performance of state industries, budg-
etary allocations, or, for example, cabinet conduct in EU affairs. What is scrutinized
Table 4.14. Parliamentary Accountability of Cabinet
Year Parliamentary questions permitted? Formal Ex post supervision of cabineta
reporting
Written Oral Urgent Other requirements
Austria
1945–60 Yesb,c No Yesb,c No Yes Ad hoc committees of investigation or inquiry; audit office(s)
1961–88 Yesb,c,d Yesb,c,e Yesb,c No Yes Ad hoc committees of investigation or inquiry; audit office(s)
1989–2000 Yesb,c,d Yesb,c,e Yesb,c Yesf Yes Ad hoc committees of investigation or inquiry; audit office(s)
Belgium
1945–88 Yesf,g,e Yesf,h,e Yesf,h,e Yesf,i,h,e No Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; audit office(s); minister can be compelled to testify
before any parliamentary committee of any type, but can refuse to answer certain
questions
1989–2000 Yesf,g,e Yesf,h,e Yesf,h,e Yesf,i,h,e Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; audit office(s); minister can be compelled to testify
before any parliamentary committee of any type, but can refuse to answer certain
questions
Denmark
1945–6 No No No No No No such supervision
1947–71 Yes j,k,e Yes j,k,e No Yesb,i,h,e No No such supervision
1972–96 Yes j,k,e Yes j,k,e No Yesb,i,h,e No Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; minister can be compelled to testify before any
committee on any matter
1997–2000 Yes j,k,e Yes j,k,e Yesb,k,e Yesb,i,h,e No Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; minister can be compelled to testify before any
committee on any matter
Finland
1945–65 Yesh No No No Yes Permanent committees to supervise ministries; minister can be compelled to
testify before any parliamentary committee of any type, but can refuse to answer
certain questions
1966–88 Yesh Yesh No No Yes Permanent committees to supervise ministries; minister can be compelled to
testify before any parliamentary committee of any type, but can refuse to answer
certain questions
1989–2000 Yesh Yesh No Yesh Yes Permanent committees to supervise ministries; minister can be compelled to
testify before any parliamentary committee of any type, but can refuse to answer
certain questions
France
1959–2000 Yes j,h,g Yes j,k,c,l Yes j,m,h,l No No Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; minister can be compelled to testify before any
parliamentary committee of any type, but can refuse to answer certain questions
Germany
1949–2000 Yesb,n,i,d Yesi,k,o Yesk No Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; audit office(s); minister can be compelled to testify
before any parliamentary committee of any type, but can refuse to answer
certain questionsGreece
Table 4.14. (contd.)
Year Parliamentary questions permitted? Formal Ex post supervision of cabineta
reporting
Written Oral Urgent Other requirements
Greece
1975–85 Yesg Yesg No Yesg Yes Ad hoc committees of investigation or inquiry
1986–2000 Yesg Yesg Yesg Yesg Yes Ad hoc committees of investigation or inquiry
Iceland
1945–2000 Yesh Yesh Yesh No No Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry
Ireland
1945–97 Yesj,g Yesb,d Yesb No Yes Ad hoc committees of investigation or inquiry
1998–2000 Yesj,g Yesb,d Yesb No Yes Ad hoc committees of investigation or inquiry; minister can be compelled to
testify before any parliamentary committee of any type, but can refuse to answer
certain questions
Italy
1948–82 Yesk,l Yesk,l Yesk,l Yesk No Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry
1983–2000 Yesk,l Yesk,l Yesk,p Yesk No Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry
Luxembourg
1945–82q Yesf,g,e Yesf,g,e Yesf,k,d,e Yesf,n,i,k,d,e Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; audit office(s); minister can be compelled to testify
before any committee on any matter
1983–9 Yesf,g,e No Yesf,k,d,e Yesf,n,i,k,d,e Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; audit office(s); minister can be compelled to testify
before any committee on any matter
1990–2000 Yesf,g,e Yesf,g,e Yesf,k,d,e Yesf,n,i,k,d,e Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; audit office(s); minister can be compelled to testify
before any committee on any matter
The Netherlands
1945–2000 Yes j,c,p Yesj,c,p Yesf,c,p Yesb,c,e Yes Ad hoc committees of investigation or inquiry
Norway
n,k
1945–8 Yes No No No Yes Permanent committees to supervise ministries; permanent committees with right
to read cabinet minutes; audit office(s)
n,k k
1949–71 Yes Yes No No Yes
1972–80 Yesn,k Yesk No No Yes Permanent committees to supervise ministries; audit office(s)
1981–4 Yesn,k Yesk No No Yes Permanent committees to supervise ministries; permanent committees with right
to read cabinet minutes; audit office(s)
1985–8 Yesn,k Yesk No Yesf,k Yes Permanent committees to supervise ministries; permanent committees with right
to read cabinet minutes; audit office(s)
k k f,k
1989–2000 Yes Yes No Yes Yes Permanent committees to supervise ministries; permanent committees with right
to read cabinet minutes; audit office(s)
Portugal
1976–2000 Yes j Yes j Yes j No Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; minister can be compelled to testify before any
committee on any matter
Spain
1977–2000 Yesb,n,m,d,e,p Yesb,n,m,d,e,p No Yesb,f,i,g,e,p Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; minister can be compelled to testify before any
parliamentary committee of any type, but can refuse to answer certain questions
Sweden
1945–90 Yes j,k,p No No No Yes Permanent committees to supervise ministries; permanent committees with right
to read cabinet minutes; audit office(s); minister can be compelled to testify
before any committee on any matter
1991–2000 Yes j,k,p Yesj,k,p No No Yes Permanent committees to supervise ministries; permanent committees with right
to read cabinet minutes; audit office(s); minister can be compelled to testify
before any committee on any matter
United Kingdom
1945–78 Yesn,c,d Yesc Yesb,c Yesb,d Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry
n,c,d c b,c b,d
1979–2000 Yes Yes Yes Yes Yes Permanent committees to supervise ministries; ad hoc committees of
investigation or inquiry; audit office(s)
a
This column refers to the ways in which the cabinet can be scrutinized by the Parliament. Note that this does not necessarily presume that the Parliament itself must perform the actual audit.
The Parliament can also use information collected by another body, such as an audit office, when it conducts this scrutiny. The scrutiny can involve the cabinet as a whole or individual ministers
and ministries. The full list of coding alternatives is the following: (1) permanent committees to supervise ministries; (2) permanent committees with right to read cabinet minutes;
(3) ad hoc committees of investigation or inquiry; (4) audit office(s); (5) minister can be compelled to testify before any parliamentary committee of any type, but can refuse to answer certain
questions; (6) minister can be compelled to testify before any committee on any matter (and there is no explicit right for the minister to refuse to answer).
b
Restrictions on the right to ask such questions.
c
The government must respond to at least some types of such questions.
d
The government must respond to some such questions within a specified time limit.
e
Parliament can force response to be given by a specified cabinet member.
f
Restrictions on the conditions under which such questions can be asked.
g
The government must respond to all such questions within a specified time limit.
h
The government must respond to all such questions.
i
Parliamentary debate and motions of no-confidence may follow some such questions.
j
No restriction on any MP’s right to ask such questions.
k
Government may ignore such questions at will.
l
Any member of government can give response.
m
Restrictions on the right of the government to ignore such questions.
n
Several types of such questions.
o
Response can even be given by member of the administration.
p
Any cabinet member can give response.
q
The right to ask questions was only written into the Chamber’s standing orders in 1965, but in practice they existed during the whole period. However, specific time limits for answers and
other similar details have existed only since 1965.
172 Torbjörn Bergman et al.
varies from country to country. Almost all countries (except Denmark, France, Iceland,
and Italy) require the cabinet to bring such reports to the parliament. Belgium initiated
such reporting practices in 1989, to combat a relaxed budget routine in which much gov-
ernment money was regularly spent before the Parliament had approved any expenditures.
Ex post parliamentary supervision of the executive includes various kinds of audits.
While audit institutions of some kind exist practically everywhere, our country spe-
cialists find audit office(s) designed to scrutinize the executive particularly politically
relevant in Austria, Belgium, Germany, Luxembourg, Norway, Sweden, and the United
Kingdom. Even these audit offices are not overwhelmingly concerned with the cabinet
or individual ministers, but tend to concentrate on administrative units and civil
servants. Yet, this scrutiny often reveals how well (or badly) the cabinet has managed
its civil servants. Pollitt et al. (1999) have analysed the audit offices of Finland, France,
The Netherlands, Sweden, and the United Kingdom and discovered a trend away from
pure financial audit to performance audit. That is to say, audit offices are increasingly
concerned with the quality of the public management rather than with mere book-
keeping issues. And they give increasing publicity to the results of their scrutiny, appeal-
ing to the media and the general public rather than exclusively to the audited
institutions and parliament. If that trend continues, the political relevance of audit
offices is likely to increase.
Parliamentary Committees
In many countries parliamentary committees have been important internal fora ever
since the inception of parliamentary government. Although their practical influence and
importance vary considerably among our seventeen countries, committees are every-
where involved in the production of public policy and of laws in particular. Parliamentary
committees also try to supervise policy implementation, and this practice has spread to
more countries over time. When they conduct this scrutiny, committees commonly ask
whether laws have been effectively implemented. The use (or misuse) of public funds is
also a common committee concern.
Table 4.14 reports whether scrutiny takes place in standing (permanent) or ad hoc
(temporary) committees. Standing committees perform ex post scrutiny in thirteen
countries, including Denmark (since 1972) and the United Kingdom (since 1979). Their
operations are guided partly by formal rules and partly by conventions. Three criteria
give some indication of committee power. The first is whether committees can assume
that cabinet ministers will testify before them, which is the case in ten countries:
Belgium, Denmark (since 1972), Finland, France, Germany, Ireland, Luxembourg,
Portugal, Spain, and Sweden. The second rule is whether ministers in principle are
obliged to answer any kind of question that the committees put to them, as is the case
in four of these countries (Denmark, Luxembourg, Portugal, and Sweden). In some of
these countries, ministers might not be legally required to testify or provide answers.
But in practice, ministers appear when they are called and do not simply refuse to
answer questions (though they may answer equivocally or evasively). The third crite-
rion is whether parliamentary committees are allowed to read and investigate cabinet
Democratic Delegation and Accountability 173
minutes, which is the case only in Norway and Sweden. This authority does enhance
the relevant committee’s status and influence (see the country chapters for details).
Cabinet minutes are not always the most exiting and informative documents, but they
may illuminate critical executive discussions and decisions.
On the basis of these criteria, we can distinguish four types of parliamentary com-
mittee systems: (a) the most powerful category, consisting of those that score positively
on three (Denmark, Luxembourg, Portugal) or all four (Sweden) of our criteria, or at
any rate give committees the most exclusive privilege: access to cabinet papers
(Norway); (b) a medium degree of committee authority, as in Belgium, Finland, France,
Germany, and Spain, which all provide for routine scrutiny and oblige ministers to
some degree of cooperation; (c) a low degree of formal accountability to standing com-
mittees, as in Iceland, Italy, and the United Kingdom, where such committee privileges
are generally lacking; (d) and finally, Austria, Greece, and The Netherlands, which do
not force cabinet members to account to standing parliamentary committees, which
instead serve the purpose of legislation exclusively.
Most countries (all except Finland, Norway, and Sweden) provide for executive
scrutiny through ad hoc investigative committees set up to investigate (alleged) mis-
management. Investigative committees often have extensive rights, including the author-
ity to call witnesses and to read government papers. Yet, setting up such a committee,
defining its task and duration, and making actual use of its investigative prerogatives
require decisions that are often highly politicized. With the partial exception of
Germany (where the vote of a quarter of the MPs suffices to launch an investigative
committee) such decisions are made by majority vote. Since such investigations tend to
be a hassle for the governing parties, they are more likely under minority cabinets, when
majority coalitions do not impose strict coalition discipline on such matters, when the
investigation relates to the conduct of a government (and parties) that have left office, or
when public pressure on the parliamentary majority becomes so great that resistance
might involve higher (electoral) costs than allowing an investigation. Despite all these
limitations, investigative committees have occasionally had great political impact, for
instance in Austria, France, Germany, and Portugal (see country chapters).
Across Western Europe the use of scrutiny procedures has increased over time.
Denmark launched a major parliamentary reform in 1972 that gave parliamentary
committees the right to scrutinize the government. Ireland in 1998 extended to com-
mittees the right to have ministers testify before them. In 1979, the British House of
Commons established a system of select committees whose main purpose is to oversee
government policy. Norway introduced parliamentary committee hearings on a trial
basis in 1995, and these procedures have since been given a more permanent status.
Table 4.15. National Scrutiny Process before a Decision in the EU Council of Ministers,
Lower House only (1993–2000)a
Country and Members in EU Affairs Ministers Degree of involvement
year b the EU Affairs Committee, appears
Committee jurisdiction includes before Other Plenary Overall estimate
EU Affairs parlia- meetings of influence on
Pillar Pillar Pillar Committee mentary cabinet
I II III committees ministers
Austria
1995 MP Yes Yes Yes Yes Weak Weak Moderate
1996–2000 MP ⫹ MEP Yes Yes Yes Yes Weak Weak Moderate
Belgium
1993–2000 MP ⫹ MEP Yes Yes Yes Yes Weak Weak Weak
Denmark
1993–2000 MP Yes Yes Yes Yes Moderate Weak Strong
Finland
1995–2000 MP Yes No Yes Yes Strong Moderate Moderate
France
1993–2000 MP Yes Yes Yes Yes Weak Weak Weak
Germany
1993–2000 MP ⫹ MEP Yes Yes Yes Yes Strong Weak Weak
Greece
1993–2000 MP ⫹ MEP Yes Yes Yes Yes Weak Weak Weak
Ireland
1993–2000 MP ⫹ MEP Yes Yes Yes Yes Weak Weak Weak
Italy
1993–6 MP Yes No No No Weak Moderate Weak
1997–2000 MP Yes Yes Yes Yes Weak Moderate Moderate
Luxembourg
1993–2000 MP ⫹ MEP Yes Yes Yes Yes Weak Weak Weak
The Netherlands
1993–9 MP Yes Yes Yes Yes Moderate Weak Weak
2000 MP Yes Yes Yes Yes Moderate Weak Moderate
Portugal
1993–2000 MP Yes Yes Yes Yes Moderate Weak Weak
Spain
1993–2000 MP Yes Yes Yes Yes Weak Weak Weak
Sweden
1995–2000 MP Yes Yes Yes Yes Moderate Weak Moderate
United Kingdom
1993–97 MP Yes No No Yes Weak Weak Weak
1998–2000 MP Yes Yes Yes Yes Weak Weak Weak
a
Norway and Iceland are excluded from this table because they are not members of the European Union.
b
Austria, Finland, and Sweden became EU members on 1 January 1995, which is why no data is reported before that date
for these three countries.
EU policies. Furthermore, over time, the EEA countries seem to have been progres-
sively marginalized in EU decision-making. At least, that is certainly the view from the
EEA states themselves, as Strøm and Narud point out in Chapter 17.
D E L E GAT I O N W I T H I N T H E C A B I N E T
The cabinet is at the heart of national policy-making under parliamentary democracy.
As Rose (1991: 9) puts it: ‘Parliament does not in any meaningful sense govern; that is
the responsibility of the executive in which a Prime Minister is primus’. Laver and
Shepsle (1996: 28–9) characterize the cabinet’s role somewhat more comprehensively:
‘(I)t is the cabinet that symbolizes the apex of political responsibility. It is the Cabinet that
is expected to guide affairs of the state by making and overseeing the implementation
178 Torbjörn Bergman et al.
of policy on important issues. It is the cabinet that is expected to react to and deal with
major crises and emergencies. And, perhaps most significantly of all, it is the cabinet
whose survival is on the line at election times’.
To understand delegation within the core executive, we must be able to identify a
principal, who at the same time is the parliament’s agent. Here is where the doctrine of
parliamentary democracy is ambiguous: is the principal the PM or the cabinet? The
collective responsibility of the cabinet to parliament would suggest that parliament del-
egates to the cabinet collectively. Moreover, many constitutions vest a set of powers,
often critical ones, in the cabinet collectively. At the same, the PM’s formal role in the
appointment and dismissal of cabinet members would suggest that he is the direct
agent of parliament and the master of his cabinet colleagues. Moreover, PMs enjoy
additional prerogatives, such as agenda control over cabinet meetings, that suggest
their supremacy vis-à-vis the other cabinet members. Yet, in coalition governments the
PM’s appointment and dismissal powers are almost exclusively formal with respect to
the cabinet members from parties other than his or her own. Often PMs are also seri-
ously constrained with respect to the appointment and dismissal of cabinet members
from their own party. But then, while the PM’s constitutional powers are ‘clipped’ in
forming a coalition government, he or she can occasionally shake off these chains once
a cabinet has assumed office, either by exploiting situations in which the coalition part-
ner(s) have little bargaining power (and mostly with regard to relatively minor issues)
or by threatening to end the coalition.38
Agency problems within the cabinet then depend on the perspective one takes on
whose agent the cabinet minister is. We shall generally consider the PM the principal
and the other cabinet members the agents. Yet, there are at least five competing forces
that pull and tug at cabinet members and cause potential agency problems in their rela-
tionship with the PM: party, coalition, faction, department, and personal (though not
necessarily political) ambition. As mentioned above, parliamentary government is in
large part party government, and nowhere do parties control political recruitment as
tightly as with respect to cabinet access (see Strøm 2000). The fusion of the executive
and the parliamentary majority, which Bagehot (1963: 65) called the ‘efficient secret’ of
parliamentary government, nowadays very much rests on cohesive political parties. The
party is relevant even in coalition governments, which account for roughly 60 per cent of
the governments covered in this volume. Here, however, each cabinet member is the
servant of two masters—his or her party and the coalition. While the respective minis-
ter’s party generally expects him or her to push party goals, the coalition partners want
the minister to be faithful to the coalition agreement (Strøm and Müller 1999). For
some government parties, factionalism complicates delegation and accountability. If
ministers adhere to factional rather than party goals, party leaders have another reason
38
For example, in 1982, when the German coalition between the SPD and the FDP clearly was close to its
end, but the alternative coalition had not yet been negotiated, Chancellor Schmidt dismissed the FDP cabinet
members and continued in office with a single-party SPD Cabinet until the deal between the FDP and the
CDU/CSU was struck.
Democratic Delegation and Accountability 179
to worry about agency loss (while for their coalition partners it may either aggravate
or diminish agency problems, depending on their policy distance to the faction relative
to the party). Yet another pull derives from the fact that most cabinet ministers are first
and foremost heads of government departments. Lack of department-specific knowledge,
or a deliberate strategy to win their civil servants’ support, may transform ministers into
department ambassadors, who promote their ministry’s interests rather than govern-
ment or party policy (Andeweg 1988). Given that the success or failure of ministers
usually is measured against their achievements as department heads, they clearly have
incentives to promote agency interests over other commitments. Moreover, agency
loss may be a function of personal ambition. Even if ministers share the bond that
co-partisanship provides, their personal career incentives may be quite different from
those of the PM or the party leader. As King (1975: 230) has noted with regard to Britain,
‘every Cabinet contains the Prime Minister’s future successors and therefore, probably,
his present rivals’. Indeed, these may be included in the cabinet precisely because they
are the PM’s rivals. As cabinet members they share collective cabinet responsibility and
cannot openly criticize government policy (Rose 1976: 356). If anything, personal rivalry
will be greater in coalitions, in particular if the coalition parties actively compete for the
premiership. Finally, agency loss can even result from a lack of ambition on the minis-
ters’ part (leisure shirking), or from their prioritization of their private interests.
Agency problems may derive not only from conflicting preferences within the
cabinet, but also from asymmetries of information. Individual cabinet members, par-
ticularly in their capacities as heads of departments, are flooded with information and
demands on their attention. Most studies of cabinet decision-making report that ‘line’
ministers largely defer to their colleagues on issues outside the jurisdiction of their
respective ministry (Andeweg 2000: 378), at least in part because they have no time and
opportunity to inform themselves about the majority of issues that reach the cabinet.
These shortages of time and asymmetries of information are only slightly less severe
for the PM. For all these reasons, the PM may have good reason to worry about agency
problems within the cabinet.
Cabinet
Cabinet decisions
Agenda
Decision
Proposals control
rules
only because of the preferences of the ministers in charge, but also because different
departmental cultures may cause different problem definitions and strategies (hence
whether universities are teamed ministerially with either research or education may lead
to either research- or teaching-orientated ministry proposals). Third, the PM has dismissal
power vis-à-vis the other cabinet members. Although this ultimate weapon is not used
very frequently in practice, ministers live and work in the shadow of potential dismissal.
They can never be safe in their office and hence have incentives to be faithful agents.
Fourth, PMs may have the formal right to instruct the ministers. These instructions can
be procedural (e.g. on the time-priorities of issues) or material (e.g. what general direc-
tion a proposal to cabinet should take). Fifth, PMs often have (negative) agenda control,
which means that they can prevent ministers from bringing issues to the attention of the
cabinet. Sixth, the cabinet decision rules may give the PM a special role in cabinet
decision-making, well beyond having one vote in a body of 10–20 people. Yet, the PM can
only use (or, at least, benefit from) many of these rules if he can draw on sufficient per-
sonnel resources to provide the required information. Below, we therefore survey not only
the relevant rules but also the support staff situation.39
This range of prime ministerial powers might suggest that under parliamentary gov-
ernment, the cabinet is a strict hierarchy, in which all the individual ‘line’ ministers are
fully and equally subordinated to the PM. And while this is a useful ideal type of
cabinet politics, reality is a good deal more complex. Certainly, constitutional rules mat-
ter and in some cases do establish a hierarchical command structure. Thus, as Giovanni
Sartori puts it, the German Chancellor is ‘first above equals’ in part because of two
institutional rules, the most important of which is the constructive vote of no-
confidence. According to Sartori (1997: 107), the premiership system receives ‘an addi-
tional reinforcement . . . from the fact that Parliament appoints only the Chancellor,
not—as in most parliamentary systems—the full government’.
39
Another power the Prime Minister may hold either unilaterally or share with the cabinet is the right to
call a confidence vote. Huber and McCarty (2001) argue that PMs who hold this right enjoy an advantage
vis-à-vis those who do not.
Democratic Delegation and Accountability 181
Not all PMs are so constitutionally privileged. But then, the power of PMs rests not
only on such constitutional foundations, but also on how secure they are in office, or in
other words, on how costly it would be to replace them. Thus, in Sartori’s typology of
Western European PMs, the German-style ‘first above equals’ is only the second-
strongest category. At least the British PM is, as a ‘first above unequals’, even more
powerful in relation to the other cabinet members (Sartori 1997: 103). Yet, the power
of the British PM rests less on institutional rules than on party and party system char-
acteristics, more precisely on the dominance of two cohesive and strong parties, which
produces stable single-party cabinets. Yet, British PMs remain in office only as long as
their respective party considers them an electoral asset (as Margaret Thatcher experi-
enced in 1990; see Jones 1995).
Thus, there are two determinants of prime ministerial power: (a) the characteristics
of party politics and (b) institutional rules (in Sartori’s account represented primarily by
the constructive no-confidence vote). Sartori considers the former factor more import-
ant. Specifically, prime ministerial power depends on ‘strong partisan discipline’ and a
two-party system. Conversely, if these conditions are not present, the PM’s constitu-
tional powers are likely to be more formal than real, and going beyond them will be
simply impossible.
Party Politics
Party politics is structured in part by the institutions that govern delegation from voters
to parliament, as well as delegation from parliament to the cabinet. In this section we
take them as given. We have identified two relevant aspects of party politics—party
cohesion and party system format. Since we have already discussed party cohesion in a
previous section, we concentrate here on the party system. More specifically it is the
party composition of the incumbent cabinet and its most likely replacement that affects
prime ministerial powers and, more generally, cabinet decision-making.
We distinguish three types of party system, defined by the modes of party competi-
tion for government office: two-party systems, two-bloc systems, and pivotal systems
(for a similar perspective see Mair 1997: 206–14). Typically countries fall into one of
these categories in stable ways, though there are shifts from one type to another. But
the important point is the different competitive logics of these respective types.
Two-party systems may have more than two parliamentary parties, but only two are
strong enough to compete for government office. They alternate in government in
single-party majority or stable minority cabinets (Sartori 1976). In such governments the
PM is the leader of the electorally victorious party, thus combining the authority of his
party office and that of PM. Elections are contests between alternative PMs, both
directly (with the voters voting for the party led by the more appealing prime minister-
ial candidate) and indirectly (through the effects of leadership on the party and/or the
cabinet) (Crewe and King 1994; King 2002). The government party must be careful not
to damage the leadership image of the PM and consequently must submit to him or her.
Two-bloc party systems are defined by competition for government office between
two blocs of (stably) aligned parties, which often form pre-electoral coalitions. At first
sight the logic of party competition seems identical to that of two-party systems. Yet,
182 Torbjörn Bergman et al.
a bloc of two or more parties is not a party. Typically, there is also electoral competi-
tion within each bloc. While the overarching goal for all parties is to win a bloc major-
ity, the individual parties are also concerned about their own strength, which, inter alia,
may influence their share in the spoils from government office. What does this imply
for PMs in two-bloc party systems? On the one hand—except when the office of PM
itself is contested among the parties of one bloc (as has occasionally happened among
the non-Socialist parties in France)—PMs are safe in office as long as the respective bloc
wins. And this safety may allow them to exploit their formal powers. On the other
hand, even in well-established blocs, the defection of one party seems more likely than
a party split in two-party systems. This is what German Chancellor Helmut Schmidt
experienced in 1982, when the FDP—after having fought three elections in alliance
with and helped by the SPD—crossed the floor, removed Schmidt from government
office, and subsequently contested five elections in alliance with the CDU/CSU.
Finally, in pivotal party systems, the party composition of government is decided not
in elections but in negotiations between party leaders. Parties fight elections individu-
ally rather than in alliance with other parties. Cabinet composition and the office of PM
are up for grabs in (mostly post-electoral) coalition negotiations. Under such circum-
stances the PM is hardly ever safe. In coalition government an incumbent PM needs not
only to appeal to the voters but also keep the coalition partner(s) happy. And the coali-
tion partners have both incentives and means to check the PM. Consequently, he or she
is severely constrained from making full use of the institutional prerogatives of office.
Some party systems are hybrids, in which the alternative to single-party government
is not a single party but either a solid bloc of parties or a loose coalition, or the alter-
native to a solid bloc of parties is not a competing bloc but a loose coalition. Such party
systems may engender different cabinet dynamics at different times. Another hybrid
type is coalition government in which one party has an absolute parliamentary major-
ity (e.g. Austria 1945–49, Germany 1957–60). In such situations, the PM can be expected
to be strong, though not as strong as in a single-party cabinet.
A final hybrid type is the weak single-party minority cabinet (the Danish, Italian,
and—much less frequently—the Belgian and Finnish cases rather than Swedish Social
Democratic minority cabinets). In some respect it resembles a coalition government in
a pivotal party system: the PM (and indeed any cabinet member) is insecure and
concessions have to be made to other parliamentary parties. In other respects the weak
single-party minority cabinet resembles the single-party majority cabinet. The PM pre-
sides over party comrades exclusively and they may have even greater incentive to sub-
mit to the leader than in majority situations, since they cannot count on majority
support in parliament and since as a rule cabinet survival depends on the skills of the
PM as negotiator. Therefore, if we consider the PM’s power in the cabinet rather than
in the political system at large, weak minority single-party cabinets come closer to
single-party majority than to pivotal system coalition cabinets.
We now turn to the empirical record. Several party systems have not changed their
format. The party systems of the United Kingdom, Spain, and—leaving aside the short
period of coalition government—Greece, have always been two-party systems, while—
again leaving aside exceptions—those of Iceland, Luxembourg, and The Netherlands
Democratic Delegation and Accountability 183
have always been pivotal ones. France has consistently been a two-bloc system in the
electoral arena, though it has occasionally produced single-party cabinets.
The party systems of Austria, Belgium, Denmark, Finland, Germany, Ireland, Italy,
Norway, Portugal, and Sweden have gone back and forth between different party sys-
tem types. Austria has been a two-party system between 1966 and 1983 and otherwise
a pivotal system. Germany has shifted between a pivotal system and a two-bloc system.
Typically, a pivotal coalition transformed itself into a bloc of government parties that
eventually broke down (as happened in 1982) or was voted out of office (as happened
in 1998). Ireland covers almost the full spectrum of party system types, it has been close
to two-party (under Fianna Fáil single-party cabinets), two-bloc (under Fine
Gael–Labour cabinets), and—since Fianna Fáil became ‘coalitionable’ and Labour has
been in government with both major parties—pivotal.
Norway and Sweden have gone back and forth between two-bloc (with the Social
Democrats in office) and pivotal systems (with various government or parliamentary coali-
tions). Given the fragility and short duration of the Swedish red–green alliance of the
1950s, we consider the party system of that time as pivotal rather than two-bloc. Even
when the choice in elections is between Social Democratic single-party government and a
competing bloc of non-socialist parties, the internal workings of the ‘bloc’ are those of a
pivotal coalition in Norway and Sweden. This is clear from the facts that it remains a mat-
ter of post-electoral negotiations which party will get the premiership (occasionally lead-
ing to PMs from relatively small parties—Bondevik and Ullsten) and that the other parties
are generally not willing to submit to the PM. Denmark and Italy in its First Republic also
have gone back and forth between single-party and coalition cabinets, but the former was
never close to a majority. In its Second Republic Italy features blocs of aligned parties con-
testing elections but less intra-bloc cohesion than other two-bloc systems.
than those pertaining to parliament, presumably due to the expectation of greater prefer-
ence convergence among members of the cabinet. One of the more formal and critical
powers within the cabinet is the ability to determine ministerial jurisdictions. If this is
in the hands of the PM, as in Denmark, France, Germany, Greece, Iceland, Ireland,
Portugal, Spain, and the United Kingdom, he or she can then shift policy areas or
administrative units between ministries in order to have a congenial minister.
Some constitutions also allow PMs to issue binding instructions to their ministers.
The best-known example is the Richtlinienkompetenz of the German Chancellor. With
the exceptions of Ireland and Italy (since 1988), which combine it with individual min-
isterial accountability, this authority to instruct ministers exists primarily in systems
where ministers are accountable to parliament only via the PM (Germany, Portugal,
Spain, and the United Kingdom). The reason may be that this authority conflicts with
the constitutional principle of ministerial responsibility and/or accountability.
Consequently, the PM’s authority is also always limited to general policy guidelines.
The PM is not the only member of the cabinet that may be more than equal. Positions
such as the Deputy PM, Minister for Foreign Affairs, and Minister of Finance (or
Economy) often carry particular status and influence. Finance ministers often have a
position second only to the PM, as there are few government decisions of any signifi-
cance that do not have financial implications. As Larsson (1993: 207–8) argues, ‘the min-
ister of finance can even be regarded as a second Prime Minister, since no other minister is
involved in all the aspects of the life of the cabinet in the way the minister of finance is’.
While other ministries, such as environmental portfolios, also may have jurisdictions
that systematically crosscut those of other line ministries, they typically are not backed
up by the institutional powers enjoyed by the PM or the minister of finance.
Table 4.16 identifies the PM’s institutional powers in cabinet deliberations. The table
employs broad categories. We refer to the country chapter for details and, in particular,
for contextual information on the practical relevance of these powers. Recall that in
coalition systems all formal powers tend to be circumscribed. As shown in Table 4.16,
Belgium, Sweden, and the United Kingdom have no formal cabinet decision rules. In
all three countries, the PM summarizes the cabinet consensus and thus has significant
discretionary authority. Denmark and Spain have the same decision rule but here it is
constitutionally mandated. No vote is actually taken. Finland, France, Germany,
Greece, Iceland, Ireland, Luxembourg, The Netherlands, and Norway have a formal
majority voting rule, even if some of these countries use it very rarely. In Austria, Italy,
and Portugal, cabinet consensus is constitutionally required. Finally, in France, the
President rather than the PM defines the cabinet consensus, provided that both belong
to the same political alliance. Otherwise, the PM prevails.
Cabinet decision rules can favour the PM over other ministers. Agenda control,
which is enjoyed by the PMs of Denmark, Germany, Ireland, Italy, Luxembourg,
Portugal, and the United Kingdom, allows the PM to decide the timing of cabinet delib-
eration and to preempt issues that would undermine cabinet cohesion. While cabinet
decision-making everywhere in reality tends to follow informal and consensual proce-
dures, the underlying rules do matter when it is possible for the PM or other cabinet
members to impose them. Thus, when these rules allow the PM to summarize and
Table 4.16. Delegation Within the Cabinet: Prime Ministers and Cabinet Ministers
Country and year Formal Actual decision PM right to PM right to Ministers’ PM formal PM steering or PM full control Administrative structure under
cabinet rule appoint a dismiss a parliamentary right to coordination over cabinet the PM’s supervision
decision accountability decide rights vis-à-vis agenda?
rule ministry ministers?
jurisdiction?
Austria
1945–86 Yes Unanimity Through act by Through Direct No Yes No Civil service staff
President act by
President
1987–2000 Yes Unanimity Through act by Through Direct No Yes No Personal staff (political appointees);
President act by civil service staff
President
Belgium
1945–2000 No Consensus No No Direct No Yes No Personal staff (political appointees);
defined by PM civil service staff
Denmark
1945–63 Yes Consensus Through Through Direct Yes No Yes None
defined by PM act by act by
Monarch Monarch
1964–2000 Yes Consensus Through Through Direct Yes No Yes Civil service staff
defined by PM act by act by
Monarch Monarch
Finland
1945–2000 Yes Majority No No Direct No No No Personal staff (political appointees);
civil service staff
France
1959–2000 Yes Consensus Through Through Via PM only Yes Yes No Personal staff (political appointees);
defined by act by act by civil service staff
President President President
Germany
1949–2000 Yes Majority Through Through Via PM only Yes Yes Yes Personal staff (political appointees);
act by act by civil service staff
President President
Greece
1949–2000 Yes Majority Through Through Direct Yes Yes No Personal staff (political appointees);
act by act by civil service staff
President President
Iceland
1945–2000 Yes Majority Through Through Direct Yes No No None
act by act by
President President
Ireland
1945–2000 Yes Majority Through Through Direct Yes Yes Yes Personal staff (political appointees);
act by act by civil service staff; PM occupies
President President specific (PM) portfolio
Italy
1948–87 Yes Unanimity Through No Direct No No Yes None
act by
President
1988–2000 Yes Unanimity Through No Direct No Yes Yes Personal staff (political appointees);
act by civil service staff
President
Luxembourg
1945–94 Yes Majority Through Through Direct No Yes Yes Civil service staff; PM occupies
act by act by specific (PM) portfolio
Monarch Monarch
1995–9 Yes Majority Through Through Direct No Yes Yes Civil service staff
2000 act by act by Civil service staff; PM occupies
Monarch Monarch specific (PM) portfolio
The Netherlands
1945–2000 Yes Majority No No Direct No No No Civil service staff
Norway
1945–8 Yes Majority Through Through Direct No No No None
act by act by
Monarch Monarch
1949–55 Yes Majority Through Through Direct No No No Personal staff (political appointees);
act by act by
Monarch Monarch
1956–2000 Yes Majority Through Through Direct No No No Personal staff (political appointees);
act by act by civil service staff
Monarch Monarch
Portugal
1976–2000 Yes Unanimity By PM alone By PM alone Via PM only Yes Yes Yes Personal staff (political appointees);
civil service staff
Spain
1978–2000 Yes Consensus Through Through Via PM only Yes Yes Yes Personal staff (political appointees);
defined by PM act by act by civil service staff
Monarch Monarch
Table 4.16. (contd.)
Country and year Formal Actual decision PM right to PM right to Ministers’ PM formal PM steering or PM full control Administrative structure under
cabinet rule appoint a dismiss a parliamentary right to coordination over cabinet the PM’s supervision
decision accountability decide rights vis-à-vis agenda?
rule ministry ministers?
jurisdiction?
Sweden
1945–63 No Consensus Through Through Direct No No No None
defined by PM act by act by
Monarch Monarch
1964–74 No Consensus Through Through Direct No No No Personal staff (political appointees);
defined by PM act by act by civil service staff
Monarch Monarch
1975–2000 No Consensus By PM alone By PM alone Direct No No No Personal staff (political appointees);
defined by PM civil service staff
United Kingdom
1945–2000 No Consensus Through By PM alone Via PM only Yes Yes Yes Personal staff (political appointees);
defined by PM act by civil service staff
Monarch
a
Through act by Head of State (Monarch or President) means that the PM/cabinet has the decisive influence, but that, as a formality only, the Head of State must also be involved.
Democratic Delegation and Accountability 189
interpret the sense of the meeting, his position will be stronger than when the rules
prescribe voting procedures in which each cabinet member may have equal weight. At
least theoretically, under a unanimity rule any single minister can prevent a cabinet
decision. Majority rule lies somewhere between ‘decision-making by interpretation’
(Steiner and Dorff 1980) and unanimity.
Prime Ministerial Staff Formal authority may not be sufficient to prevent agency prob-
lems within the cabinet. On any particular issue, the respective line minister is likely to
have much more detailed knowledge of policy than the PM. In order to offset such
problems of asymmetric information, PMs need independent access to information
and staff (see Peters et al. 2000). Nowadays all PMs except the Icelandic one have their
own supporting office and staff. As Table 4.16 shows, significant prime ministerial staffs
are a relatively recent development in the Nordic countries. Until the mid-1950s, with
a few exceptions, the PMs in these countries had essentially no support staff of their
own, except for a custodian and a couple of secretaries. By the mid-1960s, the Danish
and Swedish PMs ensured that they had a civil service staff of their own. And by the
mid-1970s, the Swedish PM’s office began to be filled also by political appointees.
Finland is an exception, as is the role of the Finnish President. Yet, relatively limited
staffing is not exclusively a Nordic phenomenon. Until quite recently even the Dutch
and Italian PMs had relatively limited staff support.
Summary: Institutional Powers Obviously, there is considerable variation in the rights
and resources of European PMs. One way to summarize their relative institutional powers
is to construct a simple index. Ours contains the two institutional variables stressed by
Sartori (1997), the six active prime ministerial powers discussed above and the personnel
resources at the PM’s disposal. While the institutional rules have remained remarkably
stable (we record only one relevant change in the case of Italy, in 1988), the personnel
resources of PMs have in many cases increased over the postwar period.
These various powers are not all equally important, as our index will reflect. The first
two variables, the constructive vote of no confidence and the parliamentary accountab-
ility of ministers via the PM only, are dichotomous. In each case we assign three points
when these rules apply, and zero otherwise. With regard to prime ministerial appoint-
ment and dismissal powers, we add one point for each of these powers. We have noted
fine institutional differences but do not consider them sufficiently important to discrim-
inate between the countries in each main category. Likewise we add one point each for
prime ministerial powers in determining ministerial jurisdictions, giving instructions to
ministers, and controlling the cabinet agenda. For cabinet decision rules we allocate two
points when the PM sums up the cabinet discussion, one point if majority-rule voting
applies, and zero if cabinet decisions are made by unanimity. Finally, with regard to per-
sonnel resources, we distinguish between two categories of staff—civil servants and
political appointees. We score one point for each type of staff resource at the PM’s dis-
posal. The maximum index value is thus fifteen. As always, the number of points allo-
cated per dimension, the scoring on each dimension, and the placement of some cases
can be discussed. However, the countries identified with institutionally powerful PMs
would remain the same if we had weighted each variable identically.
190 Torbjörn Bergman et al.
Applying these counting rules to our seventeen countries, we can identify particularly
institutionally powerful PMs in Spain (which scores a perfect fifteen points), Germany
(fourteen) and the United Kingdom (twelve), followed by the Portuguese (ten), French
(nine), and Irish (eight) PMs. The PMs of Belgium and Denmark (who have increased
their scores from six to seven points over the postwar period) and Greece and
Luxembourg (six) follow. Thanks to increased staffing, the institutional powers of the
Swedish PM have increased (from four to six points). Towards the lower end of the spec-
trum, the powers of the Finnish PM have remained stable (five points), while those of
the Italian and Norwegian PMs have increased (from three to five points). The powers
of the Icelandic and Dutch PMs have remained unchanged (at four points), whereas
those of the Austrian Chancellor, the institutionally weakest PM, have moved up mar-
ginally (from three to four points). Recall that formal prime ministerial authority and
resources alone do not give us a realistic assessment of their power. As noted above,
under favourable conditions PMs may be able to go well beyond their formal powers
while under unfavourable conditions (including coalition government) they may be
severely constrained.
Assessing Prime Ministerial Power We can now bring the two determinants of prime
ministerial power, party system cohesion and institutional authority, together. Figure 4.4
places each country according to our discussion in the text and Table 4.16. The vertical
dimension represents the prime ministerial powers derived from party and party
system characteristics, reflecting the three major categories distinguished above (single-
party cabinets, ‘bloc’ coalition cabinets, and coalition cabinets in pivotal party systems). We
again use a fifteen-point scheme. This is not meant to indicate that the party and institu-
tional dimensions are of equal importance. We tend to agree with Sartori that the former
is more important than the latter. In the party dimension we have given fifteen points to
majority and thirteen points to minority single-party cabinets. Countries with a mixed
record are placed in-between according to the time spent under the respective type (thus a
country is scored at fourteen if there is an even split between majority and minority single-
party cabinets). We have given ten points to ‘bloc’ coalition cabinets, and five points to coali-
tion cabinets in pivotal party systems. Finally, we have adjusted the placement on the basis
of party system characteristics by moving systems with less cohesive parties down by two
points. Note that the numbering of the respective countries does not indicate a temporal
sequence but the relative strength of the PM within different party system configurations.
Figure 4.4 covers the entire post-war period. Yet, some of these cabinets were clearly excep-
tions in the countries’ overall record and hence have been introduced with shaded rather
than bold symbols.40 In our discussion we will leave these cases aside.
The horizontal dimension of Fig. 4.4 represents institutional powers. To keep the figure
readable we have refrained from reporting temporal variation. The PMs are placed at
40
In Belgium, five cabinets were single-party, two minority, and three majority. Of the former, one
(Spaak) lasted for a mere 7 days and the latter all occurred in the early 1950s under the only single-party major-
ity in postwar Belgium. In Finland, four of thirty-seven partisan cabinets were single-party, one in each
decade from the 1940s to the 1970s. In Germany three minority single-party cabinets lasted only for a few
days and the only majority one dates back to 1960–61.
Democratic Delegation and Accountability 191
GR UK
Strong 15
A1 B1 G1
IR1 PORT1
N1 S1 SP
13 DK
FIN1 F1
Powers from party system format and party cohesion
11
I1
IR2 G2
9
I2 F2
7
LUX, S2
5 FIN2
A2 IR3 PORT2 G3
B2, DK2
3 ICE, NL, N2
I3
Rare types of government
Frequent types of government
1
Weak 0
0 1 3 5 7 9 11 13 15
Weak Institutional powers Strong
the mean within the ranges reported above. None of these temporal variations
amounts to more than two out of fifteen points, and the change has always strength-
ened prime ministerial powers. Hence, Fig. 4.4 slightly underrates the current institu-
tional powers of the Austrian, Belgian, Danish, Italian, Norwegian, and Swedish PMs.
Overall prime ministerial power thus increases along the diagonal from the lower
left-hand corner of Fig. 4.4 to the upper right-hand corner. Of course our scheme is a
reduction of the real world’s complexity. We do not, for example, take into account the
different personalities of PMs. Given these simplifications, our classification fits remark-
ably well with qualitative assessments of prime ministerial power in Western European
democracies.
According to Fig. 4.4, the British and Spanish PMs are thus the strongest ones. The
Greek PM has fewer institutional powers, but single-party government and cohesive
192 Torbjörn Bergman et al.
parties have lifted those incumbents who have been party politicians to the apex of
prime ministerial power. These placements are in line with most of the empirical liter-
ature. King (1991: 43), for example, claims that ‘the British Prime Minister is probably
able to be more powerful inside his or her own government than any other head of
government anywhere else in the democratic world’.41 Likewise, the placement of
Spain can easily be accommodated with Heywood’s (1991) account of the Spanish PM’s
power and Bar’s (1997) classification of Spain as a case of ‘prime ministerial govern-
ment’. Finally, our inclusion of Greece among the particularly strong PMs conforms to
Koutsoukis’ (1994) characterization of the Greek PM as the ‘primus solus’, the ‘singu-
larly powerful’ head in a ‘system of “Prime Minister centrism” ’ (1994: 273, 280). Those
PMs in Portugal, Ireland, France, Denmark, Sweden, Norway, Austria, and even Italy
who headed single-party cabinets have more in common with this premier league of
PMs than with their compatriots who led coalition governments.
The PM of ‘bloc’ coalition cabinets in France and—for some time—the German
Chancellor and the Irish Taoiseach fall in the upper right quadrant of Fig. 4.4, indicating
considerable institutional powers and a cohesive bloc of government parties. Yet, Irish
PMs have done better or worse than this placement but German Chancellors only worse.
In the lower right hand quadrant of Fig. 4.4 we find PMs who are institutionally relatively
strong but who are constrained by coalition government. The German Chancellor stands
out here, being more powerful than the Portuguese and Irish PMs under coalition rule
and Finnish PMs, who invariably serve in coalition governments. The Italian PM in the
Second Republic also ends up in the upper half of Fig. 4.4, but the lack of party cohesion
and formal powers do not allow him to play a particularly strong role.
Accounts of the German system all stress the Chancellor’s enormous institutional
powers. According to Müller-Rommel (1994: 162) ‘the Chancellor plays a dominant
role in Cabinet decision making’, not so much because of direct interference in domes-
tic policy but because this job is effectively done at the pre-cabinet stage by the
Chancellor’s 450 staff members. The constraints resulting from multi-party cabinets
are also generally noted, though the literature lends support to our distinction between
pivotal and bloc politics: ‘the FDP dares not play fast and loose with a coalition it has
established, because a substantial proportion of its own electoral support came from
the voters either favouring its formation in the first place or subsequently attracted to
the alignment. Why jeopardize that support?’ (Smith 1991: 53). In such situations, we
have argued, the Chancellor is less restricted than in pivotal systems in the exercise of
formal powers. Yet, several German Chancellors have not chaired their respective par-
ties or have maintained rather distant relationships with them (Padgett 1993; Helms
2001), which has increased their appeal to the general public but reduced their grip on
their party comrades.
Empirical accounts of the French case stress the role of the President in cabinet
decision-making, given the fact that he chairs official cabinet meetings and is involved
41
Three years later, however, King is somewhat more equivocal: ‘The British prime minister is, within his
own domestic sphere, one of the most powerful heads of government in the Western world, far more pow-
erful than most prime ministers in Western Europe’ (King 1994: 210).
Democratic Delegation and Accountability 193
Austria 12 3 100 0 50 50
Belgium 28 3 100 0 32 68
Denmark 14 4 100 0 100 0
Finland 33 3 97 3 58 42
France 14 4 100 0 0 100
Germany 20 6 0 100 0 100
Iceland 17 1 100 0 100 0
Ireland 10 2 20 80 0 100
Italy 32 2 0 100 0 100
Luxembourg 8 3 0 100 0 100
The Netherlands 23 3 61 39 0 100
Norway 7 3 100 0 86 14
Portugal 4 3 50 50 0 100
Sweden 7 2 100 0 100 0
Source: Müller and Strøm (2000b) and data on Iceland provided by Indridi H. Indridason.
Notes: A government is counted with any general election or change in party composition.
Luxembourg: data for the 1964–99 period.
distinguish between routine conflicts, which are likely to emerge frequently, and seri-
ous conflicts, which are hopefully more rare but have the potential to lead to a break-
down of the coalition. Furthermore, we distinguish conflict management mechanisms
that are cabinet-internal—such as inner cabinets or cabinet committees—and conflict
management devices that include party representatives from outside the cabinet, and
which in extreme cases can be composed entirely of the latter (coalition committees,
party summits). Coalition-internal mechanisms beyond the cabinet itself are generally
not backed up by the constitution or indeed by law. Yet, in practice they can be more
consequential than the constitutional rules.
The cabinet-dominated systems consist exclusively of Nordic countries. The party-
dominated cluster consists of Germany, Ireland, Italy, and Luxembourg. The four
mixed cases, which apply both cabinet-internal and cabinet-external conflict manage-
ment mechanisms, can be subdivided into two categories. Austria and Belgium lean
more towards the cabinet-dominated type, while The Netherlands and Portugal display
more similarities with the party-dominated type.
On the whole, compared to the previous links in the parliamentary chain of delega-
tion and accountability, within-cabinet delegation is less structured by formal rules.
Parliamentary constitutions are largely silent about intra-cabinet relationships, and
most countries do with a minimum of other formal rules, relying largely on conven-
tions that allow for more flexibility (Blondel and Müller-Rommel 1997). Hence, parties
and individual leaders have ample room to adapt the rules of the game to their needs.
The need for flexibility may result in part from the many forces that tug and pull on
196 Torbjörn Bergman et al.
cabinet ministers, and particularly from the demands for careful management of the
preference divergence in multiparty coalition governments. While the doctrines of par-
liamentary government still largely reflect the Westminster heritage of cohesive single-
party governments, constitutional practice must meet the demands of contemporary
Western European reality.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
The final step in the chain of parliamentary delegation is between cabinet members as
head of departments and their civil servants. In the US context, scholarship employing
principal–agent models has examined such delegation more than any of the previous
links. In Western Europe, the information, as well as the analysis, is much more scant.
Civil servants responsible for policy implementation are not only the last players in
the chain of delegation. They often also have a crucial and earlier influence, in particu-
lar through their role in drafting legislation. And while on politicized issues the impe-
tus for law-making is likely to come from politicians, with respect to many
non-politicized topics, civil servants are the ultimate agenda setters vis-à-vis not only
their respective minister but also cabinet and parliament. While in each step of delega-
tion the agents are the agenda setters and the principal has the power to accept or reject
their proposals, here the agents de facto skip one or two steps of the chain and make
proposals to the principal of their principal (the cabinet) and even the cabinet’s princi-
pal (the parliament). Also, very often civil servants are well placed to protect bills
through the various stages of the political decision-making process: with the help of
‘house-trained’ (captured) ministers they may win the cabinet’s support and even be
invited to parliamentary committees to give expert testimony and thereby explain the
virtues of their proposals (e.g. see Suleiman 1984; Page 1985: ch. 7). For these reasons,
it has been claimed that in Western Europe ‘the top professional administrators cer-
tainly wield more influence in policy-making than do average members of Parliament’
(Self 1972: 150). Consequently, Self (1972: 161) maintains that political control over the
administration ‘does not mean simply (or primarily) making sure that policies which
have been met are faithfully executed . . . It means also establishing an atmosphere in
which administrators will continually be aware of political guidelines and constraints’.
Agency Problems
In the classic civil service model that has provided the reference point for Western
Europe bureaucracies for more than a century, civil servants are the servants of the
state, not of the political masters of the day. They are bound by law and obliged to
impartiality. The bureaucratic system of record-keeping, if properly done, allows all
who have access to it to pin down political interventions in routine decision processes
and hence to establish specific accountabilities. To allow civil servants to resist unrea-
sonable demands from their political masters, they enjoy tenure, a guaranteed (min-
imal) career based on seniority, and other protections. These features, of course, may
Democratic Delegation and Accountability 197
exacerbate preference diversity between ministers and civil servants. Civil servants are
obliged by (and, it seems, often firmly devoted to) formal and often inefficient admin-
istrative procedures. To politicians who ‘live and die’ within the confines of the elec-
toral cycle, the delay caused by these procedures is often unreasonable and disturbing.
Moreover, of all the agents with whom we are concerned in this study, civil servants are
most likely to leisure-shirk. If they content themselves with an average career, it is not
hard to see why this might be the case. As early as 1944 the House of Commons Report
of the Committee on the Training of Civil Servants neatly summarized the criticism of the
bureaucratic official: ‘The faults most frequently enumerated are over-devotion to
precedent; remoteness from the rest of the community, inaccessibility and faulty han-
dling of the general public; lack of initiative and imagination; ineffective organization
and waste of manpower; procrastination and unwillingness to take responsibility or to
give decisions’ (quoted from Strauss 1961: 43).
Preference diversity between politicians and civil servants may also lead to the more
serious forms of agency loss, namely dissent-shirking and sabotage (see Huber 2000).
Why so? In contrast to the other actors in the chain of delegation, civil servants come
to office without an electoral mandate. Most are already there when politicians with a
mandate from the voters assume public office. Ideally, civil servants are expected to be
perfectly neutral towards such a mandate and, within the confines of the law, loyally
serve their political masters of the day. In practice, civil servants have their own opin-
ions that may diverge from those of their political superiors on at least two dimensions:
the feasibility and the intent of policies.
Civil servants are considered experts in making things happen, that is, in knowing
which lever to pull to achieve specific policy outcomes. At least, their ministry or
agency will have established procedural wisdoms and routines, and there may be con-
siderable institutional knowledge about the effectiveness of different policy instru-
ments (e.g. see Morstein Marx 1963: 86–90). If politicians are not willing to listen to
their civil servants with regard to the choice of policy instruments, this may cause the
civil servants to ‘undersupply’ relevant advice.
Since it is their job to prepare decisions of broad social importance, civil servants are
also likely to have their own policy preferences. One source of preference divergence
between politicians and (top) civil servants lies in their class background. If the latter
overwhelmingly originate from the upper or middle classes, while the former come
from or appeal to lower classes, such preference diversity may well result.44 One class-
ical study of the attitudes of both politicians and bureaucrats has indeed demonstrated
relevant differences in ideology and visions of government (Aberbach et al. 1981: chs. 5
and 6). Civil servants are (or, at least, were) somewhat to the right of MPs, and they
were generally less willing to respond to popular demands and more inclined to defend
established policies.
44
For a strong claim, see Miliband (1969: ch. V), who sees the bureaucracy as a major obstacle to socialist
reform. Peters (2001: 112–13) shows that 70–80% of senior civil servants in Western Europe have an upper-
or middle-class social background.
198 Torbjörn Bergman et al.
Administrative Governance
While the civil service in most parliamentary systems is classically hierarchical, differ-
ent countries have their peculiarities that sometimes constrain the ability of the cabinet
member to exercise her will. In most parliamentary systems, civil servants are at least
constitutionally subordinated to their respective department heads. Yet, in some coun-
tries direct control of civil servants rests with the PM, or with the cabinet as a whole.
These are conditions that clearly weaken the authority of line ministers over the civil
servants in their respective departments. Moreover, departmental and other civil serv-
ice staffs are rarely uniformly hierarchical (Page and Wright 1999). In most real-world
political systems, the public administration is a very large and complex organization.
Regulations concerning personnel appointment and dismissal, for example, are rarely
uniform across all hierarchical levels or specialized functions. Thus, there may be mul-
tiple categories of public administrators, some with explicitly partisan appointments,
others with strictly non-partisan ones. Even among non-partisan personnel, some (e.g.
judges) may have much stronger protections against interference or dismissal than oth-
ers. Some civil servants have highly specialized tasks for which extensive professional
preparation is necessary. It is here that we would expect particularly large information
asymmetries between bureaucrats and their political superiors. Recognizing the
scarcity of their skills, some systems of public administration go far in granting such
policy specialists autonomy, whereas others retain a greater degree of subordination
and hierarchy. Thus, systems of public administration differ both in the complexity of
their hierarchical arrangements and in their promotion of specialization in some devi-
ation from hierarchical principles.
Austria
1945–73 n.a. Yes No No n.a. Minister– n.a. n.a. n.a.
Cabinet–
President
1974–2000 Privacy; national security; Yes No No n.a. Minister– n.a. n.a. n.a.
resources; third person Cabinet–
President
Belgium
1945–93 n.a. Yes No Yes Hierarchical distinctions; Minister– Cabinet– Minister n.a.
distinctions by terms Monarch Monarch
of appointment
1994–2000 National security; third Yes No Yes Hierarchical distinctions; Minister– Cabinet– Minister n.a.
person; levels of distinctions by terms Monarch Monarch
government of appointment
Denmark
1945–63 n.a. Yes No No n.a. Minister n.a. n.a. n.a.
1964–70 Privacy Yes No No n.a. Minister n.a. n.a. n.a.
1971–2000 National security; Yes No No n.a. Minister n.a. n.a. n.a.
recourses
Finland
1945–2000 National security Yes No Yes Hierarchical distinctions Minister– Cabinet– n.a. n.a.
President President
France
1959–2000 Third person; levels of Yes No Yes Distinctions by terms of Cabinet– Minister n.a. n.a.
government appointment President
Germany
1949–2000 n.a. Yes No No n.a. Minister n.a. n.a. n.a.
Table 4.18. (contd.)
Country Restrictions on Agencies under the Are high civil Categories of high Who appoints high civil servants?
and year freedom of control of servants civil servants b
information lawa appointed in Category 1 Category 2 Category 3 Category 4
Minister Cabinet more than one
and/or PM way?
Greece
1975–85 n.a. Yes No Yes Distinctions by terms of Minister– Cabinet President n.a.
appointment Cabinet
1986–2000 Privacy; national security; Yes No Yes Distinctions by terms of Minister– Cabinet President n.a.
resources; third person; appointment Cabinet
levels of government
Iceland
1945–96 n.a. Yes No No n.a. Minister n.a. n.a. n.a.
1997–2000 National security; Yes No No n.a. Minister n.a. n.a. n.a.
levels of government
Ireland
1945–96 n.a. Yes No Yes Hierarchical distinctions Minister Cabinet n.a. n.a.
1997–2000 National security; Yes No Yes Hierarchical distinctions Minister Cabinet n.a. n.a.
resources; third
person
Italy
1948–2000 n.a. Yes No No n.a. Minister– n.a. n.a. n.a.
Cabinet
Luxembourg
1945–2000 Privacy Yes Yes Yes Hierarchical distinctions; Cabinet– Minister– n.a. n.a.
distinctions by terms of Monarch Monarch
appointment
The Netherlands
1945–2000 National security Yes No Yes Hierarchical distinctions Minister– Minister n.a. n.a.
Cabinet
Norway
1945–70 n.a. Yes No Yes Hierarchical distinctions; Cabinet– Minister n.a. n.a.
distinctions by terms of Monarch
appointment
1971–2000 Privacy; national security; Yes No Yes Hierarchical distinctions; Cabinet– Minister n.a. n.a.
resources; third person distinctions by terms of Monarch
appointment
Portugal
1976–90 n.a. Yes No Yes Hierarchical distinctions Minister– Minister–PM n.a. n.a.
Cabinet
1991–2000 National security; third Yes No Yes Hierarchical distinctions Minister– Minister–PM n.a. n.a.
person Cabinet
Spain
1978–2000 National security; third Yes Yes Yes Hierarchical distinctions; Minister Minister– Minister– Minister–
person functional distinctions; Cabinet Monarch Cabinet–
distinctions by terms of Monarch
appointment
Sweden
1945–74 Privacy; national security; No Yes Yes Hierarchical distinctions; Cabinet– Minister n.a. n.a.
third person distinctions by terms of Monarch
appointment
1975–2000 Privacy; national security; No Yes Yes Hierarchical distinctions; Cabinet Minister n.a. n.a.
third person distinctions by terms of
appointment
United Kingdom
1945–2000c n.a. Yes No No Ministers may appoint Minister n.a. n.a. n.a.
up to two advisors
45
Note, however, that we are concerned with national-level policy implementation, and not with the
implementation conducted by states, regions, or other subnational communities. For example, given
the complexities of German federalism, delegation from ministers to civil servants is more complicated than
in most other European countries. If accountability is to include scrutiny of the implementation of all
national policy, then in the German case—and many others—subnational bureaucracies must be taken into
account.
204 Torbjörn Bergman et al.
moving ministers around between different departments. If ministers are moved fre-
quently, they miss the chance of becoming specialists ‘on the job’. Rose (1975: 20) has
estimated ‘the minimum time required for a minister to carry a major new policy
within his department’ as ‘at least three years’—‘one year for the minister to understand
the techn-ical intricacies of his department sufficiently well to settle the administrative
and not unimportant details of a bill, a year to push through Parliament and/or secure
Treasury authorization for expenditure, and a year to establish guidelines for adminis-
tering the resulting policy so that administration can subsequently be routinized and
thus delegated to others’. Given that Rose refers to the British case of cohesive single-
party governments, 4 years may be more realistic for other political systems. This may
even apply to ministers who happen to stay in one particular ministry, if ex ante they
had reason to expect to be moved around, because under such circumstances they have
little incentive to invest in specialization in the jurisdiction of their department.
Unfortunately, we do not have good comparative data on this aspect of ministerial
careers. Yet, Table 4.19 contains data on ministerial duration and portfolio stability for
fourteen of our seventeen countries. If we assume that a minister must be in charge of
one and the same department for 3 or 4 years to become effective, then only Austria,
Germany, The Netherlands, and Sweden seem to meet the conditions for effective min-
isterial governance, though Denmark and Norway also come close. At the low end of
the spectrum we find Italy, Finland, Belgium, and France. Note that, except for Finland,
these countries also show high levels of bureaucratic politicization.
To the best of our knowledge the literature has not produced a comprehensive cross-
national ranking of the specialization of civil servants, based on their training and
careers. We expect that civil servants everywhere display a much higher degree of spe-
cialization than their political masters.
Most parliamentary democracies have legal regulations that aim to prevent political
parties from intervening in the delegation between cabinet ministers and civil servants
(Müller 2000a). Yet, as the relevant literature (e.g. Page and Wright 1999; Bekke and van
der Meer 2000) and the following chapters show, European countries vary not only in
the partisan penetration of the civil service (see also Chapter 22), but also with regard
to the use of political appointees as a buffer between professional politicians and civil
servants. Partisan ties between ministers and their subordinates can help overcome
some of their agency problems. Yet, they create problems of their own. Ministers may
be too trusting of subordinates with whom they share a party affiliation. Even more
critically, they may be loath to intervene when such servants misbehave. There is cer-
tainly no reason to expect corrupt practices to be limited to those civil servants that
have no party affiliation or one different from the minister.
In practice, the relationship between ministers and civil servants is much less one of
command than what the constitution typically suggests. Successful ministers need to
convince civil servants of the urgency and relevance of their projects. In so doing they
need to provide material and non-material incentives in order to tie civil servants to the
policies pursued by the political leadership. In all likelihood, the civil service is not
monolithic and ministers may need to get involved in internal disagreements and rival-
ries among civil servants. In other words, a ministry is an arena that taxes the minister’s
skills at bargaining as much as command. The New Public Management (NPM)
designs, which were introduced in Britain in the 1980s and later diffused throughout
Western Europe, greatly enhance the potential for ministerial control. Under NPM
schemes ministers control more financial and career incentives and can by tying
rewards to outcomes more effectively align the preferences of civil servants with their
own. Note, however, that ministers have not been made more accountable to
parliament for the acts of their civil servants.
Freedom of Information
Information is a key to the mitigation of civil service agency problems. Since we suspect
information asymmetries within governmental organizations, as well as between gov-
ernment officials and citizens, any mechanism that can offset those discrepancies in
knowledge may help contain agency problems. Freedom of information laws do precisely
that: they lower the cost of information and obligate politicians and civil servants to make
information about public decisions available to, in most cases, any citizen. Freedom of
information rules thus are not mechanisms that serve the principal in this last stage of
parliamentary delegation—cabinet ministers—as much as the ultimate principal, the peo-
ple. Indirectly, however, such rules impinge on the minister–civil servants relations. They
allow third parties access to information and hence increase the likelihood of reliable ‘fire
alarms’ (McCubbins and Schwartz 1984). Thus, they provide incentives to civil servants
to behave according to the rules.
Table 4.18 provides information on freedom of information rules regarding citizen
and media access to government documents. There is considerable cross-national
206 Torbjörn Bergman et al.
variation in such access, and the countries that have freedom of information laws
differ substantially in their scope and interpretation. While, for instance, the
information rights of Swedish citizens are extensive, in Austria they are much
more restricted due to concerns for potentially conflicting goals such as the protection
of privacy and government efficiency (resource constraints). Thus, many existing
laws providing ‘open’ access to government documents come with considerable
restrictions, as reported in Table 4.18. Such caveats notwithstanding, some countries
(Finland, France, Luxembourg, The Netherlands, Spain, and Sweden) have long
had freedom of information statutes. Beginning in the 1960s, others have opened up
such access. Denmark did so in 1964, Norway in 1971, Austria in 1974, Greece in
1986, Portugal in 1991, Belgium in 1994, and, finally, Iceland and Ireland in 1997.
The UK adopted such a law in 2000. The remaining countries do not have laws
that in this way guarantee freedom of information. Yet, even with the quite restrictive
British laws in force until 2000, for example, much government information was
provided voluntarily.
C O N C LU S I O N
In this chapter we have mapped out the real-world chain of parliamentary delegation,
in Western European societies, that begins with the citizens (the ultimate principals)
and ends with civil servants (the ultimate agents). Our data covers the entire period of
democracy between 1945 and 2000 in seventeen states (in France since the start of the
Fifth Republic). Since constitutions change fairly slowly, these are in almost all cases
also the mechanisms that exist early in the twenty-first century. In the chapters that
follow we shall discuss how these mechanisms interact in particular polities. But before
we delve into this richer picture country-by-country, it is valuable to summarize the
main cross-national patterns that emerge.
Below we summarize the empirical patterns that this chapter has uncovered and
compare the real world of European parliamentary democracy with our ideal–typical
conceptions. Table 4.20 identifies seventeen mechanisms of democratic delegation and
accountability, from voters all the way to civil servants, in their ideal–typical form. In
the next two columns, Table 4.20 matches these ideal–typical features with the
most common real-world mechanisms. The final column identifies the contemporary
trend in institutional design and evolution. Note that the most common
mechanism often is different from the ideal–typical mechanism, which means that a
trend in favour of the most common mechanism actually can be a trend away from the
ideal type.
The ideal–typical parliamentary democracy implies a chain of delegation that is indi-
rect, singular, and monopolistic. It is indeed the presumption of hierarchy and control
from below that makes parliamentary government democratic. It is also because of the
essential role of hierarchy that the principal–agent approach is particularly suited for our
study. The principal–agent approach directs attention towards existing elements of
hierarchy in representative democracies. If democratic governance ceased to have
hierarchical elements, the principal–agent approach might no longer be very relevant,
but then neither would the ideals of representative democracy.
To this constitutional design we must add political parties. The degree to which par-
liamentary democracies function as hierarchies is to a large extent determined by how
well political parties function as bonding and informational economizing devices.
Chapter 3 pointed out that parliamentary systems are generally decisive and character-
ized by administrative as well as incentive efficiency. They are therefore less prone to
leisure shirking than other regime types. When centralized and cohesive, political
parties can reinforce hierarchy and fusion and promote efficient delegation. In the clas-
sical Westminster model, centralized parties and single-party cabinets all reinforce the
hierarchy and fusion of majoritarian democracy.
Though the four stages of parliamentary delegation that we have discussed have
many common features, the agency problems that they present are by no means ident-
ical. In the first and final step, delegation involves large sets of citizens that are political
‘amateurs’ and much smaller (relatively speaking) sets of professional politicians.
Of course, in the first of these stages, the amateurs are the principal and the professional
politicians the agents, whereas in the final stage (cabinet members to civil servants) the
Table 4.20. Ideal-Typical and Actual West European Mechanisms of Delegation and Accountability: Summary
Delegation and/or Ideal-typical Most common (modal) Major empirical Cross-national trend,
accountability mechanism under mechanism in Western exceptions 1945–2000
mechanism party-based Europe, 2000
parliamentarism
Step I: Voter–MP
Legislative cameral Unicameralism Bicameralism (9 countries), Unicameralism (the Weakening
structure which, with the exception of Nordic countries, Greece, bicameralism, with
Italy, is asymmetrical Luxembourg, and Denmark and Sweden
(i.e. the Upper House is much Portugal) moving to
weaker) unicameralism
Locus of MP By party leaders at the By party leaders (elite- By party leaders at the Gradual weakening of
candidate selection constituency level dominated) at the national level (France, party leaders in favour
(elite-dominated/ constituency (district) level Greece); open primaries of more membership
decentralized) (11 countries) (Iceland); membership participation
votes (Finland); advisory
membership votes
(Ireland and Sweden)
Preference voting Closed party lists Intraparty (7 countries) or Closed party list (Iceland, Moderate increase in
on individual interparty (Ireland and Norway, Portugal, and preference voting
candidates Luxembourg) preference Spain); SMD (France and
voting the UK); mixed systems
(Germany and Italy)
Step II:
Parliament–PM
Inauguration rule Positive Negative parliamentarism, Positive parliamentarism No significant trend
parliamentarism, that is, that is, parliamentary consent is (Belgium, Germany, (although Finland
cabinet must win tacit (10 countries including Greece, Ireland, Italy, switched to positive
majority vote Finland) Luxembourg, and Spain) parliamentarism in
March 2000)
Parliamentary Some forms of Urgent, written, and/or oral Urgent, written, and/or Urgent, written, and
questions directed questions permitted, but questions. Restrictions apply oral questions. No oral questions have
at PM and cabinet restricted use in terms of time limits, restrictions on individual spread to more
ministers conditions, and/or number of MPs (Denmark, France, countries and become
tabled written questions per Ireland, The Netherlands, more frequently used
MP (11 countries) Portugal, and Sweden)
Ex post scrutiny of Limited scrutiny rights Standing committees with No ex post scrutiny by Increase in the right
cabinet affairs rights to scrutinize standing committees, and ability to scrutinize
government documents and which have legislative government affairs.
to hold hearings. Possibility roles only (Austria, Increasing use of
of setting up ad hoc Greece, and the ad hoc committees.
committees (11 countries) Netherlands). Scrutiny by
standing committees only
(Finland, Norway, and
Sweden)
No-confidence No-confidence motions No-confidence motions can No-confidence motions No significant trend
vote: Who can be can be directed against be directed against the PM can be directed against the
targeted the PM only or against individual PM only (France,
(with implications for ministers (11 countries) Germany, Ireland,
the whole cabinet under Portugal, Spain,
collective and the UK)
responsibility)
No-confidence Absolute majority vote, Simple majority Absolute majority Trend towards higher
vote: Decision rule which counts absentees (9 countries) (Belgium, France, thresholds; Sweden
in favour of the PM Germany, Greece, switched to absolute
Iceland, Portugal, majority in 1971 and
Spain, and Sweden) Belgium in 1995
Parliamentary None or purely Potential involvement Dissolution not permitted Decreasing
dissolution: ceremonial beyond pure formality (Norway); no role for involvement; Denmark
Involvement of (13 countries) Head of State (Denmark, and Sweden have
Head of State Sweden); purely formal completely abolished
role (Spain) (Monarch)
involvement
Parliamentary PM can dissolve PM/cabinet can dissolve Dissolution not permitted Weak trend in favour
dissolution: PM parliament parliament (counterbalances (Norway); dissolution of greater PM power
power parliamentary accountability powers placed primarily over dissolution, as in
power) and call an early (or with the Head of State Finland in 1991
extra) election. In most (Austria, Finland, France,
countries this at least Greece, Iceland, Italy, and
formally requires the Portugal)
consent of the Head of State
(9 countries)
Table 4.20. (contd.)
Delegation and/or Ideal-typical Most common (modal) Major empirical Cross-national trend,
accountability mechanism under mechanism in Western exceptions 1945–2000
mechanism party-based Europe, 2000
parliamentarism
Note: The most common mechanism often is different from the ideal-typical mechanism, which, for example, means that a trend in favour of the most common mechanism actually can be a
trend away from the mechanism suggested by the ideal type.
212 Torbjörn Bergman et al.
relationship is the reverse. At any rate, the large numbers of citizens and civil servants
involved implies serious issues of coordination, and both of these sets of delegation rela-
tionships tend to be fairly heavily and formally regulated. At the same time, however, the
relative ex post sanctions are fairly weak. Once politicians have been elected to office,
voters cannot easily oversee them, and many electoral systems allow no easy sanctions.
And while civil servants typically face very detailed contracts and extensive formal
screening, there is relatively little that their masters can do to punish them for non-
performance ex post, even though New Public Management techniques have increased
the politicians’capacities vis-à-vis their civil servants (see Peters and Pierre 2001).
The two intervening stages of delegation involve much smaller and more ‘intimate’
sets of professional politicians. In these relationships, continuous oversight is much
more feasible, and the immediate sanctions are much more dramatic and effective.
That does not mean that ex ante screening is unimportant (e.g. it clearly plays a critical
role in the recruitment of cabinet members) but it does imply a greater repertoire of
sanctions by which principals can keep their agents in line. At these stages of delega-
tion, principals are further helped by the existence of ever-present, and in most cases
cohesive, political parties. Thus, the most formidable challenges in contemporary
parliamentary delegation would seem to arise towards the beginning and end of the
chain.
Parliamentary democracy in Western Europe is not always what it appears to be, nor
does it necessarily conform to the principles of hierarchy, singularity, and indirect con-
trol. Certainly, many parliamentary constitutions contain a number of archaic features,
so that identifying their real operations may be far from straightforward. Yet, the issues
of political delegation and accountability are clearly visible through the patterns that
we have identified, and it is also clear that parliamentarism contains strong elements of
centralization and hierarchy. At the same time, no such picture can be timeless, and
nothing is constant but change.
From a bird’s-eye perspective, a number of well-known features of parliamentary
democracies appear to be weakening. While constitutional texts are slow to change,
there are a number of important changes in the working constitutions. As Table 4.20
shows, some institutional changes favour more direct involvement by the voters
(people) and their directly elected representatives (MPs). This is exemplified by the
opening up of candidate nomination processes, an increase in preference voting, more
scrutiny by parliamentary committees, and more use of parliamentary questions, as
well as by the spread of freedom of information laws.
Yet, other forces work in the opposite direction. On most of our seventeen indica-
tions, parliamentary democracies exhibit more fusion and hierarchy that they used to
a few decades ago. The cross-national trend favours unicameralism, positive parlia-
mentarism, higher thresholds for no-confidence votes, a more critical role for PMs (rel-
ative to the Head of State) in parliamentary dissolution, more staff and resources for
the PM to monitor ministers, but a lessening of the control of line ministers over civil
servants.
The evidence of party decline is more decisive, and particularly striking with respect
to the party-in-the-electorate (Chapter 22, see also Dalton and Wattenberg 2000).
Democratic Delegation and Accountability 213
But even the institutional features of parliamentary politics nowadays are less con-
ducive to centralized party control. Once on the ballot, contemporary candidates for
office are less certain that they will actually be chosen as MPs than was the case early
in the postwar period. Not only have the voters become less predictable with regard
to their party choice, more candidates are now subject to preference voting. The cross-
national tendency towards a more direct link between voters’ preferences and candi-
dates for office is the most important change in the relationship between voters and
representatives.
A continued erosion of political parties could weaken the chain of parliamentary
delegation. Weak and factionalized parties could diffuse authority and muddle account-
ability. Yet, so far, while voters have gained some influence over canditate selection,
while MPs have secured new vehicles of executive scrutiny, and while the public at least
potentially has been granted greater access to government documents, parties have
strengthened their finances and retain control of the core executive functions.
We now turn to an in-depth look at the parliamentary chain of delegation and
accountability in each of seventeen countries. This chapter has shown that even though
they share common core features, parliamentary systems are by no means all alike. A
serious assessment of their merits and demerits clearly requires more detailed descrip-
tion and a fuller analysis of how they work in practice. That is indeed the purpose of
the next seventeen chapters, which collectively will lay out country features, describe
the constraints on national policy-making, and discuss the agency relations embodied
in their living constitutions. In the final two chapters, we discuss two devices that may
protect against agency loss in the long chain of parliamentary delegation, namely polit-
ical parties and external constraints.
REFERENCES
Aberbach, Joel D., Putnam, Robert D., and Rockman, Bert A. (1981). Bureaucrats and Politicians in
Western Democracies. Cambridge, MA: Harvard University Press.
Andeweg, Rudy B. (1988). ‘Centrifugal Forces and Collective Decision-Making: The Case of the
Dutch Cabinet’. European Journal of Political Research, 16: 125–51.
—— (1991). ‘The Dutch Prime Minister: Not Just Chairman, Not Yet Chief ?’ West European
Politics, 14(1): 16–32.
—— (2000). ‘Ministers as Double Agents? The Delegation Process Between Cabinet and
Ministers’. European Journal of Political Research, 37: 377–95.
—— and Nijzink, Lia (1995). ‘Beyond the Two-Body Image: Relations Between Ministers and
MPs’, in Herbert Döring (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt am
Main: Campus and New York: St. Martin’s Press.
Arter, David (1991). ‘One Thing Too Many: The Shift to Unicameralism in Denmark’, in
Lawrence D. Longley and David M. Olson (eds.), Two Into One. The Politics and Processes of
National Legislative Cameral Change. Boulder: Westview Press.
Bagehot, Walter (1963 [1867]). The English Constitution. London: Fontana/Collins.
Bakema, Wilma E. (1991). ‘The Ministerial Career’, in Jean Blondel and Jean-Louis Thiébault
(eds.), The Profession of Government Minister in Western Europe. Houndmills: Macmillan.
214 Torbjörn Bergman et al.
Baldwin, Nicholas D. J. and Shell, Donald (eds.) (2001). Second Chambers. London: Frank Cass.
Bar, Antonio (1997). ‘Spain: A Prime Ministerial Government’, in Jean Blondel and Ferdinand
Müller-Rommel (eds.), Cabinets in Western Europe. Houndmills: Macmillan.
Bekke, Hans A. G. M and van der Meer, Frits M. (eds.) (2000). Civil Service Systems in Western
Europe. Cheltenham: Edward Elgar.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
—— (1995). Constitutional Rules and Party Goals in Coalition Formation. Umeå: Umeå University.
—— (1997). ‘National parliaments and EU Affairs Committees: Notes on Empirical Variation and
Competing Explanations’. Journal of European Public Policy, 4: 373–87.
—— (2000). ‘The European Union as the Next Step of Delegation and Accountability’. European
Journal of Political Research, 37: 415–29.
—— and Erik Damgaard (eds.) (2000). Delegation and Accountability in European Integration: The
Nordic Parliamentary Democracies and the European Union. London: Frank Cass (Also available as
a special issue of Journal of Legislative Studies, 6(1) ).
Bille, Lars (2001). ‘Democratizing a Democratic Procedure: Myth or Reality?’ Party Politics,
7: 363–80.
Blondel, Jean (1991). ‘The Post-Ministerial Careers’, in Jean Blondel and Jean-Louis Thiébault
(eds.), The Profession of Government Minister in Western Europe. Houndmills: Macmillan.
—— and Müller-Rommel, Ferdinand (eds.) (1997). Cabinets in Western Europe, 2nd edn.
Houndmills: Macmillan.
Bogdanor, Vernon (1984). ‘The Government Formation Process in the constitutional Monarchies
of North-West Europe’, in Dennis Kavanagh and Gillian Peele (eds.), Comparative Government
and Politics. London: Heinemann.
—— (ed.) (1988). Constitutions in Democratic Politics. Aldershot: Gower.
Bowler, Shaun, Farrell, David M., and Katz, Richard S. (eds.) (1999). Party Discipline and
Parliamentary Government. Columbus: Ohio State University Press.
Brazier, Rodney (1997). Ministers of the Crown. Oxford: Clarendon Press.
Brusewitz, Axel (1929). ‘Vad menas med parlamentarism.’ Statsvetenskaplig Tidskrift, 32:
323–34.
Choe, Yonhyok (1997). How to Manage Free and Fair Elections. Göteborg: Göteborg University.
Chubb, Basil (1982). The Government and Politics of Ireland, 2nd edn. London: Longman.
Cole, Alistair M. (1989). ‘Factionalism, the French Socialist Party and the Fifth Republic: An
Explanation of Intra-Party Divisions’. European Journal of Political Research, 17: 77–94.
Cowley, Philip (2000). ‘Can Sheep Bark? British Labour MPs and the Modification of Government
Policy’. ECPR paper, Copenhagen, 14–19 April.
Cox, Gary W. (1997). Making Votes Count. Cambridge: Cambridge University Press.
Cotta, Maurizio (1997). ‘Italy: A Fragmented Government’, in Jean Blondel and Ferdinand
Müller-Rommel (eds.), Cabinets in Western Europe. Houndmills: Macmillan.
Crewe, Ivor and King, Anthony (1994). ‘Are British Elections Becoming More “Presidential”?’, in
M. Kent Jennings and Warren E. Miller (eds.), Elections at Home and Abroad. Ann Arbor:
University of Michigan Press.
Criscitiello, Annarita (1994). ‘The Political Role of Cabinet Ministers in Italy’, in Michael Laver
and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge:
Cambridge University Press.
De Winter, Lieven (1991). ‘Parliamentary and Party Pathways to the Cabinet’, in Jean Blondel
and Jean-Louis Thiébault (eds.), The Profession of Government Minister in Western Europe.
Houndmills: Macmillan.
Democratic Delegation and Accountability 215
De Winter, Lieven (1995). ‘The Role of Parliament in Government Formation and Resignation’,
in Herbert Döring (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt am Main:
Campus and New York: St. Martin’s Press.
Diermeier, Daniel and Feddersen, Timothy J. (1998). ‘Cohesion in Legislatures and the Vote of
Confidence Procedure’. American Political Science Review, 92: 611–21.
Dogan, Mattei (1989). ‘How to Become a Cabinet Minister in Italy: Unwritten Rules of the
Political Game’, in Mattei Dogan (ed.), Pathways to Power. Boulder: Westview Press.
Dunleavy, Patrick with Bastow, Simon (2001). ‘Modelling Coalitions that Cannot Coalesce: A
Critique of the Laver–Shepsle Approach’. West European Politics, 24(1): 1–26.
Dunleavy, Patrick and Rhodes, R. A. W. (1990). ‘Core Executive Studies in Britain’. Public
Administration, 68: 3–28.
Duverger, Maurice (1980). ‘A New Political System Model: Semi-Presidential Government’.
European Journal of Political Research, 8: 165–87.
Elgie, Robert (ed.) (1999). Semi-Presidentialism in Europe. Oxford: Oxford University Press.
—— and Machin, Howard (1991). ‘France: The Limits of Prime-ministerial Government in a
Semi-presidential System’. West European Politics, 14(2): 62–78.
Elklit, Jørgen and Svensson, Palle (2001). ‘What Makes Elections Free and Fair?’, in Larry
Diamond and Mark F. Plattner (eds.), The Global Divergence of Democracies. Baltimore: Johns
Hopkins University Press.
Eriksen, Svein (1997). ‘Norway: Ministerial Autonomy and Collective Responsibility’, in Jean Blondel
and Ferdinand Müller-Rommel (eds.), Cabinets in Western Europe. Houndmills: Macmillan.
Farrell, Brian (1994) ‘The Political Role of Cabinet Ministers in Ireland’, in Michael Laver and
Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge:
Cambridge University Press.
Finer, Herman (1946). The Theory and Practice of Modern Government, 2nd edn. London: Methuen.
Frognier, André-Paul (1997). ‘Belgium: A Complex Cabinet in a Fragmented Polity’, in Jean Blondel
and Ferdinand Müller-Rommel (eds.), Cabinets in Western Europe. Houndmills: Macmillan.
Fysh, Peter (1996). ‘Candidates and Parties of the Right’, in Robert Elgie (ed.), Electing the French
President. The 1995 Presidential Election. Houndmills: Macmillan.
Gallagher, Michael and Marsh, Michael (eds.) (1988). Candidate Selection in Comparative Perspective.
London: Sage.
Gallagher, Michael (1988). ‘Conclusion’, in Michael Gallagher and Michael Marsh (eds.),
Candidate Selection in Comparative Perspective. London: Sage.
Gerlich, Peter, Müller, Wolfgang C., and Philipp, Wilfried (1988). ‘Potentials and Limitations of
Executive Leadership: The Austrian Cabinet Since 1945’. European Journal of Political Research,
16: 191–205.
Golden, Miriam A. and Chang, Eric C. C. (2001). ‘Competitive Corruption. Factional Conflict
and Political Malfeasance in Postwar Italian Christian Democracy’. World Politics, 53: 588–622.
Headey, Bruce (1974). British Cabinet Ministers. London: George Allen & Unwin.
Helms, Ludger (2001). ‘The Changing Chancellorship: Resources and Constraints Revisited’.
German Politics, 10: 155–68.
Hine, David (1993). Governing Italy. Oxford: Oxford University Press.
—— and Finocchi, Renato (1991). ‘The Italian Prime Minister’. West European Politics, 14(2):
79–96.
Huber, John D. (1996a). ‘The Vote of Confidence in Parliamentary Democracies’. American
Political Science Review, 90: 269–82.
—— (1996b). Rationalizing Parliament. Legislative Institutions and Party Politics in France.
Cambridge: Cambridge University Press.
216 Torbjörn Bergman et al.
Huber, John D. (2000). ‘Delegation to Civil Servants in Parliamentary Democracies’. European
Journal of Political Research, 37: 397–413.
Huber, John D. and McCarty, Nolan (2001). ‘Cabinet Decision Rules and Political Uncertainty in
Parliamentary Bargaining’. American Political Science Review, 95: 345–60.
—— Shipan, Charles R., and Pfahler, Madelaine (2001). ‘Legislatures and Statutory Control of
Bureaucracy’. American Journal of Political Science, 45: 330–45.
Hug, Simon (2001). Altering Party Systems. Ann Arbor: University of Michigan Press.
Jensen, Torben K. (2000). ‘Party Cohesion’, in Peter Esaiasson and Knut Heidar (eds.), Beyond
Westminister and Congress: The Nordic Experience. Columbus: Ohio State University Press.
Jones, G. W. (1995). ‘The Downfall of Margaret Thatcher’, in R. A. W. Rhodes and Patrick
Dunleavy (eds.), Prime Minister, Cabinet and Core Executive. Houndmills: Macmillan.
Katz, Richard S. (1997). Democracy and Elections. Oxford: Oxford University Press.
—— (1999). ‘Representation, the Locus of Democratic Legitimation and the Role of the National
Parliaments in the European Union’, in Richard S. Katz and Bernhard Wessels (eds.), The
European Parliament, the National parliaments, and European Integration. Oxford: Oxford
University Press.
—— and Mair, Peter (eds.) (1992). Party Organizations. London: Sage.
—— —— (1995). ‘Changing Models of Party Organization and Party Democracy’. Party Politics,
1: 5–28.
King, Anthony (1975). ‘Executives’, in Fred I. Greenstein and Nelson W. Polsby (eds.), Handbook
of Political Science. Volume 5. Governmental Institutions and Processes. Reading, MA: Addison-
Wesley.
—— (1991). ‘The British Prime Ministership in the Age of the Career Politician’. West European
Politics, 14(2): 25–47.
—— (1994). ‘Ministerial Autonomy in Britain’, in Michael Laver and Kenneth A. Shepsle (eds.),
Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge University
Press.
—— (ed.) (2002). Leaders’ Personalities and the Outcomes of Democratic Elections. Oxford: Oxford
University Press.
Knapp, Andrew (2004). ‘Two Pyrrhic Victories? France’s Governing Parties, the Ecologists, and
the Far Right’, in Peter Mair, Wolfgang C. Müller, and Fritz Plasser (eds.), Political Parties and
Electoral Change: Party Responses to Electoral Markets. London: Sage Publications.
Koutsoukis, Klemenis S. (1994). ‘Cabinet Decision Making in the Hellenic Republic, 1974–1992’,
in Michael Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government.
Cambridge: Cambridge University Press.
Krehbiel, Keith (1993). ‘Where’s the Party?’, British Journal of Political Science, 23: 235–66.
Larsson, Torbjörn (1993). ‘The Role and Position of Ministers of Finance’, in Jean Blondel and
Ferdinand Müller-Rommel (eds.), Governing Together. New York: St. Martin’s Press.
—— (1994). ‘Cabinet Ministers and Parliamentary Government in Sweden’, in Michael Laver and
Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge:
Cambridge University Press.
—— (1997). ‘Sweden: The New Constitution—An Old Practice Adjusted’, in Jean Blondel and
Ferdinand Müller-Rommel (eds.), Cabinets in Western Europe. Houndmills: Macmillan.
Laver, Michael and Schofield, Norman (1990). Multiparty Government: The Politics of Coalition in
Europe. Oxford: Oxford University Press.
—— and Shepsle, Kenneth A. (1996). Making and Breaking Governments. Cambridge: Cambridge
University Press.
Lijphart, Arend (1999). Patterns of Democracy. New Haven: Yale University Press.
Democratic Delegation and Accountability 217
Linton, Martin and Southcott, Mary (1998). Making Votes Count. The Case for Electoral Reform.
London: Profile Books.
Loewenberg, Gerhard and Patterson, Samuel C. (1979). Comparing Legislatures. Boston: Little,
Brown.
Longley, Lawrence D. and Olson, David M. (eds.) (1991). Two Into One. The Politics and Processes of
National Legislative Cameral Change. Boulder: Westview Press.
Lupia, Arthur W. and Strøm, Kaare (1995). ‘Coalition Termination and the Strategic Timing of
Parliamentary Elections’. American Political Science Review, 89: 648–65.
Lutz, Donald S. (1994). ‘Toward a Theory of constitutional Amendment’. American Political
Science Review, 88: 355–70.
Mair, Peter (1997). Party System Change. Oxford: Oxford University Press.
Mann, Thomas E. and Orren, Garry R. (eds.) (1992). Media Polls in American Politics. Washington,
DC: Brookings.
Mayhew, David (1974). Congress: The Electoral Connection. New Haven: Yale University Press.
McCubbins, Mathew D. and Schwartz, Thomas (1984). ‘Congressional Oversight Overlooked:
Police Patrols Versus Fire Alarms’. American Journal of Political Science, 28: 165–79.
Mény, Yves and Knapp, Andrew (1998). Government and Politics in Western Europe. Oxford: Oxford
University Press.
Miliband, Ralph (1969). The State in Capitalist Society. New York: Basic Books.
Mitchell, Paul L. (2000). ‘Voters and their Representatives: Electoral Institutions and Delegation
in Parliamentary Democracies’. European Journal of Political Research, 37: 335–51.
Molina, Ignacio (1999). ‘Spain: Still the Primacy of Corporatism’, in Edward C. Page and
Vincent Wright (eds.), Bureaucratic Élites in Western European States. Oxford: Oxford University
Press.
Morstein Marx, Fritz (1963). ‘The Higher Civil Service as an Action Group in Western Political
Development’, in Joseph LaPalombara (ed.), Bureaucracy and Political Development. Princeton:
Princeton University Press.
Mueller, Dennis C. (1989). Public Choice II. Cambridge: Cambridge University Press.
Müller, Wolfgang C. (1993). ‘The Relevance of the State for Party System Change’. Journal of
Theoretical Politics, 5: 419–54.
—— (1994). ‘Models of Government and the Austrian Cabinet’, in Michael Laver and Kenneth A.
Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge
University Press.
—— (2000a). ‘Political Parties in Parliamentary Democracies: Making Delegation and
Accountability Work’. European Journal of Political Research, 37: 309–33.
—— (2000b). ‘Patronage by National Governments’, in Jean Blondel and Maurizio Cotta (eds.),
The Nature of Party Government. Houndmills: Palgrave-Macmillan.
—— (2002). ‘Parties and the Institutional Framework’, in Kurt Richard Luther and Ferdinand
Müller-Rommel (eds.), Political Parties in the New Europe. Oxford: Oxford University Press.
—— Jenny, Marcelo, Steininger, Barbara, Dolezal, Martin, Philipp, Wilfried, and Preisl-Westphal,
Sabine (2001). Die österreichischen Abgeordneten. Vienna: WUV Universitätsverlag.
—— and Strøm, Kaare (2000a). ‘Conclusion: Coalition Governance in Western Europe’, in
Wolfgang C. Müller and Kaare Strøm (eds.), Coalition Governments in Western Europe. Oxford:
Oxford University Press.
—— —— (eds.) (2000b). Coalition Governments in Western Europe. Oxford: Oxford University Press.
Müller-Rommel, Ferdinand (1994). ‘The Role of German Ministers in Cabinet Decision Making’,
in Michael Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government.
Cambridge: Cambridge University Press.
218 Torbjörn Bergman et al.
Müller-Wirth, Moritz (1992). Die Debatte um die Parlamentsreform in Italien 1971–1988. Frankfurt am
Main: Peter Lang.
Nassmacher, Karl-Heinz (ed.) (2001). Foundations for Democracy. Approaches to Comparative Political
Finance. Baden-Baden: Nomos.
Nassmacher, Karl-Heinz (2002). ‘Die Kosten der Parteitätigkeit in westlichen Demokratien.’
Österreichische Zeitschrift für Politikwissenschaft, 31: 7–20.
Nexø Jensen, Hanne and Knudsen, Tim (1999). ‘Senior Officials in the Danish Central
Administration: From Bureaucrats to Policy Professionals and Managers’? in Edward C. Page
and Vincent Wright (eds.), Bureaucratic Élites in Western European States. Oxford: Oxford
University Press.
Nordhaus, William A. (1975). ‘The Political Business Cycle’. Review of Economic Studies, 42:
169–90.
Nousiainen, Jaakko (1994). ‘Finland: Ministerial Autonomy, Constitutional Collectivism, and
Party Oligarchy’, in Michael Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and
Parliamentary Government. Cambridge: Cambridge University Press.
—— (1997). ‘Finland’, in Jean Blondel and Ferdinand Müller-Rommel (eds.), Cabinets in Western
Europe. Houndmills: Macmillan.
Norris, Pippa (1997). ‘Conclusions: Comparing Passages to Power’, in Pippa Norris (ed.), Passages
to Power. Cambridge: Cambridge University Press.
—— (2001). ‘The Twilight of Westminster? Electoral Reform and its Consequences’. Political
Studies, 49: 877–900.
O’Leary, Brendan (1991). ‘An Taoiseach: The Irish Prime Minister’. West European Politics, 14(2):
133–62.
Olsen, Johan P. (1983). Organized Democracy: Political Institutions in a Welfare State—The Case of
Norway. Bergen: Universitetsforlaget.
Padgett, Steven (1993). ‘The Chancellor and his Party’, in Stephen Padgett (ed.), Adenauer to Kohl.
London: Hurst.
Page, Edward C. (1985). Political Authority and Bureaucratic Power. Brighton: Wheatsheaf Books.
—— and Wright, Vincent (eds.) (1999). Bureaucratic Élites in Western European States. Oxford:
Oxford University Press.
Paltiel, Khayyam Zev (1981). ‘Campaign Finance: Contrasting Practices and Reforms’, in David
Butler, Howard R. Penniman, and Austin Ranney (eds.), Democracy at the Polls. Washington,
DC American Enterprise Institute.
Patterson, Samuel C. and Mughan, Anthony (eds.) (1999). Senates. Bicameralism in the
Contemporary World. Columbus: Ohio State University Press.
Peters, B. Guy (1997). ‘Bureaucrats and Political Appointees in European Democracies: Who’s
Who and Does it Make Any Difference?’, in Ali Farazmand (ed.), Modern Systems of
Government. Thousand Oaks: Sage.
—— (2001). The Politics of Bureaucracy, 5th edn. London: Routledge.
—— and Pierre, Jon (eds.) (2001). Politicians, Bureaucrats, and Administrative Reform. London:
Routledge.
—— Rhodes, R. A. W., and Wright, Vincent (eds.) (2000). Administering the Summit. Administration
of the Core Executive in Developed Countries. Houndmills: Macmillan.
Pierre, Jon (1995). ‘Governing the Welfare State: Public Administration, the State and Society in
Sweden’, in Jon Pierre (ed.), Bureaucracy in the Modern State: An Introduction to Comparative Pubic
Administration. Cheltenham: Edward Elgar.
—— Svåsand, Lars, and Widfeldt, Anders (2000). ‘State Subsidies to Political Parties: Confronting
Rhetoric with Reality’. West European Politics, 23(3): 1–24.
Democratic Delegation and Accountability 219
Pollitt, Christopher, Girre, Xavier, Lonsdale, Jeremy, Mul, Robert, Summa, Hilkka, and
Waerness, Marit (1999). Performance of Compliance? Performance Audit and Public Management in
Five Countries. Oxford: Oxford University Press.
Polsby, Nelson W. (1975). ‘Legislatures’, in Fred I. Greenstein and Nelson W. Polsby (eds.),
Handbook of Political Science. Volume 5. Governmental Institutions and Processes. Reading, MA:
Addison-Wesley.
Poguntke, Thomas (1999). ‘The Winner Takes All: The FDP in 1982–1983: Maximizing Votes,
Office, and Policy?’, in Wolfgang C. Müller and Kaare Strøm (eds.), Policy, Office, or Votes?
Cambridge: Cambridge University Press.
Rasch, Bjørn Erik (1995). ‘Parliamentary Voting Procedures’, in Herbert Döring (ed.), Parliaments
and Majority Rule in Western Europe. New York: St. Martin’s Press and Frankfurt am Main:
Campus.
—— (2001). ‘Parlamentariske instrumenter og minoritetsparlamentarisme’, Vedlegg 3, Norges
offentlige utredninger (NOU) 2001: 3. Velgere, valgordning, valgte. Avgitt til Kommunal- og
regionaldepartementet 30. januar 2001.
Raunio, Tapio and Hix, Simon (2000). ‘Backbenchers Learn to Fight Back: European Integration
and Parliamentary Government’. West European Politics, 23: 142–68.
—— and Wiberg, Matti (2000). ‘Does Support Lead to Ignorance? National parliaments and the
Legitimacy of EU Governance.’ Acta Politica, 35: 146–68.
Riker, William H. (1986). The Art of Political Manipulation. New Haven: Yale University
Press.
Rose, Richard (1975). ‘The Making of Cabinet Ministers’, in Valentine Herman and James E. Alt
(eds.), Cabinet Studies. London: Macmillan.
—— (1976). The Problem of Party Government. Harmondsworth: Penguin.
—— (1991). ‘Prime Ministers in Parliamentary Democracies’. West European Politics, 14(2): 9–24.
Rose-Ackerman, Susan (1999). Corruption and Government. Cambridge: Cambridge University
Press.
Rush, Michael (2001). The Role of the Member of Parliament Since 1868. Oxford: Oxford University
Press.
Russell, Meg (2000). Reforming the House of Lords. Lessons from Overseas. Oxford: Oxford University
Press.
Saalfeld, Thomas (1995a). ‘On Dogs and Whips: Recorded Votes’, in Herbert Döring (ed.),
Parliaments and Majority Rule in Western Europe. New York: St. Martin’s Press and Frankfurt am
Main: Campus.
—— (1995b). Parteisoldaten und Rebellen. Opladen: Leske ⫹ Budrich.
Sartori, Giovanni (1976). Parties and Party Systems. Cambridge: Cambridge University Press.
—— (1997). Comparative Constitutional Engineering. Houndmills: Macmillan.
Schmidt, Manfred G. (1996). ‘Germany. The Grand Coalition State’, in Josep M. Colomer (ed.),
Political Institutions in Europe. London: Routledge.
Schou, Tove Lise (1997). ‘Denmark’, in Jean Blondel and Ferdinand Müller-Rommel (eds.),
Cabinets in Western Europe. Houndmills: Macmillan.
Self, Peter (1972). Administrative Theories and Politics. London: George Allen & Unwin.
Shugart, Matthew Soberg and Wattenberg, Martin P. (eds.) (2000). Mixed-Member Electoral
Systems. Oxford: Oxford University Press.
Smith, Gordon (1991). ‘The Resources of a German Chancellor’. West European Politics, 14(2):
48–61.
Somit, Albert, Wildenmann, Rudolph, Boll, Bernhard, and Römmele, Andrea (eds.) (1994). The
Victorious Incumbent: A Threat to Democracy? Aldershot: Dartmouth.
220 Torbjörn Bergman et al.
Steiner, Jürg and Dorff, Robert H. (1980). A Theory of Political Decision Modes. Chapel Hill:
University of North Carolina Press.
Strauss, E. (1961). The Ruling Servants. London: George Allen & Unwin.
Strøm, Kaare (1990). Minority Government and Majority Rule. Cambridge: Cambridge University
Press.
—— (1994). ‘The Role of Norwegian Cabinet Ministers’, in Michael Laver and Kenneth A.
Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge
University Press.
—— (2000). ‘Parties at the Core of Government’, in Russell J. Dalton and Martin P. Wattenberg
(eds.), Parties without Partisans. Oxford: Oxford University Press.
—— and Müller, Wolfgang C. (1999). ‘The Keys to Togetherness: Coalition Agreements in
Parliamentary Democracies’. Journal of Legislative Studies, 5(3/4): 255–82.
—— and Swindle, Stephen M. (2002). ‘Strategic Parliamentary Dissolution’. American Political
Science Review 96: 575–91.
Suleiman, Ezra N. (ed.) (1984). Bureaucrats and Policy Making. New York: Holmes & Meier.
Szarka, Joseph (1999). ‘The Parties of the French “Plural Left”: An Uneasy Complementary’.
West European Politics, 22(4): 20–37.
Taagepera, Rein and Shugart, Matthew S. (1989). Seats and Votes. Cambridge: Cambridge
University Press.
Thiébault, Jean-Louis (1994). ‘The Political Autonomy of Cabinet Ministers in the French Fifth
Republic’, in Michael Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary
Government. Cambridge: Cambridge University Press.
Tsebelis, George and Money, Jeannette (1997). Bicameralism. Cambridge: Cambridge University
Press.
Tufte, Edward R. (1978). Political Control of the Economy. Princeton: Princeton University Press.
Verzichelli, Luca (2003). ‘Much Ado About Something?’. Journal of Legislative Studies, 9(2): 35–55.
Verzichelli, Luca and Cotta, Maurizio (2000). ‘Italy: From “Constrained” Coalitions to
Alternating Governments?’ in Wolfgang C. Müller and Kaare Strøm (eds.), Coalition
Governments in Western Europe. Oxford: Oxford University Press.
Welan, Manfried (1976). ‘Die Kanzlerdemokratie in Österreich’, in Andreas Khol, Robert
Prantner, and Alfred Stirnemann (eds.), Um Partei und Parlament. Graz: Styria.
Wiberg, Matti (1995). ‘Parliamentary Questioning: Control by Communication?’, in Herbert
Döring (ed.), Parliaments and Majority Rule in Western Europe. New York: St. Martin’s Press and
Frankfurt am Main: Campus.
Willett, Thomas D. (ed.) (1988). Political Business Cycles. Durham: Duke University Press.
Woodhouse, Diana (1994). Ministers and Parliament. Accountability in Theory and Practice. Oxford:
Clarendon Press.
5
I N T RO D U C T I O N
Though the Austrian polity deviates from the ideal type of parliamentary democracy
as defined in this volume (see Chapter 1), in practice it has nevertheless functioned as
a parliamentary system. This is largely due to the strength of party politics, with cohes-
ive political parties being the real actors, using and abusing political institutions. This
chapter will argue that political parties ensure that delegation in the democratic chain
is successful (because all principals are better off than in a situation without parties), yet
they account for agency losses in each stage of the delegation process.
Parliamentarism has been introduced stepwise in Austria (Widder 1979). After the
aborted revolution of 1848 and the period of neoabsolutism, a process of gradual
parliamentarization began in 1861, though it was not completed during the Habsburg
monarchy. In the period 1918–20, assembly government prevailed. The 1920
Constitution established a regime that was almost an ideal-type parliamentary demo-
cracy. In 1929, the Constitution was substantially amended as a consequence of the
emergence of a strong anti-parliamentary, anti-party spirit. At the end of Second World
War, the United States and the Soviet occupation powers as well as the Austrian
Communists pressed for a new Constitution, while the Socialist Party called for a
return to the Constitution of 1920. Nonetheless the Constitution as amended in 1929
was reintroduced in December 1945, thus replacing the provisional Constitution of
April 1945. Since then, the formal framework in which politics take place has been
modified by numerous constitutional amendments and constitutional laws and clauses.
Its basic architecture has not, however, been substantially changed.1
Thanks to Cynthia Kite, Franz Fallend, Michael Schaden, Hubert Sickinger, Emmerich Tálos, and, of course,
my fellow editors for comments on earlier versions.
1
In addition to the Bundes-Verfassungsgesetz, which in this chapter is refered to as the Constitution, there
are other constitutional laws and clauses with constitutional status within ordinary statute. All ot these have
the same legal status. They are all equally binding. In 1993 constitutional lawyers counted no less than 63 con-
stitutional amendments, 225 separate constitutional laws, 427 clauses with constitutional rank in ordinary
legislation, and 181 state treaties with constitutionally relevant clauses enacted since 1945.
222 Wolfgang C. Müller
Constitutionally, Austria has since 1929 been semi-presidential, though in practice
the parliamentary side has continued to prevail (Müller 1999a). Parliamentarism is
constrained not only by a constitutionally strong President, but also by federalism,
a strong Constitutional Court, direct democracy, consociationalism, and corporatism.
The latter two features of postwar Austrian politics are not, however, embedded in the
constitutional framework. The parliamentary majority decides the extent to which it
will allow these arrangements to tie its hands.
Despite its small size Austria is a federal country. Nine Länder constitute a separate
level of government between national and local politics. The Constitution contains a
general competence clause in favour of the Länder, giving them jurisdiction over every-
thing that is not explicitly reserved for the national level, the Bund. Yet, the Constitution
has reserved the bulk of competencies for the national level (Werndl 1984). After the
transfer of competencies to the European Union, there is, de facto, hardly anything left
for the Länder. However, much national public administration is carried out through
the Land administrations. With regard to these matters Land administrations are sub-
ordinated to the federal ministries (Weber 1987).
With the exception of the budget, all types of draft legislation can be introduced in
parliament by either chamber, the cabinet, or the people (i.e. 100,000 citizens support-
ing a People’s Initiative). In practice, government proposals and private members’ bills
from (a minimum of five) Nationalrat MPs prevail.2 In order to be enacted, constitu-
tional amendments require a quorum of 50 per cent of MPs3 and the support of a
two-thirds majority of the votes cast. The same applies to the Nationalrat’s rules of
procedure (Article 30) and, since 1962, to school legislation (Article 14 and 14a), which
do not have constitutional status. Constitutional amendments are subject to a
referendum if one is demanded by one-third of the MPs in one of the two chambers
(Article 43). This has, however, never occurred. In contrast, a referendum is obligatory
for a ‘total revision’ of the Constitution (Gesamtänderung)4 (Article 43). This has
happened only once. In 1994 a referendum was held on the constitutional amendments
required for Austria’s accession to the European Union.
To pass, ordinary legislation requires a quorum of one-third of MPs and a majority
of the votes cast (50 per cent ⫹ 1). A referendum on a law is held before it comes into
force if the Nationalrat so decides (Article 43). Even this use of the referendum
has occurred only once: a 1978 referendum on the use of nuclear energy (Müller 1998).
The process of budget-making differs from normal law making in that it requires the
2
Before 1988 the minimum number of MPs was eight. Bundesrat proposals in parliamentary practice are
completely irrelevant. With few exceptions, the people’s initiative (Volksbegehren) has also been irrelevant with
regard to actual law-making. However, it has been quite relevant as a means for party competition (Müller
1998, 1999b). The bulk of laws originate from private member initiatives and government proposals. In prac-
tice, the distinction between these two ways of initiating a law is blurred by the fact that many private mem-
ber bills are de facto worked out in government departments.
3
If not otherwise stated, all references to MPs refer to those of in the Nationalrat.
4
The Constitution itself does not indicate what constitutes a ‘total revision’. Constitutional lawyers and
the Constitutional Court have interpreted this clause as referring to amendments that involve changes in one
or more of the guiding principles of the 1920 Constitution. These unquestionably include the principles of
democracy, federalism, the rule of law, the separation of powers, and liberalism (see Walter 1972: 101–13).
Austria 223
government to submit a proposal at least ten weeks before the end of the budget year.
MPs may propose a budget only if the cabinet fails to do so (Article 51).
Although Austria has a two-chamber system, which reflects its federal constitution,
it is a very asymmetrical one (Kathrein 1986; Schäffer 1999; Fallend 2000a). The upper
chamber, the Bundesrat, represents the Länder. Its members are elected by the Land
diets. The representation system in the Bundesrat takes into account the size of each
Land’s population and party strength. With regard to ordinary legislation passed by the
Nationalrat, the Bundesrat can issue only a suspensory veto. This veto can be overrid-
den by the Nationalrat with a quorum of 50 per cent of MPs and a majority of votes
cast. By vetoing legislation, the Bundesrat can delay it for, at most, about 2 months.
With regard to constitutional amendments, the Bundesrat is more powerful for two
reasons. First, since 1984, changes in the distribution of competencies between the
national level (Bund) and the Länder become effective only if the Bundesrat accepts
them with a quorum of 50 per cent and a two-thirds majority (Article 44). Second, it
can enforce a referendum on any constitutional amendment, if this is demanded by a
third of its members. In its own right, however, the Bundesrat cannot block constitu-
tional amendments that do not fall under Article 44. Moreover, there are important
decisions that do not involve the Bundesrat at all. In particular, it does not particip-
ate in budget-making and other financial acts (Article 42). The Nationalrat and the
Bundesrat are politically largely ‘congruent’ (Lijphart 1999: 207), in the sense
that the same parties or coalitions held majorities in both chambers. This is true despite
the fact that the representational systems are different enough to have produced different
plurality parties in the two chambers (during 19.5 years in the period 1945–2000, including
11.4 years when two parties in the Bundesrat were of equal strength). From 1945 to
2000 the two chambers were politically ‘incongruent’ only for 5.1 years. Nevertheless,
the Bundesrat’s limited political relevance is much more a consequence of its lack of
formal powers than a result of political congruence. In practice, political parties and
non-constitutional bodies consisting of Land government representatives are much
more important national channels of political influence for the Länder than is the
Bundesrat (Luther 1986, 1997; Dachs 1996).
At the lowest level of general rules in Austria is the government decree (Verordnung).
Such decrees are issued by individual ministers or, in rare cases, by the cabinet. Some
of them require the consent of Parliament’s main committee. According to the
Constitution (Article 18), the content of a government decree must, as a rule, already
be determined by law. However, the decree may reduce the degrees of freedom con-
tained in the law. It is also possible to introduce new rules (which add to those already
contained in the law) by decree, however, only if this is specifically authorized by con-
stitutional law (Koja 1986: 141–64).
PA RT I E S
It is hard to overstate the relevance of parties in Austria. All elections, with the partial
exception of local elections in small communities, are structured by parties. All office
holders in governmental institutions are party nominees, and, in almost all cases, they are
224 Wolfgang C. Müller
also members and functionaries of their respective parties. Moreover, the major parties
have built up huge mass organizations and have colonized large parts of the civil society.
Despite this, the Austrian Constitution mentions political parties only briefly and
primarily in a negative sense (by excluding their functionaries and employees from
being members of the Constitutional Court). Since 1945, however, official constitutional
doctrine as elaborated by academic lawyers and upheld by the courts is that, though the
Constitution does not prescribe the role of political parties, it proceeds from the
assumption that parties exist and are essential for the functioning of the constitutional
system. A 1975 party law contains a constitutional clause stating that the existence and
variety (‘Vielfalt’) of political parties comprise major elements of the democratic order
of the Republic of Austria, and defines their tasks very generally. While arguably these
constitutional provisions serve to legitimize party democracy, MPs and other politicians
are not bound constitutionally to their parties.
Before 1975 the actual legal form of the existing parties differed and depended,
de facto, on prevailing political power relations. Three parties, the Socialist (Social
Democratic since 1991) Party (SPÖ), the Austrian People’s Party (ÖVP), and the
Communist Party (KPÖ)—the parties that founded the Second Republic in 1945—
could afford to have an extra-constitutional status by not falling under one of the organ-
izational forms recognized by Austrian law. As opposed to these ‘historic’ parties,
younger parties like the League of Independents (VdU) and its successor, the Freedom
Party (FPÖ), assumed the legal form of an association (Verein) in order to avoid any risk
(such as being treated as an illegal organization by the authorities) created by their
extra-constitutional status. Since 1975, however, ‘political party’ is a legal form sui
generis. Party status requires the fulfilling of only minimal formal conditions and pro-
vides substantial advantages. To begin with, party status prevents state intervention in
internal affairs. Only the courts have the right to ban political parties,5 and they can do
so only for reasons set out in constitutional law. Currently, only Nazi activities come
under this rubric. Party status is also a necessary though not sufficient condition for ben-
efiting from major parts of Austria’s very generous system of public party finance.
However, whether parties get public money, and how much they get, depends entirely
on their electoral performance. With the exception of election years, in which subsidies
are generally higher, only parties with parliamentary representation receive public
funds (for the party organization, the parliamentary party, and their political education
institutions). In 1999, an election year, support amounted to more than 720 million ATS
(c. 52 million euro) at the national level. Over the typical 4 years’ electoral cycle, even
more money is given to parties at the subnational levels (Sickinger 1997, 2000a). The
allocation between parties is largely proportional, with a bias in favour of small parties.
This has, however, been reduced in recent years.
Throughout the postwar period the SPÖ and ÖVP have dominated political life.
With the exception of the immediate postwar period (in which the KPÖ was included
5
Strictly speaking the law considers unconstitutional parties as non-existent and public authorities treat
them accordingly. If this behaviour is challenged, the courts decide over the existence or non existence of
these parties.
Austria 225
in the cabinet) and the 1983–7 period (during which the FPÖ participated in government)
only these two parties held government office until 2000. They did so either in ‘grand
coalition’ governments—which were supported by a parliamentary base of up to 95 per
cent of the MPs—or in single-party governments (1966–83) (Müller 2000c). Even dur-
ing the period when the party system essentially functioned as a two-party system, high
levels of parliamentary cooperation between the major parties were maintained
(Müller 1993, 1997a). Nonetheless, the SPÖ and ÖVP have always competed vigorously
with one another, more often checking each other and competing with each other for
votes than working together in competition against opposition parties. Since 1986,
SPÖ–ÖVP dominance has been reduced considerably by the establishment of new par-
ties (the Greens in 1986 and the Liberal Forum in 1993) and in particular by the spec-
tacular rise of the FPÖ, which won fifty-two seats in 1999 and hence temporarily
gained equal status with the ÖVP (but fell back to a mere 18 seats in 2002). Following
the 1999 election the ÖVP and FPÖ joined forces in government, causing a major
change in the structure of party competition (Müller 2000d).
While the SPÖ has been a very cohesive, disciplined, and centralized party, the ÖVP
is more factionalized and decentralized. However, this difference has directly affected
only internal decision-making processes. In the parliamentary and governmental
arenas both parties have acted cohesively. Parliamentary voting cohesion is generally
very high. Indeed, no parliamentary battle has ever been won or lost as a result of a lack
of party voting cohesion in parliamentary parties.
The major interest groups traditionally have constituted another set of highly important
actors in Austrian politics. Some are public institutions (called ‘Chambers’). They are
established by law and automatically enrol all potential members. Chambers are legally
entitled to issue opinions on proposed legislation. However, the real influence of the
major interest groups on public decisions stems from their close link to the traditional
major parties, and, indeed, their ability to commit them to their positions in socio-
economic policy-making. Thus the major interest groups have as much influence as the
parliamentary majority is willing to grant them. This willingness has been considerably
reduced since the 1970s, reaching a new low with the ÖVP–FPÖ governments in office
since 2000.
D E L E GAT I O N F RO M VOT E R S TO M P S
Rules of Electoral Competition
Voters delegate to MPs under a set of rules which favour party politics and already
established parties. A minimum length for campaigns of roughly 1 month (20 days
before 1969) results from various clauses in the electoral law. However, in practice even
early elections that were called on short notice (e.g. in 1995) allowed for a campaign of
2 months. There are no restrictions on the publication of opinion polls during campaigns.
There has never been a legal limit on campaign spending. Free broadcasting time (abol-
ished in 2002) and financial subsidies from the state are available only to parties
that already have established themselves in Parliament. Only in elections years do
226 Wolfgang C. Müller
non-parliamentary parties which have won at least 1 per cent of the votes get a refund
towards their campaign costs. Thus the parties with parliamentary representation enjoy
a competitive advantage.
6
Other constitutional means to shorten the term of the Nationalrat exist, but have not been used in practice.
Austria 227
From 1945 until 1970, twenty-five electoral districts existed. One hundred and sixty
five seats were distributed among the parties participating in the election in two stages,
the first of which operated within electoral districts. In the second round the electoral
districts were aggregated into four units, and parties’ unused votes were counted,
provided that they had received at least one seat in the first stage (a Grundmandat). Three
features tended to make this electoral system a relatively ‘strong’ one (cf. Sartori 1997:
42–4). First, the electoral system worked against small parties since it established a sub-
stantial threshold for parliamentary representation—the requirement that parties must
win a seat in at least one of the twenty-five relatively small electoral districts. Second,
apportionment among electoral districts was (and still is) based on the number of
citizens rather than those eligible to vote, thus favouring rural areas (and hence the
ÖVP), since they—as a rule—include a higher proportion of children. Third, the fact
that the seats distributed in the first stage were cheaper in votes than those distributed
in the second stage7 favoured the big parties, because they received a higher proportion
of their seats in the first stage. Thus this electoral system served the interests of the two
major parties in general, and the ÖVP in particular, effectively shielding them against
competitors. More specifically, the ÖVP’s plurality in 1953 and 1959 and its majority in
1966 were manufactured by the electoral system.
In 1970, the SPÖ and FPÖ joined forces to pass a new electoral law, which came into
force in 1971. The number of electoral districts was reduced to nine, whose boundaries
were identical with the Länder. The number of parliamentary seats was increased from
165 to 183, and in the second stage of seat distribution the electoral districts were aggreg-
ated into only two units. Not surprisingly, this reform benefited the two parties that
introduced it as well as new parties (which faced the same problem as the FPÖ). First,
it lowered the threshold of representation by reducing the number of electoral districts
and increasing the number of seats. Second, it reduced the impact of distributing seats
according to population rather than voters by uniting rural and urban areas in larger
electoral districts. Third, it removed the advantage of parties that receive a larger
proportion of their seats in the first round of seat distribution (i.e. the big parties) by
making these seats more expensive in votes.8
In 1992, SPÖ and ÖVP—which had been governing together since 1987—introduced
a new electoral law. It divided the nine electoral districts into forty-three regional electoral
districts (Regionalwahlkreise). Seats are distributed in three stages. First, at the regional
electoral district level. In the second stage, at the Land electoral district level. The cost of
seats in these two stages is identical within each Land. In the third stage, seats are distrib-
uted at the national level. Only parties that win either one or more seats at the regional
electoral district level or at least 4 per cent of the valid votes nationwide participate
in the second and third stages. Seats already distributed at the regional and Land levels
are deducted from the number of seats calculated at the national level. The 4 per cent
barrier has raised the threshold for parliamentary representation. In 1999 the Liberal
7
Until 1970, in the first round in each electoral district the number of votes required for a seat was
calculated by the Hagenbach–Bischoff system, while the d’Hondt system was used in the second round.
8
Since 1971, seats in the first round are distributed by the Hare system, while the d’Hondt system
continues to be used in the second (or, since 1992, third) round.
228 Wolfgang C. Müller
Forum missed the required 4 per cent (by 0.35 per cent) and hence lost parliamentary
representation, while it would have won several seats under the previous electoral system.
9
Vranitzky was much more popular than his party and leading his party’s lists in all electoral districts he
did not depend on preference votes for election to Parliament. Eventually he received enough preference
votes to give him seventeen seats (Müller 1990; Müller and Plasser 1992).
Austria 229
candidate was elected by preference votes. Nevertheless, under the 1992 electoral
system there has been an increase in intraparty competition, with the preference votes
counting in the next candidate selection.
The empirical record of preference voting indicates that Austrian voters think very
much in terms of parties rather than individual candidates. Survey evidence compiled
since 1980 confirms this. Nonetheless, the increased use of preference voting and the
candidates’ competition for preference votes has been accompanied by an increase in
the role that candidates play in influencing voters’ decisions.
In sum, until 1992 the party organizations were virtually free to determine who
would obtain a seat in parliament. The 1992 reform has not substantially reduced that
freedom. Thus, intraparty candidate selection remains much more decisive than the
general election for parliamentary representation.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
Cabinet Installment
According to the Constitution the directly elected federal President appoints the federal
Chancellor and, at the proposal of the Chancellor, the other cabinet members and
secretaries of state (junior ministers). The Constitution contains no restrictions what-
soever as to who should be invited to form a government or on the government’s political
composition. However, de facto, the President has little discretion. The government
needs parliamentary support in order to survive and govern effectively. The President
therefore refrains from appointing a cabinet that is opposed by a parliamentary major-
ity. This, in theory, might still leave some leeway for the President. Potentially, presi-
dential influence increases when alternative majority coalitions can be formed and
the coalition preferences of the parties are not (yet) fixed. In practice, however, the
President largely has limited himself to ratifying whatever the parties agree to. This
means that the President is occasionally pre-empted by public party declarations or
even pre-electoral commitments. It is well-established convention that the task of
forming a government goes first to the leader or ‘Chancellor candidate’ of the largest
parliamentary party (Welan 1997: 47). With the single exception of 2000, the largest
party (and, with one exception, also the first person officially designated the Chancellor
candidate by the President) has always succeeded in forming a viable government. On
several occasions the President intervened in the government formation process; while
exercising some influence, he was not decisive (Müller 1999a). The most remarkable
presidential intervention occurred in the 1999/2000 government formation. It ended by
the President swearing in a cabinet he had worked hard to prevent. Indeed, he had
never given the formal task of forming a government to the new Chancellor Wolfgang
Schüssel. However, the President exercised some influence. He pushed for a preamble
to the coalition agreement that obliged the coalition parties to some principles he
wanted the government to observe and he rejected two FPÖ ministerial nominees,
which no President had done since 1949.
Once it is formed and sworn in, the cabinet must present itself to Parliament (the
Nationalrat) within one week after assuming office (Article 70, see Ringhofer 1977:
229). The cabinet does not have to go through an investiture vote; however, it must pres-
ent an outline of its programme and, by convention, this is followed by a debate. This
occasion can also be used to introduce a vote of no confidence. To unseat a government
requires a parliamentary majority against it (for details see below). Austria is thus an
example of ‘negative’ parliamentarism, as the burden of proof rests with the opposi-
tion (cf. Bergman 1993). Since 1945 only one government—the ÖVP–FPÖ government
formed in 2000—had to face a vote of no confidence in connection with its first appear-
ance before Parliament. In addition, only once was an individual minister challenged—
unsuccessfully—on such an occasion (Table 5.1).
Table 5.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Austria 1945–2000 a
No Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the Who dissolved
formation votes (iv) votes (cv) cabinet end Parliament in this
n.a. No. of Cabinet Cabinet under specific with an early case?b
ncv removed resigned to institutional election?
by ncv pre-empt ncv mechanism
only
n.a.
postwar period. For the parliamentary party groups of the government parties an informal
ex ante mechanism is equally important: their policy experts’ participation in negotiating
draft legislation in the coalition (Müller 1993; Sickinger 2000b; Müller et al. 2001; Tálos
and Kittel 2001).
A reform of the Parliament’s rules of procedure in 1975 made it possible for
a parliamentary minority to launch an appeal to the Constitutional Court in order
to exercise abstract norm control. One-third of the MPs can bring any law before the
Constitutional Court, which has the power to invalidate laws that violate the principles
of the Constitution. The decisions that subsequently come from the Court are made
according to a decision rule that, by its very nature, is quite different from Parliament’s.
The possibility of an appeal to the Constitutional Court has actually proven to be
a quite effective means of control. Not only have several laws been invalidated, the
parliamentary majority now tries to anticipate potential appeals by the minority and to
pre-empt them by keeping them in mind during the process of drafting legislation.
Given tight party discipline in Parliament, appeals to the Constitutional Court, with a
single exception, were only launched when the parliamentary opposition held a minimum
of a third of the seats. Thus this instrument was not important under grand has already
faced government. In contrast, the incumbent ÖVP–FPÖ coalition has already faced
more opposition appeals to the Constitutional Court than any previous government.
Parliament has the authority to pass resolutions demanding specific action on the
part of cabinet or cabinet members (Article 52). Although these are not actually legally
binding, cabinet members’ parliamentary accountability makes them, de facto, bind-
ing. Between 1971 and 1999, Parliament passed 730 such resolutions, some of which
were of great importance—for example, the resolution that fixed the course of
Austria’s European Union (EU) integration policy (Fischer 1997: 110–1).
Cabinet Accountability
The political accountability of the cabinet is explicitly regulated in the Constitution
(Article 74) and in the rules of procedure. If the Nationalrat passes a vote of no confidence,
the cabinet (or individual members who have lost Parliament’s confidence) must be
dismissed by the President (Welan 1967). A motion of no confidence must be sponsored
by five MPs (eight before 1988). To pass a vote of no confidence requires a quorum of
50 per cent of MPs and a majority of the votes cast. The cabinet is not accountable to
the second chamber, the Bundesrat.
The opposition has regularly introduced motions of no confidence since the
mid-1960s to criticize general government policy, particular policies, or individual cabinet
members. Table 5.1 summarizes the behavioural record of no-confidence votes (directed
against the cabinet or the Chancellor) and parliamentary dissolution.10 In light of the
fact that all but one of the cabinets were majority governments, that coalition
10
Note that two no-confidence notions were directed against the Vice-Chancellor and leader of the second
government party under cabinets Sinowatz and Klima, respectively.
236 Wolfgang C. Müller
discipline guides voting in parliament (Müller 2000c), and that parliamentary parties are
highly cohesive, it is not surprising that all motions of no confidence have failed.
Cabinet members can be held accountable through the Constitutional Court for
contravening the law (Ministeranklage) (Articles 142 and 143). The act of invoking the
Constitutional Court in such cases is political, but the procedure itself judicial (Korinek
1986; Schambeck 1997: 72–80). To bring a minister before the court, a quorum of
50 per cent of MPs must be present in plenary session and a majority of votes cast must
support doing so. Given these requirements, it is not surprising that this instrument has
had no practical relevance in the postwar period.
In a party democracy, it is not surprising that ‘sanctioned’ controls have been less
effective than ‘unsanctioned’ mechanisms. The cabinet can be held accountable ex post
through a range of parliamentary instruments, including several types of questions,
formal reporting requirements, committees of investigation, checks by the audit office,
and reports by the ombudsman office.11
Over the postwar period the instruments by which cabinets can be held accountable
have been reformed several times (Auracher-Jäger 1997; Atzwanger and Zögernitz
1999). The rules of parliamentary procedure originally restricted the opposition’s
opportunity to effectively criticize the government and present alternatives in a way that
would attract broad public attention. This has changed since 1961 due to the introduc-
tion of new parliamentary instruments, primarily ones of importance to the opposi-
tion, and because the number of MPs needed to invoke these instruments has been
reduced over time. This trend has continued increasingly, but it has been accompanied
by efforts to ration the use of these instruments (Schefbeck 1997).
Written questions, which must be sponsored by five MPs, have throughout the post-
war period proved the most important means for scrutinizing the executive. The
number of parliamentary questions has increased enormously, particularly as a result
of the termination of ‘grand coalition’ governments in 1966, an event which—for the
first time—led to the emergence of a strong parliamentary opposition. Interestingly,
the return to grand coalition government in 1987 was also followed by a major increase
in parliamentary questions. Compared to the 1971–86 period, the number of written
questions per month increased by 183 per cent between 1986 and 1996. In the 1996–9
term more than 6,700 written questions were asked.
The second type of question that has existed throughout the postwar period is
the urgent question (Dringliche Anfrage), a written question that is motivated orally by
its main sponsor during the plenary session in which it is introduced. Following
the minister’s oral defence, the question is debated. An urgent question originally required
the support of twenty MPs, but this was reduced to five in 1988. At the same time, a
limit on the number of urgent questions was established: each MP could support two
per year until 1996 and one per year thereafter, with an additional minimum annual
quota of four per parliamentary party since 1996. Typically the opposition parties
exhaust their contingent, and plan to do so irrespective of the political agenda. This led
to sixty-nine urgent questions in the 1996–9 term. Urgent questions from the opposition
11
See Gerlich (1973), Neisser (1986) and Nödl (1995).
Austria 237
typically aim to challenge government policy and to highlight its own alternative pro-
posals. Since 1996, an urgent question can be substituted by an urgent proposal, allow-
ing for a more constructive approach.
In 1961 a question hour was introduced. This is the only time when each individual
MP has the right to question ministers before Parliament (without being sponsored
by other MPs); however, question time is a scare resource, allowing, for instance, only
255 questions in the 1996–9 term (to be followed by 927 ad hoc questions). In 1988 the
Parliament introduced an hour’s debate on current subjects. One such debate can be
held each week if five MPs (with an alternation of the proposing parliamentary parties)
so demand. However, this debate is cancelled if an urgent question is introduced.
Generally, the effectiveness of these and other parliamentary instruments has been
increased by reforms aimed at giving them more publicity.12
Parliamentary questions must relate to executive business (thus excluding legislative
and private affairs). Since 1961 cabinet members have been required to answer written
questions within 2 months (no time limit existed before.) At a minimum, parliamentary
questions require a formal response.13
Parliament has always had the power to order the audit office (Rechnungshof ) to scru-
tinize a particular branch of the executive. The audit office is formally an auxiliary insti-
tution of Parliament charged with exercising financial oversight of the executive
branch of government. It is designed to be independent of the parliamentary majority
(Widder 1997)14 and much of its investigative activity is carried out on its own initiative.
However, regardless of whether an investigation was conducted at Parliament’s
request, the audit office always reports its findings and suggestions to Parliament.
While it is up to the parliamentary majority to decide whether or not Parliament will
act on the basis of audit office reports (e.g., by changing laws, passing resolutions to
force the executive to take certain actions, or censuring a minister), reports also have a
more indirect impact. Because they provide a wealth of arguments that can be used by
the opposition and receive considerable public attention, they put pressure on the
government (Müller 1996a: 43–5). A 1975 reform gave any minority of one-third of
MPs the right to appeal to the audit office. In 1988 the necessary number of MPs was
12
Since 1996 the number of urgent questions has been limited to one per day. If more than one is put for-
ward, the party that had not asked an urgent question for the longest time is given preference. Since the gov-
ernment parties do not nearly exhaust their contingent they are more likely to be in that position and hence
have the chance to buy time, if they find this appropriate.
13
A cabinet member can respond but not provide a substantive answer for a variety of reasons. While
these have not been legislated, parliamentary practice has established the following reasons: (1) if the ques-
tion goes beyond the questioned minister’s jurisdiction, (2) if the information needed to answer the question
is impossible or very costly to gather, (3) if the answer to a question would reveal information which must
be kept secret in the national interest. However, only in exceptional cases can a minister refuse to an answer
by referring to the bureaucracy’s obligation to secrecy (Czerny and Fischer 1982: 297; see Neisser 1986:
678–83 for a partly deviating interpretation).
14
The audit office is linked to Parliament only by the election of its President, who since 1986 is elected
for one non-renewable term of 12 years. Although the President—like cabinet members—is accountable to
Parliament and can—in principle—be removed using the same procedures (see below), this is unlikely. The
parliamentary majority’s removal of an inconvenient President would be seen as an illegitimate use of power
and would probably be more politically damaging than criticism from the audit office.
238 Wolfgang C. Müller
reduced to twenty. Minority access to the audit office however has remained rationed.
Initially a new minority order required the completion of the previous investigation,
later that contingent was increased to two, and since 1993 three, investigations carried
out simultaneously.
A parliamentary majority can also set up special investigative committees (Fallend
2000b). Each party represented in the Parliament’s main standing committee automatically
has representation in investigative committees. These committees work under the
same rules as criminal courts. They can force witnesses to testify before them, and they
have access to government papers. Until 1988 investigative committees worked behind
closed doors. The President of Parliament can now grant access to media representa-
tives during testimony of witnesses and experts, which in fact had been done in all
cases. In practice, investigative committees have been established only to look into alle-
gations of major abuses of executive power. This happened once in the 1940s, twice in
the 1960s, five times in the 1970s, and five times in the 1980s (not counting the re-estab-
lishment of investigative committees after elections). Investigations that have been cov-
ered by the media have had a particularly important political impact, one that has
damaged governing parties. This is one reason why the government majority has not
approved new investigative committees in the 1990s. Only after the formation of the
ÖVP–FPÖ government in 2000, a new investigative committee was set up in order to
look into the business of a ministry previously headed by SPÖ ministers.
In addition to the Parliament’s instruments for monitoring the government, the
latter is subject to many reporting requirements. Schambeck (1997: 105–8) lists more
than sixty reports that the cabinet must regularly submit to Parliament. In many cases
a specific law demands a regular report to Parliament—for example, on social security,
research, arts, and public debt.
Finally, parliamentary oversight can make use of a weak, that is, unsanctioned, ver-
sion of institutional checks (cf. Kiewiet and McCubbins 1991). The activities of the
audit office that stem from its own initiative (rather than investigations ordered by
Parliament) fall under this rubric. Furthermore the ombudsman office
(Volksanwaltschaft), which was established in 1977 as another auxiliary institution of
Parliament, receives citizens’ complaints about the administration. In addition to trying
to solve these problems by working with the administration, it reports to Parliament
and suggests general remedies (Widder 1997). However, both the ombudsman and the
audit office can only bark but not bite. That is, they report their findings to Parliament,
which must then decide whether to act.
In short, Parliament has many mechanisms to hold the cabinet accountable, but in
practice party cohesion and coalition discipline undermine their immediate effectiveness:
parties and coalitions continue to support their cabinet members even when convincing
criticism is tabled against them (Müller and Bubendorfer 1989). It is only to the extent
that parliamentary accountability mechanisms raise public attention to unpopular policies
and maladministration that they can produce indirect effects. Either these factors enter
the voters’ evaluations of the government parties or, anticipating such effects, the
government parties remove cabinet members who have become an electoral liability or
reverse their policies.
Austria 239
D E L E GAT I O N W I T H I N T H E C A B I N E T
Division of Labour Between the Cabinet and Individual Ministers
Once the cabinet is appointed, there is a clear division of labour between the cabinet
and individual ministers. All tasks of the government rest with individual ministers
unless they are explicitly reserved to the cabinet by constitutional or ordinary law. The
number of cabinet tasks is relatively small. The most important are:
(a) to propose legislation and the budget to Parliament (Articles 41 and 51);
(b) to provide reports to Parliament;
(c) to make proposals about appointments (including top civil servants’ posts and
constitutional judgeships) to the President (Articles 65, 67, 147);
(d) to nominate candidates (after agreement with the Nationalrat’s main committee)
for Austria’s most important positions in EU bodies (Article 23c);
(e) to table issues before the Constitutional Court (Article 138–42);
(f ) to propose the dissolution of Parliament to the President (Article 29);
(g) to exercise the government’s (very limited) emergency decree powers15
(Article 18);
(h) to veto Land legislation (Article 98);
(i) to empower the minister of defence to dispose over the army (Article 80).
15
The cabinet and the permanent sub-committee of the Parliament’s main committee jointly propose
emergency decrees to the President. The cabinet also countersigns these decrees.
240 Wolfgang C. Müller
16
See Welan and Neisser (1971), Gerlich, Müller, and Philipp (1988), Gerlich and Müller (1997), and Müller
(1997b).
17
With few short-lived exceptions, the leaders of government parties have always been in cabinet and have
held the position of Chancellor or Vice-Chancellor.
18
The two FPÖ leaders who served in government until 2003 actually fell in disgrace with their party and
lost office in less than one term in government.
Austria 241
Cabinet Coordination
Coordination of cabinet work can be accomplished using either negative or positive
instruments. The fact that decisions in cabinet have to be unanimous constitutes an
important means for negative coordination. Proceeding from an informal norm of reci-
procity, ministers tend not to interfere in the policies of their cabinet colleagues, as long
as these policies have no direct impact on their own departments (Müller 1994).
There are both formal and informal positive forms of coordination. The draft
legislation review procedure (Begutachtungsverfahren) is a formal, permanent means of
coordination. Draft bills from individual departments are sent to other ministries (as
well as to a variety of interest groups and administrative units), which comment on
them. They might also recommend—or even demand—revisions (Fischer 1972; Tálos
and Kittel 2001).
Implementation of laws is not usually the responsibility of the cabinet, but that
of individual ministers. Coordination of individual ministers’ (or parties’) implementa-
tion activities can, however, be enforced by introducing an ‘accord’ clause
(Einvernehmens-Klausel) into a particular law. Such a clause specifies that a particular
group of ministers must agree on administrative measures to implement the law
(Barfuß 1968). To the chagrin of administrative reformers, coalition governments
frequently insert ‘accord’ clauses into their legislation in order to guarantee interparty
consensus in the implementation of relevant laws.
Coexisting alongside these formal instruments are many informal, bilateral contacts
between ministers. The cabinet has no authority to make decisions over matters that fall
under the competence of individual ministers. In practice, however, non-constitutional
bodies such as coalition committees and meetings of the party leaders have done so.
Prominent issues are often coordinated—which is not to say formally decided—by the
Chancellor (and Vice-Chancellor). Policies that require additional financial expenditures
need to be coordinated with the minister of finance.
Formally, all portfolios are equal. Ministers are therefore formally equal—both
as portfolio-holders and as members of the cabinet. The exception to this equality is
the minister of finance, who is ‘more equal’ than others in the implementation phase.
The finance minister has always had a strong position, one that was further strengthened by
a 1986 constitutional amendment (Article 51a–c) and new budget law. These changes
gave the minister the authority to spend the reserve budget if economic developments
suggest that doing so would be wise. In addition, the ministry of finance closely super-
vises budget implementation. In this capacity it has the authority to prevent the spend-
ing of money already granted by Parliament, if changing economic conditions warrant
doing so and if the cabinet agrees.
242 Wolfgang C. Müller
The Constitution does not require that position-taking in EU bodies by cabinet mem-
bers or civil servants be preceded by a cabinet decision. In practice, however, government
parties and departments have agreed on a more comprehensive coordination mechanism
which, by and large, seems to work quite well. There are also provisions for regular coor-
dination on EU matters at the technical and political levels (Müller 2000a; 2001).
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
Federal administration is the responsibility of ministers and all government agencies
are subordinated to specific ministers (Article 69). Ministries are established by law and
are strictly hierarchically organized. In principle, all decisions are to be made by the
minister. However, since it would be rather cumbersome and ineffective to bring all
matters to the minister and wait for him or her to make a decision, the minister is
entitled to delegate specified groups of tasks to civil servants who then make decisions
autonomously, but in the minister’s name. A minister may reverse prior decisions
delegating power to civil servants, and may also reserve the right to make particular
decisions himself/herself. A minister is also entitled to issue a formal order (Weisung) to
civil servants directing them to act in accordance with his or her will. Furthermore, the
Constitution stipulates that civil servants must carry out such an order as long as it does
not violate criminal law (Article 20).19 As a consequence, the minister is constitutionally
responsible for all the acts of his or her ministry (Brandtner 1992).
As is true elsewhere, however, the Austrian bureaucracy is actually much more
influential than the Constitution suggests (Kneucker 1973, 1981; Neisser 1974, 1997). In
general, a minister depends on the cooperation of his or her bureaucracy. If it does not
readily provide its expertise, if it ‘over informs’ or ‘under informs’ the minister, or if it
follows a policy of deliberately limiting the minister’s decision-making alternatives to
those that follow the department’s conventional wisdom, then the minister will have
little policy impact and might even suffer political problems. Given this, ministers often
try to avoid issuing orders. Indeed, some former ministers claim that they have never
issued a formal order. If they want to implement policies, they critically depend on civil
servants acting supportively even when they do not intervene.20 Indeed, the capacity of
ministers to actually monitor the behaviour of their civil servants is rather limited. The
‘spoils system’ is restricted to the positions of minister and secretary of state. Civil
servants retain their positions regardless of what cabinet is in office. Ministers may
appoint a few personal secretaries and staff members in their personal cabinets that
have increased from a total of about thirty in the early 1970s to more than eighty in the
1990s (Liegl and Müller 1999). This is still a quite small number of persons, particularly
in light of the fact that many staff members are not concerned with running the depart-
ment at all, but act as the minister’s press secretaries. Nevertheless, some ministers have
19
Constitutional lawyers continue to discuss the precise meaning of ‘criminal law’ (strafgesetzliche
Vorschriften) in this context (Koja 1986), but it is nonetheless clear that the limits for not observing ministerial
orders are narrowly drawn.
20
Sometimes, however, formal orders cannot be avoided since some civil servants ask for them in order
to avoid responsibility (Neisser 1974: 236).
Austria 243
managed to shelter themselves from their civil servants, heavily relying on their
personal staff in making politically important decisions. If anything, these ministers
have been less successful than those who tried to work closely with the permanent civil
servants.
Despite the absence of a true ‘spoils system’, partisan loyalty has traditionally helped
ensure civil servant support for ministers’ policies. Political parties have colonized the
civil service, and several departments for decades have been one-party dominions
(Müller 1989). Although a civil service reform in 1989 introduced an objective civil
service recruitment system, promotions to key positions still reflect political affiliation
(Liegl and Müller 1999). And although ties of partisan loyalty do not work in every case,
ministers still rely heavily on them. Of course, if ministries change hands from one party
to another, the party-politicization of the civil service may aggravate the delegation
problem. In such situations ministers might make full use of their organizational
powers in order to reorganize the ministry and fill key positions with people who
are partisan loyal. This is eased by the fact that since 1996 all appointments to leader-
ship positions in the civil service are made for 5 year terms only. Reversing the politi-
cization of a ministry, of course, requires partisan candidates waiting in the wings, and this
has often not been the case when ministries dominated by one political party have been
handed over to another. Because there is no tradition of civil servants moving among
ministries, ministers must then turn to other strategies in order to exert control over
the ministry. These include exploiting personal rivalries and career interests within the
civil service, and finding allies among the civil servants to promote specific goals. In any
case, contrary to the constitutional prescription of hierarchy and submission, the
minister must behave within the ministry very much as he or she would in a genuinely
political environment.
Not surprisingly, not all ministers have been equally good at directing their depart-
ments. They have faced different challenges in terms of the intransigence or support of
their ministries, the policy agenda, and—perhaps most importantly—differences with
regard to personal skills. While it would be easy to compile a list of ministers widely
considered to have been failures, it would be hard to argue that lack of control of their
ministries was the single most important factor in their failure. Rather, their behaviour
in political arenas outside of their ministries has given them the reputation of being
unsuccessful.
E X T E R NA L C O N S T R A I N T S
Parliamentarism in Austria is constrained by the presidency, federalism (which was
discussed in the Introduction), the Constitutional Court, direct democracy, consocia-
tionalism and corporatism, and membership in the EU.
The President
Under the terms set out in the Austrian Constitution, the country is about as ‘semi-
presidential’ as France. In practice, however, the President has never exercised his
244 Wolfgang C. Müller
strongest powers—cabinet dismissal and dissolution of Parliament. The President has
exercised only moderate influence in the government formation process. In day-to-day
politics, the President can distort the parliamentary chain of delegation with regard to
administrative and judicial appointments and in foreign policy. The Constitution allows
the President to reject government proposals in these matters, though it does not give
him the power to impose alternatives. In a worst case scenario this could lead to
stalemate. The Constitution thus requires that the President and government cooperate
if the country is to be governed successfully. This has indeed been the norm in postwar
Austria, with the President generally accepting the predominance of Parliament
(Welan 1997; Müller 1999a).
parliamentary government. This is even more the case under the ÖVP–FPÖ coalition,
which does not command a two-thirds majority.
Direct Democracy
Austrian parliamentary democracy is constrained by three instruments of direct
democracy: the people’s initiative, the consultative referendum, and the referendum
(Müller 1998). These instruments have not, however, undermined the primacy of
parliamentary democracy or its institutions. In response to an initiative from the people
(the Volksbegehren) the Parliament is obliged only to consider the bill. There is no direct
path from an initiative to a referendum. This is because only the Parliament has the
authority to call for a legally binding or a consultative referendum. Formally speaking,
only the former is binding. In practical terms, however, this is also true of the consultative
referendum, the results of which are difficult for Parliament to ignore. As has already
been mentioned, the Constitution also provides for an automatic referendum in particular
circumstances. A referendum must be held whenever the constitutional undergoes a
‘total revision’ (Article 44). A referendum must also be held before a federal President
can be removed from office (Article 60). Even in these cases, however, the initiative rests
with the Parliament. It can always avoid a referendum, though perhaps only at the price
of sacrificing major policy or political ambitions.
21
In the case of matters that need the consent of the other chamber of Parliament—the Bundesrat—
(Article 44 [2]), this applies to positions taken by both chambers.
Austria 247
lower than those for Austria’s first year of membership and they do not indicate
that Parliament has had an important impact on EU policy. Moreover even when the
Parliament has acted, interviews with civil servants and interest group representatives
suggest that it has not necessarily done so on the most important issues (Falkner and
Müller 1998). The policy specialists of the government parties, however, can exercise
some influence within the mechanisms of coalition governance.
C O N C LU S I O N
The terms of its Constitution make Austria an imperfect parliamentary democracy. To
begin with, voters delegate to two agents: MPs and the President. The cabinet then is
appointed by the President and formally accountable to both Parliament and the
President. Thus, it has two principals. Individual cabinet ministers are accountable to
Parliament and the Chancellor (and indirectly also to the President whose cooperation
is required with regard to ministerial appointments and dismissals). These relations are
all defined primarily by an initial expression of trust through elections or appointment
and by the possibility that incumbents can be removed if they do not satisfy
the expectations of their principals. MPs can only be removed in subsequent elections.
The President can, in principle, be removed between elections, but in practice this is not
particularly likely.22 In contrast, the cabinet and individual ministers can be held
accountable by their principals at any time. However, the Constitution does not give
ministers’ principals the formal power to issue directives to their agents. The final step in
the delegation chain, that from ministers to civil servants, differs with regard to both
dimensions. Civil servants cannot be removed from their positions, yet they are
obligated to follow the instructions of their political masters.
Political parties and constitutional conventions have made the delegation process
more straightforward and akin to parliamentary democracy. The role of the President
has largely been limited to ratifying the outcome of the party game. The existence of
strong and cohesive parties makes Austria a party democracy par excellence. In reality,
voters delegate authority to political parties rather than to individual MPs. Parties then
maintain control over MPs (Müller and Jenny 2000). The making and breaking of gov-
ernments is a party game, as is the process of governing. Individual parties are cohe-
sive. When in government, the party and its representatives in the cabinet are very
closely connected to one another. There is fusion of top party and government offices.
In some cases the direction of interaction between the two has turned around—that is,
rather than giving government positions to party officials, government members were
given party offices. As noted above, party is also important for delegation from minis-
ters to civil servants. In short, political parties help make democracy work. At the same
time, however, political parties undermine accountability as specified by the design of
political institutions. The specific mechanisms to hold politicians accountable largely
22
It would require first a two-thirds majority in the Nationalrat, then a majority in the
Bundesversammlung (that is the Nationalrat and Bundesrat combined), and, finally, a majority in a referen-
dum. If such an attempt failed, however, the Nationalrat would automatically be dissolved.
248 Wolfgang C. Müller
become ineffective because of the dominance of party loyalty over the formal tasks of
political office. It is only when parties anticipate that such behaviour would inflict dam-
age on their electoral prospects—their ‘brand name’—that they are ready to hold
accountable their members in public office. In effect, this results in delegation being
successful (since this form of party democracy leaves the citizens better off than a
world without parties), but also in considerable agency losses.
The public is fully aware of the transformation of the process of delegation and
accountability due to the existence and activities of political parties. Therefore the
agency problems the public discusses are of a very general nature. To the extent that
‘shirking’ occurs, it is first and foremost attributed to political parties (and to politicians
in general). Since the 1980s, parties have increasingly been criticized. Although they are
still accepted as a necessary part of a democratic polity, citizens have become much
more critical about their actual performance in the delegation process. The two major
political parties have responded by implementing organizational and institutional
reforms, for instance the introduction of primaries and the 1992 electoral reform aimed
at tying voters and MPs closer together. Opposition parties have characterized them-
selves as ‘movements’ rather than parties. Institutional reforms that impact on the
process of delegation have partly been responses to the new dissatisfaction with party
politics. Thus, the numerous reforms of the parliamentary rules of procedure have
been prompted in part by the majority’s desire to demonstrate its willingness to make
possible effective parliamentary scrutiny. Likewise, the establishment of ‘objective’ civil
service recruitment practices in 1989 aimed at demonstrating the government
parties’ respect for the political neutrality of the civil service. Although the relevance
and long-term impact of some of these reforms should not be underestimated, they
could not solve the problems caused by the party system. From 1987 until 2000, Austria
was locked into a ‘grand coalition’ government. On the one hand, there was no alter-
native that a majority of voters and parties found acceptable. On the other hand, elec-
toral responsiveness remained low. Voters had to rely on intraparty and intra-coalition
mechanisms of political accountability, to which there were obvious limits. The end of
‘grand coalition’ government in 2000 has changed the structure of party competition
(Müller 2000d). Basically, a two block system of ÖVP–FPÖ versus SPÖ–Greens has
emerged, which, in principle, increases electoral accountability.
REFERENCES
Adamovich, Ludwig (1971). Handbuch des österreichischen Verfassungsrechts. Vienna: Springer.
—— (1973). ‘Die Koordinationskompetenz des Bundeskanzlers in verfassungsrechtlicher Sicht’.
Juristische Blätter, 95: 234–42.
—— and Funk, Bernd-Christian (1980). Allgemeines Verwaltungsrecht. Vienna: Springer.
Aldrich, John H. (1995). Why Parties? Chicago: University of Chicago Press.
Auracher-Jäger, Barbara (1997). Die Mechanismen im Nationalrat. Wien: Verlag Österreich.
Atzwanger, Konrad and Zögernitz, Werner (1999). Nationalrat-Geschäftsordnung. Vienna: Manz.
Barfuß, Walter (1968). Ressortzuständigkeit und Vollzugsklausel. Vienna: Springer.
Austria 249
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
Brandtner, Werner (1992). ‘The Organization of State Administration—An Overview’, in Federal
Chancellery (ed.), Public Administration in Austria. Vienna: Federal Chancellery.
Czerny, Wilhelm F. and Fischer, Heinz (1982). Kommentar zu Geschäftsordnung des Nationalrates.
Vienna: Österreichische Staatsdruckerei.
Cox, Gary W. and McCubbins, Mathew D. (1993). Legislative Leviathan. Berkeley: University of
California Press.
Damgaard, Erik (1995). ‘How Parties Control Committee Members’, in Herbert Döring (ed.),
Parliaments and Majority Rule in Western Europe. New York: St. Martin’s Press.
Dachs, Herbert (1996). ‘The Politics of Regional Subdivisions’, in Volkmar Lauber (ed.),
Contemporary Austrian Politics. Boulder: Westview.
De Winter, Lieven (1991). ‘Parliamentary and Party Pathways to the Cabinet’, in Jean Blondel
and Jean-Louis Thiébault (eds.), The Profession of Government Minister in Western Europe.
London: Macmillan.
Duverger, Maurice (1959). Political Parties. London: Methuen.
Falkner, Gerda and Müller, Wolfgang C. (eds.) (1998). Österreich im europäischen Mehrebenensystem.
Vienna: Signum.
Fischer, Heinz (1971). ‘Wahlpunkte und Vorzugsstimmen’. Österreichische Zeitschrift für
Öffentliches Recht, 22: 95–112.
—— (1972). ‘Zur Praxis des Begutachtungsverfahrens im Prozeß der Bundesgesetzgebung’.
Österreichische Zeitschrift für Politikwissenschaft, 1: 35–54.
—— (1997). ‘Das Parlament’, in Herbert Dachs et al. (eds.), Handbuch des Politischen Systems Öster-
reichs. Vienna: Manz.
Fallend, Franz (2000a). ‘Der Bundesrat in Österreich’, in Gisela Riescher, Sabine Ruß, and
Christoph M. Haas (eds.), Zweite Kammern. München: Oldenbourg.
—— (2000b). ‘Demokratische Kontrolle oder Inquisition? Eine empirische Analyse der parla-
mentarischen Untersuchungsausschüsse des Nationalrates nach 1945’. Österreichische Zeitschrift
für Politikwissenschaft, 29: 177–200.
Gerlich, Peter (1973). Parlamentarische Kontrolle im Politischen System. Vienna: Springer.
—— and Müller, Wolfgang C. (1997). ‘Austria: Routine and Ritual’, in Jean Blondel and Ferdinand
Müller-Rommel (eds.), Cabinets in Western Europe. London: Macmillan.
—— ,——, and Philipp, Wilfried (1988). ‘Potentials and Limitations of Executive Leadership: the
Austrian Cabinet since 1945’. European Journal of Political Research, 16: 191–205.
—— and Ucakar, Karl (1981). Staatsbürger und Volksvertretung. Salzburg: W. Neugebauer.
Kathrein, Irmgard (1986). ‘Der Bundesrat’, in Herbert Schambeck (ed.), Österreichs
Parlamentarismus. Berlin: Duncker & Humblot.
Khol, Andreas (1995). ‘Demokratieabbau durch EU-Regierungsgesetzgebung?’, in Günther
Schefbeck (ed.), 75 Jahre Bundesverfassung. Vienna: Verlag Österreich.
Kiewiet, D. Roderick and McCubbins, Mathew D. (1991). The Logic of Delegation. Chicago:
University of Chicago Press.
Kneucker, Roul F. (1973). ‘Austria: An Administrative State’. Österreichische Zeitschrift für
Politikwissenschaft, 2: 95–127.
—— (1981). ‘Public Administration: The Business of Government’, in Kurt Steiner (ed.), Modern
Austria. Palo Alto: SPOSS.
Koja, Friedrich (1986). Allgemeines Verwaltungsrecht. Vienna: Manz.
Korinek, Karl (1986). Ministerverantwortlichkeit. Vienna: Orac.
250 Wolfgang C. Müller
Körner, Margit (1994). ‘Das EU-Begleit-Bundesverfassungsgesetz—die Mitwirkung der
Parlamente von Bund und Ländern bei der Schaffung von neuem EU-Recht’. Österreichisches
Jahrbuch für Politik, 1994: 513–41.
Liegl, Barbara and Müller, Wolfgang C. (1999). ‘Senior Civil Servants in Austria’, in Edward C. Page
and Vincent Wright (eds.), Senior Civil Servants in a Service State. Oxford: Oxford University Press.
Lijphart, Arend (1999). Patterns of Democracy. New Haven: Yale University Press.
Luther, Kurt Richard (1986). ‘The Revitalization of Austrian Federalism’, in Michael Burgess
(ed.), Federalism and Federation in Western Europe. London: Croom Helm.
—— (1997). ‘Bund–Länder Beziehungen: Formal- und Realverfassung’, in Herbert Dachs et al.
(eds.), Handbuch des Politischen Systems Österreichs. Vienna: Manz.
McCubbins, Mathew D. and Schwartz, Thomas (1984). ‘Congressional Oversight Overlooked:
Police Patrols Versus Fire Alarms’. American Journal of Political Science, 2: 165–79.
Michalitsch, Martin (1989). ‘Die Geänderte, Realverfassungsändernde Rechtsprechung des
Verfassungsgerichtshofes’. Österreichisches Jahrbuch für Politik, 1989: 197–207.
Michels, Robert (1963). Political Parties. New York: Free Press.
Müller, Wolfgang C. (1983). ‘Direktwahl und Parteiensystem’. Österreichisches Jahrbuch für Politik,
1983: 83–112.
—— (1989). ‘Party Patronage in Austria’, in Anton Pelinka and Fritz Plasser (eds.), The Austrian
Party System. Boulder: Westview.
—— (1990). ‘Persönlichkeitswahl bei der Nationalratswahl 1990’. Österreichisches Jahrbuch für
Politik, 1990: 261–82.
—— (1992). ‘Austria (1945–1990)’, in Richard S. Katz and Peter Mair (eds.), Party Organizations.
London: Sage.
—— (1993). ‘Executive–Legislative Relations in Austria: 1945–1992’. Legislative Studies Quarterly,
18: 467–94.
—— (1994). ‘Models of Government and the Austrian Cabinet’, in Michael Laver and Kenneth
A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge
University Press.
—— (1996a). ‘Political Institutions’, in Volkmar Lauber (ed.), Contemporary Austrian Politics.
Boulder: Westview.
—— (1996b). ‘Wahlsysteme und Parteiensystem in Österreich 1945–1995’, in Fritz Plasser, Peter
A. Ulram, and Günther Ogris (eds.), Wahlkampf und Wählerentscheidung. Vienna: Signum.
—— (1997a). ‘Das Parteiensystem’, in Herbert Dachs et al. (eds.), Handbuch des Politischen Systems
Österreichs. Vienna: Manz.
—— (1997b). ‘Regierung und Kabinettsystem’, in Herbert Dachs et al. (eds.), Handbuch des
Politischen Systems Österreichs. Vienna: Manz.
—— (1998). ‘Party Competition and Plebiscitary Politics in Austria’. Electoral Studies, 17: 21–43.
—— (1999a). ‘Austria’, in Robert Elgie (ed.), Semi-Presidentialism in Europe. Oxford: Oxford
University Press.
—— (1999b). ‘Plebiscitary Agenda-Setting and Party Strategies’. Party Politics, 5: 305–17.
—— (2000a). ‘Austria’, in Hussein Kassim, B. Guy Peters, and Vincent Wright (eds.), The National
Co-ordination of EU Policy: The Domestic Level. Oxford: Oxford University Press.
—— (2000b). ‘Political Parties in Parliamentary Democracies: Making Delegation and
Accountability Work’. European Journal of Political Research, 37: 309–33.
—— (2000c). ‘Austria: Tight Coalitions and Stable Government’, in Wolfgang C. Müller and
Kaare Strøm (eds.), Coalition Governments in Western Europe. Oxford: Oxford University Press.
—— (2000d). ‘Das Österreichische Parteiensystem: Periodisierung und Perspektiven’, in Anton
Pelinka, Fritz Plasser, and Wolfgang Meixner (eds.), Die Zukunft der Österreichischen Demokratie.
Vienna: Signum.
Austria 251
—— (2001). ‘Ministerial Government at the European Level: The Case of Austria’, in Hussein
Kassim, Anand Menon, B. Guy Peters, and Vincent Wright (eds.), The National Coordination of
EU Policy: The European Level. Oxford: Oxford University Press.
—— and Bubendorfer, Heidemarie A. (1989). ‘Rule-Breaking in the Austrian Cabinet: Its
Management and Its Consequences’. Corruption and Reform, 4: 131–45.
–––– and Jenny, Marcelo (2000). ‘Abgeordnete, Parteien und Koalitionspolitik: Individuelle
Präferenzen und Politisches Handeln im Nationalrat’. Österreichische Zeitschrift für
Politikwissenschaft, 29: 133–52.
—— ,——, Steininger, Barbara, Dolezal, Martin, Philipp, Wilfried, and Preisl-Westphal, Sabine
(2001). Die Österreichischen Abgeordneten. Individuelle Präferenzen und politisches Handeln. Vienna:
WUV Universitätsverlag.
—— , Phillip, Wilfried, and Steininger, Barbara (1996). ‘Austria: Party Government within Limits’,
in Jean Blondel and Maurizio Cotta (eds.), Party and Government. London: Macmillan.
—— and Plasser, Fritz (1992). ‘Austria: The 1990 Campaign’, in Shoun Bowler and David M.
Farrell (eds.), Electoral Strategies and Political Marketing. London: Macmillan.
—— and Scheucher, Christian (1994). ‘Persönlichkeitswahl bei der Nationalratswahl 1994’.
Österreichisches Jahrbuch für Politik, 1994: 171–97.
—— and Steininger, Barbara (2000). ‘Not Yet the Locus of Power: Parliamentary Party Groups in
Austria’, in Knut Heidar and Ruud Koole (eds.), Parliamentary Party Groups in European
Democracies. London: Routledge.
—— and Ulram, Peter A. (1995). ‘The Social and Demographic Structure of Austrian Parties,
1945–1993’. Party Politics, 1: 145–60.
Neisser, Heinrich (1974). ‘Die Rolle der Bürokratie’, in Heinz Fischer (ed.), Das Politische System
Österreichs. Vienna: Europaverlag.
—— (1986). ‘Die Kontrollfunktion des Parlamentes’, in Herbert Schambeck (ed.), Österreichs
Parlamentarismus. Berlin: Duncker & Humblot.
—— (1997). ‘Die Verwaltung’, in Herbert Dachs et al. (eds.), Handbuch des Politischen Systems Öster-
reichs. Vienna: Manz.
Nödl, Andreas (1995). Parlamentarische Kontrolle. Vienna: Böhlau.
Pelinka, Anton and Welan, Manfried (1971). Demokratie und Verfassung in Österreich. Vienna:
Europaverlag.
Perthold-Stoitzner, Bettina (1993). Die Auskunftspflicht der Verwaltungsorgane. Vienna: Manz.
Pfeifer, Helfried (1964). ‘Über die Beschlußfassung der Regierung und die Verantwortlichkeit
ihrer Mitglieder’. Juristische Blätter, 86: 485–500 and 541–51.
Ringhofer, Kurt (1977). Die österreichische Bundesverfassung. Vienna: Verlag des Österreichischen
Gewerkschaftbundes.
Sartori, Giovanni (1997). Comparative Constitutional Engineering. Houndmills: Macmillan.
Schäffer, Heinz (1999). ‘The Austrian Bundesrat’, in Ulrich Karpen (ed.), Role and Function of the
Second Chamber. Baden-Baden: Nomos.
Schambeck, Herbert (1997). Regierung und Kontrolle in Österreich. Berlin: Duncker & Humblot.
Schefbeck, Günther (1997). ‘Der Nationalrat Zwischen Rechtssetzungsorgan und Tribüne’.
Journal für Rechtspolitik, 5: 117–35.
Schlesinger, Joseph A. (1991). Political Parties and the Winning of Office. Ann Arbor: University of
Michigan Press.
Sickinger, Hubert (1997). Politikfinanzierung in Österreich. Thaur: Kulturverlag.
—— (2000a). ‘Parteien- und Wahlkampffinanzierung in den 90er Jahren’, in Fritz Plasser, Peter A.
Ulram, and Franz Sommer (eds.), Das Österreichische Wahlverhalten. Vienna: Signum.
—— (2000b). ‘Die Funktion der Nationalratsausschüsse im Prozess der Bundesgesetzgebung’.
Österreichische Zeitschrift für Politikwissenschaft, 29: 157–76.
252 Wolfgang C. Müller
Tálos, Emmerich and Kittel, Bernhard (2001). Gesetzgebung in Österreich. Vienna: Wiener
Universitätsverlag.
Walter, Robert (1972). Österreichisches Bundesverfassungsrecht. Vienna: Manz.
Weber, Karl (1987). Die mittelbare Bundesverwaltung. Vienna: Braumüller.
Welan, Manfried (1967). ‘Das Mißtrauensvotum nach der Österreichischen Bundesverfassung’.
Österreichische Juristen-Zeitung, 22: 561–69.
—— (1974). ‘Der Verfassungsgerichtshof—eine Nebenregierung?’, in Heinz Fischer (ed.), Das
Politische System Österreichs. Vienna: Europaverlag.
––––, (1988). ‘Constitutional Review and Legislation in Austria’, in Christine Landfried (ed.),
Constitutional Review and Legislation. Baden-Baden: Nomos.
—— (1997). Das Österreichische Staatsoberhaupt. Vienna: Verlag für Geschichte und Politik.
—— and Heinrich Neisser (1971). Der Bundeskanzler im Österreichischen Verfassungsgefüge. Vienna:
Hollinek.
Werndl, Josef (1984). Die Kompetenzverteilung Zwischen Bund und Ländern. Vienna: Braumüller.
Widder, Helmut (1979). Parlamentarische Strukturen im Politischen System. Berlin: Duncker &
Humblot.
—— (1986). ‘Der Nationalrat’, in Herbert Schambeck (ed.), Österreichs Parlamentarismus. Berlin:
Duncker & Humblot.
—— (1997). ‘Rechnungshof und Volksanwaltschaft’, in Herbert Dachs et al. (eds.), Handbuch des
Politischen Systems Österreichs. Vienna: Manz.
Winckler, Georg (1989). ‘Geld und Währung’, in Hanns Abele et al. (eds.), Handbuch der Österre-
ichischen Wirtschaftspolitik. Vienna: Manz.
6
I N T RO D U C T I O N
The National Congress, the constituent assembly that defined the institutions of the
new State, laid down the main features of Belgian parliamentary democracy in 1831.
The founding fathers opted for a representative constitutional monarchy with a—for the
time—quite liberal Constitution that protected most modern civil rights and liberties.1
Apart from the expansion of suffrage (the last significant one occurred in 1948,
granting suffrage to women), until 1993 the constitutional form of political representa-
tion did not undergo important modifications, despite several waves of reform
launched in the 1970s and 1980s. The 1993 reform that completed the transformation
of the unitary State into a complex federal system also modified the structure, composi-
tion, and functioning of Parliament and the competencies and operation of the
federal executive. As these changes were implemented only after the 1995 general
elections, it is too early to evaluate their effects. This chapter therefore presents the
main features of the postwar representative system and major recent changes, with
most behavioural analyses referring to the pre-1995 situation.
The Belgian Parliament has two chambers, a feature that deviates from the singularity
principle of delegation. From the beginning, both chambers were symmetrical both
functionally (government-making, legislation, control) and organizationally. Due to the
democratization of suffrage, their composition has also become very similar.
Until 1995, representatives as well as senators could introduce legislative proposals.
To become law, a bill had to go through numerous stages in each chamber: submission
to the speaker of the chamber concerned; formal decision by the Plenary Assembly
that the proposal will be ‘taken into consideration’; committee referral; detailed
committee examination(s); amendment and vote; written report by one committee
member summarizing the committee debate; referral of the proposal plus the report
to the Plenary Assembly of the chamber; plenary debate of the committee version of
the bill; amendment by individual MPs and/or the government; vote on the bill—first
1
Basic overviews of Belgian constitutional law and representative institutions include Alen, Tillemans,
and Peeters (1992), Alen and Meerschaut (1998), Delpérée and Renders (1998), François (1997). For an
English-language introduction to Belgian politics, see Fitzmaurice (1996).
254 Lieven De Winter and Patrick Dumont
article by article and then in its entirety. If the bill won majority support (of valid votes,
excluding blank votes, with half of the members of the chamber present), it was then
sent to the other chamber. The latter then referred the bill to its relevant committee,
and repeated all the steps in the committee and plenary stages. If the version of the bill
approved by the ‘first’ chamber eventually was also adopted by the second chamber, it
was sent to the monarch for royal approval, published by the Official Record, and
received the force of law. If the ‘second’ chamber approved an amended bill, then it was
referred back to the ‘first’ chamber, where all committee and subsequent plenary stages
had to be repeated. This process continued until both chambers had adopted the same
version of the bill. Hence, there were no rules for stopping the navette (shuttle). To save
time, the coalition parties exerted strong pressure on their MPs not to modify a bill that
had already been accepted by the other chamber. Government bills, including budget
bills (see below for detail), were subject to the same procedure, but did not require the
bill taken into consideration by the Plenary Assembly.
Amongst the instruments of delegated legislation we find ‘royal decisions’ (govern-
mental decisions aimed at implementing a law approved by Parliament), ‘ministerial
decisions’ (like the previous type but taken by single ministers 2 ), ‘cadre laws’ (which
explicitly define only the general principles, to be complemented by specific measures),
and ‘task laws’ (that give the government the power to regulate a matter that normally
falls under the competence of the legislature). Finally, in the case of ‘special powers’,
Parliament accords the government large competencies in a broad policy area for a
fixed period of time, and decisions taken under such a mandate cannot be altered by
Parliament during the same legislative session.3
Since the 1960s, Belgium has followed a difficult path toward federalism. A 1993
Constitutional amendment transformed Belgium into a federal country based on three
linguistic communities (Flemish, French, and German, a tiny minority) and three
regions (Flanders, Brussels, and Wallonia). Since 1995, the Senate has lost several com-
petencies. As of 2003 only the House of Representatives can invest or dissolve a
government, approve the budget, and conduct interpellations of ministers. All govern-
ment bills are introduced in the House. Despite this, the Senate and the House remain
equally competent as regards constitutional and other institutional reforms, regional
and linguistic matters, ratification of treaties, and the organization of the judiciary. The
Senate is exclusively competent for conflicts between the national and regional/
community legislatures.4 The Senate has the right to discuss and amend bills approved
by the House, but the latter has the final word, even on bills initiated by the Senate.5
2
In the 1960–85 period, the Belgian Parliament issued 4,058 laws, while the executive issued 40,044 ‘royal
decisions’ and 11,951 ‘ministerial decisions’ (Dewachter 1992: 18).
3
In the postwar period, special powers were accorded in 1965, 1967, 1978, 1982, 1986, 1993, and 1996.
They were often valid for several years. In all, the government ruled with special powers for about 15% of
the 1926–86 period (Alen 1986).
4
Conflicts between chambers are solved through navette or conference committee.
5
Except for matters in which the Senate is (equally) competent, the Senate examines bills adopted by the
House only if fifteen members request it within one week. The Senate then has one month to examine it and
propose amendments. The House can accept or reject the Senate’s amendments. If the House adds amendments
Belgium 255
Thus, the Senate has evolved into a ‘reflection chamber’ and a meeting place between the
federal and regional/community levels of government.
To protect the rights of regional and linguistic minorities, some bills require qualified
majorities of not only a two-thirds majority of the valid votes cast, but also by a majority
of valid votes in each linguistic group in each chamber (with a majority of members of
each linguistic group present).6 The laws that are subject to this special procedure are
those that implement specific constitutional rules regarding the communities and
the regions (i.e. modifying borders, competencies, and/or statutes of the provinces, the
communities, and regions). Constitutional amendments have to be approved by a
two-thirds majority of valid votes in each chamber, with a quorum of half of the MPs
present.
Federalization not only modified the role of the Senate, it created an entirely
new chain of delegation in the regions and communities, each of which has its own
directly elected legislature, an executive headed by a PM, and civil service. This level
of government controls about one-third of overall public spending. After the 1993
reforms, regional competencies include urban planning, environment, local government,
housing, and parts of various other sectors. The communities’ competencies include
nearly all educational matters, culture, parts of health policy, and assistance to families,
the disabled, elderly, youth, etc. Almost all these competencies are exclusive. We can
therefore focus on federal agents since regional and community actors, in principle, do
not interfere in federal policies.
Finally, although laws and royal decisions are formally signed by the Head of State
(but always countersigned by at least one minister), his role in legislation and the day
to day running of the executive has since the Second World War become redundant.
Informally, and especially during the reign of King Baudouin (1950–92), however,
the coalition and policy preferences of the King have sometimes proven significant,
especially with regard to moral or ethno-regional issues.
P O L I T I C A L PA RT I E S
The main particularity of the chain of delegation in Belgium, which like Italy is often
labelled a partitocracy (De Winter, Della Porta, and Deschouwer 1996), is the over-
whelming presence of disciplined political parties. From a two party system during
most of the nineteenth century, when the Catholic and Liberal parties opposed each
of its own, the Senate has three days to examine them and send the amended bill back to the House. The
House then makes a final decision on the bill, if necessary after further amendments. Similar navette proce-
dures apply to bills first introduced in the Senate. Again, the House makes the final decision.
6
Since 1970, linguistic minorities also have the power to delay or block legislation threatening their
interest, through the alarm-bell procedure. In both chambers, if at least three quarters of the members of a
linguistic group introduce an alarm bell motion (after the committee report on a bill and before the final
plenary vote), then parliamentary procedure is suspended. The Council of Ministers must then within thirty
days propose to the relevant chamber either the original bill or an amended version.
256 Lieven De Winter and Patrick Dumont
other across the religious divide, the Belgian party system turned to a three-party type
with the Socialist breakthrough at the end of the century. It remained that way until at
least 1965. The parties’ relative strength varied considerably throughout the period, but
most of the time the three ‘traditional’ parties (Christian Democrats, Socialists, and
Liberals) together received more than 90 per cent of the total vote and alternately
shared governmental offices in different coalition combinations. In the 1960s and
1970s, the number of parties represented in Parliament rose dramatically. The linguistic–
regional parties (Volksunie in Flanders, Rassemblement Wallon in Wallonia, and Front
Démocratique des Francophones in the Brussels region) came first and even entered
governmental coalitions in the 1970s. The growing saliency of the linguistic and regional
cleavage divided the Christian-Democrat, Liberal, and Socialist parties, and each tradi-
tional party split into two organizationally and programmatically independent Flemish
and French-speaking branches (in 1968, 1972, and 1978 respectively). At the end of the
1970s came the ultra-Flemish nationalist and anti-migrant Vlaams Blok, the poujadist
Union Démocratique pour le Respect du Travail, and the Green parties (AGALEV in Flanders
and ECOLO in the Francophone areas). Thus, by 1981 fourteen parties were represented
in Parliament. During the 1980s, the Communists, the Reassemblement Wallon (RW),
and the UDRT lost their last representatives in Parliament. The 1991 general election
introduced the latest newcomers, the Francophone extreme-right Front National, and
the libertarian ROSSEM (which ran only in the 1991 elections). Despite this, until the
law on public party finance was passed in 1989, parties did not have any legal status.
Even parliamentary parties were not recognized in the parliamentary rules until 1962
(De Winter and Dumont 2000). The extreme fragmentation of the party system (the Rae
index reached 0.91 in 1999) expresses a multitude of cleavages and policy dimensions:
left–right, clerical–anticlerical, regional–linguistic, materialist–postmaterialist, system–
antisystem. There are no national parties left in Belgium, that is, parties that run
candidates in all twenty constituencies. All parties are homogeneously Flemish or
Francophone, and contest only Flemish or Francophone constituencies (with the
exception of the bilingual Brussels-Halle-Vilvoorde constituency).
Due to this fragmentation, since the 1970s cabinets have been rather unstable coali-
tions of between four and six parties (thirty-four cabinets 1946–2000 ). The pivotal
Christian Democrats (the largest party family until 1987, in power from 1958 to 1999)
usually chose either the Socialists or the Liberals as coalition partners (De Winter,
Timmermans, and Dumont 2000).
To guarantee a minimal degree of cohesion and stability, governmental parties
minimize the interference of other political actors: voters, party rank-and-file, MPs,
individual ministers, civil servants, and even the judiciary. Policy-making within parties
is highly centralized in the hands of the party executive—mainly the party president
and his entourage. Within the government, power is concentrated in the hands of the
Prime Minister (PM) and vice-PMs (the government’s party heavyweights). Given the
instability of multiparty coalitions, parties have invented and to some extent institu-
tionalized particular coalition maintenance mechanisms, such as the inner cabinet
(the Kernkabinet), the watchdog role for junior ministers and ministerial cabinets, party
summits, and central patronage committees.
Belgium 257
D E L E GAT I O N F RO M VOT E R S TO M P S
Electoral Rules and Electoral Responsiveness
The 150 members of the House of Representatives are elected through proportional
representation in twenty multimember constituencies (in the 1948–94 period, there
were thirty constituencies and 212 Representatives). Parties in each constituency
normally nominate a number of candidates equal to the number of Representatives to
be elected (two to twenty-two). In practice, voters only decide the number of seats a party
will receive, not who will fill the seats. Voters cannot add names to the lists of candi-
dates. More importantly, the electoral code fixes the way in which seats are distributed
among each party’s list candidates. There are two alternative ways of casting votes
(for the House as well as the Senate): a list vote, endorsing the order of candidates on
the list that is presented, and a preference vote. Every candidate whose number of
preference votes reaches the eligibility figure (calculated by dividing the party’s total
constituency vote by the number of seats it won, plus one) receives a seat. Usually, only
candidates at the top of party lists do. If the head of the list receives fewer preference
votes than this, list votes are added to his preference votes until the required number is
reached. This procedure is repeated until all the party’s seats have been allocated.
However, if all list votes were needed before all the seats have been assigned, then the
remaining seats are given to those remaining candidates who have the largest number
of preference votes. In practice, voters rarely manage to alter the ordered list (this has
happened for less than 1 per cent of all MPs elected since the First World War; Rihoux
1996) in spite of the fact that preference voting has increased from 16 per cent in 1919
to 60 per cent in 1999.
Nonetheless, preference votes are not entirely irrelevant to a candidate’s political
career. Selectors do take into account a candidate’s previous electoral performance.
Still, many lists give safe places to privileged candidates—pressure groups representat-
ives, parliamentary specialists, national party leaders, or even ministers—who receive
few preference votes, at the expense of candidates with a larger personal vote. The
wide use of preferential voting thus does not make MPs electorally accountable to their
voters. There are no term limits and there is no mandatory retirement age, or legally
enforced recall mechanism: formally an MP only has to resign when he finds himself in
a position that violates the criteria of illegibility and of incompatibility of mandates.7
The co-optation of thirty-one of seventy-one senators also perverts the simple chain
of delegation and accountability. The parties decide whom they will back as candidates
for co-optation, and to what extent their electoral popularity will serve as a criterion.
Most co-opted senators are candidates who failed to win a seat in the general election,
but some co-opted senators do not run in any type of election.
7
In the case of incompatibility (for instance if an MP takes a civil service job) resignation is automatic. In
case of post-electoral illegibility violations (for instance when an MP loses his civic rights, a measure
usually reserved for a court conviction for a serious crime) the chamber may have to initiate expulsion
proceedings. To date, this has never happened, because in every case the MP decided to resign before his
chamber could expel him.
258 Lieven De Winter and Patrick Dumont
These structural constraints undermine the MPs’ ability to represent the policy
preferences of their respective voters. First, the fragmented party system obliges MPs to
position themselves programmatically on more dimensions and issues than ever before.
Given their strict obligation to the national party programme, it is not easy to adopt a
profile that best suits their specific constituencies. In addition, majority MPs are obliged
to honour the coalition agreement. Many MPs try to circumvent this structural handicap
by generating electoral support through non-policy related activities—for example, case
work, pork barrel politics, and symbolic representation (De Winter 1997).
Candidate Selection
Belgian parties have active and relatively autonomous constituency and local organiza-
tions. The constituency parties make most candidate selection and deselection decisions.
Interference by national party headquarters is rare. All traditional parties—except for
the Francophone Socialists (PS)—gradually shifted away from the poll procedure
(a type of primary in which all party members could participate) in the 1960s and early
1970s (De Winter 1980). In most parties, the process is now in the hands of local and
constituency party activists, while in the CVP (Flemish Christian Democrats) it is
controlled by the three intraparty factions (workers’, farmers’, and middle classes’
organizations).8 Most parties in the late 1960s and 1970s set an age limit of 65 years for
elected mandates, allowing rather generous exceptions. Since the flashlight success of
Pensioners Parties in the early 1990s, some parties relaxed or even dropped these
limits (Smeets, 1994). Selectors’ preferences are influenced by background characteristics
as well as the candidates’ behaviour inside and outside Parliament. Because of the
threat of deselection, selectors can usually sanction an MP who does not live up to their
expectations (De Winter 1988).
8
In ECOLO and AGALEV, as well as in the VLD since 1992, candidates are selected by the regular rank-
and-file members. In the Communist Party and the FDF, a national committee decides.
Belgium 259
(Iker 1998). In the past, tax-deductible donations could only be made to the party
research centres and, since 1985, to specially created party treasury funds. The 1989
legislation on public financing of parties provided all parties represented in both cham-
bers a lump sum plus an additional amount per vote in the most recent elections. New
regulations were introduced in order to reduce the costs of campaigning. A number of
campaign techniques were also prohibited. Tax deductible gifts were first limited, and then
prohibited altogether in 1993. Responsibility for enforcing these campaign restrictions rests
with a parliamentary committee. Also the 1993 law almost quadrupled subsidies.
State subsidies are all given to central party offices, which decide how to allocate
these funds to different candidates. Hence, even in terms of campaign resources, MPs
have grown more dependent on their respective national parties. Other traditional
sources of party income are membership dues (still about one out of ten voters is a
party member) and contributions from elected officials and parliamentary parties—
most of whom transfer the bulk of their resources to central party office (De Winter
and Brans 1999; De Winter and Dumont 2000). Finally, government parties can lay
claim to a considerable number of the members of the so-called cabinets ministériels.
These are the personal staffs of ministers and junior ministers (secretaries of state),
though the appointments are controlled by parties rather than by individual ministers.
In the 1980s, there were roughly 3,000 ‘cabinetards’; however, their numbers have
decreased drastically in the 1990s (partly due to increased public financing of central
party offices).
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
Since the early 1970s Belgium has undoubtedly had the most complex coalition bar-
gaining system in Western Europe, and government formation has become its most
crucial policy decision-making stage. The large number of relevant parties increases
uncertainty and shirking potential, thus undermining cabinet stability. Hence, in order
to reduce agency loss, parties have developed an elaborate set of ex ante and ex post
delegation control mechanisms, which tend to reduce cabinet ministers to pure
party agents. Governmental decision-making—since the introduction of coalition
government after the First World War (Perin 1960)—is extremely collective
(Timmermans 1994: 116–9).
Table 6.1 summarizes the behavioural record with respect to cabinet investiture,
no-confidence and confidence votes, and parliamentary dissolution. While an unsuccessful
investiture vote occurred against Spaak in 1946, there have been no additional instances
since then. Furthermore, though no inventory is kept in Belgium regarding the num-
ber of no-confidence votes introduced in the chambers, at no time has a cabinet been
removed by such a vote. Nonetheless, eight cabinets have resigned in order to pre-empt
a no-confidence vote, for example, Martens VI in 1987. In eleven instances the
Parliament has been dissolved and the cabinet ended in an early election, for example,
Dehaene I in 1995.
Formally, the making of a new cabinet starts with the King’s consultations with the
main parliamentary parties. The King can then appoint a formateur, though he usually
appoints an informateur first, or sometimes a mediator if the political situation is excep-
tionally difficult. In practice, the largest party usually takes the initiative if it is willing
to do so.
The electoral responsiveness of coalition bargaining is very low, with parties taking
office more often after electoral losses than after gains. Electoral accountability is low
as well, as voters are hardly ever called upon to indicate which parties should govern
after a government breaks down (De Winter, Timmermans, and Dumont 2000). This
lack of electoral responsiveness is due not only to the formal formation rules, but also
to two behavioural regularities that stem from the party system:
1. Political and linguistic symmetry in composing coalitions. Traditional parties belonging
to the same ideological family have always been together in government or opposition,
despite the fact that since the 1970s they have become fully autonomous. Thus coali-
tions often include a relatively small surplus party, because it is the sister party of a
major coalition party from the other linguistic camp.
2. Federal/regional symmetry. As far as numerically possible, coalitions at the
regional/community level contain the same parties as the national-level coalition.
Because the negotiators at both levels are often identical and, until 1999, elections at
both levels were held simultaneously, these negotiators usually managed to link the
formation process at the two levels.
The result of interparty negotiations is embodied in the coalition policy programme,
which since the late 1960s is usually published as an appendix to the official ‘government
declaration’ delivered in Parliament.12 It has grown over time (often more than 100
pages) and is frequently referred to as the ‘coalition bible’ (De Winter, Timmermans,
and Dumont 2000). While they are essentially packages of policy agreements, coalition
agreements sometimes also contain statements about competencies and behavioural
rules the coalition members are expected to follow. The 1992 and 1995 agreements
12
Until 1958 the government declaration was the only written statement presented in both chambers of
Parliament. In the early 1960s, the party leaders began to write down the results of the negotiations, though
these reports were only published after the cabinet had been sworn in or not at all (Neels 1975). In the 1970s,
policy negotiations became more institutionalized and coalition agreements were published immediately
after the negotiators had signed them.
Table 6.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Belgium 1946–2000a
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv) No. of Did the Who
formation confidence cabinet end dissolved
No. of Voting results of No. Cabinet Cabinet votes under with an early Parliament in
unsuccessful successful iv of removed resigned spec. inst. election? this case?c
investiture (1: pro gov., 2: against gov., ncvd by ncve to mech.b
votes 3: abstentions and 4: other) pre-empt
1 2 3 4 n.a.
13
There are also ‘specific rules of the game’ concerning the way in which parties will deal with controversial
policy subjects through postponements or referral to some external committee of experts. Governments
may also stipulate that certain delicate policy issues (like abortion or institutional reform) should be decided
by Parliament autonomously, without government initiatives or interference. On other issues, the
government may reserve for itself the right to initiate legislation (De Winter, Frognier, and Rihoux 1996).
Belgium 265
their parties’ supreme decision-making bodies. Thus, while some majority MPs do crit-
icize the government programme during the debate preceding the vote of confidence,
they rarely dare to cast a negative vote.14
14
In the entire postwar period, an average of only 0.6% of the majority MPs voted against the govern-
ment in the investiture vote (De Winter, Timmermans, and Dumont 2000).
266 Lieven De Winter and Patrick Dumont
15
The success rate of private member’s bills has gone down by about a half during the past three decades.
On the other hand, the proportion of private member’s bills increased from less than 20% of all approved
bills in the 1960s to about one third in the 1990s (De Winter 1998b).
16
There are no formal requirements for the government as a whole to report to Parliament. The cham-
bers can require the presence of ministers and ask them questions and hold interpellations.
Belgium 267
and answered orally), and urgent questions (introduced and answered orally).17 All
involve demands for information. They can expose neglect, abuse, or incorrect application
of the law. They can force a minister to voice an opinion on delicate matters or to
suggest improvements and reform. Yet despite their substantial control potential, about
half of the questions are simple demands for information, often inspired by mere vote-or
publicity-seeking motives. Moreover, because ministers’ responses to questions are not
followed by debate, the usefulness of questions for controlling the government and
ministers is limited.
Interpellations are the classic and most powerful tool of parliamentary control. They
are questions put to the government or a minister on a matter of general importance.
The initial question and response can be followed by a motion for debate between the
interpellant and the government, which can include a proposal not to grant the
government the confidence of the House (since 1995 the Senate no longer holds inter-
pellations). Although in principle matters of local or special interest are not legitimate
matters for interpellations, in practice these have become more numerous. By now most
interpellations focus on a minor aspect of a particular governmental policy and are
inspired by a member’s personal electoral concerns. Given their particularistic nature,
the House increasingly tends to relegate them to public committee meetings. Only inter-
pellations of a general nature or of exceptional political importance are still held in
plenary session. In about one of three interpellations, the interpellant (or another MP)
introduces a motion. Members of the opposition usually introduce a motion of censure.
The House votes on such motions one week after they are introduced, which—if
necessary—gives the government ample time to re-establish the confidence of majority
MPs, who thus need not express themselves immediately. This is a face-saving device in
case the government’s or minister’s reply has failed to satisfy the House.
Before reforms in 1989, departmental budget bills (indicating planned expenditures
for the coming year) were usually submitted quite late—after the budget year had
already begun. Thus, most of the money was spent before the final budget was
approved!18 In addition, budget bills were usually rushed through the chambers by the
government majority. Parliament’s budgetary oversight of the executive therefore
became nothing but a farce.19 Since then, the rules governing the government’s budget
bill have been changed to strengthen parliamentary oversight. The budget process ends
with parliamentary approval of the Law on Accounts, which includes a financial report
from the Audit Office. This requires the introduction of a bill to that effect during
October following the end of the fiscal (i.e. calendar) year. The vote on this law was
originally intended to provide an opportunity to assess governmental policy. Yet
17
House committees started holding weekly sessions devoted to questions concerning current events. In
addition, most questions are now asked in committee meetings rather than in plenary. Committees can
summon the minister, ask all questions, and hear any kind of person or group for advice. Committees do not
have the right to read cabinet minutes.
18
Of all departmental budgets approved during the 1976–88 period, only 6.8% were approved on time
(Dewachter 1992: 121).
19
Until the early 1990s, there were over 300 ‘budget funds’ in different ministerial departments. These allowed
them to accumulate funds (often by loans) and make expenditures that did not appear in the official budget.
Government expenditures therefore tended to expand unnoticed and uncontrolled (Lammens 1993: 101–4).
268 Lieven De Winter and Patrick Dumont
because passage of the Law on Accounts often occurs several years later, it has become
a mere formality, not least because the responsible government is usually no longer in
power. In addition, the Audit Office’s report evaluates the regularity and legality of
governmental expenditures, rather than their efficacy and necessity.
Parliamentary committees of investigation have the same powers as an examining
magistrate. Until the 1980s, such committees were quite uncommon because majority
MPs—in solidarity with the government—were reluctant to allow investigations that
could embarrass a specific minister or the entire government. In addition, even when a
committee of inquiry suggested that a minister was politically responsible for certain
failures, the minister did not resign as long he was supported by his party. However,
while the House created only nine such committees in the entire 1880–1988 period,
since then it has established roughly one each year. They have developed into an
alternative instrument of parliamentary oversight and get considerable publicity. In
addition, regular committees can—in their legislative function—gather advice and
documentation from persons or institutions outside the Parliament. However, the
number of such hearings remains extremely low.20
Finally, since 1993, at the beginning of every parliamentary year the PM holds a ‘state
of the union’ speech, which usually launches a general debate of government plans.
Since 1989, before the annual debate (in autumn) on each departmental budget, the
minister has to present a policy note (policy goals, budgetary means, timing) which is
followed by a debate for his sector for the coming year(s).
20
Parliamentary ombudsmen can serve as an additional tool for overseeing the government bureaucracy.
In spite of the introduction of ombudsmen in other public institutions, the House did not establish a parlia-
mentary ombudsman until 1997. One reason for this reluctance was the fact that MPs themselves are very
active in constituency casework. The prominence of this role is related to a clientelistic political culture, to
the lack of other sources of administrative redress, and to the wide scope of party patronage until the early
1990s (De Winter 1996, 1997). To some extent, this casework has informational spin-offs for legislative
oversight, as it reveals the quality of government services.
Belgium 269
Formal and informal party constraints further enhance voting discipline. First, the
statutes of all parties give supreme authority to the party’s national congress, and
between congresses to the party executive. In most party executives a majority of
members are MPs, but they do not generally consider themselves delegates of the
parliamentary party (De Winter and Dumont 2000).
In addition, most party statutes define the role of their public office holders in
agency terms: office holders are agents of the party, and they are obliged to carry out
the party programme. In almost all parties, MPs must ask permission from their group
(leader) to introduce a private member bill or amendment, hold interpellations, or sup-
port a bill sponsored by another party. The parliamentary group can explicitly sanction
voting rebellions in a variety of ways—from a simple warning to expulsion from com-
mittees, the parliamentary party group, and/or the party. The fact that most ministers
(until 1995) and party presidents have been members of Parliament, and therefore usu-
ally attend parliamentary caucus meetings, has been an additional source of party con-
trol over the group’s decision-making (De Winter 1993).
The small staffs of individual MPs and the relatively well subsidized party research
centres add to the MPs dependency on his party organization. MPs often rely on
information provided by their research centre in drafting bills, amendments, and inter-
pellations. For most policy sectors, a group of paid experts and volunteer specialists
associated with the party research centres prepare the party’s proposals in collaboration
with the MPs who specialize in these areas. Thus MPs to a large extent depend on their
party’s brain trust.21
D E L E GAT I O N F RO M C A B I N E T TO I N D I V I D UA L M I N I S T E R S
Rules Concerning Cabinet Composition and Operation
There are no legal texts concerning internal cabinet rules, except for a few articles in
the Constitution, such as the ‘parity’ rule requiring the cabinet to have equal numbers
of Dutch and French-speaking members, not taking the PM and the junior ministers
into account. Furthermore, ministerial and royal decrees on regional and linguistic
matters require the cabinet’s consent. Otherwise, there are no fixed rules specifying the
types of decisions the cabinet can take.
Yet, at least in some cabinets, a text called ‘Practical instructions of the Council of
Ministers’ has been distributed. The 1985 version (Alen and Dujardin 1986: 532–4),
which is still operational, stipulates that the cabinet not only debates and decides on
‘overall governmental policy’ but also on all bills ministers want to introduce in
Parliament, all ministerial decisions with ‘important budgetary repercussions’, and all
matters that could ‘jeopardize government solidarity’. The latter rule brings most
21
In addition, MPs depend heavily on the ministers’ cabinet ministériel in their individual and collective
constituency casework. Ministerial cabinets play an important role in pushing the administration to consider
an MPs client’s case benevolently. More importantly, they control the administration’s handling of pork
barrel benefits. In addition, parties have increasingly centralized patronage with regard to the recruitment
and promotion of civil servants (De Winter 1981). Hence, MPs depend on the ministers and national party
headquarters for access to electorally beneficial favours.
270 Lieven De Winter and Patrick Dumont
matters decided by individual ministers under collective cabinet responsibility and gives
coalition parties effective veto power over policies proposed by ministers of other par-
ties. In fact, the objection of one party or minister to a certain policy proposal is suffi-
cient to force a discussion by the full cabinet, followed by a decision taken by consensus.
In addition, for each matter that may ‘involve the government as a whole’, ministers
must consult the relevant cabinet committee or the cabinet. Also, ministers are not sup-
posed to make declarations about matters that fall under their colleagues’ competencies
and that might embarrass them. Even in their own jurisdiction, they must be ‘extremely
discreet’ about making public statements as long as the government has not yet decided
the matter! Finally, ministers should not make any declaration or take any action
expressing a personal point of view that challenges the government agreement.
By constitutional convention, cabinet decisions are taken by consensus formulated
by the Prime Minister. The 1985 principles state that a minister, when he does not agree
with a cabinet decision supported by a ‘fundamental’ majority, has two alternatives:
accept it without public dissent, or resign.22
These operational rules suggest that Belgian ministers are not at all policy dictators
in their jurisdictions.23 Every policy initiative of any significance must be scrutinized
and approved by the cabinet. However, the vague formulations of these rules leave the
PM and his Vice-PMs with some discretion (Senelle 1983; Timmermans 1994).
22
Formally the King dismisses ministers. In practice, a PM can only dismiss ministers with the agreement
of their party leaders.
23
Formally the right to determine ministerial jurisdictions belongs to the King, or thus to the government
as a whole. Hence, in practice, competencies are decided during the formation negotiations.
24
In addition the Vice-PMs can rely on a cabinet for each ministerial portfolio they have. Thus, some of
them had over two hundred personal staff members.
Belgium 271
Until 1992, when they were abolished, cabinet committees prepared its decisions.
Cabinet committees included not only the relevant ministers but usually also each
coalition party, regardless of the actual departmental competencies. Therefore, each
party was fully informed about the decisions that were under preparation. Coordination
was assured by the PM, who chaired most of the committees, and by the personal staffs
of the ministers.
Junior Ministers
The number and the competencies of junior ministers (a position that has existed since
1960) are decided collectively by the formation negotiators. Until 1991 junior ministers
often did not belong to the same party as their senior, or—in the days of the unitary
traditional parties—often came from a different linguistic wing of the party (De Winter,
Timmermans, and Dumont 2000). In these cases, junior ministers tended to operate as
watchdogs over their ministers. The junior ministers’ decision-making autonomy was
often relatively large and rested on the overall balance of power between coalition
parties. This watchdog system led to so many conflicts between ministers and their jun-
iors that since 1991 they have always belonged to the same party. In addition, the most
recent governments have reduced the number of junior ministers dramatically.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
The so-called ministerial cabinets25 are the main link between the minister and his civil
servants. The excessive use of ministerial cabinets is partially due to the politicization
of the civil service, which often creates tensions when top civil servants belong to
25
The number of cabinetards involved in policy-making grew from 108 in 1953 to 325 in 1986 (and
decreased to 235 in 1995), while their overall number increased from 750 in 1960 to 2000 in 1987 (Dewachter
1992: 238; Hondeghem 1996: 55). Current estimates of the overall number including all levels of government
are at 4,000 (Suetens and Walgrave 2001).
272 Lieven De Winter and Patrick Dumont
a different party than their minister. In fact, about three out of four higher civil servants
are party members (Dierickx and Majersdorf 1993: 151), of which six of ten are
Christian Democrats (Tegenbos 1990).
This high politicization of the civil service is a traditional feature of Belgian con-
sociationalism, which became more institutionalized and pervasive during the 1970s due
to the fragmentation of the party system (De Winter 1981, 1996). Two types of inter-
party negotiations govern the distribution of civil service promotions between govern-
mental parties. First, the distribution of top positions (director general and secretary
general) is decided by the government collectively. Second, promotions of lower ranks
of university-trained civil servants are decided by an unofficial interparty committee, in
which each coalition party is represented. For each vacancy, the committee considers the
candidates for promotion and the support each one has from a coalition party. In prin-
ciple, each coalition party has a quota of nominations (approximately proportional to its
parliamentary strength) that is usually fixed in a secret protocol annexed to the coalition
agreement. Only when there is no partisan candidate does the committee nominate the
candidate proposed by the board of directors of the ministerial department.
Given this encompassing politicization, ministers feared that they could not
unconditionally rely on the loyalty of their civil servants. Hence, until the early 1990s,
ministerial cabinets carried out most policy development, verification of policy imple-
mentation, and mediation with interest groups (Hondeghem 1996: 51–8).26 The civil
service, including high civil servants, basically only provided the cabinet with the
information necessary for policy planning and policy and department management
(except for personnel policy, which was also controlled by the ministerial cabinets).
Many cabinetards go back and forth between the ministerial cabinets and the civil
service, with cabinet service quickening promotion in the civil service, and the civil
service serving as a waiting room for the politically active whose party is in opposition
(Dewachter 1992: 237).27 A similar exchange operates between ministerial cabinets on
the one hand, and the judiciary and the public enterprises on the other.
However, since the late 1980s, and especially after the breakthrough of the extreme
right in the 1991 ‘Black Sunday’ elections and the subsequent calls for a ‘new political
culture’, there has been a general move towards accountability and depoliticization
of the civil service. As a consequence, at the present time, about five of six higher civil
servants are appointed on the proposal of the departmental board of directors
(Hondeghem 1996: 60). In addition, a code on the relationship between ministers and
their administration has expanded the role of the latter. The ministerial staffs are now
required to meet regularly with their top civil servants. In the past, ministerial staffs
tended to work with civil servants loyal to their party, often bypassing the departments’
hierarchy. Higher civil servants have been made more responsible for the management
of their departments, are regularly evaluated, and can now be sanctioned (or even fired)
26
Parties (together with trade unions) interfere with the promotion (and often also with the initial recruit-
ment) of personnel at different levels (ranging from the janitor in a public kindergarten to the chairman of the
board of directors of SABENA airlines) in other employment sectors too, including public radio and television,
public education, (semi-) public and quasi-autonomous enterprises and services, and local government.
27
Other incentives are the political power these positions give and their better remuneration.
Belgium 273
for poor performance. The role of the departmental secretary-general has been enhanced
in budgeting, personnel policy, and policy implementation. A college of secretaries-
general advises the government and coordinates matters that concern several departments
(De Ryck 1994: 473–5). As a consequence, ministers now tend to have much more direct
contact with their top civil servants. The emancipation of the higher civil service has been
enhanced by the reduction of the size of the ministerial cabinets by half,28 basically due
to the introduction of a constitutional amendment limiting the number of ministers to
fifteen, and by the transfer of competencies to regions and communities.29
The (lack of ) delegation from ministers to civil servants has varied across time and
policy sectors. Some departments (finance, defence, and foreign affairs) traditionally
managed to maintain a relative high level of autonomy vis-à-vis the ministers and their
cabinets. As up to the late 1970s the latter two were not policy sectors of strong partisan
conflicts, parties felt less inclined to politicize recruitment.
Finally, in the 1990s several important public enterprises were given more autonomy
(airways, telephone, railroad, ferries, post, some public banks and credit institutions,
etc.). Management contracts between the minister and ‘public managers’ heading
these enterprises have been concluded. They stipulate the objectives of the service and
the financial means allocated by the minister in charge. Managers themselves can
decide how to allocate these resources and achieve their goals. However, a government
commissioner supervises the manager to ensure that the contract is respected. If it is
not, he can ask the minister to annul the manager’s decisions. Parliament has no super-
visory authority vis-à-vis these enterprises insofar as the latter are only required to submit
their annual reports and budgets to Parliament during the following year. This rules out
any parliamentary debate on the policy and budgetary options open to these enterprises
(Bock and Debroux 1995).
In many of the public enterprises, public stockholding has been reduced to 51 per cent
or less (De Ruyter, Michielsen, and Mortelmans 1994). Private sector participation
puts a check on direct ministerial or party interference in policy-making and, in particular,
personnel recruitment.
E X T E R NA L C O N S T R A I N T S
The Head of State and the Referendum
The democratic accountability of the monarch has grown over time. Constitutionally,
legislative power is shared between the two chambers and the monarch—in practice,
the cabinet. Every bill signed by the monarch requires the countersignature of a minister
28
The figures reported in note 25 suggest that there has only been a transfer from the federal government
to the executives of the regions and communities.
29
These reforms are part of a wider effort to modernize and democratize the civil service, including
a move towards openness. Since 1991, administrative decisions have to contain motivation. Since 1994, citizens
can consult administrative documents, even when they do not have a personal interest in them. Still, many
restrictions remain. Civil servants are bound by professional secret. Administrative documents deposited in
the governmental archives can only be consulted after 100 years. Ministers can decide to make them accessible
after 50 years! Thus, even after 1991 one cannot not freely obtain information on the preparation of
governmental decisions.
274 Lieven De Winter and Patrick Dumont
who can be held accountable. The impact of the monarch on cabinet decision-making
has gradually become insignificant. The consultative referendum over the King’s
Question (1950) concluded this process.30 In addition, the effect of the referendum was
so devastating that the potential (re-)introduction of this particular instrument of direct
democracy completely disappeared from the political agenda of governments and
parties until the early 1990s. By now, the only visible political role of the monarch
occurs during cabinet formation and crises.31
30
The King’s Question whether or not Leopold III (1934–50) should remain Monarch given his compromis-
ing attitude towards the Nazis during Second World War created a major and violent conflict—Socialists,
Communists, and some Liberals were pitted against the Catholic Party, the Church, and other conservative
circles. In the 1950 referendum, called by the governments of the day to resolve the issue, the North and the South
voted differently: 72% of the Flemings voted in favour of the King, while only 42% of Walloons did so. Overall,
57.6% voted in favour of the King. After violent protests, Leopold abdicated in favour of his son Baudouin.
31
This is especially true with regard to his role in the nomination of government formateurs and informateurs.
Also, when a governmental crisis is imminent, the King can take time to decide whether or not to accept the
resignation of the cabinet. The King has weekly meetings with the PM and regularly receives ministers.
Preferences expressed by King Boudouin did carry some weight in ministerial decision-making. His moral
judgement was generally appreciated, apart from his refusal in 1988 to sign the bill liberalizing abortion. In
order to prevent a major constitutional crisis, the government (with the King’s approval) declared the King ‘tem-
porarily unable’ to exercise his powers. Then, given the presumed incapacity of the King, all the ministers
signed the bill, which was instantly promulgated. After 36 hours, the King was found fit to govern again.
32
In Grilli, Masciandaro, and Tabellini’s (1991) classification of central bank political independence,
Belgium obtains the lowest score of eighteen countries. In other surveys its position is less dramatic, but still
quite low (Eijffinger 1997).
33
The Council of State consists of thirty councillors nominated by the minister of interior.
Belgium 275
government must be submitted to the Council of State by the relevant ministers. The
initiative to submit a draft bill may also come from Parliament. In other cases asking
for this court’s advice is optional. Hence, out of all rules for which an opinion might
be requested (bills, amendments, ministerial rules issued by federal, regional, or comm-
unity executives or parliaments, collective labour agreements, etc.), only 20 per cent are
submitted to the Council of State. Although it has no formal legislative power, its
advice has gained moral weight since 1970, when constitutional reform began to dominate
the political agenda. Nonetheless, when there was a solid consensus between coalition
parties, the governmental majority has on occasion not hesitated to pass a bill or
amendment that the Council had judged unconstitutional.
2. The Court of Arbitration was established in 1984 in order to determine whether
legislative rules issued by federal, regional, and community assemblies were enacted in
compliance with the Constitution and its enabling legislation. In 1988 its competencies
were enlarged to include reviewing the compliance of legislative rules with fundamental
rights enumerated in three constitutional articles: the principles of equality, non-
discrimination, and the right to and freedom of education. The Court exercises its
jurisdiction through annulment proceedings that may be initiated by the executive or
legislative bodies of the federal, regional, or community levels of government or by
any legal or natural person. In such cases, the Court’s rules are binding. This is not the
case, however, when the Court expresses an opinion on prejudicial rulings initiated by
ordinary courts. In practice, the Court has interpreted its new competence broadly,
including all cases of discrimination and violation by federal, regional, and community
legislative assemblies of constitutional rights and other rights granted in international
treaties. It has therefore evolved into a genuine constitutional court, though it has
limited policy competence.
Recruitment to the courts, and especially the constitutional ones, is heavily subject to
party patronage (De Winter 1981). Although there is no strong evidence that judges
recruited through party patronage are easily influenced by party pressure, their parti-
san recruitment and career dependence raise questions about the independence and
political neutrality of the ‘third branch’.
34
By 1991 trade unions and employers organizations had concluded nearly 30,000 social agreements.
Since 1968, these agreements have been legally binding. However, following the economic recession of the
276 Lieven De Winter and Patrick Dumont
there appears to be an extensive and often secret network of informal negotiations.
Clearly, these consultative and (co-)decisional networks undermine Parliament’s and
the executive’s competence in many policy sectors, insofar as the former ratifies and the
latter executes decisions taken in the neo-corporatist network (De Winter 1999).
These pressure groups reflect the consociational or pillarized character of Belgian
society. In fact, the predominance of traditional political parties (Christian Democrats,
Socialists, and Liberals) is partly due to the fact that they aggregate, articulate, defend,
and implement the interests of a ‘pillar’ or ‘social world’ that includes trade unions,
farmer and ‘middle class’ organizations, socio-cultural organizations, educational and
health service networks, press networks, etc. Their strength is due to the fact that policy
implementation in many sectors of public life is subcontracted to pillar organizations.
In practice, the Committee most often discusses texts that have already been ratified
by Community authorities. This is not surprising given the fact that the government
usually transmits European documents to Parliament very late in the legislative
process. Therefore, as executive decisions can only be challenged ex post facto,
Parliament is reduced to a mere rubber-stamping body, and is forced to transpose
Community rules into domestic law more or less automatically.
Different proposals have been suggested to correct this situation. In 1993, an article
was added to the constitution specifying that before the government signed any document
the chambers were to be kept informed, from the opening round of negotiations
intended to revise the EU treaties, about the progress and results of negotiations.
Contrary to classic international treaties, the EU Treaties require the preliminary
approval of Parliament. Second, a new legislative article states that proposals sent to
the European Council of Ministers on regulations and directives from the European
Commission are to be transmitted to the chambers. This tool is increasingly used and
enables Parliament more regularly to examine EU proposals.
Still, although these rules afford some measure of political control, if the government
violates its obligations the only applicable sanction is to inform Parliament. Hence,
while the current House and Senate Advisory Committees in charge of European
Matters have become more prominent, they remain very dependent on governmental
goodwill. Their dependency is due to the executive’s monopoly on information and to
the lack of transparency that characterizes European decision-making. The Advisory
Committees frequently cannot meet their obligations in a timely fashion, usually
because the government does not in due course transmit relevant proposals for directives,
regulations, or other texts produced by the European Commission.
C O N C LU S I O N
The Belgian partitocracy violates the ideal-type chain of parliamentary delegation in
many ways, insofar as political parties play a predominant role at each stage. They
channel the delegation of power from voters to MPs, from Parliament to the cabinet,
from the collective cabinet to individual ministers, and from ministers to their civil
servants. Hence, they might be considered the effective principals in the polity. The voters
only determine the relative weight of the parties. The parties themselves choose the
individuals who will occupy Parliament and—since electoral responsiveness in coalition
formation is very low—they also decide how and with whom to exercise their power.
In order to make coalitions function and to maintain their status as principals, they have
developed a large number of ex ante and ex post control and monitoring mechanisms.
Therefore, many actors in the parliamentary chain of delegation (including MPs,
ministers, and civil servants) have been reduced to mere party agents.
To some extent, partitocracy is unavoidable and even functional for political system
stability. Given the extremely high fragmentation, a minimum level of government
stability and coherence requires power to be concentrated in the hands of a small number
of actors. Autonomous MPs, ministers, top civil servants, and to some extent even judges
and courts could ‘spoil’ the extremely complex and delicate process of delegation, and
even contribute to its collapse.
278 Lieven De Winter and Patrick Dumont
On the other hand, the costs of partitocracy are high, and partitocratic practices
carried to their extreme have undermined political legitimacy and governability.35 Since
the 1991 ‘Black Sunday’ elections, institutional reform and democratic refinement have
been high on the political agenda. This has produced a large number of institutional
reform proposals (De Winter and Brans 1999). The working group coordinated by the
Speaker of the House reached a consensus in 1997—but most of its provisions have not
yet been enacted and implemented. They include restricting the accumulation of local
executive offices by MPs, closing loopholes with regard to private party financing,
lowering the ceiling on campaign spending for individual candidates, allowing groups
of citizens to introduce a ‘popular legislative initiative’ to Parliament, re-evaluating and
simplifying existing legislation, increasing access to public administrators, increasing
the transparency, responsiveness, responsibility, and efficacy of public administration,
depoliticizing civil service recruitment and promotions, and drafting a code of conduct
for MPs that would curtail their clientelistic intervention in public administration.36
The more radical proposals of the opposition parties include electoral system
reforms. These include reducing the impact of list votes on the election of MPs; the
introduction of a (more) majoritarian system, direct election of members of the executive
(from mayors to the PM), and the introduction of a binding referendum. There are also
calls for an absolute prohibition against holding multiple public offices, limits on outside
occupations for public officials, and term limits.
Hence, mainly since the early 1990s, Belgium has witnessed a gradual decline in the
informal system of partitocratic delegation. Thus, some formal agents (the cabinet, top
civil servants, some MPs) are gradually reacquiring autonomy. Some of these evolu-
tions will facilitate the restoration of the simple and stepwise chain of delegation, while
other aim at introducing direct links between parts of the chain that, in the ideal–typical
model, are not adjacent.
REFERENCES
Alen, André (1986). ‘De “Bijzonder Machten”: Een Nieuwe “Besluitenregering” in België?’
Tijdschrift Voor Bestuurswetenschappen en Publiek Recht, 41: 123–59.
—— and Dujardin, Jean (1986). Casebook Belgisch Grondwettelijk Recht. Brussels: Story-Scientia.
—— and Meerschaut, Frank (1998). La Constitution de la Belgique fédérale. Diegem: Kluwer.
—— Tillemans, Bernard and Peeters, Patrick (1992). International Encyclopaedia of Constitutional
Law: Belgium. Deventer: Kluwer.
Arcq, Etienne (1993). ‘Le Taux de Syndicalisation 1982–1991’. Courrier Hebdomadaire du CRISP, 1386.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
35
In the 1990s Belgium had the highest level of public debt of the fifteen EU members (still 114% of GNP
at the end of the decade). During the Dutroux affair, it also scored the lowest with regard to citizen satisfac-
tion with the functioning of democracy (Eurobarometer, 47.1).
36
The Flemish Parliament has in 1997 promulgated such a code, drafted by Dr. De Winter, which has
gradually become the standard for similar codes being drafted by other national, provincial, and local assemblies
(De Winter and Dumont 2000).
Belgium 279
Bock, Isabelle and Debroux, Philippe (1995). Le Contrat de Gestion, ses Limites et ses Potentialités.
Bruxelles: Services Fédéraux des Affaires Scientifiques, Techniques, et Culturelles.
Delpérée, Francis and Renders, David (1998). Code Constitutionnel. Bruxelles: Bruylant.
De Ruyter, Karin, Michielsen, Stefaan, and Mortelmans, Johan (1994). België Verkoopt. De Stille
Privatisering van de Belgische Overheidsbedrijven. Groot-Bijgaarden: Scoop.
de Ryck, Etienne (1994). ‘Ministeries’, in Mark Deweerdt, Clem De Ridder, and Roger Dillemans
(eds.), Wegwijs Politiek. Leuven: Davidsfonds.
Dewachter, Wilfried (1987). ‘Changes in a Particracy: The Belgian Party System from 1944 to
1986’, in Hans Daalder (ed.), Party Systems in Denmark, Austria, Switzerland, The Netherlands and
Belgium. London: Pinter.
—— (1992). Besluitvorming in Politiek België. Leuven: Acco.
De Winter, Lieven (1980). ‘Twintig Jaar Polls, of de Teloorgang van een Vorm van Interne
Partijdemocratie’. Res Publica, 22: 563–85.
—— (1981). ‘De Partijpolitisering als Instrument van de Particratie. Een Overzicht van de
Ontwikkeling Sinds de Tweede Wereldoorlog’. Res Publica, 23: 53–107.
—— (1988). ‘Belgium: Democracy or Oligarchy’, in Michael Gallagher and Michael Marsh (eds.),
Candidate Selection in Comparative Perspective. London: Sage.
—— (1991). ‘Parliamentary and Party Pathways to the Cabinet’, in Jean Blondel and Jean-Louis
Thiébault (eds.), The Profession of Government Minister in Western Europe. London: Macmillan.
—— (1993). ‘The Links Between Cabinets and Parties and Cabinet Decision-Making’, in Jean
Blondel and Ferdinand Müller-Rommel (eds.), Governing Together. London: Macmillan.
—— (1995). ‘The Role of Parliament in Cabinet Formation and Resignation’, in Herbert Döring
(ed.), Parliaments and Majority Rule in Western Europe. New York: St. Martin’s Press.
—— (1996). ‘Party Encroachment on the Executive and Legislative Branch in the Belgian Polity’.
Res Publica, 43: 325–52.
—— (1997). ‘Belgian MPs between Party and Voters’. Journal of Legislative Studies, 3(3): 128–54.
—— (1998a). ‘The Belgian Parliament’, in George Kurian (ed.), World Encyclopedia of Parliaments
and Legislatures, Vol. I. Washington, DC: Congressional Quarterly.
—— (1998b). ‘Parliament and Government in Belgium: Prisoners of Partitocracy’, in Philip
Norton (ed.), Parliaments and Governments in Western Europe. London: Frank Cass.
—— (1999). ‘Belgium: Insider Pressure Groups in an Outsider Parliament’, in Philip Norton (ed.),
Parliaments and Pressure Groups in Western Europe. London: Frank Cass.
—— and Brans, Marleen (1999). ‘Belgien. Berufspolitiker und die Krise des Parteienstaates’, in
Jens Borchert (ed.), Politik als Beruf. Opladen: Leske ⫹ Budrich.
——, Della Porta, Donatella, and Deschouwer, Kris (1996). ‘Comparing Similar Countries’. Res
Publica, 38: 215–36.
——and Dumont, Patrick (1999). ‘The Belgian Party System(s) on the Eve of Disintegration’, in
David Broughton and Mark Donovan (eds.), Changing Party Systems in Western Europe. London:
Cassell.
—— –––– (2000). ‘Belgium: Subjects of Partitocratic Dominion’, in Knut Heidar and Ruude
Koole (eds.), Parliamentary Party Groups in European Democracies. London: Routledge.
——, Frognier, André-Paul, and Rihoux, Benoît (1996). ‘Belgium: Still the Age of Party Government?’,
in Jean Blondel and Maurizio Cotta (eds.), Party and Government. London: Macmillan.
—— and Thierry, Laurent (1995). ‘The Belgian Parliament and European Integration’. Journal of
Legislative Studies, 1(3): 75–91.
——, Timmermans, Arco, and Dumont, Patrick (2000). ‘Belgium: On Government Agreements,
Evangelists, Followers, and Heretics’, in Wolfgang C. Müller and Kaare Strøm (eds.), Coalition
Government in Western Europe. Oxford: Oxford University Press.
280 Lieven De Winter and Patrick Dumont
Dierickx, Guido and Majersdorf, Philippe (1993). La Culture Politique des Fonctionnaires et des
Hommes Politiques en Belgique. Brugge: Vanden Broele.
Eijffinger, Sylvester C. W. (1997). Independent Central Banks and Economic Performance.
Cheltenham: Edward Elgar.
Fitzmaurice, John (1996). The Politics of Belgium: A Unique Federalism. London: Hurst.
François, Aymé (1997). De l’Etat Unitaire à l’Etat fédéral. La Dynamique Institutionnelle de la Belgique.
Louvain-la-Neuve: Ciaco.
Frognier, André-Paul (1993). ‘The Single Party/Coalition Distinction and Cabinet Decision-Making’,
in Jean Blondel and Ferdinand Müller-Rommel (eds.), Governing Together. London: Macmillan.
—— (1997). ‘Belgium: A Complex Cabinet in a Fragmented Polity’, in Jean Blondel and
Ferdinand Müller-Rommel (eds.), Cabinets in Western Europe, 2nd edn. London: Macmillan.
Grilli, Vittorio, Masciandaro, Donato, and Tabellini, Guido (1991). ‘Political and Monetary
Institutions and Public Financial Politicies in Industrial Countries’. Economic Policy, 13: 341–92.
Hondeghem, Annie (1996). ‘De Politieke en de Ambtelijke Component in het Openbaar
Bestuur’, in Rudolf Maes and Kathleen Jochmans (eds.), Inleiding tot de Bestuurkunde. Brussel:
Studiecentrum Open Hoger Onderwijs.
Iker, Laura (1998). ‘Evolution des Règles de Financement et de Contrôle des Partis Politiques’.
Courrier Hebdomadaire du CRISP: 1607–9.
Lammens, André (1993). Het Bankroet van België. Antwerpen: Darbo.
Langerwerf, Etienne (1980). ‘Het Stemgedrag in het Parlement. Onderzoek in de Kamer van
Volksvertegenwoordigers Voor de Periode 1954–1965’. Res Publica, 22: 177–88.
Mattson, Ingvar and Strøm, Kaare (1995). ‘Parliamentary Committees’, in Herbert Döring (ed.),
Parliaments and Majority Rule in Western Europe. Frankfurt am Main: Campus and New York:
St. Martin’s Press.
Neels, Leo (1975). ‘Regeringsverklaringen en Regeerakkoorden als Documenten van Toenemend
Publiekrechterlijk Belang. Een Nieuwe Vorm van Publiek Recht?’ Rechtskundig Weekblad,
38: 2370–410.
Perin, François (1960). La Démocratie Enrayée. Essai Sur le Régime Parlementaire Belge de 1918 à 1958.
Bruxelles: Institut de Science Politique.
Rihoux, Benoît (1996). ‘Electoral Reform and Electoral Behaviour in Belgium: Change Within
Continuity . . . Or Conversely’. Res Publica, 38: 255–78.
Senelle, Robert (1983). De Ministerraad in België. Deventer: Kluwer.
Smeets, René (1994). Ouderen en Politiek. Leuven: Davidsfonds.
Strøm, Kaare, Budge, Ian, and Laver, Michael (1994). ‘Constraints on Cabinet Formation in
Parliamentary Democracies’. American Journal of Political Science, 38: 303–35.
Suetens, Mik and Walgrave, Stefan (2001). ‘Belgian Politics Without Ministerial Cabinets?’ Acta
Politica, 36: 180–205.
Tegenbos, Guy (1990). ‘Partijnota: Zes op Tien Hoge Ambtenaren Zijn Kristendemokraat’.
De Standaard, May 9, 3.
Timmermans, Arco (1994). ‘Cabinet Ministers and Parliamentary Government in Belgium: The
Impact of Coalitional Constraints’, in Michael Laver and Kenneth A. Shepsle (eds.), Cabinet
Ministers and Parliamentary Government. Cambridge: Cambridge University Press.
Verminck, Mieke (1986). ‘Concensus en Oppositie in het Belgisch Parlement Tijdens Een
Verkiezingsjaar’. Res Publica, 28: 475–87.
7
I N T RO D U C T I O N
After three decades of constitutional struggles between liberals and conservatives—the
latter more or less aligned with the constitutional monarchy established by the consti-
tution of 1849—the cabinet’s accountability to the lower, or popular, chamber
(Folketinget) of Parliament was finally accepted in 1901. Danish parliamentary democracy
was consolidated in the early twentieth century, but the principle of cabinet accountability
was not formally written into the constitution until 1953. The same reform abolished the
upper chamber, which historically had served to protect conservative interests.
The prevalent Danish Constitution of 1953 is very short. It does not say much
about the formation, operation, and resignation of governments. It does not establish
a constitutional court, and while the Supreme Court has claimed the power of judicial
review, it has actually been reluctant to invalidate legislative decisions. The Danish
Monarch is a prominent figure in the constitutional text, but the functions actually
reserved for the Queen or King are only ceremonial. By law the Danish National Bank
is an independent institution with responsibility for important aspects of monetary
policy, but it is not without links to the world of politics. One-third of the Bank’s
Council members are MPs elected by Parliament, and the minister of the economy
‘oversees’ (as it is formally expressed) its activities.
Denmark is basically a small, unitary state with only about 5 million inhabitants.
Thus, federalism does not complicate the chain of political delegation. Yet, the Danish
public sector is one of the most decentralized in the world, and the government annu-
ally bargains with the organizations of local and regional governments (‘Kommunernes
Landsforening’, ‘Amtsrådsforeningen’) in order to coordinate macroeconomic policy
with the activities of individual local and regional governments (Blom-Hansen 1998).
According to the Constitution, legislative power is shared by Parliament and govern-
ment (‘the King’). In reality the head of state (the Monarch) no longer has political influ-
ence. However, as will be discussed below, the electorate may, in some instances, have
a direct say in the adoption of laws and other decisions. The constitution prescribes that
bills must go through three parliamentary readings before a final vote can be taken. The
Standing Orders of the Folketing state that a bill may be referred to a committee after
282 Erik Damgaard
the second reading. Committees may receive advice and evidence from various
interested parties. The committee in question—usually one of the permanent, special-
ized committees that reflect the party composition of full Folketing—then reports
to the floor, which makes the final decision at the end of the third reading. More than
one-half the MPs must be present for a decision to be made, and the decision rule is
50 per cent ⫹ 1 of all valid votes cast (ignoring abstentions). The budget proposal is
dealt with in the same way as other bills. The government may issue more detailed
rules than those adopted in the form of law (delegated legislation), provided that such
government decrees are authorized by laws. They are known as ‘bekendtgørelser’, that
is, statutory instruments.
The procedures for amending the Constitution are strongly biased in favour of
the status quo. While three readings are required before ordinary bills can be enacted,
a much more demanding procedure is required for constitutional changes. A bill to
amend the Constitution must first be passed by the Folketing. If the government
favours the amendment, it must then call for the election of a new Folketing, which
must adopt the identical amendment once again. Finally, the amendment must be
submitted to a referendum for approval or rejection within 6 months after its final
passage. The amendment is adopted if a simple majority of those voting in the referendum
favour it, provided that this majority constitutes at least 40 per cent of the electorate.
As these rules obviously give veto power to several players, a high-level of consensus is
required to change the Constitution.
Given these observations, Denmark would appear to be rather close to an ideal–typical
conception of parliamentary democracy in which voters, Parliament, and government
are the crucial actors. However, at least five general features of Danish politics complicate
this simple picture. First, Denmark is a small state, heavily constrained by international
ties. It is a member of NATO (since 1949) and the European Union (EU) (since 1973) as
well as a multitude of other international organizations. These memberships clearly
limit the discretion of Danish authorities, but they also create opportunities for security
and influence that would not otherwise exist (cf. Dahl and Tufte 1974).1
Second, Denmark holds the postwar world record with regard to minority govern-
ment rule in a parliamentary democracy. Except for 20 months during 1993–4, minor-
ity governments have held power in Denmark since 1971 (Damgaard 2000c). Minority
governance is bound to affect the nature of parliamentary delegation and accountabil-
ity in various ways. What does delegation mean when the government does not have
the powers of a majority?
Third, although representative democracy and cabinet accountability are core
elements in the political system, the Constitution allows for—and in some situations
1
It might be argued that the chain of delegation and accountability only breaks down if a particular
instance of delegation is not matched by an appropriate accountability relationship. If delegation, in
Robert A. Dahl’s terms, is ‘a revocable grant of authority’ (Dahl 1989:112–4), there need not be any problem.
However, if delegation turns out to be an ‘alienation’ of power, there certainly is a democratic problem. The
same reasoning can be applied to domestic state–local/regional government relations, although in this case
the Parliament (if not the responsible minister) can at least recall delegated powers. In Kiewiet and
McCubbins’ (1991) terminology, this delegation is certainly not ‘abdication’.
Denmark 283
P O L I T I C A L PA RT I E S
The existing Danish Constitution, like its predecessors from 1920, 1915, 1866, and 1849,
does not mention political parties at all. Nevertheless, parties are key actors in mass
politics, in Parliament, and in government. They must therefore be integrated in realistic
models of parliamentary delegation and accountability.2 As the parties are very cohesive,
or disciplined, one could perhaps argue that they structure delegation and accountabil-
ity between voters, Parliament, and government, without affecting the parliamentary
model. On the other hand, the inclusion of parties probably violates the singularity
principle that each agent has only one principal. An individual MP, for example, may act
as an agent for his voters as well as for his/her party organization.
Constitutionally, Danish MPs are Burkean free agents insofar as they are bound
solely by their own conscience and not by any directives from their voters. In practice,
however, they are also party agents ( Jensen 1996; Damgaard 1997). MPs generally think
that they are perfectly able to represent their party and constituency simultaneously,
although these interests might sometimes conflict. While some MPs spend a lot of time
on constituency-related matters, on the whole they are party representatives, even if
some claim that their own convictions are those that ultimately matter. While this
might be true, candidates must nonetheless choose a party in order to become a Member
of Parliament in the first place.
That the electoral law recognizes political parties is revealed by the fact that it
presupposes the existence of such organized bodies. In fact, the basic structure of party
2
Gunnar Sjöblom (1995) thus describes Danish representative democracy in terms of a ‘party government’
model.
284 Erik Damgaard
organization is adapted to the administrative structure of the election law in the sense
that parties have national, regional, constituency, and local organizations (Bille 1994).
The high degree of organizational uniformity has not, however, been forced upon
parties. No law regulating the structure and internal affairs of parties has ever been
proposed. Parties are regarded as voluntary, private associations, and—although this
has changed—for a very long time they were not interested in direct public subsidies
(Bille 1994).
A distinction should be made between the (extra-parliamentary) membership party
and the parliamentary party groups (of MPs). The latter are recognized in the Standing
Orders of Parliament. Public subsidies for parliamentary party groups were introduced
in the mid-1960s in order to improve the working conditions of MPs and party groups
in terms of staffing and facilities (Pedersen and Bille 1991). This support has sub-
sequently been expanded. It was not until 1986, however, that a law was passed
providing public support to finance the activities and electoral campaigns of membership
parties.3 The parties do not have to account for how the money is spent, but since 1990
they have been required to provide information about their various sources of income
(public support, membership fees, organizations, private enterprises, etc.) in gross
figures. Since 1996 the names of private donors must also be revealed if the donation is
more than DKK 20.000 per year (Elklit and Pade 1996). Finally, while the difference
between the two branches of the parties is important, it is worth keeping in mind
that in practice they are closely interconnected with respect to staffing and financing
(Bille 1994).
Danish political parties and the characteristics of the party system have been
described elsewhere (see e.g. Pedersen 1987; Bille 1994; Damgaard 2000c). No single
party has commanded a majority since the early twentieth century, and in the postwar
period ‘only’ seven parties have been in government, either alone or in coalition. These
seven parties include four ‘old parties’ and three ‘new parties’. The four old ones are the
Social Democrats (‘Socialdemokratiet’, SD), the Social Liberals (‘Det Radikale Venstre’,
RV), the Liberals (‘Venstre’, V), and the Conservatives (‘Det Konservative Folkeparti’,
KF). The new parties that have participated in government are the Center Democrats
(‘Centrum-Demokraterne’, CD), the Christian People’s Party (‘Kristeligt Folkeparti’,
KRF), and the Justice Party (‘Danmarks Retsforbund’, DR). Parties of the extreme Left
and Right have never been in government.
D E L E GAT I O N F RO M VOT E R S TO M P S
The Constitution states that no more than 175 MPs (apart from two elected in
Greenland and two in the Faroe Islands) are to be elected to a 4-year term using a
proportional representation formula to be set out in the election law. It also states that
3
According to the rules introduced in 1987, each party (or independent candidate) obtaining at least 1,000
votes was entitled to 5 DKK per year per vote obtained in the latest Folketing election. In 1996 the amount per
year per vote was increased to 19,50 DKK. Thus, for parties that obtain at least 1,000 votes public funding is
exactly proportional to the distribution of the national vote, regardless of whether or not the party wins seats.
Denmark 285
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
Cabinet Formation
According to the Constitution, the Monarch formally appoints the Prime Minister and
other ministers. There are no positive constitutional rules governing the cabinet forma-
tion process, only informal norms that have gradually evolved during the twentieth
century. On the other hand, the Constitution explicitly states that no minister can
remain in office if the Parliament has passed a motion of no confidence against him/
her. The implication is that Danish cabinets are formed according to the principle of
‘negative parliamentarism’ (Rasmussen 1972), which does not require a vote of investi-
ture. A government can stay in office as long as it is tolerated by Parliament, that is, if
there is no explicit majority against it. Negative parliamentarism generally facilitates
minority cabinets (Bergman 1993, 1995), and Denmark is a clear case in point.
Denmark’s informal rules governing cabinet formation (Damgaard 2000c) require
that the parliamentary parties formally advise the Monarch. If their advice unambigu-
ously points to a majority cabinet or a cabinet with the assured support of a majority
in Parliament, then the Monarch is obligated to appoint such a cabinet. If no such
majority emerges, the goal is to appoint the minority cabinet most likely to survive.
The whole process is one of bargaining among parties, which often prefer that in the
first stage a leading parliamentarian is appointed by the Monarch to serve as a ‘royal
informateur’. When advice from the parties is unclear, it is the acting Prime Minister,
and not the Monarch, who is responsible for interpreting it.
4
The delegation/accountability relationships become even more complex if one considers MPs to be
neither voters or parties agents nor principals, but political representatives with a complex role involving
interaction with voters, the interests of which should be taken care of, and independent judgement as to the
best decision to make (cf. Pitkin 1967; Eulau and Karps 1977). Note that Esaiasson and Holmberg (1996;
cf. Holmberg 1997) talk about ‘representation from above’ in the similar Swedish case.
Denmark 287
proposal under debate, that the government will resign or call elections if the proposal
is rejected by Parliament. Governments may also receive a vote of confidence during
debates on interpellations and government policy. The procedure is the same for votes
of no confidence.
Technically speaking, such matters of confidence are normally settled using a
procedure that the Standing Orders, until 1997, termed a ‘proposal to pass on to the
next item on the agenda (proposal for a resolution on the order of business)’. It is now
called a proposal for a ‘decision’. This important instrument is very flexible because it
allows a parliamentary majority to state its opinion on the government and on policies
in any area. It can be used to instruct the government without necessarily phrasing the
request in terms of confidence. Unless no confidence is explicitly mentioned, such
parliamentary declarations are not legally, though usually are politically, binding. The
instrument was frequently used against minority governments between 1982 and 1988
by the so-called ‘alternative majority’ on defence and security matters, and led to
Danish ‘footnotes’ in NATO documents.
There are no statistics on the number of no-confidence votes taken. They would be dif-
ficult to produce insofar as one cannot simply count the number of votes on ‘motions for
a resolution on the order of business’ or ‘decisions’. Many of these do not concern confi-
dence as such, and it is often not easy to decide the extent to which they do. What can be
said, however, is that since 1945 only two governments have been removed by a
no-confidence vote—in 1947 and 1975. In addition, the rejection of the government’s
budget bill for 1984 was interpreted as a no-confidence vote, and the government called
new elections. In 1993 Prime Minister Schlüter resigned, probably anticipating a defeat in
Parliament because of a ‘scandal’ (‘Tamilsagen’). However, (minority) governments more
commonly resign, or call elections, when they are unable to obtain parliamentary sup-
port. Cabinets dissolving Parliament (which cannot dissolve itself ) are not required to
resign from office. Often they may in fact be able to stay in power after the election.
Although no individual minister has ever been forced to resign by the passage of a
motion of no confidence, there are some cases in which ministers have resigned
because they anticipated a vote of no confidence ( Jens Peter Christensen 1997: ch. 24).
Schlüter V in 1993 is one such instance. However, it is often difficult or impossible to
establish whether a minister who resigned would actually have been forced out by a
no-confidence vote had he not left office voluntarily.
The Prime Minister has de facto power to dissolve Parliament at almost any time,
although it is the Monarch who formally dissolves it on the proposal of the PM. The
only restriction on the power of dissolution is that a new cabinet cannot dissolve
Parliament before it and Parliament have met in a public meeting. In deciding whether
or not to dissolve Parliament, the PM must pay attention to the views of coalition
partners. For example, in the autumn of 1997 the PM chose not to call elections—
although he and his party actually wanted to do so—because of the views of his
coalition partner. In all, twenty cabinets have ended in early elections, eighteen of
which were dissolved by the PM. Table 7.1 summarizes the behavioural record with
respect to cabinet investiture, no-confidence and confidence votes, and parliamentary
dissolution.
Table 7.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Denmark 1945–2000 a
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the cabinet Who
formation votes (iv) votes (cv) end with an dissolved
n.a. No. Cabinet Cabinet under early election? Parliament
of removed resigned to specific in this
ncv by ncv pre-empt ncv institutional case?b
mechanism
n.a.
with which ministers must comply, including the duty to ‘speak the truth in front of
parliament’, to ‘comply with the rules concerning the disbursement of state funds’, to
‘comply with acts of parliament’, and the duty to ‘take action as head of the government
administration’ ( Jens Peter Christensen 1997: 604).
D E L E GAT I O N W I T H I N C A B I N E T
The Danish Prime Minister is more than a primus inter pares in relation to the cabinet
colleagues from his own party. Although the formal act is carried out by the head of
state, the Prime Minister alone appoints and dismisses ministers from his party, and he
must approve, at least formally, the ministers proposed by coalition partners because
they subsequently become members of ‘his’ cabinet. Since 1964 the Prime Minister has
a small secretariat (cf. Statsministeriet 1988) staffed by civil servants who monitor the
various government departments, though they rarely directly interfere in department
activities because individual ministers are quite autonomous. Furthermore, since coalition
292 Erik Damgaard
governments have become the rule in Danish politics, there are limits on the Prime
Minister’s authority over ministers from other parties.
The PM chairs the weekly meetings of the cabinet, which has about twenty
members. All politically important topics are dealt with in cabinet meetings, if only to
make sure that there is agreement. Interministerial differences will normally be sorted
out before an item reaches the cabinet. The agenda is set by the PM, who receives pro-
posals from individual ministers. Items discussed in cabinet meetings include government
and private member’s bills and draft resolutions, ministerial reports to Parliament,
interpellations, questions that might be controversial in Parliament, ongoing bargaining
with other parties, important appointments, the establishment of committees, etc.
(Olsen 1988).
Except for the special status of the Prime Minister and the important role of the
minister of finance over the budgets of spending ministries, Danish cabinet ministers
are formally equal and quite powerful within their respective jurisdictions, which in
principle are mutually exclusive. However, because irrespective of formal jurisdiction
issues are often of interest to more than one minister, interministerial conflicts of inter-
est are not unknown. The decision rule applied in cabinet meetings under such
circumstances might best be described as ‘consensus as defined by the PM’.
The constitution dictates that ministers’ jurisdictions are to be decided by the
Prime Minister (formally the Monarch). Therefore, Parliament cannot interfere with
the distribution of offices and jurisdictions among the government ministers, although
parliamentary majorities have in fact tried to do so. Ministers do not have junior ministers
or secretaries of state at their disposal (though several of them have found ways to
appoint informal political assistants). ‘Cabinet’ and ‘government’ are therefore synonymous
terms in Denmark.
Cabinet ministers are the heads of their departments but not dictators within their
jurisdictions. They form part of a team directed by the Prime Minister, who also needs
to take into account the views of coalition partners. Danish governments are equipped
with various mechanisms to coordinate and, if necessary, settle interdepartmental
disputes. In addition to regular cabinet meetings, governments have some kind of inner
cabinet and a varying number of issue-specific cabinet committees (Olsen 1978;
Christensen 1985; Olsen 1988). One cannot say that the cabinet may, if it wants, take
decision-making power away from individual ministers, but a minister needs the approval
of the Prime Minister and cabinet colleagues for politically important decisions.
It has been argued that longevity in ministerial position and previous parliamentary
experience affect the performance of ministers. With the exceptions of the Prime
Minister and the minister of foreign affairs, recent analysis has shown a decline in indi-
vidual ministerial longevity since 1945, and particularly since 1960. Thus, the average
period in office between 1945 and 1960 was 3 years and 8 months, whereas it was
2 years and 11 months during 1960–78 (and 2 years and 10 months 1978–94). These
figures can be compared to 5 years and 3 months in 1920–40 (Kirk and Knudsen 1996).
Kirk and Knudsen argue that as a consequence an increasing proportion of a minister’s
time in office is spent ‘learning the job’. Their proposition that this development has
weakened ministers is doubtful, however, because of other factors that also play a role
in political leadership. For example, there is no corresponding decline in the proportion
Denmark 293
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
Traditional Danish administrative policy states that a government minister is responsible
for his/her governmental department. While he or she might not be personally respons-
ible for mistakes made by civil servants (cf. below), ministerial accountability is the basic
norm. There is a direct line of command from the minister downwards through the
administrative bureaucracy. Higher-level civil servants are appointed by the minister, with
the support of his/her cabinet colleagues in the cabinet’s committee on appointments.
Ministers are therefore in a position to control the careers of top civil servants.
Danish civil servants are usually recruited on the basis of merit. This usually requires
that they hold a university degree in law, economics, or political science, although
special agencies may require more technical degrees. Civil servants are supposed to be
politically neutral and to report only to their own minister. Promotions are also
granted on the basis of merit.
A large number of problematic cases about the duties and responsibilities of civil
servants as opposed to ministers have led to a series of official investigations. A recent
summary of the results of these investigations states that civil servants are obliged to
‘inform and counsel the minister’, to ‘warn the minister (the “duty to put the foot
down”)’, and to ‘refuse to carry out clearly illegal orders by the minister’. Similarly, the
minister’s responsibility vis-à-vis civil servants include the duty to ‘question his or her civil
servants in such a manner that the minister does not deliberately deprive him or herself
of knowledge about important factors’. The minister is also obligated to ‘respect the
division of labour between the minister and the civil servants—including the duty to
respect the fact that the professional background and responsibilities of the civil service
establish limits on the right of the minister to give instructions’. Finally, the minister must
‘respect general principles of administrative law’ ( Jens Peter Christensen 1997: 604–5).
The relationship between ministers and civil servants is a hotly disputed issue in
contemporary Danish government. Because there are no partisan junior ministers or
secretaries of state, civil servants are more or less required to perform genuinely political
tasks. Civil servants have consistently opposed the idea of introducing junior ministers,
294 Erik Damgaard
and they do in fact willingly give ministers political advice. However, they do not
appear publicly in the capacity of political advisors. They might sometimes get into
trouble as a result of the dilemma between serving the minister and being an adminis-
trative expert of the classical type, but normally the cooperation between civil servants
and ministers works to their mutual satisfaction. Only in purely partisan matters will
civil servants resist requests to assist their ministers.
The integration of organized interests into the making and implementation of public
policy has important consequences for the traditional, judicial, and parliamentary insti-
tutions of accountability. ‘Delegated administration’ in the Danish labour market,
agriculture, and cooperative housing, for example, has weakened relationships of political
accountability and led to quite a few scandals involving the misuse of public funds during
the 1980s and early 1990s. However, none of the cases involved outright corruption
(Christensen 1993).
There is really no tradition of state-owned enterprises and nationalized industries in
Denmark. However, public corporations organized in accordance with Danish company
law, with the central government owning all or part of the shares, do exist, and new
ones have increasingly been established over the past two decades. The same is true of
public enterprises with their own boards of directors (which are exempted from
unrestricted departmental supervision).5 The whole development may be interpreted
as a shift from hierarchical governance towards market governance. Parliamentary and
governmental control are thus clearly reduced insofar as there is now less direct
accountability to political authorities. Still, note that when the international privatisation
movement started in the early 1980s, Denmark had few nationalized industries to
privatise (Damgaard 1989; Christensen and Pallesen 1997).
Another development in the past 25 years is a steady process of delegation from polit-
ical executives to civil servants at lower administrative levels. For example, the reformed
budgetary system makes it possible for public institutions to transfer resources from one
spending item to another, and to transfer resources from the present year to the next
one (Christensen and Pallesen 1997). This obviously reduces financial control from
the top.
In sum, ministers are in charge of their departments and ultimately responsible to
Parliament. Civil servants are accountable to their minister. As indicated above,
however, it is not quite clear whether a minister always controls or directs the behaviour
of civil servants. The civil servant may also to some extent represent the interests of the
bureaucracy. At the same time, administrative developments in recent decades imply
that a number of areas are no longer directly controlled by government ministers.
Consequently, they cannot be held accountable for all relevant activities.6
5
Thus, in 1975 there were forty-one organizations of one of these two types. By 1995 this number had
more than doubled, to eighty-five. In addition some genuine privatization has occurred during the same
period (involving a life insurance company and a mining company with a number of associated companies).
Furthermore, several existing government organizations have acquired a semi-autonomous status
(Christensen and Pallesen 1997).
6
Moreover, the freedom of information act of 1964 entitled citizens to see government documents
in cases to which they were parties. In 1971 a new law introduced a general right (with some exceptions,
Denmark 295
E X T E R NA L C O N S T R A I N T S
As mentioned by way of introduction, it has become accepted that Danish courts can
rule on the constitutionality of specific pieces of legislation, although the courts have
been cautious in their exercise of this authority (cf. Alivizatos 1995). This caution
might, however, change in the future. A small step in the direction of judicial activism
was recently taken when the Supreme Court (surprisingly) allowed a group of citizens
to sue the Prime Minister for having signed the Maastricht Treaty, which, they claimed,
violated the Constitution. The PM was acquitted by a High Court in 1997 and by the
Supreme Court in 1998. However, there will almost certainly be more future cases,
because the Supreme Court also ruled that it alone has the authority to decide on the
constitutionality of laws. In 1999 the Court even overruled a law. Consequently, the
Danish court system will become more involved in political matters than it has been in
the past.
The Danish National Bank must also be taken into account by the government,
particularly if its director propounds the opinions of the bank in the media. While it is
difficult to measure the Bank’s influence, it seems to have been quite important in the
making of Danish monetary policy, not least policy with regard to liberalizing interna-
tional capital flows (i.e. deregulation of foreign exchange control) in the 1980s ( Jensen
1989). One study concluded that, from a governmental point of view, the National
Bank is ‘one of the most influential partners in economic policy’ (Bent Christensen
1985: 172). It appears, however, that its influence varies across issues and over time.
A third point to be made is that Danish voters do not delegate all their power to
Parliament. The constitution of 1953 requires the electorate to make the final decisions
in some matters—including changes in voting age, constitutional changes, and, under
certain circumstances, delegations of power to international authorities (in particular
when the Parliament passes a bill to that effect by a majority that is smaller than a
qualified absolute majority of five-sixths). In addition, decisive referendums on bills
passed in Parliament, with certain exceptions (including taxation), can be organized at
the request of one-third of MPs. Thus, defeated parliamentary minorities can, in some
instances, appeal to the electorate to reverse a parliamentary decision. Finally,
Parliament can order consultative referendums. Danish voters, however, have no right
to initiate referendums (see Svensson 1996).
Since 1945 there has been one referendum on amending the constitution (1953), five
referendums about changing the voting age (1953, 1961, 1969, 1971, 1978), six on the
EC/EU (1972, 1986, 1992, 1993, 1998, 2000)7, and four in 1963 about bills passed by
the Parliament—all of which were rejected in referendums. While fifteen referendums
in 45 years may not seem to be all that many, the very possibility of a referendum
influences the behaviour of parties and MPs, because even legislative majorities must
consider the likelihood of its use in any particular situation.
including cases involving the personal and economic relations of individuals, state security matters, and
internal working documents) for citizens to see documents in cases dealt with by government authorities.
7
The 1986 referendum on the Single European Act was consultative but actually binding.
296 Erik Damgaard
In the modern world, a small state like Denmark cannot possibly in any meaningful
and realistic sense be ‘sovereign’. Therefore, the 1953 constitution introduced
a section (§ 20) stating that: ‘Powers vested in the authorities of the Realm under this
Constitutional Act may, to such extent as shall be provided by statute, be delegated to
international authorities set up by mutual agreement with other states for the promotion
of international rules of law and cooperation.’ This is the constitutional basis for
Danish membership in the EC/EU.
Since the very first days of Danish membership (1973), a major issue has been how
the Folketing can make sure that the government, acting in the EU Council of
Ministers, does not commit the country in ways that the parliamentary majority does
not approve. The solution to this problem is the European Affairs Committee (EAC)
(previously named the Market Relations Committee). Its role in Danish EU policy-making
is that the minister in question obtains a bargaining mandate from the committee, and
that the government provides the committee with timely and full information on EU
developments. The specific details with regard to the latter have changed over time (see
e.g. Arter 1995; Dag Arne Christensen 1997; Sidenius 1997; Sidenius et al. 1997), mainly
in an effort to further improve parliamentary control, openness, and legitimacy. The
Maastricht Treaty and the associated Edinburgh Agreement did not fundamentally
alter the parliamentary control system.
The EAC is involved in all EU matters, and it is the only Folketing committee that can
give the minister a bargaining mandate. A mandate requires that there is no majority (in
Parliament as a whole) against the position proposed by the government. Thus, the
Committee delegates power to the government, which is accountable to the Committee
and the whole Parliament. Other standing committees also receive information from
the government on matters within their jurisdiction (Sidenius et al. 1997), for example
the committees on the environment, legal affairs, and taxation, but so far only the
Committee on the Environment seems to have influenced the decision-making process
very actively. Attempts to use plenary sessions for debates on proposed EU-directives and
other measures have not been very successful.
C O N C LU S I O N
It is perfectly legitimate to claim that Danish voters delegate power to MPs, who again
delegate power to the government, the ministers of which in turn delegate power to
civil servants. In each link there is a reversed relationship of accountability: MPs are
accountable to the electorate (at the next election in any case), cabinets and ministers
are accountable to Parliament at all times, and civil servants are accountable to their
minister. Ex ante mechanisms of control exist in all these relationships. Voters have a
choice among several candidates with widely divergent policy preferences, a majority
of MPs decides which government is acceptable, and ministers direct the work of civil
servants. Similarly, ex post control mechanisms are found in all relationships. Voters
may refuse to re-elect MPs, MPs may force cabinets and ministers out of office, and
ministers may order changes in administrative rules and decisions.
In all the delegation/accountability relationships mentioned, agents try to anticipate
the reactions of their principals. MPs think a lot about how the voters will react at the
Denmark 297
next election (relevant information is constantly provided by public opinion polls), and
they adjust their behaviour and decisions accordingly. Cabinets and individual ministers
shape their proposals according to their expectations about how Parliament will react to
them, and civil servants adjust their advice and proposals according to their knowledge
of the preferences of the ministers.
All of this sounds as if we are talking about an ideal type of parliamentary democracy.
Looking at the real world of politics, however, this description leaves out some important
aspects. Keep in mind that voters are party voters, and that MPs and ministers are party
representatives. Furthermore, MPs and ministers are not only ultimately agents of the
voters. The case can be made that MPs and ministers as political leaders play a very
active role in relation to voters, when the latter do not actually have clear preferences or
goals. On the other hand, Danish voters do not delegate all their power to Parliament.
The electorate sometimes decides issues in referendums. As of 2000, voters have actu-
ally rejected six of the fifteen proposals that have been submitted to referendums.
The links between Parliament and government in Denmark are more complex than
assumed by the ideal model of parliamentary democracy. The cabinets formed are almost
always minority cabinets, which implies that the delegation of power to cabinets is very
conditional. The parliamentary instruments of control are generally highly effective. This
also applies to the very important EU policies of the Danish government. The European
Union is certainly a significant environmental constraint on Danish policy-making, and
the EAC attempts to preserve national parliamentary influence over EU policy.
While such national actors as the courts and the central bank play important roles in
specific incidents, they are generally not as politically important as their counterparts
in, say, Germany. The role of organized interests (including local and regional govern-
ments) should not be neglected. A recent study on Denmark and Norway has shown a
decline in the use of traditional corporatist institutions (interest organization member-
ships in various boards, councils, and commissions) in the past two decades, but during
the same period there has been an increase in the frequency and intensity of informal
interest group contacts with public authorities. In particular interest organizations have
increased their contacts with parliamentary committees, party groups, and MPs—
presumably because the parliaments of the two countries since the 1970s have increased
their power vis-à-vis the administration (Christiansen and Rommetvedt 1997).
Though the delegation and accountability perspective is a simplification of reality, it
has proved to be useful for a general understanding of Danish parliamentary demo-
cracy. Within the limits established by the present analysis a few conclusions can be
drawn on possible agency problems in Denmark. First, Danish voters appear to be well
represented by MPs through ex ante and ex post control mechanisms as well as anticipated
reactions. The proportional representation election system—with its low barriers
for party representation and preferential voting within parties—assures a high level of
representativeness. But it does not always assure a congruence of popular and elite policy
positions. Although on some occasions the electorate is called upon to make the final
decision in a referendum, most decisions must be made by responsible MPs regardless
of whether a majority of voters are in favour. After all, MPs are not only agents of the
electorate, they are also expected to govern the country.
298 Erik Damgaard
Second, Danish MPs decide who should be allowed to form the cabinet. Very often
the outcome is a minority cabinet in need of external parliamentary policy support.
Furthermore, a variety of control instruments is available to MPs and party groups in
relation to any cabinet, but these instruments are particularly serious if applied to
minority cabinets. In short, the Danish Parliament is a strong principal vis-à-vis the cabinet
as agent. On the other hand, government ministers usually have much better access to
relevant information than MPs, which they might conceivably use to their own advantage
by hiding or misrepresenting information. In the important area of EU policy-making
a standard complaint has been that the EAC is informed too late to really influence the
Danish policy position. However, various steps have recently been taken to involve the
EAC at an earlier stage of EU policy-making.
Third, individual cabinet members are relatively autonomous within their jurisdic-
tions, although they are all collectively responsible to Parliament. However, it is often
discussed whether ministers are able to control their own departments, which often
work closely with organized and institutionalized interests. The evidence is mixed
and the discussion so far inconclusive, but the need for ministers to appoint political
assistants has finally been acknowledged (Betænkning 1998).
REFERENCES
Alivizatos, Nicos C. (1995). ‘Judges as Veto Players’, in Herbert Döring (ed.), Parliaments and
Majority Rule in Western Europe. Frankfurt am Main: Campus, and New York: St. Martin’s Press.
Arter, David (1995). ‘The Folketing and Denmark’s “European Policy”: The Case of an
“Authorising Assembly”?’. Journal of Legislative Studies, 1 : 110–23.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
—— (1995). Constitutional Rules and Party Goals in Coalition Formation. Umeå: Department of
Political Science.
Bille, Lars (1994). ‘Denmark: The Decline of the Membership Party?’, in Richard S. Katz and
Peter Mair (eds.), How Parties Organize. London: Sage.
—— (1993). ‘Candidate Selection for National Parliament in Denmark 1960–1990’, in Tom Bryder
(ed.), Party Systems, Party Behavior and Democracy. Copenhagen: Copenhagen Political Studies
Press.
——(1992). ‘Denmark’, in Richard E. Katz and Peter Mair (eds.), Party Organizations. London: Sage.
Blom-Hansen, Jens (1998). ‘Macroeconomic Control of Subcentral Governments: Experience
from America and Scandinavia.’ Government and Policy, 16: 323–40.
—— (2000). ‘Still Corporatism in Scandinavia? A Survey of Recent Empirical Findings’.
Scandinavian Political Studies, 23: 157–81.
Betænkning (1998). Forholdet mellem minister og embedsmand. Copenhagen: Statens Information
(Betænkning no. 1354).
Christensen, Bent (1985). Nationalbanken og forvaltningsret. Copenhagen: Jurist- og
Økonomforbundets Forlag.
Christensen, Dag Arne (1997). ‘Europautvala i Danmark, Sverige og Noreg: Sandpåstrøningsorgan
eller politiske muldvarpar?’. Nordisk Administrativt Tidsskrift, 78: 143–62.
Christensen, Jens Peter (1997). Ministeransvar. Copenhagen: Jurist- og Økonomforbundets
Forlag.
Denmark 299
I N T RO D U C T I O N
The birth of parliamentary democracy in Finland is closely linked to the struggle for
independence. In the aftermath of the 1905 revolution in Russia and a general strike in
Finland, the Constitution was rewritten in 1906. The previous Constitution was based
on the 1772 Swedish Constitution. The new Constitution established—for the first time
in a European country—universal suffrage. At the same time the old four-estate assembly
was replaced by a unicameral Parliament, the Eduskunta. The first general nationwide
parliamentary elections were held in 1907. In the summer of 1917 the Constitutional
Committee approved a regulation on parliamentary governance. Following the revolutions
in Russia, Finland declared independence in December 1917. Parliamentary governance
was consolidated in the 1919 Constitution Act.1 Section 2 of the Form of Government
Act read:
Sovereign power in Finland shall belong to the people, represented by Parliament convened in
session. Legislative power shall be exercised by Parliament in conjunction with the President of
the Republic. Supreme executive power shall be vested in the President of the Republic. In
addition, for the general government of the State there shall be a Council of State comprising the
Prime Minister and the requisite number of Ministers. Judicial power shall be exercised by
independent courts of law, at the highest instance the Supreme Court and the Supreme
Administrative Court.
The inter-war period saw the gradual consolidation of parliamentarism and the main
political parties. The short but bloody civil war of 1918 had bitterly scarred the defeated
Reds. Of twenty-three cabinets that held office between the declaration of independence
and the outbreak of Winter War in late 1939, all but two were built around bourgeois parties.
We would like to thank Heikki Paloheimo (University of Turku) for his assistance in compiling the
government data and Voitto Helander (Åbo Akademi) for his encyclopaedic wisdom. For a general overview
of the Finnish political system, see Pesonen and Riihinen (2002).
1
Finland used to have four different laws which together constituted the Constitution. These were the Form
of Government Act (1919), the Parliament Act (1928), the Act Concerning Ministerial Responsibility (1922), and
the High Court of Impeachment Act (1922). There have not been any major constitutional reforms. A new
unified Constitution entered into force in March 2000. Note that this chapter presents primarily the situation
under the old constitution. See www.eduskunta.fi for the current version.
302 Tapio Raunio and Matti Wiberg
The Second World War changed the political situation in Finland, and the leftist parties
were brought into the government. Since the war, Finnish parliamentary democracy
has been influenced by a number of different constraining factors. The imperative of
what—in popular language—was referred to as the ‘electoral district of Moscow’, cast
a long shadow on almost all aspects of Finnish politics for several decades. More
recently, membership of the European Union (EU) has since 1995 added a new dimen-
sion to Finnish politics. During the 1970s and 1980s, Finland was also a strongly corpo-
ratist country (Paloheimo 1990). Another constraint on Finnish parliamentarism is the
role of the country’s President.
The Finnish system combines parliamentary democracy with a strong presidency
and is therefore characterized as semi-presidential (Arter 1987; Nousiainen 1992b;
Jansson 1993; Hidén and Saraviita 1994). The President has wide-ranging powers,
especially in foreign policy. In its 1917 report, the constitutional drafting commission
stated that the executive branch should have a fairly independent position based on the
authority of the President. Compared with other Western European chief executives,
the Finnish President is a strong political player—provided the person who occupies the
post chooses to exercise his powers. The post-Second World War Presidents have used
their powers in varying degrees, with Urho Kekkonen (1956–81) testing the limits of
presidential power more than the others—sometimes even overstepping the formal
boundaries of presidential authority.
The legislative initiative is vested in the President, the Council of State—that is, the
government, consisting of the Prime Minister (PM) and other ministers—and individ-
ual MPs. Private members’ bills are commonly proposed, but less than 1 per cent are
successful. New laws generally originate in legislative proposals from the Council of
State. A government proposal may contain one or more laws or amendments to
existing laws. It is the President who decides, in a plenary sitting of the Council of State
and at the suggestion of the cabinet, that the government will present a proposal to the
Parliament. Following a submission debate, legislative matters require three readings. In
the first reading the plenary sends the bill to a special committee for preparation.
No decisions are taken concerning the form or content of the law in this reading. These
matters are settled in the second reading on the basis of statements by the standing
committee. Finally, the bill must either be accepted or rejected by a simple majority in
the third reading. This can take place no earlier than the third day after the second
reading. Until September 1992 a bill could also be postponed until after the next
election if one-third (i.e. sixty-seven) of MPs so decided. Amendments to the
Constitution are left in abeyance by a majority of the votes cast. In the first parliamen-
tary session after the elections, the new Parliament must adopt the bill without mate-
rial alterations with two-thirds of the votes cast. Constitutional amendment can also be
declared urgent by five-sixths of the votes cast. In such cases, the bill can be adopted by
the same Parliament with two-thirds of the votes cast (see Wiberg 1994b and Rasch
1995 for details).
There is also one category of general rules below laws: The decisions of the Council
of State (Valtioneuvoston päätös). These are given and implemented without parliamentary
involvement. Their role is to specify the implementation of laws and decrees, and they
therefore contain only technicalities.
Finland 303
P O L I T I C A L PA RT I E S
Finland’s party system has traditionally been characterized by a high degree of frag-
mentation. As measured by the effective number of parties, Eduskunta is one of the
most fragmented parliaments in Europe (see Lane, McKay, and Newton 1997: 163).
Multiparty democracy has historically been rooted in the permanence of five main
parties (Sundberg 1999).
There is no extreme right-wing party in Finland. The moderate right is represented
by the National Coalition (KOK, Kansallinen kokoomus), the Swedish People’s Party
(RKP, Ruotsalainen kansanpuolue/Svenska folkpartiet), the Centre Party (KESK, Suomen
Keskusta), and the Christian Union (SKL, Suomen Kristillinen Liitto; since 2001 Christian
Democrats, Kristillisdemokraatit, KD). Among the leftist parties, the centre-left Social
Democratic Party (SDP, Suomen Sosialidemokraattinen Puolue) is traditionally the
strongest, while the Left Alliance (VAS, Vasemmistoliitto) is a collection of socialists and
communists of various colours. The rural and peripheral areas are mainly represented
by the Centre Party, the Christian Union, and the True Finns (PS, Perussuomalaiset, the
political successor of the Rural Party, SMP). The Left–Right dimension is the dominant
political cleavage in Finland. Other salient and partly overlapping cleavages are
centre–periphery, urban–rural, national–international, language (Finnish–Swedish),
environmentalism, and secularism–religion. During the 1990s the national–interna-
tional dimension increased in significance, which also highlighted the centre–periphery
cleavage (Raunio 1999). The National Coalition, the Social Democratic Party, the
Green League (VIHR, Vihreäliitto), and the Left Alliance draw their support mainly
from the urbanized southern parts of the country. In popular support, the Green
League is one of the strongest environmental parties in Europe, and in 1995 it became
the first green party to gain a cabinet seat in Europe. In practice, the small Christian
Union and the True Finns have been politically marginalized. No political party has
ever gained a parliamentary majority of its own. The all-time maximum Shapley-
Shubik voting power of a parliamentary party group in Finland is 0.4, meaning that no
single party has ever had more than 40 per cent of the voting power in the Eduskunta
(Wiberg and Raunio 1997).
Although parties are not mentioned in the Constitution, in reality they dominate
politics. Finnish political parties are cohesive in their parliamentary voting behaviour
(Wiberg 2000). Political parties were first legally recognized in the 1969 Party Act,
which gave them a privileged status in elections and in the allocation of public funds.
Both extra-parliamentary and parliamentary party organizations receive public funding
based on the share of seats won in the most recent parliamentary election. Parties that
are represented in Parliament have been publicly funded since 1968 and parliamentary
party groups since 1967. Parties without parliamentary representation do not get
public funding (see Wiberg 1991a). The public support of parties takes two main forms:
Money is allocated to the extra-parliamentary party organization (called party support
proper) and also to the parliamentary party groups (called public support for
parliamentary groups). Paragraph 9 of the Party Law defines the legal basis for Party
support as follows: ‘Within the limits set by the state budget, a political party
represented in Parliament may be allocated financial assistance in order to support the
304 Tapio Raunio and Matti Wiberg
public activity that has been defined in its rules and general programme. Such support
must be divided in proportion to the parliamentary seats won by the parties in the most
recent parliamentary elections’ (Wiberg 1991a: 73). Various extra-parliamentary party
organizations—such as students’, youth, and womens’ organizations—get public
money both from the state and from (some) municipalities. These kinds of public
support amount to more than five times the sums provided directly to the party organ-
izations (party support proper) (see Wiberg 1991a and 1997 for details). The Finnish
system of public party support is transparent: all relevant documents are available
to anyone who asks for them (no need even to be a Finnish citizen). Candidate
recruitment is decentralized and based on primaries in district organizations.
Otherwise, however, parties are centralized (see Sundberg 1997). Intraparty politics is
run by the elected party elite. Responsibility for day to day party business and bargaining
with others is vested in the party executive organs.
There is little variation among the parties with regard to internal decision-making
and organization.2 According to the Party Law (1969) and the Law on the Election of
the Parliament (1975) all parties must be democratically organized. The national level
organization typically includes the party executive committee, the party council,
various working committees, and youth, women’s, and students’ organizations. The
chairperson (party leader) is the overall head of the party. He chairs the party executive
and has almost always simultaneously held a seat in the Parliament. However, he has
seldom been the head of the parliamentary party group. The leader is selected by the
party congress, normally for 2–3 years, with no formal term limits. A party congress
brings together delegates from throughout the country and is the main forum for
approving political programmes. Congress delegates make active use of their right to
submit motions. While party congresses usually meet only once every 2 or 3 years, the
other national level party organs meet more regularly—at least a few times a year
(Sundberg 1994, 1996).
The parliamentary party group is relatively independent from the rest of its party
during the legislative term. The internal autonomy of the parliamentary party is very
strong because no other party organ meets on a continuous basis. The parliamentary
party presents annual reports about its activities to the party’s central organs, and party
secretaries and/or other staff from the central office attend parliamentary party meetings.
However, the ability of party leaders to influence the reselection of MPs is limited
because candidates are selected by district branches. Therefore, major parliamentary
party groups always include some troublesome MPs with whom the party leadership
is not entirely happy. The independence of the parliamentary party also depends upon
whether or not the party is in government. The decision to join the cabinet is made
jointly by a leading party unit and the parliamentary party. The party executive is more
closely involved in the work of the parliamentary party if a party is in the government
than if it is part of the opposition (Sundberg 1994; Wiberg 2000).
2
The structure of Finnish parties is analysed in Sundberg (1994, 1996) and Sundberg and Gylling (1992).
The internal organization and decision-making of the parliamentary party groups is examined in Wiberg
(2000).
Finland 305
Altogether there have been forty-five cabinets since 1945, most of them including
non-partisan ministers. The average number of cabinet parties is three. There have also
been six caretaker cabinets, the most recent in 1975. Cabinets have normally been over-
sized. During the Kekkonen presidency (1956–81), the usually short-lived governments
were mainly based on coalitions between the Social Democrats and the Centre. This
coalition has occasionally included the Left Alliance and more often the Swedish
People’s Party and other small parties. The conservative National Coalition was kept
out of the government by the President from 1966 to 1987 due to ‘general causes’—
read ‘Moscow’. While this was not publicly stated, it was understood by all relevant
players. Since mid-1980s all parties have been regierungsfähig, and governments have
stayed in office for the whole 4-year legislative term.
D E L E GAT I O N F RO M VOT E R S TO M P S
Except for some adjustments carried out in 1952, 1955, and 1969, the electoral system
used in parliamentary elections has remained stable throughout the 80 years that have
lapsed since Finland became independent.3 Electoral turnout, however, has now fallen
to below the West European average.
In 1906 the minimum voting age was 24. It was lowered to 21 in 1944, 20 in 1969, and
finally to 18 in 1972. Section 14 of the Constitution states that ‘Every Finnish citizen
who has reached eighteen years of age has the right to vote in national elections and
referendums.’ With the exception of soldiers, ‘everyone with the right to vote and who
is not under guardianship can be a candidate in parliamentary elections’ (Section 27).
The 200 members of the Eduskunta are elected for a 4-year term. The country is
divided into one single-member and fourteen multimember electoral districts. Åland is
entitled to one seat regardless of its population. The remaining 199 seats are allocated
to the fourteen multimember districts according to their populations. In the 1995 elec-
tions district magnitude ranged from 7 to 31 seats per constituency. Finland’s traditional
provinces form the electoral districts. Each district is a separate subunit and there is no
national pool of supplementary seats. The formula used for allocating seats to districts
is the method of largest remainders, and the procedure to allocate seats to parties is the
d’Hondt method (Kuusela 1995).
Parties, electoral alliances, and other groups present a list of candidates in the
constituencies. The voter casts one vote for one candidate. This means that Finnish
voters are not restricted to choosing among political parties. They also have the right
to choose between candidates. The electorate chooses 200 MPs from about 2,000
candidates. Helander (1997: 75) describes the situation as follows:
‘In spite of the decentralised nature of the Finnish electoral system the recruitment process
results . . . in a relatively elitist parliament in terms of social background. . . . [S]trong contempt for
3
The electoral system used in both parliamentary and presidential elections is explained in detail in
Kuusela (1995) and Nurmi (1990), and Tarasti and Taponen (1996). Developments until the 1960s are covered
in Törnudd (1968). Helander (1997) provides a thorough analysis of the recruitment of candidates for par-
liamentary elections in Finland. For a bibliography of Finnish electoral studies, see Wiberg (1995).
306 Tapio Raunio and Matti Wiberg
parties and politicians has not been able to change the picture. The professional politicians with
skills and other kinds of resources constitute the pool from which the legislators are recruited.
The selection process in all its stages results in a more and more elitist outcome. The party
members at the first step in the recruitment process are conservative: . . . usually prefer to select
male, highly educated, urban persons . . .. The voters are, if possible, even more conservative . . . ’.
The parties have considerable power in selecting candidates, while voters have the
power to choose among these candidates. There is no such thing as a safe seat in the
Finnish Parliament (Ruostetsaari 2000).
Candidate selection within parties is based on district-level primaries, with the local
branches playing a dominant role.4 Registered political parties and voters’ associations
with at least 100 persons have the right to nominate candidates, but only political
parties can enter into electoral alliances, technical coalitions that present a list of
candidates containing no more names than will be chosen from the electoral district.
Alliance partners profit electorally from pooling their resources. Yet, although there are
some in every election, such alliances have not been particularly prevalent. Prior to the
1969 electoral law the party leadership was able to control the nomination process
in the districts. One consequence was that parties tended to nominate the same
candidates in several constituencies. (On Åland, political parties have no right to
nominate candidates.)
The 1969 law requires parties to use membership voting as a means of selecting
candidates. Voting must be undertaken in constituencies where the number of nominees
exceeds the official upper limit of candidates the party has the right to nominate. The
parties themselves decide how the primaries are to be conducted, but in practice their
regulations tend to closely resemble those found in the election law. If party statutes do
not include regulations for this purpose, the directives set out in the law must be
followed. The law states that a local branch or, alternatively, that a group of at least
fifteen members has the right to nominate candidates. A group of at least thirty mem-
bers from different branches has the same right. The district executive body of the
party has the right to replace a maximum of a quarter of the candidates put forward by
local branches (Sundberg 1995, 1997). Local primaries are rarely used in small parties,
whereas the large parties use them in most electoral districts (see Sundberg and Gylling
1992; Tarasti and Taponen 1996; Helander, Kuitunen, and Paltemaa 1997, for campaign
regulations). The minimum length of a campaign is 50 days according to the electoral
law. Parties receive no free broadcasting time for their campaigns, but are free to buy
time commercially. The publicly owned channels organize the main political debates at
no monetary cost to the parties.
During the campaign, the national-level party organization and leadership primarily
act as a background resource, providing the local branches with necessary propaganda
and giving the party a public face. The actual work of collecting funds and spreading
the message is the responsibility of candidates’ support groups. While parties must
4
The role of parties in candidate selection is analysed in Sundberg (1995). The study by Helander et al.
(1997) includes a thorough investigation of candidate selection in the 1995 parliamentary elections.
Finland 307
report to the Ministry of Justice on their use of state subsidies (Wiberg 1991a), there is
no legal limit on campaign spending. Though this is very difficult to estimate, it
appears that total campaign spending increased more than fourfold during the 1980s
and 1990s.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
The Constitution Act of 1919 was virtually silent on the issue of government formation.5
The government was required to enjoy the confidence of the Eduskunta and the
President was ‘to appoint citizens of Finland known for their honesty and ability to
serve as members of the Council of State’. The current wording of Section 60 is:
The Government consists of the Prime Minister and the necessary number of Ministers. The
Ministers shall be Finnish citizens known to be honest and competent. The Ministers are responsible
before the Parliament for their actions in office. Every Minister participating in the consideration of
a matter in a Government meeting is responsible for any decision made, unless he or she has
expressed an objection that has been entered in the minutes.
The selection of ministers is wholly controlled by party leaders. The parliamentary
party group has no formal responsibilities in ministerial selection, and has historically
played only a minor role. Government formation decisions have been transferred from
the party executive to a larger body of representatives from both the party executive
and the parliamentary party group, but it still takes place within the party machinery
(see Nousiainen 1996: 113).
Before the new Constitution entered into force in March 2000, the President had a
central role in government formation. After the outgoing cabinet had submitted its
resignation, the President invited the Speaker of the Parliament and the representatives
of the parliamentary parties—starting with the biggest group and ending with the
smallest—to bilateral discussions. During this round of consultations, the party groups
had an opportunity to express their views on the future government, and it was the
duty of the President to assess which was the most suitable partisan solution.
Taking into account these consultations, the President then appointed either a prime
ministerial candidate or a third person whose task it is to continue negotiations—
primarily with those parties that she or he would like to see included in the cabinet. These
negotiations had three aims: to decide (1) which parties would form the government,
(2) the government programme, and (3) the allocation of portfolios. In practice the parties
first agreed on the distribution of portfolios, and then the parties’ parliamentary groups
decided which persons would fill them. However, it was not uncommon for the President
to influence the selection of individual ministers. The programme was written, or at least
formally accepted, only after these negotiations were completed.
After these three issues had been settled, the President—who had constantly been
updated on the progress of the negotiations—appointed the incoming cabinet in the
5
The history, legal norms, and political dynamics of government formation are discussed by Jansson
(1993) and Nousiainen (1988, 1992b, 1996, 1997). The study by Tiihonen (1990) provides an extensive analysis
of the development of governmental power in Finland.
308 Tapio Raunio and Matti Wiberg
last full plenary meeting of the resigning cabinet. The President’s influence should not
be underestimated. In 1987, the Centre Party and the National Coalition tried to form
a coalition government, but President Mauno Koivisto (1982–94) overruled this
bourgeois option, indicating that a coalition between the National Coalition and the Social
Democrats was preferable. He got his way due to his formal Constitutional powers.
If government formation negotiations failed and building a partisan cabinet was thus
impossible, the President had the right to appoint a caretaker cabinet consisting of civil
servants. Since 1945 Finland has had six caretaker cabinets, most recently the Liinamaa
cabinet in 1975. Moreover, about two-thirds of the remaining postwar cabinets have
included professional (i.e. non-partisan) ministers (Tiihonen 1991).
The new Constitution parliamentarized government formation, reducing the role of
the President to that of giving the formal seal of approval to the result of the
Eduskunta vote. According to Section 61 of the Constitution:
The Parliament elects the Prime Minister, who is thereafter appointed to the office by the President
of the Republic. The President appoints the other Ministers in accordance with a proposal made by
the Prime Minister. Before the Prime Minister is elected, the groups represented in the
Parliament negotiate on the political programme and composition of the Government. On the
basis of the outcome of these negotiations, and after having heard the Speaker of the Parliament
and the parliamentary groups, the President informs the Parliament of the nominee for Prime
Minister. The nominee is elected Prime Minister if his or her election has been supported by
more than half of the votes cast in an open vote in the Parliament. If the nominee does not
receive the necessary majority, another nominee shall be put forward in accordance with the
same procedure. If the second nominee fails to receive the support of more than half of the votes
cast, the election of the Prime Minister shall be held in the Parliament by open vote. In this event,
the person receiving the most votes is elected.
ordered the cabinet to step down and new elections to be held. Since 1983, all govern-
ments have survived the whole electoral period of 4 years.
At times, cabinet resignation has been caused by the action of the Eduskunta—either
a vote of no confidence following the tabling of an interpellation, or a parliamentary
decision against the wishes of the cabinet. Other factors that have led to cabinet resigna-
tion include dissolution by the President, the holding of new parliamentary or presi-
dential elections, internal disputes within the cabinet, and foreign policy conditions
(Nousiainen 2000).
Until 1991, the President alone had the right—without even consulting the cabinet
or the Parliament—to dissolve the Eduskunta and order new elections. During the
postwar era the President has exercised this right four times (1953, 1962, 1971, 1975).
A 1991 amendment to the Constitution strengthened parliamentarism by requiring
explicit prime ministerial consent for dissolving the Eduskunta: ‘On the basis of a rea-
soned initiative by the Prime Minister, the President may, after consulting the Speaker
of Parliament and the various parliamentary factions, and at a time when Parliament is
in session, dissolve Parliament by ordering that new elections be held.’ (Section 27, 22
July 1991/1074). Section 26 of the new constitution consolidated this practice:
The President of the Republic, in response to a reasoned proposal by the Prime Minister, and
after having heard the parliamentary groups, and while the Parliament is in session, may order
that extraordinary parliamentary elections shall be held. Thereafter, the Parliament shall decide
the time when it concludes its work before the elections.
The most common cause of cabinet termination has been the holding of new
parliamentary elections. Prior to the Second World War it was not customary for the
cabinet to resign following an election. Since then, however, cabinets in office—regardless
of the election result—have always submitted their resignation. Similarly, caretaker
cabinets have resigned as soon as new elections have been held. Reflecting the strong
role of the President, the cabinets also resigned following each presidential election,
even if the President was re-elected. During the past two decades, Presidents have
allowed the existing cabinets to continue in office after the presidential elections.
Internal disputes have occasionally led to cabinet termination. As a result of such
conflict, the cabinet has either resigned or the party composition has changed because
a party has left the governing coalition. The latest example of the latter occurred in
1994 when the Christian Union left the Centre-led bourgeois coalition. Foreign policy
imperatives refer to situations in which developments abroad lead to cabinet resignation.
The only postwar example is from 1958, when pressure from the Kremlin forced the
Fagerholm cabinet to step down (Nousiainen 1998: 252).
Table 8.1 shows the composition and duration of Finnish cabinets since 1945. While
the average lifespan of Finnish governments is quite short, cabinet duration has
increased considerably since the early 1980s. Since then all governments have survived
the whole 4 year electoral term. Governments have as a rule been formed by two of the
three main parties: SDP, National Coalition, and the Centre. The cabinets have almost
without exception been coalitions, and recently have brought together parties from
Table 8.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Finland 1945–2000a
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the cabinet Who dissolved
formation votes (iv) votes (cv) end with an early Parliament in
n.a. election? this case?b
No. of Cabinet Cabinet No. of Cabinet
ncv removed by resigned to cv removed
ncv pre-empt ncv under by failed
spec. cv
inst.
mech.
n.a.
6
There is relatively little empirical research on control instruments. Wiberg (1994a) contains a thorough
examination of the rules and use of parliamentary questions, while the interpellation mechanism has been
researched by Helander and Isaksson (1994).
Finland 313
D E L E GAT I O N W I T H I N T H E C A B I N E T
The lack of effective parliamentary opposition to recent governments has increased
individual ministers’ freedom of action. Delegation of authority to individual ministers
has flourished, while their accountability has not been seriously examined on a regular
basis.
Finland 315
The PM is the head and chairman of the cabinet. Another minister is deputy PM,
and takes her or his place when the PM for some reason is prevented from attending to
his duties as the chairman. If the deputy is also unavailable, then the most senior min-
ister acts as chairman. The Prime Minister takes on the duties of the President when
the latter is prevented from doing so. The constitution stipulates that the dissolution of
the Eduskunta is now dependent on a ‘reasoned proposal’ by the PM. If the PM resigns,
the whole cabinet is dissolved.
In cabinet meetings five members constitute a quorum. The Chancellor of Justice
also attends. Decisions are taken by simple majority rule, with each minister having
one vote. When the vote is tied, the PM decides the outcome. Since Finnish postwar
governments have tended to be broad coalitions, the PM needs to have good bargaining
skills because decisions are often based on deals between the coalition parties. Besides
these formal weekly plenary meetings, the cabinet also convenes regularly on an
informal basis. Such so-called ‘evening school’ (iltakoulu) meetings date from 1937–9
when PM Cajander sought a forum for more relaxed negotiations, with the goal of
making it easier to resolve potential problems.
Finland currently has thirteen ministries, each headed by a minister. The PM cannot
appoint or dismiss individual ministers without the consent of the government parties.
The cabinet used to include at most seventeen persons, but the cabinets appointed since
1995 have included eighteen ministers. Ministers with a portfolio but without a ministry
of their ‘own’ are considered equal to their departmental colleagues. There are no junior
ministers in Finland. The first secretary of state (poliittinen valtiosihteeri) was introduced
in the Prime Minister’s office in 1995. He is not directly responsible to the Parliament.
Since 1970, all ministers have had their own special political advisers (erityisavustaja),
who are to be distinguished from the technical secretaries in the ministries.
Individual ministers have strong influence over their fields of competence. While
‘matters of wide importance or matters that are significant for reasons of principle, as
well as matters whose significance so warrants’ are decided in ordinary sessions of the
cabinet, the ministers in practice have wide executive powers (Constitution, Article 67).
The autonomy of ministers has increased since the 1970s. Moreover, ministers may
delegate executive authority to their civil servants. This delegation of authority from
the PM and the cabinet as a collegiate body to the individual ministers and their civil
servants reflects the increasing workload of the government and the growing
complexity of Finnish society.
While individual ministers have gained important powers inside the cabinet, they
also bear a collective responsibility for cabinet actions. The Constitution makes this
clear: ‘The Ministers are responsible before the Parliament for their actions in office.
Every Minister participating in the consideration of a matter in a Government meeting
is responsible for any decision made, unless he or she has expressed an objection that
has been entered in the minutes.’ (Section 60).
Besides plenary meetings, the work of the cabinet is coordinated through four
statutory ministerial committees (ministerivaliokunta). These do not have formal
competence to make decisions on the behalf of the whole cabinet. Their role is to
speed up cabinet decision-making by preparing initiatives and putting together political
316 Tapio Raunio and Matti Wiberg
package deals, which the cabinet then, in practice, simply rubber-stamps. The titles and
composition of these committees are:
1. Cabinet Foreign and Security Policy Committee (ulko- ja turvallisuuspoliittinen
valiokunta): PM, Foreign Minister, and six other ministers.
2. Cabinet Finance Committee (raha-asiainvaliokunta): PM, Finance Minister, and
five other ministers.
3. Cabinet Economic Policy Committee (talouspoliittinen ministerivaliokunta): PM and
a maximum of eight other ministers.
4. Cabinet EU Committee (EU-ministerivaliokunta): PM, Minister of Foreign Affairs,
and seven other ministers.
The EU-committee is a newcomer, established following Finland’s accession to the EU
in 1995. All committees are chaired by the PM. The composition of the committees
may change depending on their agendas. The cabinet may also decide to set up addi-
tional ad hoc committees for the preparation of specific matters.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
Due to a general rise in media coverage and the gradual decline of the domestic role of
the President, the cabinet and especially the PM have become the focus of public atten-
tion and scrutiny. In practice, however, individual ministers have strong authority over
their respective fields of competence. The minister controls directly agencies in his/her
jurisdiction.7
By and large the administration is loyal to the government of the day. Civil servants
are career bureaucrats. The administration’s legalistic tradition produces an ideology of
neutrality and impartiality. Measured in terms of the number of people on the payroll,
the state is the biggest single employer in Finland. Public administration is divided into
three main levels: national, provincial, and municipal.
The national level administration consists of ministries and other central state
agencies. ‘While holding the office of a Minister, a member of the Government shall
not hold any other public office or undertake any other task which may obstruct the
performance of his or her ministerial duties or compromise the credibility of his or her
actions as a Minister.’ (The Constitution of Finland, Section 63.)
Ministers are aided by a permanent secretary (kansliapäällikkö), whose task is to oversee
the functioning of the ministry.
Ministries are divided into various sections and offices.8 Since the preparation of
issues and actual decision-making is delegated downwards within ministries, the heads
of various sections and offices are influential players. Civil servants can be held
accountable for actions they undertake in their official duties:
A civil servant is responsible for the lawfulness of his or her official actions. He or she is also
responsible for a decision made by an official multi-member body that he or she has supported
7
For detailed analyses of cabinet ministers and parliamentary government in Finland, see Nousiainen
(1994, 1997).
8
For information on ministers and ministries, see Nousiainen (1992a).
Finland 317
as one of its members. A rapporteur shall be responsible for a decision made upon his or her
presentation, unless he or she has filed an objection to the decision. (The Constitution of Finland,
Section 118)
As a result of the authority that has been delegated to civil servants, control over civil
service appointments, especially top-ranking positions, has become a sensitive matter.
Excluding those positions filled by the President, appointments are made by the cabinet,
individual ministries, and central state agencies.
The President traditionally enjoyed strong appointment powers until the new
Constitution entered into force. Following a proposal from the relevant minister, she or
he appointed the top civil servants, including the Chancellor and the Assistant
Chancellor of Justice; the Archbishop and the Bishops; and the Presidents and Justices
of the Supreme Court and the Supreme Administrative Court; army and navy
officials; the directors-general of the central state agencies; and—upon nomination by
the Council of State—the provincial governors. Based on the Council of State’s
recommendation, the President appointed other senior civil servants of the central
state agencies; the head and the presenting officials in the President’s office; and—upon
nomination by the competent authority—the highest civil servant in the Council of
State, the Supreme Court, and the Supreme Administrative Court. Up to August 1998
the President also appointed university presidents and professors. With regard to his
appointment authority, the President is not formally bound by the recommendations
of ministers. The new Constitution reduced presidential powers by narrowing the list
of persons the president appoints.
Party politics penetrates most levels of administration. Membership in a particular
party can facilitate access to influential, well-paid positions. This trend is particularly
visible in top jobs within state-owned companies, central state agencies, and ministries,
but also exists at the regional and municipal levels. Patronage is not unknown, although
it is clearly not a core element of the political system or a major factor in the relationships
between the cabinet and supporting parties. Rather, the political culture is primarily
legalistic. The spoils system has never been characteristic of the Finnish administration,
and administrative formalism sets limits to partisan advantage. Despite this, the allocation
of jobs on partisan grounds, both at the national and provincial levels, is a visible form
of patronage (Nousiainen 1996: 123).
Many party-based, ad hoc committees—whose overall composition reflected the
proportional strength of the traditional political parties—have been replaced by
non-partisan policy advisers (selvitysmies) appointed by the ministers. These advisors
are responsible for formulating policy advice and preparing reforms. This had led to
a decline in the connection between partisan politics and policy formulations, while at
the same time strengthening the technocratic administrative apparatus, which is less
directly accountable to political parties and the MPs.
E X T E R NA L C O N S T R A I N T S
Due to geopolitical realities, Finland has always been subject to external constraints.
Although the impact of the Soviet Union often has been overestimated in the Western
318 Tapio Raunio and Matti Wiberg
literature, it is true that Finns found it wise to anticipate the reactions of the country’s
eastern neighbour during the Cold War.
Domestically, organized interests, such as labour, industry, and farmers, have
traditionally not played the same significant role as in Finland’s Scandinavian neighbours.
Nonetheless, economic and social policy matters are largely handled in a corporatist
fashion. The cabinet is the central clearing house for public policies. As Nousiainen
(1996: 126) puts it, the cabinet is the only body with enough resources and information
to keep the corporatist conglomerate together. It is true, however, that big corporations,
both private and public, sometimes control the government rather than the other way
around. This occurred particularly in the 1970s and 1980s when Finland was strongly
corporatist (Paloheimo 1990).
Among other actors, the Bank of Finland has had a strong and rather independent
position, which, of course, has changed with the progressive implementation of
Economic and Monetary Union (EMU). While Finland does not have a constitutional
court, there are other legal checks on politicians and bureaucrats. The duties of the
Parliamentary Ombudsman include controlling the activities of the courts of law and
other authorities. It is also the Ombudsman’s responsibility to ensure that civil servants
and other persons employed by public corporations, as well as other persons responsible
for public affairs, uphold the law and fulfil their responsibilities. The Chancellor of
Justice is responsible for supervising the upholding of the Constitution and other legi-
slation. He/she supervises public authorities, civil servants, and other persons
employed in public corporations and elsewhere who carry out public functions. The
purpose of this supervision is to ensure that all these persons uphold the law and carry
out their responsibilities without infringing on anyone’s legal rights.
National referendums, which are only consultative, have only been used twice: in
1931 on the prohibition of alcohol, and in 1994 on EU membership. There are two
other features of political life that have a greater impact on Finnish parliamentary
democracy: the President and membership in the EU.
The President is elected for no more than two consecutive 6-year terms. Until 1982,
the President was elected by an electoral college of 300 members (301 in 1982), elected
by the same proportional system as MPs. A one-time experiment was conducted in the
1988 election involving a mixed two-ticket system of direct and indirect voting. To be
elected by a direct vote, a candidate needed to receive 50 per cent of the votes cast.
Since no candidate reached this share, the election was passed on to a simultaneously
elected electoral college (Nurmi 1990; Kuusela 1995).
A new electoral system for choosing a President was first used in 1994. If a candidate
receives more than half of the votes cast, she or he is elected President. If no one
receives the majority of the valid votes, a new election is held on the third Sunday after
the first election. In the second round, the two persons who received the most votes in
the first election run against each other. The candidate who receives the majority of
votes is elected President. In the unlikely event of a tie, the election is decided by the
drawing of lots.
The President is closely involved in law-making. First, it is up to the President to
decide whether a government initiative will be passed on to the Eduskunta for further
Finland 319
action. Second, once the Parliament has adopted a proposal, the law must be signed and
ratified by the President. The President can refuse to ratify the proposal, or may choose
not to sign it within 3 months.
An Act adopted by the Parliament shall be submitted to the President of the Republic for
confirmation. The President shall decide on the confirmation within three months of the
submission of the Act. The President may obtain a statement on the Act from the Supreme
Court or the Supreme Administrative Court. If the President does not confirm the Act, it is
returned for the consideration of the Parliament. If the Parliament readopts the Act without
material alterations, it enters into force without confirmation. If the Parliament does not readopt
the Act, it shall be deemed to have lapsed. (The Constitution of Finland, Section 77)
Since the proposal can become a law without the approval of the President, she or he
has only a suspensive veto (Wiberg 1994b). While the President’s role is formally strong,
in practice the cabinet proposals and parliamentary decisions are honoured by the
President. Presidents have supported cabinets that they have appointed. It is very rare
that conflicts between these two state organs have become public. The President and
the cabinet are engaged in a continuous web of negotiations: they both anticipate each
other’s moves.
The President makes her or his decisions in meetings of the Council of State. These
weekly meetings are normally held on Fridays and chaired by the President. Ministers
introduce their proposals, and their ministerial colleagues have the right to participate
in debate about the initiatives. However, the final decision always rests with the
President, who is not legally bound by the cabinet position, even when the cabinet is
unanimous.
The President also has independent executive powers that fall outside the remit of
the Parliament. These are regulated by Section 58 of the Constitution. The President is
the Commander-in-Chief of the defence forces, though in time of war she or he may
turn the command over to another person (Section 128). Foreign policy used to be the
exclusive domain of the President. Section 33 of the Constitution Act stated:
The relations of Finland with foreign powers shall be determined by the President. Treaties
concluded with foreign states shall, however, be approved by Parliament if they contain
provisions within the legislative sphere or if the consent of Parliament is otherwise required by
the constitution. Decisions concerning war and peace shall be made by the President with the
consent of Parliament.
Membership in the EU has, however, significantly changed the situation.9 According to
Section 93 of the Constitution:
The foreign policy of Finland is directed by the President of the Republic in co-operation with
the Government. However, the Parliament accepts Finland’s international obligations and their
denouncement and decides on the bringing into force of Finland’s international obligations in so
9
For analyses of national policy formulation on EU issues, see Murto et al. (1996), and previous work by
the authors which focuses on the Parliament’s involvement in EU matters (Wiberg and Raunio 1996; Raunio
and Wiberg 1997, 2000).
320 Tapio Raunio and Matti Wiberg
far as provided in this Constitution. The President decides on matters of war and peace, with the
consent of the Parliament. The Government is responsible for the national preparation of the
decisions to be made in the European Union, and decides on the concomitant Finnish measures,
unless the decision requires the approval of the Parliament. The Parliament participates in the
national preparation of decisions to be made in the European Union, as provided in this
Constitution.
The cabinet is the main actor with regard to EU policy formulation and coordination.
According to the Consitution, the Parliament and its committees are formally involved
in this process, but the cabinet now acts in a European arena that involves many policy
areas and this has made effective scrutiny by the Parliament more difficult.
A related issue is that, with regard to EU matters, the respective power of the cabinet,
the Eduskunta, and the President is by no means a settled issue. Particularly controver-
sial is the question of representation in EU Summits. President Martti Ahtisaari
(1994–2000) indicated that he did not want the President to be sidelined in the making
of national EU policy. Under the compromise reached between the government and
the President in 1995, the PM always attends EU Summits and the President attends
them whenever he or she so wishes. This practice will continue during the presidency
of Tarja Halonen, elected in 2000.
The Eduskunta has been an active and influential actor in the formulation of national
EU policy. Instead of setting up a special European affairs committee, the Parliament
decided to turn a pre-existing committee, the Grand Committee (Suuri valiokunta), into
a body responsible for coordinating the Eduskunta’s positions on EU issues. The Grand
Committee handles first and third pillar issues, while the Foreign Affairs Committee
deals with second pillar matters. The composition of the Grand Committee (25 MPs
and 13 alternates) reflects the parties’ strength in the chamber. The minister appears in
the Committee in person before and, when required, after the Council meeting. While
the Committee does not give legally binding voting instructions to the ministers, it is
extremely rare for a minister to act against its wishes. The Finnish system benefits from
the delegation of authority to specialized committees. The standing committees are
closely involved in the formulation and scrutiny of national policies from an early stage
in the policy process, and the final position of the Grand Committee is based on guide-
lines from the standing committees (Wiberg and Raunio 1996; Raunio and Wiberg
1997, 2000).
C O N C LU S I O N : F O R M A L RU L E S A N D I N F O R M A L P R AC T I C E S
It would be wrong to claim that Finnish constitutional regulations are intolerably far
away from reality. Finland is, in fact, a rather well-functioning liberal democracy. While
the Finnish Constitution does not even recognize political parties, in reality the Finnish
political system is heavily party-centred and government-dominated. The chain of del-
egation and accountability in Finnish democracy is influenced by a number of con-
straints relating to political parties, the party system, and the power of various kinds of
external agents—for example interest organizations and international actors like the
Finland 321
EU. The main problems for delegation and accountability can be summarized as
follows:
1. The direct link from the voters to the legislators is affected by the electoral system.
The parties play a major role in selecting the candidates and organizing the elections.
The electoral system favours large parties, and the entry of new parties has been made
less than easy by the introduction of formal criteria regulating what kind of organization
can be registered as a party. Public funding of parties only extends to those parties
represented in the Parliament, which similarly supports the cartel of the established
parties and narrows the voters’ freedom of choice. Declining turnout can also be seen
as a manifestation of voters’ lack of confidence in the link between the electorate and
the representatives.
2. Compared to an ideal type of parliamentary democracy, the link between the
legislature and the cabinet was until the enactment of the new Constitution weakened
by the fact that Parliament did not have full control over the formation of cabinet, but
rather competed with the President for this control. Although cabinets have enjoyed
the support of the parliamentary majority during the last decades, Finland has also had
a number of non-partisan cabinets who do not run for re-election and are therefore less
concerned with ex post accountability.
3. With strong majority cabinets, the executive controls the Parliament rather than
the other way around. Majority cabinets can have their way since party cohesion and
group discipline in the Parliament are strong. This is sometimes also the case with
bureaucrats and ministers. The legalistic and somewhat hierarchical Finnish society
gives top level civil servants a strong position that can be used to exercise discretion in
implementation, as well as to initiate and propose much of the cabinet’s—and thereby
the parliament’s—agenda. Finland has a particularly legalistic political culture which
highlights the role of civil servants, who have to some extent run the entire country as
an ideologically homogenous and rather closed circle of pragmatic technocrats. While
the country has numerous associations, only trade unions have been able to recruit
large numbers of rank and file members.
In sum, while it is fair to call the Finnish political system a well functioning
representative democracy, parliamentary democracy in Finland is not without its short-
comings. It has existed somewhat in the shadow of a strong presidency and important
neighbours. Throughout the postwar period, agency problems seem to have been
rather significant in the sense that all actors in the delegation game have had at their
disposal many resources to aid them in strategic information concealment. More
recently, constitutional amendments have strengthened parliamentarism and reduced
the powers of the President. Were it not for EU membership, this process would have
increased Parliament’s ability to ensure cabinet accountability. However, the development
is actually, at least in part, itself a consequence of Finland’s membership in the EU, which
has begun to reshape Finnish democracy.
322 Tapio Raunio and Matti Wiberg
REFERENCES
Arter, David (1987). Politics and Policy-Making in Finland. Brighton: Wheatsheaf.
Helander, Voitto (1997). ‘Finland’, in Pippa Norris (ed.), Passages to Power. Cambridge: Cambridge
University Press.
——and Isaksson, Guy-Erik (1994). ‘Interpellations in Finland’, in Matti Wiberg (ed.), Parliamentary
Control in the Nordic Countries: Forms of Questioning and Behavioural Trends. Jyväskylä: The Finnish
Political Science Association.
——, Kuitunen, Soile, and Paltemaa, Lauri (1997). ‘Kansalaisesta aktiiviksi, aktiivista edustajaksi—
Tutkimus vuoden 1995 eduskuntavaaleista ja niiden ehdokasasetteluista.’ Valtio-opillisia
tutkimuksia no. 51. Turku: Turun yliopisto.
Hidén, Mikael and Saraviita, Ilkka (1994). Valtiosääntöoikeuden pääpiirteet. Helsinki: Lakimiesliiton
kustannus.
Isaksson, Guy-Erik (1997). ‘Interpellationsmönstret i Finlands riksdag’, in Voitto Helander and
Siv Sandberg (eds.), Festskrift till Krister Ståhlberg. Åbo: Åbo Akademis förlag.
Jansson, Jan-Magnus (1993). Hajaannuksesta yhteistoimintaan—Suomalaisen parlamentarismin vaiheita.
Helsinki: Gaudeamus.
Kuusela, Kimmo (1995). ‘The Finnish Electoral System: Basic Features and Developmental
Tendencies’, in Sami Borg and Risto Sänkiaho (eds.), The Finnish Voter. Tampere: The Finnish
Political Science Association.
Lane, Jan-Erik, McKay, David, and Newton, Kenneth (1997). Political Data Handbook. 2nd edn.
Oxford: Oxford University Press.
Ministerial Responsibility Act (25 November 1922/274).
Murto, Eero, Väänänen, Pekka, and Ikonen, Raimo (1996). Sisäpiirit EU-Suomessa: unioni ja uudet
eliitit. Helsinki: Edita.
Nousiainen, Jaakko (1988). ‘Finland’, in Jean Blondel and Ferdinand Müller-Rommel (eds.),
Cabinets in Western Europe. Macmillan: Houndmills.
—— (1992a). Politiikan huipulla. Ministerit ja ministeristöt Suomen parlamentaarisessa järjestelmässä.
Juva: WSOY.
—— (1992b). Suomen poliittinen järjestelmä. Porvoo: WSOY.
—— (1994). ‘Finland: Ministerial Autonomy, Constitutional Collectivism, and Party Oligarchy’,
in Michael Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government.
Cambridge: Cambridge University Press.
—— (1996). ‘Finland: Operational Cabinet Autonomy in a Party-Centered System’, in Jean
Blondel and Mauricio Cotta (eds.), Party and Government. Macmillan: Houndmills.
—— (1998). Suomen poliittinen järjestelmä. Porvoo: WSOY.
——(2000). ‘Finland: The Consolidation of Parliamentary Governance’, in Wolfgang C. Müller and
Kaare Strøm (eds.), Coalition Governments in Western Europe. Oxford: Oxford University Press.
Nurmi, Hannu (1990). ‘A Theoretical Review of the Finnish Parliamentary and Presidential
System’, in Jan Sundberg and Sten Berglund (eds.), Finnish Democracy. Jyväskylä: The Finnish
Political Science Association.
Paloheimo, Heikki (1990). ‘Between Liberalism and Corporatism: The Effect of Trade Unions and
Governments on Economic Performance in Eighteen OECD Countries’, in Renato Brunetta
and Carlo Dell’ Aringa (eds.), Labour Relations and Economic Performace. Houndmills: Macmillan.
Pesonen, Pertti and Riihinen, Olavi (2002). Dynamic Finland: The Political System and the Welfare
State. Helsinki: Finnish Literature Society.
Rasch, Bjørn Erik (1995). ‘Parliamentary Voting Procedures’, in Herbert Döring (ed.), Parliaments
and Majority Rule in Western Europe. New York: St. Martin’s Press.
Finland 323
Raunio, Tapio (1999). ‘Facing the European Challenge: Finnish Parties Adjust to the Integration
Process’. West European Politics, 22(1): 138–59.
—— and Wiberg, Matti (1997). ‘Efficiency Through Decentralisation: The Finnish Eduskunta
and the European Union’, in Matti Wiberg (ed.), Trying to Make Democracy Work. Stockholm:
Gidlunds.
—— (2000). ‘Building Elite Consensus: Parliamentary Accountability in Finland’. Journal of
Legislative Studies, 6(1): 59–80.
Ruostetsaari, Ilkka (2000). ‘From Political Amateur to Professional Politician and Expert
Representative. Recruitment of the Parliamentary Elite in Finland 1863–1995’, in Maurizio
Cotta and Heinrich Best (eds.), The European Representative. 150 Years of Parliamentary Recruitment
in Comparative Perspective 1848–1998. Oxford: Oxford University Press.
Sundberg, Jan (1994). ‘Finland: Nationalized Parties, Professionalized Organizations’, in Richard
S. Katz and Peter Mair (eds.), How Parties Organize. London: Sage.
—— (1995). ‘Organizational Structure of Parties, Candidate Selection and Campaigning’, in Sami
Borg and Risto Sänkiaho (eds.), The Finnish Voter. Tampere: The Finnish Political Science
Association.
—— (1996). Partier och partisystem i Finland. Esbo: Schildts.
—— (1997). ‘Compulsory Party Democracy: Finland as a Deviant Case in Scandinavia’. Party
Politics, 3: 97–117.
—— (1999). ‘The Enduring Scandinavian Party System’. Scandinavian Political Studies, 22: 221–41.
—— and Gylling, Christel (1992). ‘Finland’, in Richard S. Katz and Peter Mair (eds.), Party
Organizations. London: Sage.
Tarasti, Lauri and Taponen, Heimo (1996). Suomen vaalilainsäädäntö. Helsinki: Edita.
Tiihonen, Seppo (1990). Hallitusvalta—Valtioneuvosto itsenäisen Suomen toimeenpanovallan käyt-
täjänä. Helsinki: VAPK-Kustannus & Hallintohistoriakomitea.
—— (1991). ‘The Professionals of Government: Caretaker Cabinets, Professional Ministers and
Full-Time Party Politicians’, in Matti Wiberg (ed.), The Political Life of Institutions. Jyväskylä:
The Finnish Political Science Association.
Törnudd, Klaus (1968). The Electoral System of Finland. London: Hugh Evelyn.
Wiberg, Matti (1991a). ‘Public Financing of Political Parties as Arcana Imperrii in Finland’, in
Matti Wiberg (ed.), The Public Purse and Political Parties: Public Financing of Political Parties in
Nordic Countries. Jyväskylä: The Finnish Political Science Association.
—— (1991b). ‘Analysis and Politics’. Studies on Political Science no. 11. Turku: University of Turku.
—— (1994a). ‘To Keep the Government on Its Toes: Behavioural Trends of Parliamentary
Questioning in Finland 1945–1990’, in Matti Wiberg (ed.), Parliamentary Control in the Nordic
Countries: Forms of Questioning and Behavioural Trends. Jyväskylä: The Finnish Political Science
Association.
—— (1994b). ‘Law Production in Finland: Strategic Considerations’, in Matti Wiberg (ed.),
Rationality in Institutions, Annales Universitatis Turkuensis B: 206. Turku: University of Turku.
—— (1995). ‘Bibliography of Finnish Electoral Research’, in Sami Borg and Risto Sänkiaho (eds.),
The Finnish Voter. Tampere: The Finnish Political Science Association.
—— (1997). ‘Suomen valtionapuvetoinen puoluetukijärjestelmä’, in Sami Borg. (ed.), Puolueet
1990-luvulla: Näkökulmia suomalaiseen puoluetoimintaan, Valtio-opillisia tutkimuksia no. 53.
Turku: Turun yliopisto.
—— (2000). ‘The Partyness of the Finnish Eduskunta’, in Knut Heidar and Ruud Koole (eds.),
Parliamentary Party Groups in European Democracies. London: Routledge.
Wiberg, Matti and Mattila, Mikko (1997). ‘Committee Careers in the Finnish Parliament,
1945–1994’, in Lawrence D. Longley and Attila Ágh (eds.), The Changing Roles of Parliamentary
324 Tapio Raunio and Matti Wiberg
Committees. Appleton: International Political Science Association, Research Committee of
Legislative Specialists.
—— and Raunio, Tapio (1996). ‘Strong Parliament of a Small EU Member State: The Finnish
Parliament’s Adaptation to the EU’. Journal of Legislative Studies, 2: 302–21.
—— —— (1997). ‘Where’s the Power? Controlling Voting Outcomes in the Nordic Parliaments
1945–95’, in Guy-Erik Isaksson (ed.), Inblickar i Nordisk parlamentarism, Meddelanden från
ekonomisk-statsvetenskapliga fakulteten vid Åbo Akademi A: 470. Åbo: Åbo Akademi.
9
I N T RO D U C T I O N
Delegation and accountability in France must be analysed in the specific context of
the ‘semi-presidential’ regime of the Fifth Republic. This regime grafts a popularly
elected President with extensive constitutional powers onto a more or less conventional
parliamentary system in which the government (the Prime Minister (PM) and the cabi-
net) are accountable to the popularly elected legislative chamber, the National
Assembly (Article 20 of the Constitution).
The uniqueness of the Fifth Republic rests on the status of the President. The
Constitution says that the cabinet is accountable to Parliament (Article 49) but that the
President is not (Article 68). However, the strength of the French presidency goes
beyond the letter of the Constitution and is only partially a result of the fact that the
President of the Republic is directly elected by the voters. Other contributing factors
are an institutionally weak Parliament, the majority system for the election of deputies,
and the considerable change in public debate brought about by the development of
television (Beaud and Blanquer 1999).
The French political system is a flexible one capable of producing a wide variety of
combinations of state and partisan power under the presidential authority. Its hybrid
rules can be interpreted to suit varying situations, a characteristic that has given the
institutions of the Fifth Republic great adaptability. If the President and National
Assembly share the same partisan loyalties, then the President has undisputed power.
In such situations, the President dominates the exercise of political power. The
predominance of the President is based on three elements: the institutional powers of
the President, the President’s electoral and popular legitimacy, and the President’s position
as the real leader of the parliamentary majority. The relationship between the President
and National Assembly is thus a hierarchical one. The parliamentary majority, the PM,
the cabinet, and the administration act under the leadership of the President.
This chain of delegation and accountability links voters to the President, with the
President controlling the PM, the cabinet, the administration, and the majority party.
De Gaulle asserted the constitutional doctrine supporting this chain of delegation and
accountability in a press conference on 14 January 1964, during which he stated:
The indivisible authority of the state is wholly confided to the President by the people, who
elected him, and [. . .] there is no other authority, be it ministerial, civil, military or judicial, which
326 Jean-Louis Thiébault
is not conferred or maintained by him. Finally, he adapts his own supreme sphere of action to
those to whom he delegates authority. Nevertheless, normally, it is essential to preserve the
distinction between the function and sphere of action of the head of state and that of the Prime
Minister (quoted from Hayward 1993a: 24).
Three types of agreement are required if the system is to work successfully. First, there
must be agreement between the President and the PM. Second, agreement between the
assembly majority and the government, based upon the principle of accountability to the
assembly, is necessary. The third type of agreement that is necessary is agreement
between the President and the parliamentary majority (Duverger 1985).
If the President and the majority in the National Assembly have different partisan
loyalties, then the PM is an important holder of executive power and the President’s
power is more limited than the authority granted him by the Constitution. This situation,
popularly known as cohabitation, occurred as a result of legislative elections in 1986,
1993, and 1997. In each case the President lost his supporting majority in Parliament.
The idealized chain of delegation and accountability in parliamentary democracy—
the one that links voters to Members of Parliament, Members of Parliament to cabinets
and the PM, the PM to cabinet members, and individual ministers to civil servants—has
been more closely observed in France during the three periods of cohabitation (1986–8,
1993–5, and 1997–2002) than when the President and the parliamentary majority were
of the same party. During periods of cohabitation the President and government proceed
from opposing political camps. The PM has undisputed control of the government. The
President retains an oversight function with regard to some particularly important matters
such as foreign and defence policy as well as some residual areas of responsibility, in
particular culture and constitutional affairs (Duverger 1985; Duhamel 1994).
The ‘residual area’ leaves the President with important constitutional powers. For
example, it includes the authority to initiate amendments to the Constitution by submit-
ting a proposal to the Parliament. After an amendment has been accepted in both cham-
bers of the National Assembly by two-thirds majority of the MPs present, he can submit
it to a joint session of Parliament at which it is finally accepted if it is supported by a three-
fifths majority of all voting MPs. A referendum is not needed for the amendment to
become law. In contrast to this, constitutional amendments of parliamentary origin must
be subject to a referendum (Article 89) (Duverger 1985; Mény 1991; Duhamel 1994).
The President also has the right to call a referendum—nominally at the ‘request’ of
the cabinet or Parliament—on non-constitutional changes in the organization of polit-
ical authority on changes in the economic and social policies and the public sector (revi-
sion of 1995), or for the ratification of treaties (Article 11). De Gaulle resuscitated the
referendum as a means of appealing to the French people over the heads of their par-
liamentary representatives. After his success with referendums on Algeria in 1961 and
1962, and the issue of the popular election of the President in 1962, de Gaulle opted for
another referendum in April 1969. In this referendum, which was about a variety of
issues including the restructuring of regional government and reforming the Senate,
de Gaulle lost. Only 47 per cent of those voting supported the proposed changes, while
53 per cent opposed. As he had promised in advance, de Gaulle interpreted the defeat
as a personal repudiation and resigned. His action serves as a good illustration of the
idea that the President should be directly accountable to the people.
France 327
P O L I T I C A L PA RT I E S
Article 4 of the Constitution of the Fifth Republic for the first time in French history
formally acknowledged the existence of political parties. However, the role of political
parties was to be confined to ‘contributing to the expression of suffrage’. De Gaulle was
determined to avoid the re-emergence of the ‘party regime’. He never openly admitted
that the operation of the French political system required a partisan parliamentary
majority supporting the President (Hayward 1993b). A trend towards party government
was associated with the emergence in 1962 of a stable majority coalition of parties in
the National Assembly. Moreover, party competition was increasingly dominated by
the presidency during the 1960s and 1970s. This was reflected in the evolution of the
system from a fragmented and undisciplined multiparty system with no stable coalition
towards a two-bloc, four-party structure with lasting coalitions (Machin 1993). In this
respect, the development of the Fifth Republic has not followed de Gaulle’s conception
of the presidency and the political system as being above parties (Reif 1987). Rather,
political parties structure the relationship between voters on the one hand and both the
President and the Parliament on the other hand.
The first step in this process was the emergence of the Gaullist party (earlier under
the RPR label, since 2002 under the UMP label) as a presidential party in Parliament
(Cole 1993). With its coalition allies, it held a stable majority in the National Assembly
between 1962 and 1981. It brought about an unprecedented consolidation of the tradi-
tionally fragmented Right. The lack of an alliance between Socialists (PS) and
Communists (PCF) meant that the Left suffered from disunity until 1972. In this
328 Jean-Louis Thiébault
situation, there was no plausible alternative to a Gaullist government. The presidential,
two-bloc system appeared to reach its peak of development in the 1978 parliamentary
elections, when the government’s Centre-Right forces and the Left opposition created
two similar-sized alliances. The Centre-Right alliance included the Gaullist RPR party and
the UDF. From 1978 to 1998 the UDF was a confederation of five centrist and moderately
conservative parties (Republican Party, Radicals, Christian Democrats, Social Democrats,
and Giscard’s club, Perspectives et Réalités). In 1998 the UDF split, now consisting of
Radicals, Christian Democrats, and Social Democrats. The new Liberal Democracy party
consists of the former Republican Party and the Giscard’s club, but does not include Mr
Giscard d’Estaing himself who stayed with the UDF. The parties of the Left were the
Socialist Party and the Communist Party (Bartolini 1984; Machin 1993).
Since 1978, the number and relative strength of the parties have changed. The
subsequent stage of the evolution, in 1981, saw the electoral defeat of the Right and
a shift in the party balance within the Left. François Mitterrand defeated Valery Giscard
d’Estaing in the 1981 presidential elections, and, in the early parliamentary elections
immediately called by Mitterrand after his victory, the Left gained a majority for the
first time in the history of the Fifth Republic. Further changes took place in the 1980s.
In opposition, the Centre-Right parties faced a new challenge on the Right from Le
Pen’s Front National (FN). On the Left, the alliance of the PS and PCF disintegrated
when the PCF left the governing coalition in July 1984. In 1986, the Centre-Right won
the parliamentary elections and cohabitation brought about a new stage of party devel-
opment. The re-election of Mitterrand in May 1988, which ended cohabitation, marked
the beginning of another phase. Early parliamentary elections held in June 1988 led to
no clear majority in the National Assembly and minority government of the Socialists
(Elgie and Maor 1992). A new cohabitation began after the victory of the Centre-Right
alliance in the 1993 parliamentary elections. In 1995 Chirac won the presidency, thus
ending the second period of cohabitation. However, a third cohabitation period began in
May 1997 after Chirac opted to dissolve the Parliament and hold early elections. The
‘plural left’ (Socialist, Communist, and Green) won the election, thus depriving Chirac
of his parliamentary majority. He regained a majority in the parliamentary elections of
2002 (after his re-election as President). The party system is now more fragmented,
and, since the 1990s, is characterized by a partisan presidency and a two-sphere party
system—one sphere of presidential contenders (PS, UDF, now UMP, and RPR), and a
non-presidential one composed of parties that do not expect to win the presidency in
the foreseeable future (Greens, PCF, FN).
Relations between the government and the parties in Parliament have undergone
substantial development since the beginning of the Fifth Republic. The periods of
cohabitation are an obvious example, but changes have occurred also during periods of
stable majority support for the President. There has been an increase in collaboration
between the two branches in that the government now consults regularly with the
party groups that form the parliamentary majority. Considerable political bargaining
between the governments and its parliamentary representatives precedes formal policy
announcements. The parliament is often encouraged to make the widest possible use
of its legislative functions and of its ‘right to control governmental policies’. The
relations between the deputies and the government are sometimes institutionalized in
France 329
D E L E GAT I O N F RO M VOT E R S TO M P S A N D T H E P R E S I D E N T
There is a dual chain of delegation and accountability from French voters to political
authorities. First, there is a link from the voters to MPs via the legislative elections.
In addition, voters delegate to the President of the Republic, who is subsequently
accountable to them, via the process of direct presidential election.
selection process shows the domination of the national party organization outside of
Parliament, which wields most of the decision-making power in this area. However, the
national organization must always take into account the strong local ties of candidates
as well as the demands of its parliamentary group. The latter wants to obtain an assurance
of renomination for all deputies facing re-election (Thiébault 1988a).
Traditionally, local connections have been less significant for the centralized and
well-organized Communist Party. In the past the nomination mechanism was subject
to the principle of democratic centralism and amounted to co-optation. The national
party organization thus controlled the selection process in its entirety (Thiébault
1988a). In recent years, however, the power of the national party organization has
declined, and it is now limited by the increased influence of the leaders of the local
party organizations.
Participation by party members in the recruitment process varies considerably
between the parties. On the Left, a large majority of members are involved, while in
the right-wing parties the percentage is much lower. In both cases, the numbers of
those who participate represent a negligible proportion of party voters. Although the
involvement of party members in the nomination of candidates for legislative elections
has increased in recent years, it remains weak. In the electoral arena, there is not much
careful ex ante screening of candidates by voters and parties. Under the two-ballot elect-
oral system, it is possible to consider the first round of voting as a ‘primary’ election
comparable with an American primary. The left-wing parties systematically use the first
round as a ‘primary’ in this fashion. The left-wing candidate with most votes becomes
the bloc’s sole candidate in the decisive second round. The parties of the
Centre-Right come to an agreement on a common candidate for the first round often
in about 80 or 90 per cent of the constituencies. Despite these variations with regard to
party practices, it is fair to say that because there is no real ‘primary’ election regulated
by the Constitution or other legislation, there is no general voter participation in the
selection process (Thiébault 1988a).
Between elections, however, parties or parliamentary party groups have power to
sanction representatives, with a rebuke or a blame, but also with the ultimate sanction
of expelling the MP from the party. MPs receive party directives coming from out-
side Parliament. While the formal institution of a whip is unknown, in most cases the
discipline of the party parliamentary groups is tight, on important as well as on
unimportant matters. Consequently, deputies have very rarely been expelled from their
party for flagrantly defying a party directive. The French Fifth Republic is therefore not
well equipped to deal with problems of ex post accountability of MPs vis-à-vis voters
and parties. It also has a weak capacity for ex post monitoring.
victory for the Left in 1997, however, he did not resign. Thus, the Fifth Republic
provides mechanisms other than presidential elections by which the President can be
held accountable. However, these mechanisms can only be invoked by the President
himself, and over the course of the Fifth Republic, the Presidents have been less and less
willing to tie their political fate to the use of these mechanisms.
D E L E GAT I O N F RO M T H E PA R L I A M E N T A N D
T H E P R E S I D E N T TO T H E C A B I N E T
The chain of delegation and accountability applicable to the French cabinet is dual in
nature—it originates from both the President and the Parliament.
1
Article 49-1 reads: ‘The Prime Minister, after deliberation in the council of ministers, may pledge the
responsibility of the government before the national assembly with regard to the program of the government
of if it be so decided with regard to a declaration of general policy.’
France 335
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence votes (cv)c Did the Who
formation votes (iv)b cabinet end dissolved
n.a. No. of Cabinet Cabinet No. of cv Cabinet with an Parliament in
ncv removed resigned to under spec. removed early this case?d
by ncv pre-empt ncv inst. mech. by failed election?
cv
1 Debré 59/01/08 2 No 4 No
2 Pompidou I 62/04/14 2 Yes 3 No Yes Head of State
3 Pompidou II 62/11/28 1 No 0
4 Pompidou III 66/01/08 1 No 3 No
5 Pompidou IV 67/04/06 3 No 0 Yes Head of State
6 Couve de Murville 68/07/10 0 0
7 Chaban-Delmas 69/06/20 1 No 0
8 Messmer I 72/07/05 1 No 0
9 Messmer II 73/04/02 2 No 0
10 Chirac I 74/05/27 2 No 0
11 Barre I 76/08/25 0 1 No
12 Barre II 78/04/03 0 1 No
13 Mauroy I 81/05/21 0 0 Yes Head of State
14 Mauroy II 81/06/22 5 No 6 No
15 Fabius 84/07/17 1 No 4 No
16 Chirac II 86/03/20 1 No 8 No
17 Rocard I 88/05/10 0 0 Yes Head of State
18 Rocard II 88/06/23 6 No 28 No
19 Mme. Cresson 91/05/16 2 No 8 No
20 Beregovoy 92/04/02 2 No 4 No
21 Balladur 93/03/29 0 1 No
22 Juppé 95/05/17 0 0 Yes Head of State
23 Jospin 97/06/02 — — — — —
a
n.d. ⫽ no data (missing data); n.a. or ‘—’ ⫽ not applicable in this case.
b
In France, there is no investiture vote for the cabinet. The PM, after deliberation in the Council of Ministers (cabinet), may (but is not obliged to) ask for a supporting
vote when he pledges the responsibility of the government or presents a declaration of general policy to the Parliament. However, these votes are not reported since the
table is designed only to record investiture votes.
c
The PM may, after deliberation in the Council of Ministers, pledge the government’s responsibility on the vote of a bill. The text shall be considered as adopted, unless
a motion of censure is voted (Article 49-3 in the Constitution).
d
The question ‘Who dissolved Parliament in this case?’ can be anwered with one of five options. This identifies the main constitutional actor that caused the early elec-
tion. We have not coded the formal signatory, but rather the person or body that made the real decision: (1) Head of State; (2) Prime Minister; (3) Cabinet;
(4) Parliamentary majority; (5) Automatic constitutional provision (e.g. if it is required when a new President takes office, or, for example, if it is required after the adoption
of a proposal for constitutional amendments).
338 Jean-Louis Thiébault
Another way of holding the government accountable by Parliament is described in
Article 49-3 of the Constitution. In this case, the government makes the adoption of
a particular bill an issue of confidence. The PM can pledge the accountability of the
government before the National Assembly in a vote on legislation for which there is
clearly no parliamentary majority. Under this procedure, if no censure motion is intro-
duced or if a censure motion fails to get an absolute majority of votes, then the bill is
said to be passed (Article 49-3 of the Constitution). This procedure thus favours the
government. It allows for the adoption of a bill for which there is no legislative support
whenever an absolute majority of representatives does not want to defeat the government.
The deputies know that a vote of censure is likely to lead to the dissolution of the
Assembly. This procedure is used only in specific circumstances, including when the
parliamentary majority is divided (the Barre government between 1978 and 1981),
when the parliamentary majority is only a relative one (the Rocard, Cresson, Bérégovoy
cabinets between 1988 and 1993), and when the government is under considerable time
pressure (the Chirac government between 1986 and 1988). Rocard used the procedure
on twenty-eight proposals in 36 months. Edith Cresson used it on eight proposals in
10 months, and Chirac employed it eight times in 24 months of cohabitation (Table 9.1).
Under the Fifth Republic, only opponents vote on confidence and no-confidence
motions (Article 49-2 and 49-3). The government does not seek explicit, positive mani-
festations of confidence from its parliamentary supporters, only implicit, negative acqui-
escence. This helps the government survive when the majority is divided (as with the
Barre governments) or when the government is a minority one (as with the Rocard,
Cresson, and Bérégovoy cabinets). This survival is ultimately made possible by a procedural
device that counts all non-hostile votes as tacit votes of support. However, it is also fair
to say that this draconian safeguard has been superfluous throughout most of the his-
tory of Fifth Republic because loyal majorities in the National Assembly have supported
stable governments (Huber 1996). For example, because of the loyalty of the left ‘plural
majority’, PM Jospin never used Article 49–3 during his five years in office 1997–2002.
The government can also be held accountable to the Parliament through the use of
particular instruments, including parliamentary questions and investigation committees.
One way in which Members of Parliament can extract information from the govern-
ment is to put written and oral questions to the ministers. The government responds
to oral questions in the National Assembly every week. Since 1974, due to the initiative
of then newly elected President Giscard d’Estaing, this question and answer session is
no longer an unproductive reading of prepared and written statements. Instead it is
characterized by much of the lively give-and-take typical of the weekly question periods
in the British Parliament. Question sessions in the National Assembly attract more
viewers than other meetings and both the majority and opposition makes good use of
time. A fundamental difference between the Fourth and Fifth Republics is that these
question sessions have not become a new ‘interpellation’ procedure—an instrument
used by deputies under the Fourth Republic to harass the government. The existing
Standing Orders of the National Assembly (1959) state that exchanges between
deputies and the government taking place during a question period may not lead to
a vote or a resolution. In the Fifth Republic, cabinets are more likely to resign because
France 339
of pressure from (or decisions by) the President than because of events that take place
in the National Assembly.
The procedure of controlling the executive machinery through the use of ad hoc
committees of investigation has rarely been used effectively. In most cases, when the
opposition calls for an investigation of something it considers scandalous or improper,
the government is able to use its numerical strength in the Parliament to prevent the
establishment of a committee of investigation. The investigative procedure is also strictly
regulated. The parliamentary majority has the exclusive right to appoint the members of
such committees. However, committees cannot ban the press from attending hearings
and committee findings must be made public. The President and the PM attempt to
prevent Parliament from subjecting their policies to careful scrutiny. Hence, these mech-
anisms designed to contain moral hazard have hardly been used in the Fifth Republic.
D E L E GAT I O N F RO M C A B I N E T A N D T H E P R E S I D E N T TO
I N D I V I D UA L M I N I S T E R S
Individual ministers head a ministry. There are no ministers without portfolio
(Thiébault 2000). In addition, ministers also collectively participate in the formulation
and management of governmental policy. In periods of cohabitation, the decision rule
in the cabinet is consensus as defined by the PM. However, at other times, when the
same party/coalition hold all branches of national government, the content and
existence of consensus is defined by the President. This is probably also the general rule
intended by the makers of the Constitution.
Since the beginning of the Fifth Republic, greater importance has been attached to
ministers’ management of departments than to their political and governmental
functions. The autonomy and therefore authority of ministers has declined as a result.
The cabinet is primarily an administrative body (Thiébault 1988b). Ministers tend to
relate to the PM, and also to the President, in a manner akin to that of civil servants.
The PM’s instructions (directives) to individual ministers are often requests to prepare
either decrees (executive decisions) to be submitted to the cabinet (Council of
Ministers), laws to be sent to the Parliament, or instructions about the preparation of the
budget. Individual ministers’ power of regulation is largely restricted to implementation
of laws, decrees, or budgetary matters (Thiébault 1994).
The President may place political and personal friends in other ministerial posts.
A number of ministers have been recruited directly from the President’s staff at the
Elysée. The President might encourage direct links between individual ministers and
the Elysée as a way to bypass the PM’s control. The PM can therefore rarely count
on the unswerving support of his ministerial colleagues. Except during periods of
cohabitation, the PM’s authority over the government is weakened by the President’s
control of the government’s agenda. The President’s programme forms the basis of
government action. The President might even intervene to impose his wishes with
regard to a policy matter without informing the government—or even against the
wishes of the ministers most concerned.
The complexities of modern government require coordination between ministries.
France has created a system of semi-permanent conflict resolution mechanisms for the
purpose of coordinating policy prior to its consideration by the cabinet (Quermonne
and Chagnollaud 1996). There are basically three types of institutions. First, there are
interministerial meetings composed of senior civil servants. These are held at the PM’s
residence at Matignon, with a member of the PM’s ministerial cabinet serving as
chairman. A member of the President’s staff represents the Elysée. These meetings end
with the PM’s representative summarizing the discussion. If agreement cannot be
reached, the proposals to be decided upon are then sent to the second level, the
interministerial committee. This committee is usually composed of ministers and
secretaries of state. Secretaries of state are junior ministers to whom the respective
minister delegates specific tasks. They do not belong to the cabinet and they are
subject to orders from their respective minister and the PM. The PM chairs the inter-
ministerial committee. It is the PM’s main instrument for resolving conflicts.
The third type of cabinet committee, the interministerial council, is the most exclusive
of these coordinating bodies. It usually consists of a few senior ministers who meet
at the Elysée under the chairmanship of the President. The President, the PM, the
ministers, and the ministerial cabinet spend a great deal of time at interministerial
council meetings, except during periods of cohabitation.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
The Fifth Republic tends to place a great reliance on ex ante accountability mechanisms,
especially the prior screening of high civil servants. The Constitution involves the
President (with the PM) in making appointments of high civil servants. A spoils system
at the summit of the French state has been developed. Because both the President and
the PM must agree on these appointments, disagreements have to be resolved during
the years of cohabitiation.
The relationship between ministers and civil servants within a ministry is a direct
chain of command. Executive agencies are not autonomous. In practice, the autonomy
of individual ministers also appears to be constrained by top bureaucrats in the
ministries, but this power does not have a real constitutional basis. Rather it stems from
the traditionally strong role of these bureaucrats who tend to have similar training and
background and enjoy high status in French politics (Rouban 1999). Nevertheless,
formally speaking, heads of departmental divisions are usually directly responsible to
342 Jean-Louis Thiébault
the department minister. The interposition of a secretary-general—a bureaucratic head
of a ministry similar to the British permanent secretary—between department heads
and the minister has been avoided (except in the ministry of foreign affairs).
The direct chain of command reinforces minister control over the political appointment
of high civil servants. The minister can make discretionary appointments—although
the cabinet formally appoints heads of divisions, and the President and/or PM some-
times get involved in individual appointments. If differences of opinion develop between
an individual minister and the head of a division within the ministry, then the situation
is usually resolved by promoting the division head or transferring him/her to an
equivalent post elsewhere.
Since the beginning of the Fifth Republic, top civil servants have prepared important
policy decisions—down to the smallest detail—without consulting Parliament or other
elective bodies. They have thus assumed de facto decision-making power over many
issues, and as a result there is little or no discussion of these matters in Parliament.
In the relationship between ministers, members of their staff, and administrative
personnel, the boundaries separating the initiation and execution of policies are necessarily
vague. Among these three groups there exists a considerable degree of homogeneity
with regard to function and mentality (Marcou and Thiébault 1996). Both ministers
and the civil servants become directly involved in establishing political priorities and
thereby in decision-making. However, in many respects, the restoration of governmental
authority has strengthened executive control of top civil servants.
Ministerial staffing has been developed in order to help individual ministers manage
their departments. All ministers have a staff (usually from ten to twenty persons) that
provides the crucial link between the political and administrative executive. The lack of
permanent secretaries in most French ministries means that the task of coordinating
and directing the work of the various divisions is the responsibility of the ministerial
staff. The fact that decision-making has become a more exclusively executive matter
under the Fifth Republic has meant that the key members of the ministerial staff
are those advisers borrowed from the civil service who are concerned with preparing,
coordinating, and supervising the implementation of policy. The overwhelming majority
of the members of ministerial staffs are recruited from the senior civil service. About
three-quarters of the holders of the most important posts of director and special
adviser are recruited from among the ‘grand corps’, notably from the group of admin-
istrative lawyers of the Council of State (see below) and the economic experts in the
finance inspectorate. This is clear evidence that these two bodies compose a parallel hier-
archy to the heads of the ministry’s divisions. In addition, it reflects the fact that in
France it is considered more important to ensure bureaucratic accountability to minis-
ters than ministerial accountability to Parliament (Beaud and Blanquer 1999).
The director of the ministerial staff has become the de facto element of administra-
tive coordination within the ministry. He receives all documents to be signed by the min-
ister. It is his task to decide which he can deal with himself and which must be referred
to the minister due to their political importance. Instructions to senior civil servants
from the director of the ministerial staff are considered to be from the minister him-
self. As well as coordination within the ministry, the director of the minister’s staff is
also responsible for preparatory coordination between ministries. When a minister
France 343
cannot attend an interministerial meeting, the director of his staff usually attends in his
place. The bulk of ministerial staff consists of specialist advisers and chargés de mission
that have responsibility for particular areas of the ministry’s work.
The staff members perform four main functions. First, they prepare memoranda for
the minister on matters to be considered in cabinet. This work might involve conducting
basic research on particular questions or actually deciding what the ministry policy
ought to be. Second, with regard to the area for which they have responsibility, they
coordinate the work of the ministry and conduct its relations with other ministries, the
latter of which involves participating in numerous meetings with members of other
ministerial staffs. They also meet with interest groups seeking to influence the ministry
as well as groups whose support the ministry wants to mobilize in support of ministry
policy. Third, they draft outlines of the minister’s speeches and interviews. Finally,
advisers and chargés de mission supervise the implementation of the ministry’s policies,
which involves maintaining close and frequent contact with senior officials in the
ministry (Quermonne and Chagnollaud 1996). The capacity of individual ministers to
actually monitor the behaviour of their civil servants, to control them at this stage of
the chain of delegation, has become more important. This, in turn, has increased the
importance of the ministers’ staff.
E X T E R NA L C O N S T R A I N T S
Institutions that are outside the parliamentary and bureaucratic chains of delegation
insert themselves in the process of national decision-making. The importance of the
President and the use of referendums have already been discussed. Moreover, the
President has considerable decree powers under Article 16 of the Constitution, which,
however, have been used only once (in 1961) (Huber 1998). Other institutions of
interest in this regard include (constitutional) courts, central banks, and international
organizations, the latter of which increasingly intervene in the delegation process and
limit actors’ freedom of action.
These developments have multiplied the constraints on the chain of delegation in
France. The influence of the Constitutional Council is now well documented. The
establishment of a Constitutional Council in 1958 was part of an attempt by the
makers of the new Constitution to protect the innovations of the Fifth Republic from
future efforts to undermine them. The champions of the traditional doctrine that the
Parliament is the repository of national sovereignty were going to be prevented from
successfully eroding the constitutional constraints imposed on Parliament. To this end,
the Constitutional Council was created. A novelty in French parliamentary democracy,
it was given the authority to decide whether laws passed by Parliament are unconstitu-
tional. A decision in 1971 expanded the Council’s power by making it a guardian of civil
liberty capable of imposing respect for constitutional rights even on the President and
the government. The Constitutional Council has thus become a constraint on the
executive organs of government, which are required to act in conformity with the
values enshrined in the Constitution and its preamble.
There is also a change in the supervision of policies in some sectors from the State to
independent administrative authorities, such as the Higher Council of Broadcasting (CSA).
344 Jean-Louis Thiébault
Government control over the Bank of France was increased by legislation in 1936 and
by nationalization of the Bank in 1945. Complete government control over the Bank of
France put ‘monetary power’ at the disposal of the political interests of the cabinet.
Relations between the Bank of France and the finance minister have often been even
closer than institutional ties suggest because the Bank’s governor has often been
a finance inspector or a former finance ministry official. However, more recently the
Bank has begun to show signs of greater autonomy. The recent more influential posi-
tion of the governor of the Bank of France can be explained by his long tenure in office
and by the fact that he now, under European monetary integration, possesses the
autonomous powers of the German Bundesbank President (Elgie and Thompson 1998).
The policy process has increasingly adopted an European dimension. The European
Union has imposed a general constraint on the cabinet and has set strict parameters
within which policies can be developed. In recent years, the French cabinet has lost
much of its power to control budgetary and monetary policies.
De Gaulle’s conception of a Europe of sovereign states was built on the notion of
EC/EU cooperation controlled by national governments. The French position turned
more pro-federal under Mitterrand, but French EU policy is primarily seen as a matter
for the President and the government and not for the Parliament.
Over the years, the French government has maintained close control over its experts
in Brussels. The general secretariat of the interministerial committee for European
cooperation (SGCI) sends instructions to Brussels early in the policy formulation and
negotiation process. The SGCI is headed by a senior civil servant that is accountable to
the PM. Highly centralized administrative coordination also takes place, with a minis-
ter especially responsible for European affairs assuming political responsibilities as well.
Under the President, the PM seeks to assert control over European policy through the
interministerial committee for European cooperation, which is charged with responsibil-
ity for high policy. The foreign affairs minister has sought to monopolize supervision of
relations between the EC commission and all French ministries. The French foreign
ministry plays a decisive role in the key EU decision-making bodies—the Council of
Ministers and the Committee of Permanent Representatives—via monthly meetings of EU
foreign ministers and the ministry’s influence over the preparation and exposition of agreed
national positions in France’s permanent delegation. The French Parliament has only
recently been given a measure of involvement in French EU policy. For years it has only
been provided with routine information about the decision-making process. However,
some measures have been adopted to better associate the French Parliament with
European policies and to introduce some possibilities for control. Accordingly, the cabi-
net must subject to Parliament’s European proposals, when they are likely to affect French
legislation. Parliamentary committees for EU affairs have been created and are consulted
on such proposals. This system has created an ex ante control, before EU rule-making.
C O N C LU S I O N
France has a flexible political system capable of producing various combinations of
state and partisan power. If the President and majority in the National Assembly come
France 345
from the same political party/group, then the relation between them is a hierarchical
one. The parliamentary majority, the PM, the government, and the administration all
act under the leadership of the President. The idealized chain of delegation and
accountability in parliamentary democracy is not observed. Of particular importance
is the fact that the President is not accountable to the Parliament. This situation is
aggravated by the weakness of parliamentary control. This is a fundamental difference
between the French and American systems, the latter of which has been sometimes
described as been a ‘congressional’ regime.
However, if the President and National Assembly majority are of different partisan
loyalties, cohabitation is conducive to the emergence of the PM as the dominant execut-
ive power. In this case the President’s power is limited to that given him in the
Constitution. The idealized chain of delegation and accountability under a parliamen-
tary regime (which links voters to Members of Parliament, Members of Parliament to
the cabinet and the PM, a PM to cabinet members, and ministers to civil servants) is
then more closely observed. The government is then collectively accountable to the
Parliament rather than to the President.
REFERENCES
Ardant, Philippe (1991). Le Premier Ministre en France. Paris: Montchrestien.
Avril, Pierre (1994). ‘Regulation of Political Finance in France’, in Herbert E. Alexander and Rei
Shiratori (eds.), Comparative Political Finance Among the Democracies. Boulder: Westview Press.
Bartolini, Stefano (1984). ‘Institutional Constraints and Party Competition in the French Party
System’. West European Politics, 7(4): 103–27.
Beaud, Olivier and Blanquer, Jean-Michael (1999). La responsabilité des élus. Paris: Fayard.
Blondel, Jean and Thiébault, Jean-Louis (eds.) (1991). The Profession of Government Minister in
Western Europe. London: Macmillan.
Carcassonne, Guy (1996). La Constitution. Paris: Le Seuil.
Chagnollaud, Dominique and Quermonne, Jean-Louis (1996). Le Gouvernement de la France sous la
5e République. Paris: Fayard.
Cole, Alistair (1993). ‘The Presidential Party and the Fifth Republic’. West European Politics,
16(3): 49–66.
Cotteret, Jean-Marie and Emeri, Claude (1994). Les systèmes électoraux. Paris: Presses
Universitaires de France.
Drysch, Thomas (1993). ‘The New French System of Political Finance’, in Arthur B. Gunlicks
(ed.), Campaign and Party Finance in North America and Western Europe. Boulder: Westview Press.
Duhamel, Olivier (1994). Droit Constitutionnel et Politique. Paris: Le Seuil.
Duverger, Maurice (1985). Le Système Politique Français. Paris: Presses Universitaires de France.
Elgie, Robert and Howard Machin (1991). ‘France: the Limits to Prime Ministerial Government
in a Semi-presidential System’. West European Politics, 14(2): 62–78.
—— —— and Maor, Moshe (1992). ‘Survival of Minority Governments: The French Case,
1988–91’. West European Politics, 15(4): 57–74.
—— and Thompson, Helen (1998). The Politics of Central Banks. London: Routledge.
Goldey, D. B. (1998). ‘The French General Election of 25 May–1 June 1997’. Electoral Studies, 17:
536–55.
Hayward, Jack (1993a). ‘From Republican Sovereign to Partisan Statesman’, in Jack Hayward
(ed.), De Gaulle to Mitterrand. Presidential Power in France. London: Hurst.
346 Jean-Louis Thiébault
—— (1993b). ‘The President and the Constitution: Its Spirit, Articles and Practice’, in Jack
Hayward (ed.), De Gaulle to Mitterrand. Presidential Power in France. London: Hurst.
Huber, John D. (1996). Rationalizing Parliament. Cambridge: Cambridge University Press.
—— (1998). ‘Executive Decree Authority in France’, in John M. Carey and Matthew Soberg
Shugart (eds.), Executive Decree Authority. Cambridge: Cambridge University Press.
Knapp, Andrew (1991). ‘The Cumul des mandats, Local Power and Political Parties in France’.
West European Politics, 14(1): 18–40.
Machin, Howard (1993). ‘The President, the Parties and Parliament’, in Jack Hayward (ed.),
De Gaulle to Mitterrand. Presidential Power in France. London: Hurst.
Marcou, Gérard and Thiébault, Jean-Louis (eds.) (1996). La Décision Gouvernementale en Europe.
Paris: L’ Harmattan.
Menon, Anand (1994). ‘Continuing Politics by Other Means: Defence Policy in the French Fifth
Republic’. West European Politics, 17(4): 74–96.
Mény, Yves (1991). Le Système Politique Français. Paris: Montchrestien.
Money, Jeannette and Tsebelis, George (1995). ‘The Political Power of the French Senate:
Micromechanisms of Bicameral Negotiations’. Journal of Legislative Studies, 1: 192–217.
Reif, Karlheinz (1987). ‘Party Government in the Fifth French Republic’, in Richard S. Katz (ed.),
Party Government: European and American Perspectives. Berlin: de Gruyter.
Rouban, Luc (1999). ‘The Senior Civil Service in France’, in Edward C. Page and Vincent Wright
(eds.), Bureaucratic Élites in Western European States. Oxford: Oxford University Press.
Schramek, Olivier (2001). Matignon, river gauche, 1997–2001. Paris: Le Seuil.
Stevens, Anne (1993). ‘The President and His Staff ’, in Jack Hayward (ed.), De Gaulle to Mitterrand.
Presidential Power in France. London: Hurst.
Suleiman, Ezra N. (1980). ‘Presidential Government in France’, in Richard Rose and Ezra
N. Suleiman (eds.), Presidents and Prime Ministers. Washington, DC: American Enterprise Institute.
Thiébault, Jean-Louis (1988a). ‘France: the Impact of Electoral System Change’, in Michael
Gallagher and Michael Marsh (eds.), Candidate Selection in Comparative Perspective. The Secret
Garden of Politics. London: Sage
—— (1988b). ‘France: Cabinet Decision-making under the Fifth Republic’, in Jean Blondel and
Ferdinand Müller-Rommel (eds.), Cabinets in Western Europe. London: Macmillan.
—— (1994). ‘The Political Autonomy of Cabinet Ministers in the French Fifth Republic’, in
Michael Laver and Kenneth A.Shepsle (eds.), Cabinet Ministers and Parliamentary Government.
Cambridge: Cambridge University Press.
—— (2000). ‘France: Forming And Maintaining Government Coalitions in the Fifth Republic’, in
Wolfgang C. Müller and Kaare Strøm (eds.), Coalition Governments in Western Europe. Oxford:
Oxford University Press.
Wright, Vincent (1993). ‘The President and the Prime Minister: Subordination, Conflict,
Symbiosis or Reciprocal Parasitism?’, in Jack Hayward (ed.), De Gaulle to Mitterrand. Presidential
Power in France. London: Hurst.
10
I N T RO D U C T I O N
The German Basic Law of 1949 was a result of historical learning. The experience of
five radical regime changes within less than a century, the collapse of the Weimar
Republic largely due to deep elite and mass rifts, and National Socialist totalitarianism
shaped the Federal Republic’s Constitution and the informal norms underpinning it
(Pulzer 1995: 7; Paterson 2000: 25). The construction of democratic delegation and
accountability was deeply influenced by the desire for political stability and consensus
and the need to stamp out dictatorship for good. The resulting constitutional design
featured a relatively strong dispersion of political power encouraging bargaining
amongst different political and economic elites. I will argue in this chapter that the
Basic Law was successful in achieving democratic regime stability, a broad democratic
consensus, and a high level of government and policy stability. The downside have been
efficiency problems (high transaction costs of policy change) and problems of political
accountability and transparency.
Article 20(2) of the Basic Law proclaims the electorate to be the ultimate democratic
principal and establishes a democratic chain of delegation and accountability in which
all state authority emanates from (but is not immediately exercised by) the people. At
the national level, referendums are restricted to decisions about boundary changes
between Länder, to be held in those Länder directly affected by these changes. As far as
the initial two links in the process of delegation and accountability are concerned—the
link between voters and Members of Parliament MPs, and the government—the Basic
Law largely conforms to the ideal type of parliamentary democracy (Strøm 2000:
268–9). Yet, there are considerable deviations, especially at the later stages of delega-
tion. The German federal government is constrained by a large number of actors with
veto powers whose authority may be derived from the Constitution or quasi-constitu-
tional norms rather than the electoral process (e.g. the Federal Constitutional Court or
the European Central Bank). While the Federal Republic largely fits Strøm’s (2000: 268)
characterization of a parliamentary system of government as ‘a single chain of delega-
tion with multiple links’ where—in each link—‘a single principal delegates to a single
348 Thomas Saalfeld
or multiple non-competing agents’, German federalism adds at least a strong element
of delegation to competing agents, which Strøm (2000: 269) considers to be more typ-
ical of presidential regimes.
The provision that each essential norm set by any political actor has to be based on par-
liamentary legislation is one of the constitutional safeguards for the democratic account-
ability of policy-makers. The federal government, a parliamentary party, or a group of 5
per cent of the Members of the Bundestag from one or more parties are entitled to intro-
duce bills in the Bundestag. The Bundesrat, the body representing the Länder executives,
can also initiate bills, but does so infrequently.1 The Basic Law distinguishes between
three types of laws with different procedures and majority requirements:
1. Constitutional amendments must be approved by two-thirds of the Members of the
Bundestag and two-thirds of the votes of the Bundesrat.2 Despite this seemingly high
threshold the Basic Law was amended 190 times between its proclamation and the
summer of 1998 (Busch 2000: 445). Certain constitutional provisions, however, are
entirely or partly protected from amendment. Article 79 (3) of the Basic Law prohibits
constitutional amendments affecting the federal nature of the German state, the par-
ticipation of the Länder in legislation, and the basic rights and principles laid down in
Articles 1 and 20 of the Basic Law. The fundamental rights in Articles 2–19 can be
amended, but their essential content may not be encroached upon.
2. A regular law is passed by the Bundestag subject to the suspensory veto of the
Bundesrat. That is, the Bundesrat can delay a bill’s passage by lodging an objection. Such
an objection can be overturned by an absolute majority of the Members of the
Bundestag. If the Bundesrat rejects a regular bill by a two-thirds majority, its veto can
only be overturned in a vote in which at least two-thirds of the Bundestag Members par-
ticipate and at least 50 per cent plus one of the Bundestag Members reject the veto. The
budget is essentially a ‘regular law’ with some qualifications. Most importantly, only the
federal government can introduce the budget bill. The Bundesrat may express an opin-
ion, but cannot veto the federal budget 3 (for a description of the budgetary process see
Ismayr 1992: 402–14).
3. By contrast, Bundesrat consent is mandatory for consent laws—that is, such laws
are subject to an absolute veto that the Bundestag cannot overturn. In effect, more than
half of all bills and most important domestic laws are ‘consent laws’ (Saalfeld 1998a: 50).
After their passage in the Bundestag bills are sent to the Bundesrat. If the Bundesrat
rejects a bill in whole or in part, it is usually referred to the conference committee with
an equal number of representatives from the Bundesrat and the Bundestag (for details
1
Most scholars do not consider the Bundesrat to be a second parliamentary chamber, but an autonomous
state organ in its own right (e.g. see Rausch 1981: 72). Given its involvement in the process of legislation,
however, the Bundesrat will be treated as a functional equivalent to a second parliamentary chamber,
although it is in reality a representation of the executive branch of the Land governments.
2
Each state has between three and six votes roughly in proportion to its population. State-government
representatives always vote in accordance with the instructions of their governments.
3
However, bills involving taxes to which—completely or in part—federal states or local authorities are
entitled require mandatory Bundesrat consent (Article 105[3] of the Basic Law). The same is true with regard
to bills concerning the distribution of taxes between the federation and the states.
Germany 349
P O L I T I C A L PA RT I E S
The political parties influence the delegation process at virtually every stage. Not only
are they privileged actors in Parliament, their influence in German politics and society
is such that Germany has often been referred to as a ‘party state’ (Grewe 1951) or ‘party
democracy’ (Gabriel, Niedermayer, and Stöss 1997). They provide the most important
screening mechanisms for parliamentary and ministerial candidates, although there
have been doubts whether they are really a safeguard against ‘adverse selection’.
Aspiring parliamentarians typically have to go through a long party career before they
are nominated for elected office, which is sometimes criticized as creating a systematic
bias in favour of career politicians and discouraging candidates with professional experi-
ence outside party politics (Scheuch and Scheuch 1993: 50–1).
One distinctive characteristic of the Federal Republic is the high degree of legal cod-
ification of party politics (Poguntke 1994). The most important legal sources deter-
mining the status of extra-parliamentary parties are the Basic Law, the Political Parties
Act, the Federal Elections Act, and the individual parties’ constitutions. Of particular
importance at the parliamentary level are the Members of Parliament Act,—which was
amended in 1994 to include a section on parliamentary parties (especially their finan-
cial accountability), the Bundestag’s rules of procedure, and the individual parliamen-
tary party groups’ rules of procedure.
Article 21 of the Basic Law formally recognizes the crucial role of political parties
in the ‘formation of the political will of the people’ and generally requires that their
internal life be organized democratically. The 1967 Political Parties Act requires a number
of specific principles of intraparty democracy. For example, it requires all party execut-
ive bodies to be elected by conferences of delegates of the rank-and-file membership
and limits the number of (not directly elected) ex officio members to a maximum of
20 per cent of the respective body’s total membership. The Act thus requires parties to
have a democratic chain of delegation and accountability parallel to the one that exists
in the polity at large.
Individually the parties contribute to the screening and accountability of elected
politicians. Perhaps more importantly, postwar changes in the party system have also
improved the scope for general electoral accountability. The 1950s and 1960s witnessed
a steady concentration of the vote in favour of the two major parties, the Christian-
Democratic CDU/CSU and the Social-Democratic SPD. The result was a bipolar coalition
system with the liberal FDP frequently in a pivotal position, at least between 1961 and
4
Presidential vetoes have been extremely rare (Saalfeld 1998b: 272–3).
350 Thomas Saalfeld
1983. The relative cabinet stability since the mid-1960s and the bipolar structure with
an increasingly competitive relationship between government and opposition have
provided a more favourable context for electoral accountability than did the ill-fated
Weimar Republic. For the first time in German history, the postwar German party
system met ‘the two key tests of sustaining stable governments and providing for
alternation of power’ (Paterson 2000: 27).
There are currently five main parties: the Christian-Democratic alliance composed
of the Christian Democratic Union (CDU) and the Christian Social Union (CSU, the
former’s more conservative Bavarian ‘sister party’), the liberal Free Democratic Party
(FDP), the Social Democratic Party (SPD), the left-libertarian and environmentalist
Green Party (Bündnis ‘90/Die Grünen), and the post-communist Party of Democratic
Socialism (PDS), whose support is concentrated in eastern Germany. CDU/CSU, SPD,
and FDP have dominated government formation for much of the postwar period. The
Green Party has been represented in the Bundestag since 1983 and formed national
government coalitions with the SPD in 1998 and 2002. The PDS was represented in the
Bundestag as a parliamentary party (Frakton) between 1990 and 2002. While the PDS
has been involved in two formal coalitions and tolerated a number of minority cabinets
at the Land and local levels in eastern Germany, it has remained ‘non-coalitionable’ fed-
erally as the most ideologically left-wing party and one that is seen to have not yet ade-
quately dealt with its past as the ruling party of the communist dictatorship in East
Germany between 1949 and 1989. It failed to overcome the 5 per cent threshold of the
Federal Republic’s electoral law in the 2002 elections.
Although constraints of space do not allow a detailed description of the development
of the German party system (Stöss 1986; Mintzel and Oberreuter 1992; Lösche 1993;
Roberts 1997), certain key characteristics should be mentioned. First, CDU and CSU
are independent parties forming a common parliamentary party in the Bundestag. The
alliance is formally renewed after each general election. As a result of the far-reaching
codification of party organization and finance, there are no fundamental organizational
differences among the main parties, apart from the fact that only CDU and SPD really
have a mass membership.5 The major German parties are not organized in a strictly
hierarchical fashion. Some authors have therefore chosen expressions such as ‘loosely
coupled anarchy’ (Lösche 1993) or ‘organised anarchy’ (Wiesendahl 2000) to describe
the organizations of the CDU and SPD and the relatively autonomous coexistence of
the different levels of party organization. One consequence is that power is dispersed
within the political parties and that governments face high transaction costs in gener-
ating the necessary intraparty consensus for policy changes. This, in turn, encourages
bargaining in informal bodies.
D E L E GAT I O N F RO M VOT E R S TO M P S
German citizens use their vote to delegate powers to elected politicians at the local,
Land, national, and European Union (EU) levels. Given the intertwined relations
5
In December 2000 the CDU had approximately 617,000 and the SPD approximately 735,000 members
(Sources: reports of CDU and SPD executive committees to the parties’ annual conferences).
Germany 351
between the Länder and federation (the national level) in German politics, the chains
between voters and Members of federal–state Parliaments on the one hand
and between voters and Members of the Bundestag on the other must be seen as com-
petitive. The Bundesrat’s crucial role in national politics, the importance of Land
bureaucracies for the implementation of national legislation, the national ambitions of
many leading Land politicians, the national importance of Land elections (which are
not considered to be second-order elections like local or EU elections), the existence of
legislative ‘joint tasks’ shared by the federation and the Länder, and the existence of
‘concurrent’ and ‘framework legislation’ where both the federation and the Länder have
powers illustrate this point. From the perspective of agency theory, therefore, the
Federal Republic has a more complicated chain of delegation and accountability than
the ideal-type of parliamentary democracy (Strøm 2000).
Given the crucial and all-pervasive role of political parties in the Federal Republic,
the direct electoral accountability of individual Bundestag Members is not as strong as
may be the case in the United States. Powerful ex ante control mechanisms such as
screening and selection of candidates are reserved for the minority of citizens who are
active members of the respective candidate’s party. Individual candidates rarely expect
significant personal electoral rewards or penalties independent of their party’s overall
fate. Like ex ante accountability mechanisms, ex post electoral accountability is mediated
by the political parties. Nevertheless, electoral competition with candidates from other
parties and, crucially, aspirants from within their own parties provides incentives
for Members of the Bundestag to share information with voters, especially through
constituency ‘surgeries’ and the local news media at constituency level.
Delegation from voters to MPs is strongly shaped by Germany’s electoral system of
proportional representation with a two-tier districting system. According to the Federal
Elections Act, 299 Members of the Bundestag were in 2002 elected at the lower level by
plurality vote in single-member constituencies. Proportionality is achieved through
adjustment at the upper level, where an additional 299 deputies were elected from Land
party lists.6 Nonetheless, the German system is not completely proportional and
discriminates against smaller and new parties. The most important violation of strict
proportionality is the 5 per cent clause of the electoral law. Parliamentary representation
requires at least 5 per cent of the national vote or three direct constituency mandates
(for more detailed accounts of the Federal Republic’s electoral system and its evolution see
Nohlen 1986; Jesse 1990). As intended by the framers of the electoral law, the 5 per cent
threshold contributed considerably to the concentration of the party system
during the 1950s. It has effectively barred small anti-democratic parties from national
parliamentary representation without completely stifling new competitors, as the
Greens since the 1980s and the PDS in the 1990s demonstrate. To some extent, the
‘additional-member’ electoral system would seem to allow for a ‘personal vote’ and give
6
Since 1949, the proportionality of the system has been gradually increased. In 1953, the ratio of lower-
level mandates (apportioned by plurality vote) to upper-level (adjustment seats to achieve proportionality)
was reduced from 60 : 40 to 50 : 50. In 1987 the d’Hondt system of apportionment of upper-level seats to each
party was replaced by the more proportional LR-Hare system.
352 Thomas Saalfeld
voters an opportunity to employ ex post facto sanctions against individual Members of
the Bundestag. In practice, however, only a small minority of German voters see the
candidates of the parties as the determining factor of their vote (Conradt 2000: 150).
Ex ante (that is, prior to the delegation from voters to Members of Parliament), voters
have no direct say in candidate selection. Those who join one of the main political
parties have a better opportunity to screen individual candidates and to hold them
personally accountable, although the process is in practice often dominated by local
and regional oligarchies. Nomination by a party is the crucial step to election: ‘a candidate
can expect to be elected to Parliament if his party places him high up on its state party
list’ (Patzelt 1995: 58). Paragraph 22 of the Federal Elections Act regulates the
nomination of both list and constituency candidates. Candidate selection is strongly
dominated by local and regional party elites, ‘and it is only in exceptional circumstances
that Land or federal party leaderships can hope to impose candidates of their own
choosing’ (Poguntke 1994: 189). Although SPD statutes require ‘consultation’ with
their national executive in the nomination of both constituency and list candidates, and
although the state or district executive committees have certain veto rights, these powers
are very rarely used. Specific ‘recommendations’ from the federal or Land executive
committee about nominations are not welcomed and frequently disregarded (Kaack
1971; Poguntke 1994). Compared to all actors in the electoral process, mid-level
regional party elites have the most effective means to screen, monitor, and sanction
candidates and incumbents.
The rules of electoral competition have not changed a great deal since 1957. There is
no legal document explicitly specifying the minimum length of a campaign. However,
provisions of the Basic Law and the Federal Elections Act indirectly point to 30 days as
the minimum. According to Article 39(1) of the Basic Law, the Bundestag is elected for
4 years (i.e. 48 months). At the earliest, regular elections are to be held 45 months after
the last elections, at the latest 47 months after them. This formulation is the result of a
1976 amendment to the Basic Law (which was first applied in 1980). Between 1953 and
1976, each Bundestag had a fixed term of 4 years unless it was dissolved by the Federal
President. It is legally and politically almost impossible and/or very costly for govern-
ments to call early elections at a politically convenient time. The constraining effect
of competition rules such as the ones mentioned above is compounded by the
constitutional provisions pertaining to the early dissolution of Parliament (see below).
There is no legal cap on campaign spending. Public funding of political parties has
been generous by international standards and has helped relatively small parties to sta-
bilize their organizations and/or cope with the ‘liability of newness’. Between 1967 and
1994, public funding was specifically designed to compensate parties for election
expenses as their ability to compete effectively in elections was seen to be in the public
interest. All parties that are permitted7 to put forward party lists (not just parliamentary
7
The Federal Constitutional Court has the power to ban parties and seize their assets if their aims are
found to anti-democratic. Parties can only put forward a Land list in elections, if their list is supported by the
signatures of one per thousand (maximum: 2,000) voters in the respective federal state. Individual
constituency candidates need the support of 200 voters in their constituency.
Germany 353
parties) are allocated free public broadcasting time during election campaigns. The par-
ties have to pay for the production costs of advertisements transmitted by
commercial stations. In both—public and commercial—systems, the norm for allocating
time is proportionality, with a minimum quota for non-parliamentary parties. It is not
possible to purchase additional time on the public broadcasting networks. There are no
restrictions on the publication of opinion polls, except that exit polls on polling day may
not be published until polling stations close.
The regulation of public subsidies to political parties is an important aspect of the
Political Parties Act of 1967. In the early 1990s, approximately two-thirds of the parties’
considerable income stemmed from public funds. Public party subsidies have been the
subject of many legal disputes, and the nature of public funding has been changed a
number of times (see Gunlicks 1988; Poguntke 1994; Saalfeld 2000c). Public financing
of political parties started with indirect subsidies in the form of tax deductions for
donations in 1954. After a number of changes, further amendments to the party laws
of 1994 stipulate that parties are entitled to a direct subsidy for each voter and year, and
for each euro they receive in membership contributions and donations, as long as the
donations originate from natural persons (rather than corporations and interest
groups) and do not exceed a certain limit per annum. The allocation formula gives a
slight advantage to smaller parties. Direct public subsidies are limited to 50 per cent of
a party’s total income (‘relative limit’) and to euro 133 million for all parties combined
(‘absolute limit’). The latter can be revised in line with the rate of inflation. In order to
be eligible for public subsidies, a party must have won at least 0.5 per cent of the vote
in the previous Bundestag or European Parliament elections or 1.0 per cent in a Land
election (Article 18, Parteiengesetz). Especially since the 1980s, a number of high-pro-
file party finance scandals have led to serious criticism of the parties as well as the sys-
tem of public funding. Nevertheless, the system is fair to smaller parties. Furthermore
it has allowed sitting Members of Parliament and their challengers to maintain a rela-
tively well-supported presence in their constituencies and thereby to share information
with interested constituents and grass-roots party members. At least for parliamentar-
ians without government office or a high-profile function in their (parliamentary) party,
regular constituency work has become a virtual precondition for renomination. It has
also allowed MPs and parliamentary parties to employ staff in the federal capital to
support their ongoing oversight activities.
Article 38 of the Basic Law underscores the independence of individual Members of
the Bundestag vis-à-vis voters, parties, or any other influences. It stipulates that
Members are to represent the people as a whole, that they are not bound by any orders
or instructions, and that they are subject only to their own conscience. Thus, in consti-
tutional theory the agency relationship between voters and MPs is deliberately
designed to grant the agent—the MP—a large degree of autonomy. The relationship
between MPs and their parties has been far more contentious. Article 21 of the Basic
Law has often been interpreted by leading constitutional lawyers (especially during the
1960s) in such a way as to make Members of Parliament virtual agents of their parties
(e.g. Leibholz 1966). Today the predominant view amongst constitutional lawyers is
that the Basic Law recognizes parties—inside and outside Parliament—as indispensable
354 Thomas Saalfeld
elements in the chain of democratic delegation and accountability (although this termi-
nology is not used). Nevertheless, the formal independence of individual Members enjoys
equal constitutional status and serves as a corrective against over-powerful parties. The ten-
sion between Articles 21 and 38 allows the Federal Constitutional Court to rule on specific
conflicts on a case-by-case basis (Arndt 1989: 652–4).
Ex post safeguards against the consequences of oligarchy and agent opportunism,
such as term limits or a mandatory retirement age, do not exist. Between 1983 and
1987, however, the Green party used a system of ‘rotation’, in which every Green
Member of the Bundestag was required to resign after 2 years and was replaced by
another candidate drawn from the respective Green Land list. This practice was con-
sidered to be of questionable constitutional status and abandoned for practical reasons.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
The agency relationship between MPs and cabinet is complicated. In reality we can dis-
tinguish at least three agency relationships, each with its specific problems:
1. Although the ‘two-body image’ of ‘the’ Parliament versus ‘the’ executive is
misleading for modern parliamentary systems of government (see King 1976), there is
a ‘residual’ collective parliamentary identity as its majority decisions are treated as
decisions of the chamber as a whole. More importantly in our context, there is a vast
information asymmetry between parliamentarians serving in government office on the
one hand and non-government members on the other. This asymmetry is largely a
result of the government ministers’ privileged access to the civil service, executive
agencies, and scientific advice. The government’s informational advantage generally
reduces the effectiveness of ongoing oversight and increases the scope for hidden action
and thus moral hazard. Like all other West European parliaments, the German
Bundestag is affected by such problems.
2. King (1976) attempts to capture the reality of modern parliamentary systems in
his distinction between ‘intra-party’, ‘inter-party’, and ‘opposition’ modes of executive–
legislative relations, indicating that there are not one but several agency relationships,
each with their own problems. The first two modes refer to the agency relationships
within each government party (‘intra-party mode’) or within a coalition of parties
(‘inter-party mode’), the third refers to those parties who do not support the government
of the day (‘opposition mode’). The most powerful ex ante (screening and selection) and
ex post accountability mechanisms (monitoring, rewards, and punishments) are effectively
(if not totally) reserved for the members of the government parties. Due to a higher
degree of incentive compatibility between government and government parliamentarians,
the Members of the governing parties tend to have easier and timelier access to accurate
executive information than Members belonging to opposition parties. The Federal
Republic differs from other West European parliamentary democracies in that the
informational disadvantage at least of the major opposition parties is usually alleviated
by their access to ministerial bureaucracies at the Land level. Major opposition parties can
often draw on support from ‘friendly’ Länder governments controlled by their own
Germany 355
regional organizations. For such governments the incentives to cooperate with the
national opposition may be strong if they compete electorally with government parties
at the Land level and/or if the head of the respective Land government has ambitions
on the national level.8
3. Legislative–executive relations in Germany are additionally influenced by the ethos
of a ‘working chamber’ contributing to a high degree of institutionalized policy special-
ization. This specialization is reflected in the importance given to committee work, but
also in the emphasis on policy formulation and monitoring through parliamentary work-
ing groups, especially in the two major parties. This creates a layer of policy experts
within the parliamentary parties with privileged access to executive information.
Although this improves the Bundestag’s overall capacity for monitoring and other ongo-
ing oversight activities, it also creates additional agency relationships (and problems)
between ordinary backbenchers on the one hand and the parliamentary parties’ policy
experts, who often act on behalf of their parliamentary parties at large, on the other.
Let us now consider some accountability mechanisms in greater detail. The Bundestag
as a whole, and the parliamentary majority parties in particular, have three crucial ex
ante devices at their disposal.
1. The Basic Law requires the election of the Federal Chancellor by an absolute
majority of the Bundestag. The formal rules of government investiture are laid down in
Articles 63, 64, and 67 of the Basic Law. The Federal President proposes to the
Bundestag a candidate for the office of Federal Chancellor. To be elected, a candidate
must be supported by an absolute majority of all Members of the Bundestag, that is,
there is an explicit support requirement of 50 per cent plus one vote of all members
entitled to vote and an implicit quorum requirement of 50 per cent plus one of all
members. Voting is done by secret ballot. The Federal President must appoint the
chancellor-candidate elected by an absolute majority of the chamber. If no candidate
receives an absolute majority in the first ballot, there may be an unspecified number of
further ballots within 14 days. If no candidate wins an absolute majority within this
period, the Federal President may appoint the candidate that has the relative majority
of votes or, alternatively, may call an election. As yet, this has never been necessary.
Once elected, the Chancellor nominates the members of his cabinet, who are sub-
sequently appointed by the Federal President without separate Bundestag approval.
The cabinet ministers’ tenure in office ends with the Chancellor’s.
2. The election of a Chancellor is only the formal completion of the process of
government and, in particular, coalition formation. Leading members of the parlia-
mentary majority parties are closely involved in this bargaining process, as are leaders
and relevant experts from the extra-parliamentary party organizations. These actors
work out the future government’s policies as well as the allocation of cabinet portfo-
lios to particular parties and politicians. The government’s policy programme is for-
mally summarized in a coalition agreement and the Federal Chancellor’s first government
8
It has to be remembered, for example, that Land Prime Ministers and ministers have the privilege to
speak in the Bundestag at any time and may be important speakers for the national opposition.
356 Thomas Saalfeld
declaration to Parliament. In effect this means that the majority parties agree on a formal
contract before the government is elected (cf. Saalfeld 2000b).
3. Although the Basic Law does not require the Federal Chancellor or cabinet
ministers to be Members of the Bundestag, all Chancellors except Kurt Georg
Kiesinger (1966–9) have been Members at the time of their election.9 To a slightly lesser
extent this is also true for cabinet ministers. The Federal Chancellor and cabinet
ministers are thus subjected to informal screening mechanisms in the form of parlia-
mentary apprenticeships. Yet, as ministers are often drawn from Länder governments or
elsewhere from outside the Bundestag, the average parliamentary seniority of German
cabinet members (1949–84: 7.7 years) is markedly lower than in the United Kingdom
(12.2), Finland (10.3), Denmark (9.7), and some other West European countries (De
Winter 1991: 48). A lengthy pre-ministerial career in party politics—inside or outside the
national chamber—provides Members of Parliament with information about a
prospective minister’s or Chancellor’s policy preferences and personal qualities. Even
when the Chancellor appoints a minister from ‘outside’ the usual career stream of pro-
fessional party politics, the latter is interested in quickly establishing a good working rela-
tionship with the relevant experts in the government’s parliamentary parties. This usually
involves becoming a Member of the Bundestag at the earliest opportunity.
Once elected, the Federal Chancellor and his cabinet are relatively well protected from
early dismissal. The President cannot dismiss them unilaterally. The obstacles for the
use of the Bundestag’s ultimate ex post accountability device, removal from office, are
relatively high. Against its own will, the government can only be collectively dismissed
by the Bundestag, and only through a ‘constructive vote of no-confidence’ (Article 67
of the Basic Law) against the Chancellor. This means that an absolute majority of the
Bundestag Members must not only be willing to depose the Chancellor and his entire
cabinet, they must simultaneously elect an alternative candidate. The right to propose
such a constructive vote of no-confidence rests with a parliamentary party or at least
25 per cent of all Members of the Bundestag. Forty-eight hours must pass between the
proposal and vote. There is no limit on the number of constructive no-confidence votes
that can be held. The decision rule for all no-confidence votes is 50 per cent of all
Members of the Bundestag eligible to vote plus one. It follows implicitly that there is a
quorum requirement of at least 50 per cent of all members of the Bundestag plus one.
Voting is by secret ballot.
The right to propose confidence votes (in which the initiative rests with the govern-
ment) rests with the Federal Chancellor. He or she alone can be the object of such a
vote. There are no unregulated confidence votes, that is, the government is under no
obligation to resign in the absence of parliamentary support for important government
bills or motions. Nevertheless, the fate of Brandt’s 1969–72 cabinet demonstrates that
the Federal Chancellor may have incentives to seek early elections by ‘manufacturing’
9
Kiesinger had served as Prime Minister of the state of Baden-Württemberg before he became
Chancellor. Nonetheless, he had considerable prior experience in the Bundestag. He was not atypical for lead-
ing German politicians, who have often gained administrative experience in a state government before
returning to national politics as a minister or Chancellor.
Germany 357
10
The members of the cabinet did not vote.
Table 10.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Germany 1949–2000a
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv) Confidence votes (cv) Did the Who dissolved
formation cabinet end Parliament in
No. of Voting results of successful ivc No. Cabinet Cabinet No. of cv Cabinet with an this case?b
unsuccessful (1 : pro gov., 2 : against gov., of removed resigned under spec. removed by early
investiture 3 : abstentions and 4 : other) ncv by ncv to inst. mech. failed cv election?
votes 1 2 3 4 pre-empt
ncv
11
Keith Krehbiel’s (1991) ‘information theory’ of committees differs from the models proposed by other
new institutionalists in that it does not primarily focus on distributional aspects of policy-making (‘gains from
trade’). Although Krehbiel does not deny that reelection-seeking Members of Parliament care about distribu-
tional benefits to their constituencies, he insists that—due to uncertainty in the policy environment—
legislators are more severely constrained in obtaining distributional benefits than is suggested in the distrib-
utional perspective. His key interest is to analyse rules and procedures providing incentives for individual
parliamentarians to develop policy expertise and then to share policy-relevant expertise with fellow legislators,
including legislators with competing distributive interests. In this context, he identifies a number of principles
that encourage the development of policy expertise and information-sharing. He calls one of his principles
the ‘heterogeneity principle’. According to this principle, specialists from opposite sides of a policy spectrum
are collectively more informative to the other members of the parliamentary chamber than specialists from
only one side of the spectrum.
Germany 361
argue that if there is public competition between politically opposing forces, then
parliamentarians have incentives to share information with the voters. In parliamentary
systems of government, the opposition can be crucial in ensuring heterogeneity and
informational efficiency in Krehbiel’s sense, if it pursues a ‘competitive’ strategy vis-à-vis
the government. Yet, although the opposition’s ‘armoury’ in the Bundestag is strong by
international comparison and has been further strengthened over the years, there have
been important countervailing forces preventing the Bundestag from being informa-
tionally efficient. The traditional hostility towards party conflict among both German
elites and voters (Grosser 1975)—which gives opposition parties little electoral incentive
to engage in constant Westminster-type adversarial politics—along with a tradition of
cooperative federalism, have favoured consensual policy-making. In this process, the
minority often has incentives to influence decision-making in the private atmosphere
of parliamentary committees, joint intergovernmental consultative bodies of the fed-
eration and the Länder, and the conference committee of the Bundestag and Bundesrat
rather than to challenge the government in public. This has repeatedly led to criticisms
of lacking transparency. Nevertheless, government–opposition relations have generally
become more adversarial since the early 1980s. This is not to say that there is no con-
flict, but that the major German opposition parties have a much wider choice of strate-
gic options than, say, their counterparts in Great Britain whose only real weapon is
publicity (Saalfeld 1998a: 62–7).
The Members of the Bundestag have a number of monitoring devices at their disposal,
which are designed to ensure ongoing public scrutiny. One distinctive characteristic of
the Bundestag’s rules of procedure is that most of these instruments can be employed
by any parliamentary party or any group of Members whose number is equivalent to
the size of a parliamentary party (5 per cent of the total Bundestag membership). The
main implication of these provisions is that the majority has less power to control the
Bundestag’s plenary agenda than is the case in more majoritarian systems of government
(Döring 1995: 225). The down side of this strong protection of minority rights through
privileging the parliamentary parties in parliamentary agenda-setting is that, if they
want to speak or ask questions, individual members de facto need the backing of their
parliamentary parties (Schüttemeyer 1994: 36–7).
Since 1983, the opposition parties have made more use of parliamentary questions
as a means of monitoring the government and forcing public debates than previously.
This increased opposition activity is largely a result of the advent of the Greens in the
Bundestag and the ensuing electoral competition not only between government and
opposition but also between SPD and the Greens as (then competing) opposition par-
ties (Saalfeld 1998a: 63–4). The Bundestag’s rules of procedure allow two main kinds of
written parliamentary questions:
1. The Große Anfrage is one of the most powerful instruments available to the
opposition parties to extract information from the government and—based on this
information—force floor debates on issues of their own choice. Only a parliamentary
party or an equivalent number of Members can request them. The government is
expected to answer such questions on current issues in writing within three weeks. If
362 Thomas Saalfeld
the government refuses, a parliamentary party or an equivalent number of Members
can still demand a full plenary debate on the issue. In the vast majority of cases the gov-
ernment responds, and the ensuing parliamentary debate is based on the government’s
written answer, which is published before the debate takes place (Schäfer 1982: 232–4).
2. Kleine Anfragen are—often quite comprehensive and detailed—written questions,
which are answered by the government in writing. They are an important means for
Members of the Bundestag to collect information, although they do not usually trigger a
major plenary debate. Only parliamentary parties or an equivalent number of members
can ask such questions. The government is expected to answer within a fortnight,
though it also has the option of refusing to answer or giving a confidential reply.
In each regular week, Members of the Bundestag will be collectively allocated up to 180
minutes for oral questions. Members of the Bundestag may ask one short question for oral
answer per day and up to two questions per week. Although such questions may be tabled
by individual members, they are expected to be cleared with the relevant party working
groups, executive committees, and whips’ offices (Arndt 1989: 670–1). The rules of pro-
cedure also allow individual members to ask ‘urgent questions’. Such questions must be
submitted to the Bundestag President’s office by noon on the day before the question is
to be asked. The President decides whether the question at stake qualifies as one of
‘urgent public interest’. Questions for oral answer do not attract the same publicity as in
the British House of Commons. Aktuelle Stunden are short, topical one-hour debates
which can be demanded by any parliamentary party or an equivalent number of
Members. Since 1983, the use of such short debates has grown considerably, facilitated by
procedural reforms and a generally more competitive opposition (Saalfeld 1998a: 63). In
1988 the Bundestag introduced so-called ‘cabinet questions’, initially on an experimental
and in 1990 on a permanent basis. Under this procedure, Bundestag Members are allowed
to ask specific ministers questions of ‘current interest’ as well as questions about the pre-
ceding cabinet meeting (Schindler 1994: 984–6). Regular formal reports from the federal
government to the Bundestag have been of growing importance to parliamentary
scrutiny and legislation. There are three main types of reports. First, the government has
a statutory duty to issue certain reports at regular intervals.12 Second, a parliamentary res-
olution can also require the government to present a report. Finally, the government
issues reports on its own initiative. The use of such formal reports has increased over time
(1949–53: 9, 1987–90: 172, total 1949–90: 1,286) (Schindler 1994: 490–1).
Despite its well-developed floor array of interpellation instruments and information
rights, the Bundestag relies heavily on executive scrutiny by committee. The function of
detailed scrutiny is thus delegated to the parliamentary parties’ policy experts serving
on the relevant committees and playing a leading role in the parliamentary parties’
parallel working groups. There are at least four different kinds of parliamentary
committees: departmental standing committees, special ad hoc committees, investigative
12
For example, this is true for the half-yearly Report on Integration in the European Communities
(Bericht über die Integration in den Europäischen Gemeinschaften), the Annual Report on the State of the
Economy ( Jahreswirtschaftsbericht), or the Annual Energy Report (Energiebericht), which are all prepared
by the Ministry of Economic Affairs.
Germany 363
committees and enquiry commissions. A residual category comprises bodies like the
Committee on Election Validation, Immunity and Rules, the Petitions Committee, and
the Senior Council. The latter, de facto, is the business (or steering) committee of the
Bundestag, which fixes its agenda and timetable, and generally ensures an efficient use
of its time, particularly, by avoiding plenary debates on rules.
Until 1969, committees were primarily agents of the chamber and could only make
recommendations to the Bundestag on matters expressly referred to them. Since 1969
the Bundestag rules of procedure have also allowed them to examine any other subject
in their remit, irrespective of an express referral. Although they do not have independ-
ent legislative powers, they have, in practice, considerable influence because their
members are the parliamentary parties’ experts in the respective policy areas. These
experts wield considerable policy influence within their parties, and when they are part
of the government majority, they cannot easily be ignored by the government. The
committees’ powers to obtain independent outside information were strengthened
considerably in 1969 when a minority of 25 per cent of the committee members were
given the authority to force public hearings (Ismayr 1999: 24).
The committee seats are allocated to the parliamentary parties in proportion to the
seat distribution in the chamber. According to paragraph 57 of the Bundestag’s rules of
procedure, the parliamentary parties nominate committee members. Hence committee
members are formally agents of their parties.13 Once party quotas are established, the
actual appointments of committee members, as well as their recall and replacement,
are left to the parliamentary parties and merely communicated to the President of
the Bundestag. Likewise, committee chairs are proportionally distributed among the
parties, and the parties fill their chairs as they wish, although they usually attempt to
reach interparty consensus on nominations. The assignment of specific committee
chairs is negotiated in the Senior Council.
The Bundestag’s principal instruments of executive oversight are its standing
departmental committees. The majority of these twenty-one committees (2003)
mirror a single government department. Only the remits of the Committees on
Appropriations, Justice, and EU Affairs cut across departmental jurisdictions. Under the
Basic Law, all committees are entitled to compel the attendance of ministers, as is the
Bundestag floor majority. Most committees regularly question ministers and civil
servants on issues within their jurisdiction. Departmental standing committees combine
a ‘watchdog function’ of ongoing scrutiny with detailed discussion of draft bills after
their first reading. Most German MPs’ parliamentary timetables are dominated by
work in standing committees and parallel working groups within their parliamentary
parties. The overlap that exists between the parliamentary parties’ working groups and
the Bundestag’s standing committees is designed to ensure that the parliamentary
parties’ agents in Bundestag committees share information with their co-partisans
(Saalfeld 2000a: 28–9).
13
Individual Members not belonging to a party are to be nominated for a committee seat by the
Bundestag President. However, they do not have the right to vote in committee.
364 Thomas Saalfeld
The characteristics and powers of the other Bundestag committees and bodies cannot
be described in detail (cf. Saalfeld 1998b: 268, 273–4, 276–7). The Bundestag’s investigative
committees can be used to inquire into instances of alleged administrative mismanagement
or government wrongdoing. Such committees must be created if at least a quarter of
the Bundestag Members request it. Commissions of Enquiry are appointed ad hoc and are
charged with the task of informing the Bundestag and the public thoroughly on policy
issues of a fundamental or long-term importance. In many ways, they resemble the
British Royal Commissions. Whereas all other committees of the House consist only
of Bundestag MPs, Commissions of Enquiry have the power to appoint outside experts
as full committee members with voting rights. The Petitions Committee can be directly
approached by citizens who have complaints about instances of administrative wrong-
doing, but it has no power to investigate matters independent of a specific complaint.
In this sense it is exclusively a ‘fire alarm’ (McCubbins and Schwartz 1984), which can
be activated by citizens in the event of implementation problems. A number of special
bodies assist the Bundestag in its ongoing scrutiny of government activities. The
Parliamentary Commissioner for the Armed Forces has been influenced by the model
of the Scandinavian Ombudsman. The Federal Audit Court, among other activities,
audits government spending supporting, in particular the work of the Appropriations
and Finance committees. The Appropriations Committee has a special Auditing
sub-committee.
D E L E GAT I O N W I T H I N T H E C A B I N E T
Delegation within the cabinet is structured by somewhat ambivalent and vague norms
in the Basic Law as well as by the political realities of party and coalition government.
This has allowed a number of interpretations ranging from a position emphasizing the
Federal Chancellor’s dominance (‘Chancellor democracy’) at one extreme to, at the
other, a position emphasizing the informational advantages of bureaucrats vis-à-vis
elected politicians and arguing that the capacity for policy development de facto rests
with the ministerial bureaucracies.
At first glance, the Federal Chancellor appears to be a very powerful principal. The
Basic Law grants him considerable ex ante controls by empowering him to determine
the general policy of the government. Although he is constrained by the Constitution,
the federal government’s rules of procedure, and the coalition agreement, he does have
certain agenda powers, such as the extent to which decisions are delegated to ministers
(important issues, including many foreign policy issues, are often unilaterally declared
to be ‘matters for the leader’), the government’s organization, the procedures that have
to be followed by the ministers, and the mechanisms for interdepartmental conflict
resolution. He bears overall responsibility for the government’s policies vis-à-vis the
Bundestag. Although cabinet ministers are individually answerable to the Bundestag in
parliamentary questions, Müller-Rommel (1994: 160)—putting it somewhat strongly—
emphasizes that ministers primarily depend on the Chancellor’s confidence: ‘ministers
are appointed and dismissed by the chancellor. They serve at his pleasure and not that
of the Parliament’.
Germany 365
According to the rules of procedure of the federal government, the Chancellor has
the right to determine the number and jurisdictions of individual departments, an
important discretionary authority. The Federal Chancellor’s Office, with a staff of more
than 500 civil servants (1994) and headed by a cabinet minister close to the Chancellor,
provides him with a considerable monitoring capacity vis-à-vis individual departments.
The Office has a system of specialized sections each liasing with, and monitoring the
activities of, a number of government departments and trying to minimize agency loss
between Chancellor and departments. According to the federal government’s standing
orders, individual ministries have the statutory duty to keep the Chancellor informed
of important policy developments in their departments. Not only does the Chancellor’s
Office attempt to coordinate the work of the departments, it also serves as the
government’s secretariat by preparing cabinet meetings and coordinates the activities
of the three German intelligence services (Busse 1997: 110–20; Rudzio 2000: 284–9).
Nevertheless, it would be inaccurate to overstate the constitutional and practical
importance of the ‘Chancellor principle’. First the constitutional norms are ambivalent
and give some weight to the ‘departmental principle’ and the ‘cabinet principle’ (see
below). Second the risk of agency loss in the form of ‘departmentalism’ is high given
the superior information available to the departmental bureaucracies. Finally, the
realities of coalition government in Germany render simple models of delegation
unrealistic. Coalition government is a constant bargaining process. Bargaining power
depends on factors such as the walk-away value of the parties rather than legal norms,
voting strength in Parliament, or the Chancellor’s formal constitutional position.
Although the Basic Law establishes a ‘Chancellor principle’, it simultaneously gives
cabinet ministers individual responsibility for all matters within their departments’
jurisdictions (‘departmental principle’). The Chancellor does not have the right to
circumvent a minister and intervene directly in a ministry, although most Chancellors
in practice have developed a strong interest in foreign and security matters and interfered
with the jurisdiction of the foreign ministry (for the Kohl era, see Korte 2000: 7). At the
domestic level, however, the experts in the departmental bureaucracies usually enjoy a
tremendous informational advantage vis-à-vis elected politicians. Laver and Shepsle
(1996: 13) therefore maintain that (generally in Western Europe) ‘only the government
department with jurisdiction over a particular policy area is effectively equipped to
develop feasible and implementable policy proposals in that area and present these to
the cabinet for decision’, strengthening the position of individual cabinet ministers vis-
à-vis cabinet and head of government. In this view, ministers would de facto almost be
the agents of ‘their’ bureaucracies. With specific reference to the (West) German
system, Mayntz and Scharpf (1975: 48) assert in a similar vein: ‘The political system’s
capacity for active policy-making is largely a capacity of its ministerial bureaucracy’.
Principals will have some control over agents if they have ex ante powers to choose
them freely from a pool of suitable competitors, if they can monitor their behaviour,
and if they can reward or punish them ex post facto. While the monitoring capacity of
the Chancellor’s office is highly developed, the Chancellor’s powers of appointment
and dismissal are limited both within his own party and, even more so, vis-à-vis other
coalition parties. Within his own party, the Chancellor has to ensure the representation
366 Thomas Saalfeld
of important ideological factions, socio-economic interests, regions, and the main reli-
gious denominations. His influence on the selection of ministerial personnel in his
coalition partner’s party is even more limited. Ex post facto, it can be extremely difficult
for a Chancellor to ‘discipline’ ministers belonging to a different coalition party if they
enjoy the support of their own party’s leadership. It may even be costly for the
Chancellor to punish defiant ministers of his own party, if they enjoy strong intraparty
support.
In addition to the ‘Chancellor’ and ‘departmental principles’ the Basic Law stipulates
a ‘cabinet principle’. Certain powers, such as the right to initiate legislation and to
appeal directly to the Federal Constitutional Court, are exercised collectively by the
federal government. Article 65 of the Basic Law states that differences between
members of the federal government are to be resolved by the cabinet as a whole. Thus,
the Federal Chancellor has no constitutional right to act as an arbiter. Also, the cabinet
must be collectively informed about the appointment of high-ranking civil servants and
political appointees to the civil service. Nevertheless, the cabinet as a collective body
has tended to be weak. It usually meets only once a week. Its agenda is prepared by the
Chancellor’s Office. Most conflicts over policy are resolved in interministerial negotiations
before an issue is allowed on the agenda by the Chancellor’s Office. The formal cabinet
decision rule is majority voting, yet such votes are highly exceptional. Differences of
opinion are usually resolved at earlier stages, for example in coalition committees or
interministerial committees of civil servants. Formal votes are therefore often an
indicator of a breakdown of the usual coordination mechanisms and an imminent
coalition crisis (Rudzio 2000: 290–3).
There are a number of organizational devices designed to alleviate the impact of
departmentalism. Direct interministerial negotiations at the bureaucratic level (includ-
ing representatives from the Chancellor’s Office) are one important device, especially
for routine matters. Cabinet committees in the British sense do exist, but are not con-
sidered to be important coordination mechanisms (Müller-Rommel 1994: 155). The
most important steering bodies are more informal coalition talks and bodies. The
reliance on such informal bodies has increased in the past two decades, especially
during the Kohl era (1982–98). They usually comprise the Federal Chancellor or his
representatives, some cabinet ministers and representatives of the extra-parliamentary
party leadership, as well as the leadership and policy experts of the parliamentary
parties. The overall direction of policy, important policy decisions, and questions of
coalition governance are dealt with in these informal bodies. The cabinet and the
parliamentary majority parties are often left with the task of implementing the decisions
made in these bodies. Routine matters tend to be dealt with in cabinet; important
policy changes are virtually always prepared in coalition talks (Korte 2002: 88–90;
Schreckenberger 1994: 334).
These bodies have been effective in reducing the scope for agency loss within the
government and the transaction costs of coordination between departments, government
parties inside and outside Parliament, the Länder, and other actors with veto powers in
the German political system. Despite a large number of such actors, this system of
informal co-ordination has been capable of responding quickly and decisively to important
Germany 367
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
As in all political systems, the information asymmetry between elected politicians and
civil servants with greater expertise and resources to generate and process information
is highly problematic. Generally, there are a number of ways for ministers to enhance
bureaucratic accountability and responsiveness in their respective departments: control
of top departmental appointments, ability to use ‘rewards and punishments’ both in
promotions and policy decisions, ability to appoint ministerial staff to top-level department
positions, and providing clear direction (Dunn 1999). For a number of reasons most
of these mechanisms are problematic in the German case.
1. The Federal Ministers often have little direct control over policy implementation.
Much of domestic federal legislation is implemented by Länder governments or para-
public bodies with considerable autonomy. Unlike the United States government, the
German federal government, in most policy areas, cannot rely on its own administrative
infrastructure at regional or local levels. The federal bureaucracy accounts for less than
one-tenth of all civil servants (for data see Rudzio 2000: 294). Since state civil servants
are primarily accountable to their state governments that are, in turn, accountable to
regional parliaments, the usual agency problems affecting unitary states are com-
pounded by multiple-principal problems if and when there is a conflict of interest
between the federation and the Länder. Even many agencies that are under direct fed-
eral jurisdiction enjoy a considerable degree of autonomy. Several parapolitical bodies
such as the Federal Bank or the Federal Employment Office, as well as a large number
of corporate bodies, foundations, and institutes, are organized under public law and
carry out important policy functions, often in cooperation with interest groups. Some
of them play a crucial role in the economic management and in the social welfare sec-
tor and ‘wield substantial power virtually unchecked by the federal government’
(Katzenstein 1987: 58, see below).
2. Despite an increasing tendency for civil servants to be party members (1970:
28 per cent, 1987: 57 per cent), ministers’ control over staffing in their departments is
limited. In 1967 the introduction of ‘parliamentary secretaries of state’ (modelled
368 Thomas Saalfeld
largely on British junior ministers) did strengthen the political control of departments
and liaison between Parliament and the top level of Federal Ministries. In addition,
there have always been political appointees to the civil service: ministers may require
permanent secretaries and heads of divisions to take ‘temporary retirement’ replacing
them with qualified persons of their own choosing. Compared to the United States or
France, however, the number of political appointees is small (1994: 161). In addition,
there are relatively few negative sanctions available to ministers. Career civil servants or
long-serving public employees cannot easily be demoted or dismissed. Ministers do
have some opportunities to use positive incentives, mainly promotions. However, civil
service law and the powerful staff councils in the ministries limit even their
powers of promotion (Rudzio 2000: 308).
3. This lack of federal control over domestic policy implementation is compounded
by two organizational features of the federal bureaucracy itself: The federal bureau-
cracy ‘is a steep hierarchy in which the apex has little control over lower levels’
(Katzenstein 1987: 19). Departments have a four-tiered organizational structure. The
minister, occasionally (depending on the size and importance of the department) aided
by one or two junior ministers, and one or two permanent secretaries are at the top of
the pyramid. The second and third levels are the divisions and subdivisions. Sections
typically have 3–5 civil servants and represent the fourth level, which is crucial in the
process of detailed policy development (Katzenstein 1987: 20; Rudzio 2000: 305).
Given the low level of control the federal government can exercise over policy
implementation, it has to rely strongly on ex ante control through the detailed
drafting of bills after early, comprehensive consultations with all veto players. This
approach involves two main sorts of costs: (a) Relatively detailed laws, leaving relatively
little discretion to local bureaucracies in the implementation process, may lack the flex-
ibility that is necessary to adjust a piece of legislation to different local conditions. (b)
The attempt to reach ex ante agreements with the Länder and interest groups involved
in the running of a number of semi-autonomous para-political bodies (such as the
Federal Employment Office or the various social insurance funds) may carry high
transaction costs for the government if and when its preferences differ from those of
important veto players. Such bargaining also leads to a lack of decision-making
transparency.
E X T E R NA L C O N S T R A I N T S
Germany fits the description of a ‘constitutional democracy’. The Basic Law is a
constitution which can be enforced by courts and which specifies particular majority
requirements for constitutional changes. It requires institutions and incumbents to act
on behalf of the constitution and certain interests, even if these diverge from the
preferences of the voters: For example, the first twenty Articles as well as Articles 101,
103, and 104 of the Basic Law guarantee fundamental and inalienable rights and liberties
which as directly enforceable law bind the legislature, the executive, and the judiciary.
Their ‘essential content’ may not be encroached upon, by either a majority of the
voters or a majority of elected Members of Parliament.
Germany 369
14
For a more detailed account see Katzenstein (1987: 58–80).
370 Thomas Saalfeld
‘enjoys a degree of autonomy from the federal government and private interest groups
which is more far-reaching than comparable institutional arrangements in the United
States, Britain, France, Japan, and Sweden’ (Katzenstein 1987: 60). The Bundesbank
must support the government’s overall economic policy, but only if that policy does not
interfere with the bank’s overriding statutory duty to safeguard the currency. With the
establishment of European Monetary Union most of the Federal Bank’s powers to set
interest rates have in practice been transferred to the European Central Bank, which
adds an additional constraint on economic and fiscal policy for all member states of the
European Monetary Union. Critics of independent central banks often focus on their
character as agents, whose relationship with democratically elected principals such as
MPs or government ministers is (by design) extremely loose, casting doubt on the demo-
cratic legitimacy of their often far-reaching decisions. Yet independent central banks
may also be important mechanisms to redress information asymmetries in economic
policy-making by reducing the government’s ability to hide unfavourable information
and providing the public with key information on economic policy (Bernhard 1998: 314).
Finally, membership of the EU is a very important constraint on at all levels of
Germany parliamentary democracy. EU policy-making involves a great deal of inter-
governmental bargaining which may entail some agency loss in the delegation from
national Parliaments to governments (Scharpf 1988; Lupia 2000; Saalfeld 2000d).
Generally the information asymmetry between the federal government and the
Bundestag is even greater in EU matters than in purely domestic affairs, making it dif-
ficult for the Bundestag to hold the federal government accountable for its role in shaping
particular EU decisions (Bulmer 1986: 222, 243). In the German case, the EU-related
agency problems are compounded by the existence of simultaneous multiple-agent and
multiple-principal arrangements. Interministerial coordination within the federal
government has traditionally been weak in EU matters, although the Schröder govern-
ment (elected in 1998) has strengthened the coordination role of the Ministry of
Finance. Yet, in practice neither the Chancellor’s Office nor any other ministry exercises
effective and detailed oversight over the negotiations between departmental experts in
the EU’s Committee of Permanent Representatives (COREPER) or other bodies
responsible for the preparation of EU legislation. It is not unheard of for different min-
istries to pursue conflicting policies.15 This trend has become more pronounced as EU
policy has increasingly become domestic in character. In their areas of exclusive juris-
diction, the German Länder are represented in the Council of Ministers and other bod-
ies by their own delegates, which adds to the multiple-agent problems. Not only may
the lack of interministerial coordination within the German government and between
federation and the Länder weaken the German bargaining position in the EU; it also
undermines government accountability vis-à-vis the Bundestag, as policy responsibility
is hard to establish. To some extent the information problems have been addressed
in a 1992 constitutional amendment which codifies the Bundestag’s and Bundesrat’s
15
For example the Ministry for Environmental Protection and the Ministry of Economic Affairs in the
field of environmental policy (see Wurzel 2000).
Germany 371
information rights in the area of the EU, and through the establishment of a new EU
Affairs Committee of the Bundestag in 1994. Nevertheless, parliamentary monitoring of
government policy in Brussels remains highly problematic (Saalfeld 1995). One of the tra-
ditional problems of parliamentary scrutiny of EU legislation, the departmental standing
committees’ strenuous defence of their traditional jurisdictions, has continued to hamper
the development of the EU Affairs Committee as an effective instrument of parliamen-
tary accountability. In practice the lack of coordination within the federal government,
leading de facto to multiple agents of Parliament with conflicting objectives, is mirrored
by competition between different scrutiny committees at the parliamentary level.
C O N C LU S I O N
The Basic Law was designed in 1948/49 to avoid the political divisions and instability
experienced under the Weimar Republic (1919–33) and to prevent the abuse of power
experienced under National Socialism (1933–45). This led to a constitutional design that
is contradictory from the perspective of principal–agent theory. On the one hand, the
Basic Law has strengthened the representative system of government, establishing a
unified chain of delegation and accountability. Compared to the Weimar Constitution,
direct democracy was reduced to a minimum, the powers of the head of state were
greatly reduced, and the powers of the Federal Chancellor were strengthened, as was
his direct accountability to the Bundestag. Yet, the Basic Law simultaneously created a
host of powerful institutional checks on the government, which, de facto, dilute the
unitary chain of delegation and accountability and lead to complex, multi-layered
agency relationships. The Land governments, the Federal Constitutional Court, the
Federal Bank, and para-political agencies make, influence, execute, delay, and veto
policy. Decisions are frequently compromises between the federal government and
some of these various actors. If no compromise is possible, the status quo is preserved
increasing the risk of a reduction of the system’s problem-solving capacity.
For much of the Federal Republic’s history, the incentives to seek consensus and
compromise have been strong and beneficial to the country’s political stability.
Nevertheless, the resulting bargaining processes often take place outside the formal
constitutional chain of delegation and accountability. Their multilevel nature and com-
plexity give the federal government additional discretion and opportunities to avoid
political responsibility. In the terminology of principal–agent theory, they have created
the potential for hidden action and ‘moral hazard’. These dangers are compounded by
the relatively infrequent alternation in office throughout the postwar period. The
revelations about former Chancellor Kohl’s party finance practices (1999/2000)
triggered the ‘biggest scandal in postwar German political history’ (Clemens 2000: 25),
and can be seen as an—admittedly extraordinary—illustration of this danger.
Given the organizational strength of the German parties, there has not been a rapid
decline of ex ante screening processes similar to the developments in other Western
countries (e.g. see Strøm 2000: 283). Nevertheless the consequences of these party-
based selection processes are in dispute as far as they favour professional career
politicians without a great deal of experience outside politics.
372 Thomas Saalfeld
Although the Bundestag has at its disposal a relatively highly developed arsenal of
ex post monitoring devices and all the usual dismissal powers, the federal government’s
negotiations in coalition committees at the federal level, in joint bodies with the Länder,
and, in particular, at in the EU are difficult to scrutinize. To be sure, parliamentarians
of all major parties are involved in these multilevel bargaining processes, but
transparency is limited and political responsibility is often difficult to establish—
especially for the voters. The virtual ‘Grand Coalition’ system (Schmidt 1996) with a
large number of veto points has ensured political stability but also increased decision-
making costs and reduced transparency. At the same time wider access to education
and a revolution in the media sector have contributed to a more educated, informed,
and critical public. The growing gap between popular expectations on the one hand and
consensual, elite-dominated decision-making ‘behind closed doors’ on the other has
arguably contributed to a growing disaffection with representative democracy. As
a result of this growing pressure, the governmental agents in the Federal Republic
(like most of their counterparts elsewhere in Western Europe) have increasingly faced
the problem of ‘diverted accountabilities’ (Strøm 2000: 284) as government ministers have
to serve several (often competing) principals simultaneously: Parliament, their parties,
and the voters.
REFERENCES
Arndt, Claus (1989). ‘Fraktion und Abgeordneter’, in Hans-Peter Schneider and Wolfgang Zeh
(eds.), Parlamentsrecht und Parlamentspraxis in der Bundesrepublik Deutschland. Berlin: Walter de
Gruyter.
Bernhard, William T. (1998). ‘A Political Explanation of Variations in Central Bank
Independence’. American Political Science Review, 92: 311–28.
Bracher, Karl Dietrich, Jäger, Wolfgang, and Link, Werner (1986). Republik im Wandel 1969–1974:
Die Ära Brandt (Geschichte der Bundesrepublik Deutschland, Volume 5/I). Stuttgart/Mannheim:
Deutsche Verlags-Anstalt/Brockhaus.
Bulmer, Simon (1986). The Domestic Structure of European Community Policy-Making in West Germany.
London: Garland.
Busch, Andreas (2000). ‘The Grundgesetz after 50 Years: Analysing Changes in the German
Constitution’. German Politics, 9(1): 41–60.
Busse, Volker (1997). Bundeskanzleramt und Bundesregierung: Aufgaben, Organisation, Arbeitsweise.
Heidelberg: Hüthig.
Clemens, Clay (2000). ‘A Legacy Reassessed: Helmut Kohl and the German Party Finance Affair’.
German Politics, 9(2): 25–50.
Conradt, David P. (2000). The German Polity, 7th edn. New York: Longman.
Dästner, Christian (1995). Die Geschäftsordnung des Vermittlungsausschusses. Berlin: Duncker &
Humblot.
De Winter, Lieven (1991). ‘Parliamentary and Party Pathways to the Cabinet’, in Jean Blondel
and Jean-Louis Thiébault (eds.), The Profession of Government Minister in Western Europe.
Basingstoke: Macmillan.
Döring, Herbert (1995). ‘Time as a Scarce Resource: Government Control of the Agenda’, in
Herbert Döring (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt am Main:
Campus and New York: St. Martin’s Press.
Germany 373
Dunn, Delmer D. (1999). ‘Mixing Elected and Nonelected Officials in Democratic Policy Making:
Fundamentals of Accountability and Responsibility’, in Adam Przeworski, Susan C. Stokes,
and Bernard Manin (eds.), Democracy, Accountability, and Representation. Cambridge:
Cambridge University Press.
Dürig, Günter (ed.) (1994). Grundgesetz mit Vertrag über die abschließende Regelung in bezug auf
Deutschland, Menschenrechtskonvention, Bundesverfassungsgerichtsgesetz, Parteiengesetz und Gesetz
über den Petitionsausschuß, 32nd edn. Munich: Deutscher Taschenbuch Verlag.
Gabriel, Oscar W., Niedermayer, Oskar, and Stöss, Richard (eds.) (1997). Parteiendemokratie in
Deutschland. Bonn: Bundeszentrale für Politische Bildung.
Goetz, Klaus H. (1996). ‘The Federal Constitutional Court’, in Gordon Smith, William E.
Paterson, and Stephen Padgett (eds.), Developments in German Politics 2. Basingstoke: Macmillan.
Grewe, Wilhelm (1951). ‘Parteienstaat—oder was sonst?’ Der Monat, 3(36): 563–77.
Grosser, Dieter (1975). ‘Die Sehnsucht nach Harmonie: Historische und verfassungsstrukturelle
Vorbelastungen der Opposition in Deutschland’, in Heinrich Oberreuter (ed.), Parlamentarische
Opposition: Ein Internationaler Vergleich. Hamburg: Hoffmann und Campe.
Gunlicks, Arthur B. (1988). ‘Campaign and Party Finance in the West German “Party State” ’.
Review of Politics, 50: 30–48.
Hasselsweiler, Ekkehart (1981). Der Vermittlungsausschuß. Berlin: Duncker & Humblot.
Ismayr, Wolfgang (1992). Der Deutsche Bundestag: Funktionen, Willensbildung, Reformansätze.
Opladen: Leske ⫹ Budrich.
—— (1999). ‘50 Jahre Parlamentarismus in der Bundesrepublik Deutschland’. Aus Politik und
Zeitgeschichte, B20: 14–26.
Jesse, Eckhard (1990). Elections: The Federal Republic of Germany in Comparison. Oxford: Berg.
Kaack, Heino (1971). Geschichte und Struktur des deutschen Parteiensystems. Opladen:
Westdeutscher Verlag.
Katzenstein, Peter J. (1987). Policy and Politics in West Germany: The Growth of a Semisovereign State.
Philadelphia: Temple University Press.
King, Anthony (1976). ‘Modes of Executive–Legislative Relations: Great Britain, France and
West Germany’. Legislative Studies Quarterly, 1: 11–36.
König, Thomas (1998). ‘Regierungswechsel ohne politischen Wandel? Ein Vergleich des
wirtschaftspolitischen Handlungsspielraums der Regierung Kohl, einer Regierung Schröder,
einer Großen Koalition und einer SPD-Alleinregierung.’ Zeitschrift für Parlamentsfragen, 29:
478–95.
—— and Thomas Bräuninger (1997). ‘Wie wichtig sind die Länder für die Politik der
Bundesregierung bei Einspruchs- und Zustimmungsgesetzen?’ Zeitschrift für Parlamentsfragen,
28: 605–27.
Korte, Karl-Rudolf (2000). ‘Solutions for the Decision Dilemma: Political Styles of Germany’s
Chancellors’. German Politics, 99(1): 1–22.
—— (2002). ‘The Effects of German Unification on the Federal Chancellor’s Decision-Making.’
German Politics, 11(3): 83–98.
Krehbiel, Keith (1991). Information and Legislative Organization. Ann Arbor: University of
Michigan Press.
Laver, Michael and Kenneth A. Shepsle (1996). Making and Breaking Governments: Cabinets and
Legislatures in Parliamentary Democracies. Cambridge: Cambridge University Press.
Leibholz, Gerhard (1966). Das Wesen der Repräsentation und der Gestaltwandel der Demokratie im 20.
Jahrhundert, 3rd edn. Berlin: Walter de Gruyter.
Lösche, Peter (1993). Kleine Geschichte der deutschen Parteien. Stuttgart: Kohlhammer.
374 Thomas Saalfeld
Lupia, Arthur (2000). ‘The EU, the EEA and Domestic Accountability: How Outside Forces
Affect Delegation within Member States’. Journal of Legislative Studies, 6(1): 15–32.
Mayntz, Renate and Scharpf, Fritz W. (1975). Policy-Making in the German Federal Bureaucracy.
Amsterdam: Elsevier.
McCubbins, Mathew D. and Schwartz, Thomas (1984). ‘Congressional Oversight Overlooked:
Police Patrols versus Fire Alarms’. American Journal of Political Science, 28: 165–79.
Mintzel, Alf and Heinrich Oberreuter (eds.) (1992). Parteien in der Bundesrepublik Deutschland,
2nd edn. Bonn: Bundeszentrale für Politische Bildung.
Müller-Rommel, Ferdinand (1994). ‘The Role of German Ministers in Cabinet Decision Making’,
in Michael Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government.
Cambridge: Cambridge University Press.
Naßmacher, Karl-Heinz (1997). ‘Parteienfinanzierung in Deutschland’, in Oscar W. Gabriel,
Oskar Niedermayer, and Richard Stöss (eds.), Parteiendemokratie in Deutschland. Opladen:
Westdeutscher Verlag.
Niclauß, Karlheinz (1999). ‘Bestätigung der Kanzlerdemokratie? Kanzler und Regierungen
zwischen Verfassung und politischen Konventionen’. Aus Politik und Zeitgeschichte, B20: 27–38.
Nohlen, Dieter (1986). Wahlrecht und Parteiensystem. Opladen: Leske ⫹ Budrich.
Paterson, William E. (2000). ‘From the Bonn to the Berlin Republic’. German Politics, 9(1): 23–40.
Patzelt, Werner J. (1995). ‘German MPs and Their Roles’, in Wolfgang C. Müller and
Thomas Saalfeld (eds.), Members of Parliament in Western Europe: Roles and Behaviour. London:
Frank Cass.
Poguntke, Thomas (1994). ‘Parties in a Legalistic Culture: The Case of Germany’, in Richard
S. Katz and Peter Mair (eds.), How Parties Organize. London: Sage.
Pulzer, Peter (1995). German Politics 1945–1995. Oxford: Oxford University Press.
Rausch, Heinz (1981). Parlament und Regierung in der Bundesrepublik Deutschland, 6th edn.
München: Landeszentrale für Politische Bildungsarbeit.
Roberts, Geoffrey (1997). Party Politics in Germany. London: Cassell.
Rudzio, Wolfgang (1991). ‘Informelle Entscheidungsmuster in Bonner Koalitionsregierungen’, in
Hans-Hermann Hartwich and Göttrik Wewer (eds.), Regieren in der Bundesrepublik II: Formale
und informale Komponenten des Regierens in den Bereichen Führung, Entscheidung, Personal und
Organisation. Opladen: Leske und Budrich.
—— (2000). Das politische System der Bundesrepublik Deutschland, 5th edn. Opladen:
Leske⫹Budrich.
Saalfeld, Thomas (1995). ‘The German Houses of Parliament and European Legislation’. Journal
of Legislative Studies, 1(1): 12–34.
—— (1998a). ‘The German Bundestag: Influence and Accountability in a Complex
Environment’, in Philip Norton (ed.), Parliaments and Governments in Western Europe. London:
Frank Cass.
—— (1998b). ‘Germany’, in George Thomas Kurian and Lawrence D. Longley (eds.), World
Encyclopedia of Parliaments and Legislatures. Vol. 1. Washington, DC: Congressional Quarterly
Press.
—— (2000a). ‘Bureaucratisation, Coordination and Competition: Parliamentary Party Groups in
the German Bundestag’, in Knut Heidar and Ruud Koole (eds.), Parliamentary Party Groups in
European Democracies. London: Routledge.
—— (2000b). ‘Coalitions in Germany: Stable Parties, Chancellor Democracy and the Art of
Informal Settlement’, in Wolfgang C. Müller and Kaare Strøm ( eds.), Coalition Governments in
Western Europe. Oxford: Oxford University Press.
Germany 375
—— (2000c). ‘Court and Parties: Evolution and Problems of Political Funding in Germany’, in
Robert Williams (ed.), Party Finance and Political Corruption. Basingstoke: Macmillan.
—— (2000d). ‘Members of Parliament and Governments in Western Europe: Agency Relations
and Problems of Oversight’. European Journal of Political Research, 37: 353–76.
Schäfer, Friedrich (1982). Der Bundestag: Eine Darstellung seiner Aufgaben und seiner Arbeitsweise,
4th edn. Opladen: Westdeutscher Verlag.
Scharpf, Fritz W. (1988). ‘The Joint-Decision Trap: Lessons from German Federalism and
European Integration’. Public Administration, 66: 239–78.
Scheuch, Erwin K. and Scheuch, Ute (1993). Cliquen, Klüngel and Karrieren: Über den Verfall der poli-
tischen Parteien – eine Studie. Reinbek bei Hamburg: Rowohlt.
Schindler, Peter (1994). Datenhandbuch zur Geschichte des Deutschen Bundestages 1983 bis 1991.
Baden-Baden: Nomos.
Schmidt, Manfred G. (1996). ‘Germany: The Grand Coalition State’, in Josep M. Colomer (ed.),
Political Institutions in Europe. London: Routledge.
Schreckenberger, Waldemar (1994). ‘Informelle Verfahren der Entscheidungsvorbereitung
zwischen der Bundesregierung und den Mehrheitsfraktionen: Koalitionsgespräche und
Koalitionsrunden’. Zeitschrift für Parlamentsfragen, 25: 329–46.
Schüttemeyer, Suzanne S. (1994). ‘Hierarchy and Efficiency in the Bundestag: The German
Answer for Institutionalizing Parliament’, in Gary W. Copeland and Samuel C. Patterson
(eds.), Parliaments in the Modern World: Changing Institutions. Ann Arbor, University of Michigan
Press.
Stöss, Richard (ed.) (1986). Parteienhandbuch: Die Parteien der Bundesrepublik Deutschland 1945–1980.
Opladen: Westdeutscher Verlag.
Strøm, Kaare (2000). ‘Delegation and Accountability in Parliamentary Democracies’. European
Journal of Political Research, 37: 261–89.
von Alemann, Ulrich (2001). Das Parteiensystem der Bundesrepublik Deutschland. Second edition.
Opladen: Leske ⫹ Budrich.
von Beyme, Klaus (1998). The Legislator: German Parliament as a Centre of Political Decision-Making.
Aldershot: Ashgate.
Wiesendahl, Elmar (2000). ‘Changing Party Organisations in Germany: How to Deal with
Uncertainty and Organised Anarchy’. German Politics, 8(2): 108–25.
Wurzel, Rüdiger (2000). ‘Flying into Unexpected Turbulence: The German EU Presidency in the
Environmental Field’. German Politics, 9(3): 23–42.
Zohlnhöfer, Reimut (1999). ‘Die große Steuerreform 1998/99: Ein Lehrstück für
Politikentwicklung bei Parteienwettbewerb im Bundesstaat.’ Zeitschrift für Parlamentsfragen,
30: 326–45.
11
I N T RO D U C T I O N
The 1975 Constitution gave the Greek President important powers such as the right to
dissolve the Parliament and to call a national referendum without the countersignature
of the cabinet. In the 1986 Constitutional revision, most of these rights were distributed
among the Prime Minister (PM), the cabinet, and the Parliament. The amendments
placed most powers in the hands of the PM. This concentration of power in one
elected representative must be understood against the backdrop of Greek historical
experience. The Second World War was followed by a bitter civil war (1944, 1945–9).
The subsequent 1952 Constitution gave the King important powers in government
formation. In 1955 King Paul pre-empted a decision by the ruling Conservative Party
by naming Constantine Karamanlis, then a senior minister, PM after the death of PM
Marshal Papagos. In 1963, without calling for new elections, the King forced
Karamanlis to resign. In 1965, a new political crisis broke out when King Constantine II
insisted on having a say in the naming of government ministers. At this time,
PM Papandreou resigned, even though his party held an absolute majority in Parliament.
The trauma that these power struggles inflicted helped bring about the 1967 military
coup d’état (Legg 1969, 1973; Katsoudas 1987).
The Colonel’s regime declared a Republic in 1973 but fell when the Turkish invasion
of Cyprus made its position untenable (1974). At this point, Karamanlis was called in
from his Paris exile and formed the so-called ‘Government of National Unity’, which
included members from the old conservative and centre parties, personalities from the
struggle against the dictatorship, and technocrats. This Government managed in a
short period to lay the bases for a liberal political system and to hold free elections
(Diamantouros 1984, 1986, 1991; Kaminis 1993). The newly elected Parliament’s main
task was to draft a new Constitution.2
1
Professor Trantas provided a chapter draft on this topic in 1998. The other authors have then revised,
updated, and completed the chapter.
2
The form of Government was settled by a referendum that decided in favour of a Republic.
Greece 377
3
If not otherwise noted, the presentation draws primarily on Manitakis (1994), Spyropoulos (1995), and
Pantélis (1979).
4
All courts can examine the constitutionality of laws. According to Article 13, § 4 of the Constitution, ‘The
Courts must not apply laws the content whereof is contrary to the Constitution’. In practice, however, the
judgements of the Supreme Civil and Criminal Court and the Council of State prevail over the decisions of
other courts. It belongs to the jurisdiction of the Special Highest Court to settle disagreement between two
courts over the constitutionality of a law (Article 100, § 1e).
378 Georgios Trantas et al.
Greece has no second chamber, no federal system, and no constitutional court. Thus, the
Greek Government is almost omnipotent, at least in institutional terms.
5
The preparation of cabinet bills was in the past mostly the work of the civil servants of the relevant
ministries. Today the relevant minister’s advisers usually prepare bills. It is an established practice that bills,
before they are sent to Parliament, are examined by the Central Committee on the Preparation of the
Legislation, a committee of experts (usually senior judges and academics) that gives its opinion on the structure
and the wording of the bill, but not on its content. The committee functions as an early warning device
concerning the constitutionality of the bill.
6
There are six standing legislative committees, each with a jurisdiction corresponding to a ministry.
According to the standing orders, a bill may not be discussed in more than three sessions of the committee.
When the deliberations are concluded, the Committee submits a report to the House.
7
Bills may be and usually are sent to the academic service of Parliament for its opinion concerning the
legal technicalities of the bill and also its constitutionality.
Greece 379
latter was common practice up until 1977 but has not been used since then.8 The bill is
first discussed on the floor on principle, then article by article. But if there is an agree-
ment among the parties, as there is with most international agreements, bills are passed
after little or no debate. Furthermore, the cabinet can introduce ‘extremely urgent’ bills
directly on the floor. Voting then takes place within one day after a limited debate. With
the assent of the Chamber, such a bill can be discussed over a period of 3–5 days.
Historically, opposition MPs commonly misused the quorum requirements to halt
government legislation. Hence, the Constitution of 1975 abolished the (formal)
quorum requirement for most parliamentary proceedings. If a vote is called, a bill passes
when a majority of the MPs present in the Chamber vote in favour. However, the
majority in favour of the bill cannot be lower than one-fourth of all MPs (and in the
sections two-fifths of the MPs that are members). These rules have led to a situation
where often only a few MPs participate in the debates. Sometimes only the minister
and one or two MPs from the opposition debate a bill. As the opposition does not
usually bother to object, in practice most bills are passed with only a handful of MPs
present.
A bill passed by the Chamber is then sent to the President of the Republic who signs
it and orders its publication in the Official Journal, and thus it becomes a law. The
cabinet, or the PM and the competent minister, must also sign the bill. Almost all acts
of the President have to be signed by a person who is responsible before the Chamber.
However, on his own, within a month, the President has the right to return the bill to
the Chamber for reconsideration and to state his reasons for this. The bill is then
discussed in plenary meeting. If a majority of all MPs of passes the bill, then the President
is obliged to publish it within 10 days. This is a very weak veto. Furthermore, the fact
that strong parliamentary majorities typically support Greek cabinets makes it highly
improbable that the veto will be sustained. Thus, the veto has never been used. As
Greece has a very strong unicameral tradition, there is no other veto player that can
stop a legislative initiative supported by a strong parliamentary majority.
The President of the Republic (in practice the proposing and countersigning minister)
has only limited autonomous rule-making powers.9 This is the right to decree the
conditions for the execution of laws. Instead the vast decree power of the Greek executive
is delegated by ordinary legislation (Spiliotopoulos 1987). The delegation is to the
President of the Republic, though for less important matters to the cabinet, the PM or
individual ministers, public bodies, or local authorities. Each law contains a vast number
of delegations, which, however, must be specific. This means that the law should
provide the general principles and procedural guidelines and specify the time limit
8
The sections are divisions of the Parliament. They are not committees but true legislative organs working
within the Parliament. By the absolute majority of the total number of deputies, the 1975/86 Constitution
allows Parliament to decide the composition and functioning of the Sections at the beginning of each
parliamentary session.
9
The autonomous decree power was sometimes abused until the early 1990s (when the courts greatly
curtailed the practice), as ministers and, sometimes, the cabinet itself set rules without delegation from
Parliament. These rules were later ratified by Parliament but until then were legally invalid though never-
theless followed by the administration.
380 Georgios Trantas et al.
within which the executive must make such authorization. In reality, the parliamentary
majorities have given Greek governments much discretion in using their decree power.
However, there seems to be a decline in the (mis-)use of this practice.
There is also an ex post legislative control mechanism. This is the ‘Council of State’,
which was created after the model of the French Conseil D’ Etat. It is the Supreme
Administrative Court of Greece. Under the Constitution the Council can annul acts
and financial decisions by administrative authorities on grounds of abuse of authority
or violation of the law. It can also ‘clarify’ the meaning of a disputed law. The Council
is also heard (ex ante) when the government issues regulatory decrees.
P O L I T I C A L PA RT I E S
On the long road to parliamentary democracy, traditional social structures and the
existing networks of clientelism were transformed and became the bases upon which
political parties were born. The Constitution of 1975 (Article 29) is the first to recognize
political parties and guarantee their freedom. Furthermore, it recommends that parties
Greece 381
10
On the right there were ultra-conservative, pro-dictatorial, and pro-monarchist parties, while PASOK
had ascended at the expense of the old Union of the Centre Party, its successor (EDIK), and the small Social
Democratic Party (KODISO). On the left, KKE was challenged by a faction that left the party after the Prague
Spring and created the Euro-communist party (‘KKE of the Interior’, later Greek Left—EAR). KKE also bat-
tled with the EDA Party, a cover party for KKE during the pre-dictatorial period when KKE was outlawed,
but later a Euro-communist Party. EDA cooperated with PASOK in the 1980s.
11
In the 2000 election it joined forces with the Conservative Party.
12
In late 1995, when Papandreou (PM and President of the Socialist Party) collapsed due to failing health,
the question of his successor as PM arose. The Constitution states that in such cases the parliamentary group
of the majority party elects a new leader who will then be named as the new PM. However, the statute of
the Socialist Party (PASOK) states that the PM should be the President of the party, elected by a party con-
gress. Because a congress requires months of preparation, it was proposed that a new PM and acting party
382 Georgios Trantas et al.
local party organizations are extremely dependent of the central party machine. These
levels are primarily an instrument for the implementation of central party decisions
and not so much a vehicle for the representation of party members. This is one reason
why Greek parties are not fractionalized.13
Party finance is a very complicated and hotly debated issue. Except in the small par-
ties on the Left, member contributions make up only a tiny proportion of the party
budget. The state pays large subsidies to the parties.14 These are mostly given to the
two big parties, but smaller parties are also financed. The Law on Party Finance pro-
vides that parliamentary representation is required for state subsidies but in 1984 ad hoc
rules were created to provide financial means to parties that used to have MPs or MEPs.
The Law regulates private contributions for candidates but not for political parties.
D E L E GAT I O N F RO M VOT E R S TO M P S
Since 1958 Greece has had a ‘reinforced proportionality’ electoral system intended to
encourage the creation of two big parties that can alternate as the majority party
(Penniman 1981). Under the present form of reinforced proportionality (1990), the
country is divided into fifty-six electorals districts (Pantélis 1991).15 The first distribution
of seats takes place on the basis of voting in these districts, using the Hagenbach–Bischoff
formula. This means that the electoral quota that a party must obtain in order to get a
seat is found by dividing the total number of votes by the total number of seats plus
President be elected by the party central committee and then ratified by the parliamentary group. This solu-
tion was rejected by the parliamentary group, which then proposed Kostas Simitis as PM while the ailing
Papandreou remained President. The party crisis continued. Simitis, seemingly with majority support in the
parliamentary group, threatened to resign. Threatened with the prospect of new elections, the party con-
gress elected Simitis President. The episode suggests that the parliamentary group does have the muscle to
impose its will on the party.
13
As the two main parties Greek parties (PASOK and ND) have few important policy differences, politics
are usually built around political tactics and personal feuds rather than policy. Personality politics dominate
the party agenda. This is no surprise considering the omnipotence of the party leader not only in the admin-
istration of the party but also in candidate selection and party finances.
14
It is well known that political parties face economic problems and that many of them are heavily
indebted to state banks. State funding was introduced in 1984. Subsidies are predominantly given to parties
represented in the Chamber of Deputies or in the European Parliament. According to the 1984 Law on Party
Financing, a party is entitled to public financing if it received at least the 3 per cent of the vote in the elec-
tion (5 per cent for an alliance of two parties, 6 per cent for an alliance of three or more parties) and ran can-
didates in at least two-thirds of the electoral districts. The law also made it possible for parties that did not
meet these qualifications to receive some support if they had met the vote requirement in one of the previ-
ous two elections. A later amendment provides public financing for parties that hold at least one seat in the
Chamber of Deputies or the European Parliament. The subsidies are divided among the parties largely
according to their vote share. Legislation also limits the electoral budgets of individual candidates. However,
as is characteristic of Greek politics, after the 1996 election the report of the control commission charged
with examining election spending was set aside. Parliament passed a bill that gave immunity to MPs who had
overspent and cancelled the heavy fines imposed under the law.
15
The electoral district is generally the Prefecture. There are five single-member districts. Only 288 MPs
are elected in electoral districts. Twelve national seats are divided among the parties on the basis of their
share of the national vote using the d’Hondt system.
Greece 383
one. Since 1990, only parties that win 3 per cent of the national vote can win seats
in this round. This makes it practically impossible for small regional parties and
independent candidates to win a seat.
A second round of seat distribution takes place in thirteen regional electoral districts.
It used to be that votes that were not used to fill seats in the first round of seat distribution
were transferred to the next round. Only parties that had won at least 17 per cent of the
vote went on to the second round (Vegleris 1983). Under the 1990 Electoral Law, not
just remaining votes, but all votes and the remaining seats are transferred to the
regional level. If after the second distribution there are still unassigned seats, there is
a third distribution. In the first phase of this third round, the party with the greatest
number of votes nationally automatically gets unassigned seats in those electoral
districts in which it got the most votes. In the second phase the total number of votes
that each party won and the remaining seats are transferred to a unique national
district. As in all previous rounds, the Hagenbach–Bischoff formula is applied. If, after
all this, there are still unassigned seats, then they are all awarded to the party that won
the largest number of votes.
Under the present electoral system, the plurality party can easily acquire a strong
majority in Parliament if its share of the national vote is 1–3 per cent higher than that
of the second-largest party. (The exact number depends on the electoral success of the
smaller parties.) Smaller parties might field a joint list of candidates in order to under-
mine the rules favouring the big parties. To avoid this, the electoral law eliminates
possible electoral coalitions of parties by limiting their participation in the second
round of seat distribution. Only parties that receive more than 3 per cent of the total
vote are allowed to participate. To aid small parties, the so-called ‘normalization’
procedure takes effect after the third round of distribution. This complex procedure,
introduced in 1990, guarantees that smaller parties will not receive fewer than 70 per
cent of the seats corresponding to their percentage of the national vote. In practice a
considerable number of seats won by the party that came in second are thus given to
parties that came in third and fourth. This procedure creates odd discrepancies at the
district level. It is now common that a small party that wins only a small proportion of
the district votes will get a district seat, while the party that comes in second may not
be represented.
Voters cast votes for parties and, if they so prefer, for individual candidates on the
party list. Candidates are listed alphabetically. The number of candidate votes allowed
depends on the district’s electoral population. In larger districts the voter has the
chance to vote for two or three candidates. Thus in the end the voters and not the
parties determine who will take a seat in Parliament, although the parties control
candidate selection. Almost all voters cast candidate votes in addition to the party vote.
The personal vote favours candidates who can create clientelist networks and those
who have the support of particular interests (Lyrintzis 1984; Tzannatos 1986).16
16
In 1985 Papandreou’s Socialist government abolished the candidate vote to promote discipline inside
the parliamentary group. However, the effort was not very successful because MPs that fell out with the
leader had nothing to lose, especially at the end of a legislative period, by coming out openly against the party
384 Georgios Trantas et al.
Campaign spending by candidates is not regulated. Electoral competition among can-
didates in larger districts is stiff, and campaign costs are considerable. Hence, success-
ful candidates must have strong financial support.17 On the other hand, the personal
vote does somewhat constrain the power of the party leader.18
Party statutes give party leaders responsibility for candidate selection. In practice,
however, all sitting MPs are chosen as candidates. Because deputies have usually built a
clientelistic network, it can be unwise for party leaders to create conflict with them just
before an election. The leader has much more discretion with respect to new candid-
ates. All parties in Parliament have internal advisory procedures. These take the form
of proposals by the party prefecture committee,19 which is usually ‘controlled’ either
by a locally powerful MP or by the party leadership. These proposals do not bind
the party leader. In many cases the leader uses such opportunities to ‘parachute’ into
a district political friends or even people without any local or party affiliation.
The Greek Constitution is rich in regulations with regard to who may be elected as
an MP. State officials may not be elected in districts in which they have exercised duties
in the last 6 months. This rule was introduced to reduce the likelihood that civil
servants would use their authority to capture the clientelist networks of local deputies.
Its practical importance today is that party officials who serve as general secretaries
in ministries or as Presidents/managers of state enterprises cannot be elected.
Furthermore, there is incompatibility between the offices of mayor or elected
prefect on the one hand and MP on the other. There exist no term limits or mandatory
retirement regulations.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
Delegation from Parliament to government is highly formalized through the
Constitution and the regulation of the Chamber of Deputies. The relevant rules tend to
promote government stability and to reinforce the rights of the Government over the
Chamber of Deputies. In addition, the party leaders appoint the party’s parliamentary
spokespersons. The latter do not function as whips but are the permanent representatives
of the parliamentary group in the plenary debates. The party leadership also appoints the
rapporteurs for legislative proposals and the members of parliamentary commissions. The
Parliamentary Rules of Procedure also prohibit private member bills that have not been
approved by the party leader or the party’s parliamentary spokesperson.
leadership. The measure seems to have been more useful to the opposition ND leader, Kostas Mitsotakis,
than to Papandreou, since Mitsotakis used it to get rid of some old party barons and entrench his position as
leader. The candidate vote was reinstated after the 1989 elections.
17
After the 1989–90 experience, a modification of the electoral law provides that, if the Chamber of
Deputies is dissolved before 2 years after the previous election, then the new MPs will be elected according
to their list positions.
18
In addition, there are twelve national deputies elected according to their position on the national party
list. The institution of national deputies was intended to facilitate the parliamentary participation of national
figures, including distinguished university professors, scientists, entrepreneurs, and also party worthies
(e.g. former ministers or party leaders).
19
In ND the proposal is formulated after a vote of the party members in the district.
Greece 385
20
The Constitution provides that, if two parties have equal number of seats, both parties will be given
a mandate to form a government. The party that won the most votes goes first. In addition, an older party
has precedence over a newer. A maximum of four exploratory mandates are given.
21
The elections of June 1989 produced a hung Parliament. After the leaders of the three big parties had
failed to form a government, President Christos Sartzetakis then summoned them to a council, where they
agreed to form a coalition government composed of the ND party and the Left party Synaspismos. The gov-
ernment was headed not by Mitsotakis, the leader of ND, but by Tzannis Tzanetakis, an ND deputy proposed
by Mitsotakis. When the Tzanetakis Government had carried out its programme, it resigned. A new series
of exploratory mandates and a party leader council failed to create a new government. The President then
gave Grivas, President of the Areios Pagos (the Supreme Civil and Criminal Court), the task of forming a
Government in order to dissolve Parliament and carry out elections. Grivas’ appointment was accepted by
all parties. The elections of November 1989 again resulted in a hung Parliament. After exploratory mandates
failed, the President called a party leader council, which agreed to a coalition government of all the parties
in Parliament, led by a former Governor of the Bank of Greece, Xenophon Zolotas. The experiment lasted
only a few months, after which the parties had their ministers resign. A new set of fruitless exploratory man-
dates were extended, followed by a council in which party leaders agreed to let Zolotas form a Government
for the purpose of carrying out the elections. The Parliament was dissolved and the election was held in April
1990. ND won the election.
Table 11.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Greece 1977–2000a
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv) Confidence Did the Who
formation votes (cv) cabinet end dissolved
with an early Parliament
No. of Voting results of No. of Cabinet Cabinet No. of cv Cabinet election? in this
unsuccessful successful iv ncv removed resigned under spec. removed case?b
investiture (1 : pro gov., 2 : against gov., by ncv to inst. by failed
votes 3 : abstentions, and 4 : other) pre-empt mechanism cv?
1 2 3 4 ncv
24
There have been only four motions of no confidence under the 1975 Constitution (1988, 1989, 1993, and
1996). All have been won, predictably, by the government. Because of the cross-national counting rule not to
report votes of no-confidence for cabinets still in power 1 January 2000, the 1996 vote is not included in
Table 11.1.
25
If two governments have resigned or been voted down by Parliament and the composition of the
Chamber does not guarantee governmental stability, the President of the Republic may dissolve the
Parliament. The rationale is that the threat of dissolution may keep the Chamber from voting a government
out of office. Yet, the President cannot dissolve the Parliament until 1 year has passed since its opening ses-
sion, except when it is not possible to form a cabinet that enjoys the confidence of the Chamber.
26
The dissolution of Parliament has practical importance because only in this case is the government
entrusted with conducting elections, which is a certain political asset during the campaign. In other cases, this
is done by a government headed by a person accepted by all the parties or by one of the Presidents of the
three Supreme Courts. In August 1993 the Mitsotakis cabinet lost its absolute majority in Parliament due to
the defection of some deputies. While the opposition was preparing a motion of censure, Mitsotakis asked
for the dissolution of the Parliament with reference to national security. His proposal was accepted. Although
at the time some criticized this as an abuse of procedure, “Mitsotakis’ ” action was technically correct.
Greece 389
debate at least once per session. The Standing Orders specify that during this debate
Greece’s relations with the EU are to be discussed. Finally, the Standing Orders give the
opposition the right to ask for four such debates per session, two for the strongest oppo-
sition party and one for each of the smaller ones.27 A variety of issues are discussed at
each session, with the opposition giving emphasis to economic or foreign policy issues.
Many existing laws require the government to submit annual reports to the Parliament
on developments in various policy areas (e.g. European integration or drug abuse). In
most cases the submission of such a report is followed by a short debate.
The usual instruments of parliamentary control include petitions, questions, and
interpellations. Each citizen has the right to petition the Chamber, either through an
MP or through the President of the Chamber. MPs can also support a petition
sponsored by another MP or the President. Ministers are required to respond to such
grievances within fifteen days of it being deposited in the Chamber. If a minister fails
to do so, the President of the Chamber makes the petition part of the agenda for debate
between ministers and deputies in Parliament.
Each deputy has the right to question ministers on any public issue within the
minister’s sphere of responsibility. Questions are written and responsible ministers are
obliged to answer in written form within twenty-five days. If a minister fails to respond
within that period, the question is submitted for debate on the floor of the Chamber,
unless the minister responds after the deadline to the deputy’s satisfaction. Deputies
can also submit ‘urgent’ questions to the government. Urgent questions must be
related to a current issue and are always discussed in the Chamber. Because there are
usually many urgent questions, only those allowed by the parliamentary leadership of
each party are submitted and only those that are subsequently drawn by lot, as
provided in the Standing Orders, are discussed. There is an effort to discuss urgent
questions from each opposition party and certain arrangements between the opposition
parties seem to facilitate that. Debates on petitions, questions, and urgent questions are
short. Debates on questions directed to the PM are longer. Only party leaders may pose
such questions.28
Parliament can also demand a deposition of files or acts from the government. In this
case, the responsible minister is obliged to provide Parliament with files or acts within
twenty days unless they are related to a military, foreign policy, or state security secret.
Deputies can also pose interpellations to ministers. In contrast to questions, these imply
a certain censure of the minister and might lead to a motion of no confidence (something
that has never occurred since 1975). An interpellation may follow a question or a
demand for a deposition, for example, in cases in which a deputy considers a minister’s
response to be unsatisfactory. Interpellations are discussed in plenary. An interpellation
debate can lead to a more general discussion, for example one that covers other inter-
pellations against the minister. The decision to broaden the debate is made by the
27
During his last period as PM, Papandreou’s government was not forced to participate in such a large
number of debates, most likely because of his health problems.
28
Questions directed at the PM were introduced in 1990 and again in 1996. They were abolished during
Papandreou’s second term as PM (1993–6).
390 Georgios Trantas et al.
Chamber after a proposal from fifteen deputies. Once a week the Chamber discusses
urgent interpellations, typically those favoured by party leaders.
Standing committees also exercise parliamentary control.29 In 1993 an addition was
made through a modification of the Standing Order. The legislative Committees can
now, in plenary or, more often, in ad hoc subcommittees, hold hearings with ministers
or under-secretaries, as well as with civil servants and officers of the armed forces. An
exception is made with regard to defence issues, on which the Minister of Defence has
the right to prohibit a hearing. Otherwise hearings are mandatory if they are supported
by two-fifths of the members of the subcommittee. In effect, this gives the biggest
opposition party the right to demand such hearings, which can be held once a month.
After a hearing the subcommittee writes a report, which reviews the issue and includes
the committee’s conclusions and recommendations. On a proposal from the
Government the Chamber can establish other committees, so called ad hoc commit-
tees. The leaders of each opposition party have the right to ask for such a committee
once a year. Such committees have been created increasingly with bipartisan support
(e.g. on drug abuse and social security reform).
An ad hoc committee on state enterprises and organizations is responsible for check-
ing the qualifications of persons nominated by the government to head the most
important state enterprises and organizations (e.g. the Public Electricity Corporation,
the Greek Organization of Telecommunications, the Bank of Greece, the National
Bank of Greece, the Commercial Bank of Greece). The committee has the authority to
call persons to a hearing and to make recommendations to the responsible minister on
appointments.
The legislation of 1990s allows parliamentary involvement in the oversight of certain
sensitive policy issues. Thus, since 1998 the Governor of the newly independent Bank
of Greece reports to the Parliament’s Standing Committee on Economic Affairs. The
newly created (1997) Citizen’s Defender (Ombudsman) is named by the government
following the proposal of the Parliamentary Committee on Institutional Affairs. Other
special oversight agencies include the Commission for the Protection of the
Inviolability of Personal Communications, the Commission for the Protection of
Personal Data, and the Commission for the Selection of Public Employees.
Some of the traditional instruments of parliamentary control seem to have fallen
into disuse. The Constitution requires the Chamber to hold a vote each year on the
state’s financial statement and the balance sheet of state finances. Although both votes
give the Chamber the opportunity to undertake a thorough examination of state
finances, in reality the Committee debate and plenary is entirely a formality, if one is
held at all. The same is true of the debate concerning the Court of Auditors’ annual
report on state finances.
29
During the preparation of a bill a standing committee has the right to hold hearings and call civil ser-
vants and experts to testify. A proposal to do so is made by the minister or one-tenth of the committee mem-
bers. Hearings take place in the presence of the responsible minister. If representatives of organizations are
called to the hearing, at least one-third must be called at the proposal of the opposition.
Greece 391
The so-called select committees of investigation have fallen into disuse. The
Constitution states that such committees can be established by a resolution supported
by two-fifths of the total number of MPs, after proposal from one-fifth of all MPs
(i.e. 120 and 60 MPs, respectively). On foreign policy and defence matters, an absolute
majority is needed (i.e. 151 MPs). Investigation committees were originally seen as
essentially a minority right, but the Socialist majority in 1986 interpreted the
Constitution as saying that an investigative committee requires the assent of the majority.
This basically rendered the institution toothless. To date the majority has agreed to
establish investigative committees only in three cases, all involving criminal charges
against ministers. However, a recent modification of the Standing Orders provides for
the possibility of investigative committees on the management of state enterprises and
organizations.
With Chamber support a Greek minister or, usually, an ex-minister can only be
brought to trial for crimes committed while in office. After an investigative committee
or judicial investigation provides evidence of such criminal acts, the Chamber votes
whether to prosecute or creates a committee of interrogation. In the latter case, the
Chamber decides by absolute majority of all MPs whether to prosecute only after
the committee has reported. If the Chamber votes to prosecute, then it elects a
committee of five MPs (plus five substitutes) to act as public prosecutors before the
Special Court that, under the Constitution, has jurisdiction in criminal cases against
ministers.30
D E L E GAT I O N F RO M C A B I N E T TO
I N D I V I D UA L M I N I S T E R S
The Constitution states that law must specify the composition and functioning of the
cabinet. By decree, the PM may appoint Vice Presidents of the cabinet (Deputy PMs),
Ministers, Substitute (or alternate) Ministers, Ministers without portfolio (now formally
known as Ministers of State), and Under-secretaries. The Law stipulates that Deputy
PM, Ministers, Alternate Ministers, and Ministers without portfolio are members of
the cabinet with full voting rights. According to the current law, Under-secretaries are
civil servants and not formally members of the cabinet. Under-secretaries are named in
almost all ministries and exercise power given them by a joint decision of the PM and
the relevant minister.
There are three models for the relations between ministers and Under-secretaries
within the ministry. The first model gives the Under-secretary exclusive competence in
certain areas. According to the second model, the Under-secretary is generally free to
administer certain policy areas but the minister has the right to oppose his decisions.
According to the third model the minister and the Under-secretary have concurrent
30
In 1989 the new coalition government brought charges against former PM Papandreou and members of
his cabinet. The two spectacular trials that followed were so acrimonious that, although three ex-ministers
were convicted, it is now a common belief among politicians that the whole system for prosecuting
ex-ministers should be depoliticized. The Socialists, back in power in 1993, tried to get revenge by prosecuting
ex-PM Mitsotakis. The charges were eventually dropped.
392 Georgios Trantas et al.
jurisdiction, meaning that in case of differences of opinion the minister’s opinion will
be imposed. The PM and the responsible minister make the choice of model, but in
practice their room for discretion varies with the political balance within the cabinet
and the governing party. The Deputy PMs are not really second-in-command in the
hierarchical chain (as their title might imply). Rather they sometimes substitute for the
PM in cabinet sessions, and they might occasionally coordinate the work of other
ministries. All these arrangements mean that governmental power below the PM is very
dispersed. No inner cabinets of senior ministers exists that can challenge the authority
of the PM.
As a constitutionally defined body, the cabinet is responsible for proposing and
implementing policy and, as mentioned above, it has limited decree power. It is also
responsible for nominating certain high officials (e.g. ambassadors, Presidents, and
Vice-Presidents of the Supreme Courts). The formal appointment takes the form of a
presidential decree. In the cabinet, voting takes place only in rare and politically
controversial cases, and there is no established voting procedure. In practice, however,
decisions are taken on the basis of majority rule, although the opinion of the PM
naturally carries particular weight.
Both in legal terms and in practice, the minister is master of his ministry and the cab-
inet has only the executive powers given to it by law. Politically, however, a minister will
rarely ignore the objections of his colleagues, in particular when there is a clear major-
ity against him in the cabinet. Hence, the cabinet as a whole can and does settle a mat-
ter politically and impose its will on an individual minister who must either implement
the decision or resign.
The Law on the Government provides for the creation of numerous cabinet com-
mittees. The Committee on Foreign Policy and Defence is responsible for most execu-
tive decisions dealing with the armed forces and the police. The Government Committees
act as inner policy-making circles. These cabinet committees’ coordinate ministerial
activities and they control the way in which ministries exercise their authority.
The Minister is the Chief Executive over the policy areas for which his ministry is
responsible. Although alternate ministers and under-secretaries can have specific
responsibilities, they are nevertheless politically responsible to the minister. Ministers
are not allowed to interfere in the administration of ministries other than their own.
Most of the rule-making powers of the cabinet members are delegated to them by the
President of the Republic. Ministers are responsible for proposing and countersigning
presidential decrees within their jurisdiction. In addition, the Constitution allows the
delegation of regulative authority to ministers and other administrative organs
(e.g. local authorities and universities) on more specific matters, matters of local interest,
or matters of a technical and detailed nature. Some of this delegation of authority (e.g. in
environmental protection) is of great practical importance. In many cases ministers
publish regulations without prior delegation of authority from Parliament. These deci-
sions are then later ratified by Parliament. Although the Council of State has declared
this behaviour unconstitutional, ministers still engage in it, though more recently in a
less provocative manner.
Some ministers enjoy particular status. Aside from the Ministers of Foreign Affairs,
Justice and Defence, and National Economy, the Minister of Finance plays an important
Greece 393
coordinating role, which dates back to the huge crises in state finances of the 1920s, and
gives him a virtual veto over all ministerial decisions with budgetary consequences. In
every ministry, the Ministry of Finance also has agents who are either civil servants or
political appointees and who are responsible for overseeing ministry spending and for
ensuring that the budget targets are maintained. No changes in ministerial budgets are
legitimate without the countersignature of the Minister of Finance. In addition, the
Minister of Finance is the only person authorized to release credits from the budget’s
reserve funds.
The Minister of Interior (who nowadays also simultaneously has the symbolic
position as the Minister of the Presidency of the Council of Ministers) is responsible for
the civil service and public administration. In addition, he has special responsibility for
the Official Gazette, which publishes all laws, decrees, or governmental and ministerial
acts. The Minister thus exercises final overall control over the legislative and adminis-
trative output of the government.
These arrangements all support the Prime Minister as the ultimate power within the
government. Although he is not formally hierarchically superior to the ministers, he
makes proposals to the President of the Republic on appointing and dismissing
ministers. (The President is bound by these proposals.) Thus, the Prime Minister is the
real master of the cabinet. The Prime Minister also has two additional mechanisms for
controlling the day-to-day business of the government: the Office of the PM and the
Secretariat of the Government. The Office of the PM, which is directed by a trusted
political friend of the PM, engages in policy planning as well as oversight of the
Government. Of particular importance is the economic adviser to the PM. The Legal
Department of the Office of the PM also plays an important role. According to estab-
lished practice, a ministerial bill or amendment cannot be introduced to Parliament
before the director of the Legal Department (usually a law professor or a judge) signs
it. This is a discreet way for the PM to control the legislative initiatives of his ministers
without getting involved in conflicts with them.
The Secretariat of the Government exists along side the Prime Minister’s Office and
works in close cooperation with it. The Secretariat is not so much a policy-making body
as an administrative staff. The Secretary General of the Government is responsible for
following the implementation of cabinet decisions. Furthermore, the Legal Office of the
Secretariat must before these are signed and published examine all drafts of presidential
decrees or regulatory decisions by the government or individual ministers. This allows
the PM to extend his control to the regulatory decisions of his government.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
The Constitution gives civil servants many guarantees concerning their careers and
independence (Papadopoulou 1989; Skouris 1994). There are no permanent Under-
secretaries in the ministries, with the exception of the Ministry of Foreign Affairs, in
which the position has not been filled since 1968. The most senior civil servants are the
Secretaries General. Because international diplomatic practice so requires, the Ministry
of Foreign Affairs has a permanent Secretary-General and one alternate permanent
Secretary-General named among its senior diplomats.
394 Georgios Trantas et al.
The Greek Civil Service used to be organized in a strictly hierarchical and bureau-
cratic manner. There were ranks connected with the exercise of certain administrative
duties. General directors, alternate general directors, and directors were the top civil
servants. In 1982, the Socialist government eliminated the positions of general director
and alternate general director along with the ranks inside the civil service hierarchy
(Tsekos 1986). These changes have enhanced party patronage (Sotiropoulos 1994;
Spanou 1996). Promotion to the position of directorship became a matter of ministerial
discretion, with ministers’ decisions being based on party affiliations. Furthermore,
they only allowed three-year appointments, thus making the highest ranks of the civil
service even more dependent on the minister. When the number of experienced top
civil servants declined, ministers instead began appointing people from outside the civil
service, either experts (e.g. lawyers, political and social scientists, economists) or party
loyalists. When the Conservatives returned to government in 1990, they re-established
the civil servant rank of general director and eliminated time limits on appointments.
In 1993, when the Socialists returned to power, they retained the civil servant rank of
general director but they reintroduced the three-year appointment period.
In addition, a curious form of neo-corporatism has developed within the civil service.
Between the minister and his personal cabinet on the one hand, and the civil service
on the other, a vacuum has emerged due to the absence of top career civil servants.
This vacuum has been filled by civil servants’ unions, which have become increasingly
powerful as the liaison between the minister and the civil service cadres.
E X T E R NA L C O N S T R A I N T S
The President of Greece has very little influence in the country’s political life. As
mentioned above, if he refuses to sign a decree on political grounds, a government that
enjoys a strong majority in the Chamber can easily override his veto. The same goes for
a proclamation of a referendum. In only one case can he override the cabinet (or more
specifically the Minister of Justice) and this concerns his right to the final word in
pardoning, commuting, altering, or reducing sentences pronounced by the courts of
law. Since 1986, the President can have an important impact on politics only by resign-
ing, since this can force dissolution of the Parliament and new elections. A new
President must be elected with the support of at least two-thirds of all MPs. If this is not
possible in two ballots, then the Chamber can elect a President with the support of a
majority of three-fifths. If the third ballot fails, Parliament is dissolved and a new elec-
tion called.31 The President is also important for maintaining the functioning of the insti-
tutions of the Republic. In this role the President has often convened councils of the
party leaders to discuss political issues of grave importance, in particular those related
to foreign affairs. The decisions of the council of the political leaders, although not
legally binding on the government, are influential.
31
The new Parliament again tries to elect a President with a three-fifths majority. If this is not possible, a
new vote is held, and this time only an absolute majority is required. Finally, if even this proves impossible,
a President can be elected with a relative majority.
Greece 395
Greek courts can have an important political impact by virtue of their authority to
exercise judicial review (Vegleris 1967; Spiliotopoulos 1983; Dagtoglou 1986; Manitakis
1988; Skouris 1988). The Council of State is responsible for examining all drafts of
presidential decrees. The control the Council exercises is limited to evaluating the
constitutionality and legality of the draft decree and its reports are only advisory.
Nonetheless, the examination procedure has acquired weight, because the practice has
developed that the President does not sign any decree that does not follow the
Council’s recommendations. The Legal Office of the Presidency of the Republic
determines whether ministers have followed the comments of the Council of State.
Furthermore, the Constitution requires that bills providing state pensions must be sent
to the Court of Auditors and to the state auditor. Although the comments of the Court
of Auditors are not binding, the minister has an incentive to follow them, because
otherwise the bill might later be struck down as unconstitutional.
A 1994 Law transformed the political organization of the regions (departments).
Rather than prefects appointed by the government, elected prefects and departmental
councils now administer departments. New legislation in 1997, which took effect with
the October 1998 communal elections, called for the fusion of the thousand[s] of small
and medium size communes and the creation of fewer and administratively more
competent metropolitan communes. This might lead to communes that are more
independent from the central government. These reforms, which are based on a French
model, have created some confusion about the competencies of the central, regional,
and local government. Handing over partial control of decentralized administrative
agencies to elected officials has produced a limited but still important institutional
constraint on the government.
Greece lacks a civil society that can balance and limit political institutions. The
Greek-Orthodox Church is the most important social institution. This is due to the
importance Greeks give to their Church as part of their national identity, but also to its
social work and financial power. Of particular political importance also is the National
Bank of Greece, a private bank company in which the majority stocks are held by the
state, the church, and public pension funds. A Governor named by the government
administers it. The Governor has in many cases exercised an important behind the
scene role bringing market realism in the development of economic policies. Greek
trade unions are under party influence. Each party has its own trade union organization.
Furthermore, trade unions are more or less financially dependent on state subsidies, a
fact that increases the government’s influence. In recent years, a degree of influence
over government economic policy has been exercised by the Union of Greek Industries
and the Union of Greek Banks. The latter is the umbrella organization of industrialists
and bankers (state and private).
Although there are no real societal forces that can challenge the parties and their
leaders (Papadopoulos 1989), these leaders are nevertheless suspicious of referendums,
which may explain why no referendum has been held since 1975. The Constitution of
1975 gave the President of the Republic the power to call a referendum. Although the
President has never exercised this power, it was viewed by many as a potential weapon
that the President could use to block government policies. Thus it is not astonishing
396 Georgios Trantas et al.
that the 1986 constitutional amendment cut the discretion of the President of the
Republic to call a referendum. The Constitution now regulates the cases and the condi-
tions under which a referendum can take place. There are two cases under which an advi-
sory referendum might take place. The first concerns a serious national matter, that is a
foreign policy question. Such a question can be put before the people when the
government proposes this and Parliament so decides by an absolute majority of its mem-
bers. The second case applies for an act of Parliament that concerns a serious question
concerning society. In that case the President can refer the bill to the people, but only if
this is supported by a three-fifths majority in Parliament. This rule permits only two such
referendums per Parliament and these referendums cannot concern state finances. It is
clear that under these rules the possibility of a referendum is extremely limited.
Greek politics is also influenced by international constraints, such as the conflict with
Turkey over Cyprus, the status of the Aegean continental shelf, and the limits of the
Greek territorial waters. The geopolitical situation became even more complicated
after the fall of Communism and during the ensuing conflicts in the Balkans. The
geopolitical situation has always made Greece look to foreign powers for protection—
primarily the United States. Otherwise, the most important international constraint is
the European one. Greece became a member of the European Communities in 1981,
and is also a member of the Council of Europe and the European Convention on
Human Rights and the European Social Charter, both of which create obligations with
regard to legislation and political practices. In recent years the legality of many Greek
laws and administrative practices have been tested before the EU and the European
Court of Human Rights.
Opinion polls show that Greeks are one of the most pro-community peoples of the
EU. Greek European policies are elaborated and implemented largely at the executive
level without the participation of Parliament. The introduction of Community law in
the Greek legal order is almost exclusively made through executive decrees. In the last
few years, the standing Special Committee on EU Issues has begun to play a still very
limited but at least more important role. This European Affairs Committee is mostly
concerned with policy issues at the European level and not with the day-to-day imple-
mentation of European policies in Greece. The Foreign Ministry regularly briefs the
committee on European developments and the committee occasionally hears experts.
It submits a report to the plenary twice a year, both of which are followed by a debate.
Otherwise, the Parliament gets involved only when a parliamentary statute is
demanded by the Constitution (i.e. tax legislation) in the implementation of
Community legislation. The fact that the implementation of Community law is dele-
gated to the executive means that it is really only when the implementation takes the
form of a presidential decree that there does exist a limited ex ante control in the form
of the required opinion of the Council of State.
C O N C LU S I O N
In Greece, the formal lines of delegation and accountability are clearly defined. The
Greek process of delegation is highly institutionalized. The institutionalization is
exhaustively regulated in the Constitution. This might suggest that Greece is an exemplary
Greece 397
REFERENCES
Alivizatos, Nicos (1979). Les institutions politiques de la Grèce à travers les crises, 1922–1974. Paris: LGDJ.
—— (1990). ‘The difficulties of “rationalization” in a polarized political system: The Greek
Chamber of Deputies’, in Ulrike Liebert and Maurizio Cotta (eds.), Parliament and Democratic
Consolidation in Southern Europe. London: Pinter.
—— and Eleftheriadis, Paulos (2002). ‘The Greek Constitutional Amendments of 2001’. South
European Society & Politics, 7:63–71.
Clogg, Richard (1992). A Concise History of Greece. Cambridge: Cambridge University Press.
—— (ed.) (1993). Greece 1981–1989. The Populist Decade. London: Macmillan.
Dagtoglou, Prodromos (1986). ‘Griechenland’, in Christian Starck (ed.), Verfassungsgerichtsbarkeit
in Westeuropa, Vol. 1. Baden-Baden: Nomos (leefloose).
Diamantouros, Nikiforos P. (1984). ‘Transition to, and Consolidation of, Democratic Politics in
Greece, 1974–1983: A Tentative Assessment’, in Geoffrey Pridham (ed.), The New Mediterranean
Democracies: Regime Transition in Spain, Greece and Portugal. London: Frank Cass.
—— (1986). ‘Regime Change and the Prospects for Democracy in Greece, 1974–1983’, in
Guillermo O’Donnel, Philippe C. Schmitter, and Laurence Whitehead (eds.), Transitions from
Authoritarian Rule. Prospects of Democracy. Baltimore: John Hopkins University Press.
—— (1991). ‘PASOK and State–Society Relations in Post-Authoritarian Greece (1974–1988)’, in
Speros Vryonis Jr. (ed.), Greece on the Road to Democracy: From the Junta to PASOK, 1974–1986.
New Rochelle: Ar. D. Caratzas.
Doukas, George (1994). ‘Papandreou in Government Again: a New “Change”?’, in Richard
Gillespie (ed.), Mediterranean Politics, 1: 227–31. London: Pinter.
Huber, John D. (1996). Rationalizing Parliament. Cambridge: Cambridge University Press.
Kaminis, Georges (1993). La transition constitutionnelle en Grèce et en Espagne. Paris: LGDJ.
Katsoudas, Dimitrios K. (1987). ‘The Constitutional Framework’, in Kevin Featherstone and
Dimitrios K. Katsoudas (eds.), Political Change in Greece—Before and After the Colonels. London:
Croom Helm.
Kentrotis, Kyriakos (1994). ‘Echoes from the Past: Greece and the Macedonian Controversy’, in
Richard Gillespie (ed.), Mediterranean Politics, Vol. 1: 85–103. London: Pinter.
Legg, Keith (1969). Politics in Modern Greece. Stanford: Stanford University Press.
—— (1973). ‘Political Change in a Clientelistic Polity: The Failure of Democracy in Greece’.
Journal of Political and Military Sociology, 1: 231–46.
398 Georgios Trantas et al.
Lyrintzis, Christos (1984). ‘Political Parties in Post-Junta Greece: A Case of Bureaucratic
Clientalism?’, in Geoffrey Pridham (ed.), The New Mediterranean Democracies: Regime Transition
in Spain, Greece and Portugal. London: Frank Cass.
Manessis, Ariostovoulos (1985). ‘L’évolution des institutions politiques de la Grèce: à la recherche
d’une légitimation difficile’. Les Temps Modernes, no. 473: 772–814.
Mantakis, Antonis (1988). ‘Fondement et légitimité du contrôle juridictionnel des lois en Grèce’.
Revue internationale de droit comparé, 40: 39–55.
—— (1994). ‘Le régime constitutionnel de la Grèce’, in Yves Guchet (ed.), Les systèmes politiques
des pays de l’union européenne. Paris: A. Colin.
Mavrogordatos, George Th. (1983). ‘The Emerging Party System’, in Richard Clogg (ed.), Greece
in the 1980s. London: Macmillan.
—— (1984). ‘The Greek Party System: A Case of “Limited but Polarised Pluralism”?’ West
European Politics, 7(4): 156–69.
Pantélis, Antoine (1979). Les grands problèmes de la nouvelle constitution hellénique. Paris: LGDJ.
——(1991). ‘Chronique: Droit constitutionnel hellénique’. Revue Européenne de Droit Public, 3: 217–32.
Papadopoulos, Yannis (1989). ‘Parties, the State and Society in Greece: Continuity within
Change’. West European Politics, 12(2): 55–71.
Penniman, Howard R. (1981). ‘The Modern Greek State and the Greek Past’, in Speros Vryonis,
Jr. (ed.), The Past in Medieval and Modern Greek Culture. Malibu: Undena Publications.
Skouris, Wassilios (1988). ‘Constitutional Disputes and Judicial Review in Greece’, in Christine
Landfried (ed.), Constitutional Review and Legislation. Baden-Baden: Nomos.
—— (1994). ‘Das Recht des öffentlichen Dienstes in Griechenland’, in Siegfried Magiera and
Heinrich Siedentopf (eds.), Das Recht des öffentlichen Dienstes in den Mitgliedstaaten der
Europäischen Gemeinschaft. Berlin: Duncker & Humblot.
Sotiropoulos, Dimitrios A. (1994). ‘Bureaucrats and Politicians: A Case Study of the
Determinants of Perceptions of Conflict and Patronage in the Greek Bureaucracy under
PASOK Rule, 1981–1989’. British Journal of Sociology, 45: 349–65.
Spanou, Calliope (1996). ‘Penelope’s Suitors: Administrative Modernisation and Party
Competition in Greece’. West European Politics, 19(1): 97–124.
Spiliotopoulos, Epaminados (1983). ‘Judicial Review of Legislative Acts in Greece’. Temple Law
Quarterly, 56: 463–502.
–––– (1987). ‘Grèce—Les délégations législatives’. Revue Française de droit administratif, 3: 729–31.
Spyropoulos, Philippos C. (1995). Constitutional Law in Hellas. The Hague/Boston: Kluwer Law
International.
Tsekos, Théodore (1986). ‘Changement politique et changement administrative: La haute fonc-
tion publique en Grèce avant et après 1981’, in Danièle Lochack et al.(eds.), La haute adminis-
tration et la politique. Paris: puf.
Tzannatos, Zafiris (ed.) (1986). Socialism in Greece—The First Four Years. London: Gower.
Vegleris, Phédon Th. (1967). ‘La Constitution, la loi et les tribunaux en Grèce’. Annales de la faculté
de droit de Liège, 12: 439–77.
—— (1983). ‘L’évolution du système et des pratiques électorales en Grèce. Leurs résultats et leurs
effets comparés—élections nationales et européennes’, in Jacques Cadart (ed.), Les modes de
scrutin des dix-huit pays libres de l’Europe occidentale. Paris: puf.
12
I N T RO D U C T I O N
Parliamentary democracy was gradually established in Iceland. Or, as Icelanders prefer
to put it, it was ‘re-established’, as the Icelandic Parliament, the Althingi (Alþingi,
founded in 930), reclaimed its central position in the Icelandic political system. The
process began with the re-establishment of the Althingi in 1845, when Iceland was still
subject to Danish rule. The parliamentary majority pressed for Iceland’s sovereignty
within a union with Denmark, while Danish authorities viewed Iceland as inseparable
part of the Danish state. After a long impasse, Christian IX, King of Denmark, came to
Iceland in 1874 and unilaterally granted the country a separate Constitution. The
Althingi received financial power, subject to the King’s veto. Executive power was placed
in the hands of one Minister and remained part of the Danish administration until
Iceland gained Home Rule in 1904. At that time the Minister became accountable to the
Althingi.
The Republic of Iceland has a written Constitution. It was adopted in a national
referendum in 1944, when Iceland declared its independence, thereby abolishing the
union with Denmark. Centuries of foreign rule, which had begun in 1262, had finally
ended. The mythological origins of parliamentary government in Iceland’s golden past
undoubtedly greatly strengthened the tiny nation’s resolve to become an independent
nation state. However, myths are seldom particularly conductive to clear thinking
about principles of good government in the present.
The principles of Iceland’s parliamentary democracy rest partly on a written
Constitution, partly on conventions. The constitutional tradition is generally stronger
than the written Constitution. Brought about in haste by the pressure of the Second
World War, without due deliberation and calm reflection (not to mention the lack of
proper consultation with Denmark, which suffered under German occupation), the
Constitution of 1944 was only intended to be temporary. The intended full-scale revision
has never been achieved, and the Constitution remains full of ambiguities. The separation
The author gratefully acknowledges the helpful suggestions on the text that have been provided by Auður
Styrkásdóttir, Indriði H. Indriasson, and the editors.
400 Svanur Kristjánsson
of powers is vague, as is the constitutional position of the President. According to the let-
ter of the Constitution, the Icelandic President has numerous political powers. In consti-
tutional practice the President’s position has varied greatly, mostly due to different
circumstances as well as the political skills and role conception of each of the five indi-
viduals who have served as Iceland’s President.
Evolving informal rules and conventions have often been a substitute for constitutional
revisions. In this, today’s Icelanders follow the political traditions of the past, when the
struggle for independence was primarily a nationalistic one rather than a battle for gov-
ernment reform. The essence of political reform was defined in terms of national sov-
ereignty rather than individual rights. This does not mean that the constitutional and
legal framework is unimportant in an analysis of parliamentary government in Iceland.
Rather, we must explain the development of the political system in terms of the way in
which a changing combination of formal and informal rules affect the action of politics.
One important tradition that sets Iceland apart from its Scandinavian neighbours is
its prevailing pattern of patronage. Politics Icelandic-style combines a personal style of
politics with a non-bureaucratic, patronage-centred mode of public administration.
This has implications for the whole chain of delegation and accountability, from voters
to civil servants. Recent efforts to develop corporatist relations have also had an impact
on Icelandic parliamentary government.
There are also environmental constraints that affect political behaviour at the
national level. The Danish ‘heritage’ is one factor that has shaped Icelandic politics.
NATO’s presence on the island has also been a crucial factor affecting the country’s
political life. Since 1993 Iceland has been formally linked to the European Union (EU)
through its membership in the European Economic Area (EEA). This relationship with
the EU is an important factor in Icelandic politics. Finally, recent developments at the
domestic local level also deserve to be mentioned here. Iceland has a strong unitary
state. In recent years, however, there has been a move towards an increasing role for
local government. The number of local governments was reduced from 204 in 1990, to
124 in the 1998 local elections. Responsibility for primary schools has been transferred
from the central government to local ones. Additional transfers of political responsibility
are planned.
National legislation is passed by the Parliament by a simple majority vote. Any MP
can propose legislation, as can standing committees and cabinet ministers, who—with
few exceptions—are also MPs. A parliamentary quorum consists of an absolute major-
ity (50 per cent ⫹ 1 of all MPs) and when voting takes place, the vote of every MP is
recorded. In the past, the Parliament’s chamber was formally divided into two quasi-
chambers. These chambers were not separately elected but in the beginning of each term
Althingi divided itself into two chambers for conducting its legislative process.
However, since 1991 the Parliament meets as a single chamber. The executive branch pro-
poses most legislation. Formally, proposals come from the cabinet; however, in reality they
are usually formulated by the government bureaucracy, interest groups, or, increasingly, in
many cases dictated by the requirements of the EEA treaty.
Parliamentary legislation is not subject to review by any constitutional court, but the
President can refuse to sign any law passed by the Parliament. Before such a law goes
Iceland 401
into effect, however, it must be put to the voters in a national referendum. The law is
invalidated if a simple majority of those participating vote against it.
Two policy areas are subject to specific procedures that apply only to them, the
budget and foreign policy. The budget proposal is processed according to special law,
which requires the cabinet to submit the following year’s budget as well as a three-year
budget plan to the Althingi at the opening of the fall session. With regard to foreign
affairs, the cabinet must confer with the foreign affairs committee before the introduction
of legislative proposals.
Proposals for amending the Constitution are introduced in the Althingi. If they are
approved by a simple majority vote—with quorum requirements identical with those
in ordinary legislation—the Althingi is dissolved and a parliamentary election is held.
The proposed constitutional amendment becomes a binding law provided that the new
Althingi passes it unaltered. An important constitutional amendment in the 1990s is the
1995 addition of a catalogue of human rights to the Constitution.
The use of executive decrees was restricted by constitutional change in 1991.
Previously executive decrees were frequently issued and only sometimes did they have
to be confirmed in the next session of Althingi. The cabinet can still, with the
President’s authorization, issue executive decrees. Such decrees must be introduced in
the Althingi at the beginning of the next meeting and are invalidated if not passed by
the Althingi in 6 weeks time. After a constitutional change in 1991, the Althingi is
permanently in session between elections. The Althingi is in recess between meetings
and can be called into session at any time. It is up to the individual minister to decide
on government regulations. Since the laws have become more open-ended and more
complex over time, this has placed more capacity to make general rules in the hands of
the ministers.
The government has various means at its disposal to conduct its business without
involving the Althingi. The government is entrusted with implementation of legisla-
tion, frequently filling the space left by open-ended legislative bills of government
decrees.
P O L I T I C A L PA RT I E S
With regard to political parties, Iceland has basically had a four-party system since the
1930’s. The Independence Party (IP) was founded in 1929 by a merger of the
Conservative Party (founded 1924) and the Liberal Party (founded 1926). It usually
receives about 40 per cent of the vote. The second largest party, the Progressive Party
(PP), is located in the centre of the political spectrum. It was established in 1916 and
since about 1980 it has received around 20 per cent of the national vote. The Social
Democratic Party was founded in 1916 as the political arm of the labour movement. It
was organizationally tied to the Icelandic Federation of Labour until 1942. Since that
time the SDP was usually the smallest of the four main parties, receiving 14–16 per cent
of the vote. Finally, the People’s Alliance (PA) is a descendant of the Communist Party
and other breakaway groups from the SDP and normally got around 15–18 per cent of
the vote, thus making it the third largest party.
402 Svanur Kristjánsson
Since the early 1980s one or two smaller parties have also been represented in the
Althingi. The Women’s Alliance (WA) held seats from 1983 (see Styrkársdóttir 1986).
Before the parliamentary elections of 1999 a new electoral alliance, the United Front,
was formed, composed of four political parties: the PA, the SDP, the WA, and the
People’s Movement (a social democratic splinter party). It received nearly 30 per cent
of the vote. A new Left-socialist party, the Left and Green Party, also fielded candidates
in all electoral districts and won close to 10 per cent of the national vote.
The main parties are formally organized as mass political parties. In reality they are
rather loose organizations. They do not collect membership dues, nor do they maintain
reliable membership lists. Since the 1970’s all traditional parties have adopted
primaries to select candidates for parliamentary elections (the WA used selection
committees for nominations). The primaries have reinforced the cadre-party character-
istics of the parties, and have contributed to frequent internal disunity. Primaries have
also reduced the parties capacity to formulate policy insofar as they encourage parties
to settle for the most general common denominator with regard to taking policy posi-
tions. But the established parties have been able to retain their central political role due
to the conservative impact of Iceland’s proportional electoral system and the very
strong tradition of majority coalition cabinets.
On the whole, and in spite of the decentralized candidate selection process and the
relatively low level of party cohesion, Icelandic political parties contest elections and
form governments as organized entities. They are considerably weaker in policy
formulation and decision-making. Thus, as we shall see, the effects of parties are much
stronger at the level of delegation from voters to MPs and in the linkage running from
Parliament to cabinet than in the delegation chain from cabinet to individual ministers
or from ministers to civil servants.
D E L E GAT I O N F RO M VOT E R S TO M P S
The electoral system is part of the Constitution. It has been changed four times in the
history of the Republic (1942, 1959, 1987, and 2000). In the period 1971–87, forty-nine
of the sixty Althingi members were elected in eight constituencies: thirty-seven in
five- and six-member constituencies, and twelve in Reykjavík, the capital. In each
constituency, seats were allotted to ranked lists put forward by the political parties
according to a simple proportional system (d’Hondt). The eleven remaining seats, the
supplementary seats, were allotted to parties that had won at least one seat, the goal
being to minimize the difference between each party’s proportion of the vote and the
party’s proportion of seats in the Althingi. The supplementary seats were filled with
candidates not elected in the constituencies, but next in line on their party’s list.
In 1987 three parliamentary seats were added and new, complex rules of allocation
were introduced. These are designed to better obtain the goal of exact proportional
representation. As before, the voter selects one of the lists put forward by the political
parties. The parties rank the candidates in advance and the voter has little chance of
altering the ranking.
The many changes made in the Icelandic electoral system in this century have helped
reduce the impact of the personal vote. This has happened in two ways: (1) By moving,
Iceland 403
1
According to Gallagher and Marsh (1988), in other West European countries the highest percent of vot-
ers participation in nominations was 2.1 % (Belgium). During elections in Iceland from 1971, the percentage
of primary elections voters relative to all eligible voters has been as follows: 1971–25%; 1978–32%; 1979–21%;
1983–26%; 1987–16%; 1991–15%; 1995–20% (Kristjánsson 1998: 177).
404 Svanur Kristjánsson
selecting candidates, nor can they use ex post mechanisms—that is, punishing MPs for
not following the party line by denying them renomination or rewarding faithful ones
by giving them safe or promising seats.
The impact of the open primary on relations between voters and MPs is complex.
Thus, it is possible to view the primaries as having strengthened the role of voters as
principals. This is of course true in a trivial sense—the winners of an open primary
must have a strong personal following; the losers do not. However, there is more to it
than this. The open primary also empowers voters by giving them an additional ex post
device. It enables local voters to press their MP into providing them with extensive and
concentrated constituency service, and offers them an opportunity to sanction the MP
at the nomination stage.
The rational MP seeking re-election will try to please voters in his electoral district,
because he/she knows that the nomination process is open to voter participation. In
fact, the nomination stage is inherently more unpredictable and dangerous to sitting
MPs than is the general election. In an open primary, there are no safe seats. Everyone
can be defeated. Turnovers at the nomination stage are therefore rather frequent and,
more importantly, sometimes very unpredictable. In contrast, there are many safe seats
on the party lists in the general election. Defeats at this stage are confined to candidates
occupying marginal places on the lists.
In short, the primary is by nature a personal vote. As mentioned, the voter can even
occasionally select candidates for more than one party. The driving force behind this
voting behaviour is not deceitful voters but rather the rules of contest. Candidates in
an open primary seek to involve all voters favourable to them in the primary, regardless
of whether they are party voters or not.
The open primary fits the prevailing pattern of patronage in Iceland. The informal,
personal, and male-dominated political network is enforced by the system of political
finance.2 It has been estimated that during the 1995 parliamentary election—when
approximately 170,000 people voted—political parties spent about 230 million Icelandic
Krona (roughly 3.3 million US Dollars).3 There are no restrictions on how much parties
can spend in elections or otherwise; nor do political parties have to reveal their income
or spending. Public support to political parties has steadily increased and during 2003 is
worth about 230 million Icelandic Krona per year. Parties not represented in the
Parliament do not get public funding.
Electoral competition is largely unregulated. There are no laws on the fundraising or
expenditures of political parties and the parties have as much availability of mass media
advertising, including television, as money can buy. Only laws prohibiting the collection
and storage of personal information limit the commission and use of public opinion
polls.
2
The origins of this informal, personal, and male-dominated political network have been traced to the
first decades of the twentieth century (Styrkársdóttir 1998).
3
Given the fact that the US population is 1,000 times that of Iceland, this is roughly equal to the amount
of money spent on all elections in the United States. In the United States in 1997, candidates and parties in
search of seats in the US House of Representatives and the Senate spent at least 660 million. The estimate for
spending in all elections exceeds 3 billion (The Economist, 8–14 February 1997).
Iceland 405
The state radio and television provide free and equal broadcasting time to parlia-
mentary parties contesting elections in all electoral districts. Other political parties
(newcomers and/or parties that do not run lists in all districts) have no guaranteed
access to mass media, although some broadcasting time is allocated to them when
contesting local or parliamentary elections.
Given the heavy financial losses parties have suffered as a result of money-losing,
now-defunct party newspapers and the high and increasing cost of electoral campaigns,
public support to parties does not by any means suffice to cover their expenditures.
Thus, most parties are heavily in debt. In 1993 all parliamentary parties except the WA
voted in favour of a change in the tax code, which allows private companies to deduct
contributions to political parties. The financing of political parties by private interests
without any public disclosure reinforces the private nature of governance in Iceland
and reinforces the link between special interests and ministers and MPs.
The gradual transformation of the Icelandic political system, from party rule to
personal politics, is also manifested at the parliamentary party group level. The MPs have
gained more freedom of action; they are agents without any strong party organization
to control them. Members of Parliament have extended their freedom of action into
new areas, including decisions about whether to support or oppose the cabinet in the
Althingi. As a result, the current PM has spoken of the need to create oversized cabinet
coalitions to protect the government from defection and opposition by government
backbenchers in the Althingi. Many events in the 1990s support the PM’s analysis.
A minimum coalition gives every government-party MP bargaining power, something
that is obviously discomforting for any cabinet.
It is important to note that the developments discussed here should not be taken as
indication of a complete collapse of party cohesion and a fatal weakening of parlia-
mentary party groups. Party groups have, in fact, retained a strong degree of cohesion
because the parliamentary system provides countervailing incentives. These incentives
have to do with the parliamentary form of the relationship between the cabinet and the
Parliament.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
Since Iceland achieved Home Rule in 1904, ministers have been accountable to the
Althingi. Formally, the Althingi is the principal, the Cabinet the agent. However, this
seemingly straightforward relationship is complicated by three interacting factors:
(1) the absence of formal rules, (2) the tradition of majority cabinet formation, and
(3) the role of the President.
The Constitution does not offer much insight into the actual processes and rules
influencing the delegation from Parliament to cabinet. The basic rules for cabinet
formation and cabinet resignation are not formally spelled out more than that the
President appoints a new cabinet. There is no vote of investiture in the Parliament.
However, a cabinet already in power, or a particular minister, can be voted out of
power by a majority (50 per cent ⫹ 1 of all valid votes) in the Parliament with quorum
requirements of 50 per cent ⫹ 1 of all members. Any individual MP can propose such
406 Svanur Kristjánsson
a vote. Formally, the President can remove the cabinet, but it would be against the basic
principles of parliamentary government to do so without prior declaration of non-
confidence on the cabinet in the Althingi. There is also a constitutional practice that
holds that a cabinet must resign if it loses a vote on a matter to which it has attached
its survival, but because of the tradition of majority coalitions the specific meaning of
this practice remains unclear.
In sharp contrast to the other Nordic countries—in particular Denmark, Norway,
and Sweden (Strøm 1986; Bergman 1995)—Iceland has no tradition of minority cabi-
nets in the post-Second World War era. The 1917–44 period was quite different in this
respect—majority coalition cabinets reigned for less than half the time during these
years.
The tradition of majority cabinet was born with the Republic in 1944. In that year,
an ‘oversized’ majority cabinet was formed. This government included the party
furthest to the right (the IP), the Social Democratic Party, and a radical Socialist party
(the United Socialist Party). The cabinet was supported by thirty-two MPs. The oppo-
sition was made up of the agrarian Progressive Party, which held fifteen seats, and five
IP MPs who had broken ranks with their party. The new cabinet was a response to a
cataclysmic event in Icelandic politics. In 1942 the Regent of Iceland—Sveinn
Björnsson, who was elected by the Althingi for a 1 year term—appointed a non-partisan
cabinet because the parliamentary parties had been unable to form a government. The
non-partisan cabinet was in power when a special session of the Althingi at Thingvellir
(the site of the old Althingi) proclaimed Iceland’s independence and the establishment
of the Republic. This historical occasion, which was attended by approximately 20 per
cent of the population and broadcast by radio all over the island, was an occasion at
which many MPs felt humiliated by their own inability to form a cabinet. Ever since,
majority cabinets have been seen as the proper result of cabinet formation. Only major-
ity cabinets are considered to be real parliamentary government. Minority cabinets are,
at best, only short-term measures. Given this, MPs are under great pressure to form
majority cabinets as soon as possible after elections.
The letter of the Constitution gives the President the right to dissolve the Parliament.
The government, however, holds executive power. Thus, in practice, the Prime Minister
(PM) proposes parliamentary dissolution to the President. It is then the President’s
decision to accept or reject such a proposal. The Constitution also gives the President
the power to appoint a non-partisan cabinet if the Althingi fails to create a cabinet tol-
erated by a majority of MPs. This provision has undoubtedly strengthened the collec-
tive identity of MPs as part of an institution, the Althingi. In recent years Icelandic
political leaders have shown themselves capable of forming cabinets quickly without
‘interference’ from the President. Given the improvements in the economy—after
experiencing great economic instability in the post Second World-War period the coun-
try now enjoys a low rate of inflation rate (2–3 per cent)—political leaders have also
been under somewhat less pressure to quickly form a cabinet. The role of the President
in cabinet formation will remain pro-forma as long as the political parties represented
in the Althingi can agree on a majority cabinet. In this sense the Althingi has firmly
established itself as the principal over the cabinet.
Iceland 407
Table 12.1 summarizes the empirical record of no-confidence votes and parliamentary
dissolutions in Iceland. Only once a cabinet had to face a no-confidence vote, Thors
IV in 1950. This cabinet was removed by this vote. In addition, seven of twenty-five cab-
inets were ended by early parliamentary dissolution, mostly by the President.
To understand the relationship between the Althingi and the cabinet, it is necessary to
put it into the context of Icelandic history—notably the origins of these two institutions.
Kristinsson (1996: 443), has observed:
Parliament has always been a major channel for political patronage. Apart from its important
role in deciding the composition of executive committees and boards, its legislative and
budgetary powers have been extensively used in the twentieth century to promote narrow
constituency interests. The constitution and the law of parliamentary procedure grant a
broader scope of manoeuvre to Icelandic parliamentarians than exists in most European
countries. There are no restrictions on the right of individual parliamentarians to propose new
legislation or amendments to the government budget. Private members bills have always been
numerous.
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence votes Did the cabinet Who dissolved
formation votes (iv) (cv) under end with an early Parliament in
n.a. No. of Cabinet Cabinet specific election? this case?b
ncv removed resigned to institutional
by ncv pre-empt ncv mechanism
n.a.
1 Thors II 44/10/21 0
2 Thors III 46/06/30 0
3 Stefansson 47/02/04 0 Yes Head of State
4 Thors IV 49/12/06 1 Yes
5 Steinthorsson 50/03/14 0
6 Thors V 53/09/11 0 Yes Head of State
7 Jonasson III 56/07/24 0
8 Jonsson 58/12/23 0 Yes Const.
9 Jonsson II 59/06/28 0 Yes Head of State
10 Thors VI 59/11/20 0
11 Thors VII 63/06/09 0
12 Benediktsson 63/11/14 0
13 Benediktsson II 67/06/11 0
14 Hafstein 70/10/10 0
15 Johannesson I 71/07/14 0 Yesc Head of State
16 Hallgrimsson 74/08/28 0
17 Johannesson II 78/09/01 0
18 Gröndal 79/10/15 0 Yes Head of State
19 Thoroddsen 80/02/08 0 Yes Head of State
20 Hermannsson I 83/05/26 0
21 Palsson 87/07/08 0
22 Hermannsson II 88/09/28 0
23 Hermannsson III 89/09/10 0
24 Oddsson I 91/04/30 0
25 Oddsson II 95/04/23 0
26 Oddsson III 99/05/28 — — — —
a
n.d. ⫽ no data (missing data); n.a. or ‘—’ ⫽ not applicable in this case.
b
The question ‘Who dissolved Parliament in this case?’ can be anwered with one of five options. This identifies the main constitutional actor that caused the early
election. We have not coded the formal signatory, but rather the person or body that made the real decision: (1) Head of State; (2) Prime Minister; (3) Cabinet;
(4) Parliamentary majority; (5) Automatic constitutional provision (‘Const.’; for example, if it is required when a new President takes office, or if it is required after the
adoption of a proposal for constitutional amendments).
c
In this particular case the Parliament was dissolved to pre-empt a no-confidence vote.
410 Svanur Kristjánsson
report. The somewhat special role of the Althingi in the field of foreign policy is
explained by several factors:
1. The heritage of the politics of independence, which calls for national unity in
defending and promoting the national interest. This emphasis was particularly
strong in ‘the new national struggle’, the extension of Iceland’s exclusive fishing
zone to 12 miles (1958), 50 miles (1972), and finally 200 miles (1975).
2. In most instances foreign policy is less tied to the politics of patronage. MPs
have less need to worry about serving local and particular interests, and can
identify more with the Althingi as an institution for promoting the common
Icelandic good.
3. The foreign policy dimension cross-cuts the Left–Right dimension of party poli-
tics, thus making possible unconventional alliances among parties. Thus the Social
Democratic Party and the IP worked closely together in supporting Iceland’s mem-
bership in NATO and the stationing of American troops in Iceland. Since 1999 the
leaders of the IP and the Left and Green Party have voiced similar arguments
against Iceland’s membership in the EU, invoking fears of the loss of national sov-
ereignty to a supranational organization as an unacceptable price of entry.
D E L E GAT I O N F RO M C A B I N E T TO I N D I V I D UA L M I N I S T E R S
The executive government in Iceland is generally characterized by three main features:
1. The separation of power between the Althingi and the executive remains unclear.
2. Individual ministers are highly independent.
3. The civil service has a low degree of autonomy.
Together these characteristics produce a highly individualized style of executive
government.
As a note of curiosity, one could say that when Iceland got Home Rule in 1904 there
was no delegation and accountability problem within the cabinet because there was
only one minister! The minister was formally a member of the Danish State Council,
but he was independent of the Danish Parliament and did not resign when Danish gov-
ernments did so.
In 1917 the number of ministers increased to three. Europe was at war and many
Icelanders felt that the executive branch should be strengthened in this time of emer-
gency. Some also wished to increase the cabinet’s capacity to direct and control the civil
service. The composition of the Althingi also played a role. No party majority existed,
which necessitated the formation of a coalition cabinet (Kristinsson 1994).
With three ministers the question of cabinet decision-making rules inevitably arose.
The Althingi debated this issue in 1920. The proposal for the new Constitution had orig-
inally contained the principle of majority voting in cabinet meetings, but the Althingi
removed this clause. The cabinet has no formal decision rule but majority voting is rarely
used and in some cabinets not at all. Generally the cabinet is not collectively responsible.
Rather, executive responsibility is held by individual ministers (Kristinsson 1994).
Iceland 411
The Oddsson cabinet formed in 1999 included twelve ministers. The mean of the
five most recent cabinets through the Oddsson cabinet is 10.4. The central government
organization is divided into twelve departments, with a total combined staff of 366
people in 1993. The largest departments are Finance (79), Education (66), and Foreign
Affairs (51) (Kristinsson 1994: 63). Each minister also has one personal assistant, who
comes to the ministry with the minister and leaves with him/her as well. Icelandic min-
isters thus have a good possibility to control their departments, if for no other reason
than because of their small size.
The formal rule for appointments and dismissals of ministers is that this is the
prerogative of the President. In practice he/she generally does so only on the initiative
of the PM. The PM also defines the jurisdictional boundaries between the ministries.
However, the leadership of coalition cabinets is inherently more complicated than that
of single-party majority cabinets. In the latter case, the PM is faced with the task of
providing leadership for a group of ministers who are not members of the same party,
but without the formal rule of collective cabinet responsibility to aid him in this task.
In Iceland the last single-party majority cabinet ruled from 1924 to 1927. Since then
majority coalitions have included ministers from two, three, and—in one case—four
political parties. The life expectancy of cabinets is generally negatively related to the
number of parties included in them. Most two-party cabinets last the full electoral
term, while no three-party cabinet has lasted the normal term of 4 years.
Some PMs have (in the 1980s and 1990s) had great success in keeping their coalitions
together, while others have failed. Two of the successful PMs were Progressive Party
leader (1979–94) Steingrímur Hermannsson and Davíd Oddsson, leader (since 1991) of
the Independence Party. Why do some party leaders succeed as PM while others fail?
On this point, available research only allows for a tentative answer. However, the
following account, based on eyewitness information, is a reasonable answer. The simi-
larities between Oddsson’s conception of the PM’s role and Hermannsson’s approach
are striking. In Oddsson’s case we find that he clearly saw the position and function of
the PM to be very different from the office of the Mayor of Reykjavík, a post he held
for 9 years (1982–91):
The job of the Prime Minister as a leader is completely different from that of the mayor. It is
mostly concerned with co-ordinating the views of individuals with different background. I do not
have the supreme power; I mean the direct power to command, which I possessed as mayor. Each
cabinet minister has full power over his department. Formally the Prime Minister can not command
him to do anything. Consequently my power as Prime Minister consists as influence rather than the
power to command, although both types clearly coexist. (Frjáls verslun nr.1 1996: 32–9)
In the same interview Oddsson also emphasized the importance of personal contact
with his cabinet ministers, of making himself accessible to them at all times. This
allows the PM to coordinate and influence. It also ‘forces’ cabinet ministers to consult
with the PM. Ministers cannot hide behind the cover of not being able to reach the PM.
The Oddsson cabinet does not use majority voting. Consensus had usually emerged
before cabinet meetings, while in the case of serious disagreement on jurisdiction
between ministers the PM made the final decision. Oddsson also made a sharp distinction
412 Svanur Kristjánsson
between the power of the PM as such and the position of the PM as his/her party
leader. The PM does not select the ministers from his partner(s) party(ies), nor is he
politically responsible for their conduct in office. In practice ministers are delegates of
their respective party, and they are not politically accountable to the PM. In his
conception of the PM’s powers and his operating style as PM, Oddson follows in the
footsteps of PM Hermannsson.
From the outside it appears to be the case that successful PMs displayed political
skills in three arenas. They were successful by (1) forming and maintaining coalition
cabinets, (2) preserving a strong position in the parliamentary party, and (3) winning
parliamentary elections. For Icelandic PMs, the first two of these skills can be the most
important. When they fail it is often because they lack in skills of coalition manage-
ment and maintenance.
The cabinet is not a collective body dominated by the PM, but rather a council of
highly independent ministers. The prevalence of coalition cabinets further complicates
the PM’s role, while at the same time absolving him of political responsibility for
ministers from his coalition partner(s). The successful PM adjusts to the formal and
informal rules of the game, not by establishing clear chains of delegation and account-
ability but by carefully learning on the job. Those who approach the PM’s role in purely
a formal fashion will quickly fail.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
In the context of delegating to civil servants, the comments of Charles Cobb, former
United States ambassador to Iceland, are worth citing:
Icelandic politics is both fast-paced and very personal in nature. In most countries political parties
are institutionalised organisations; a bureaucracy controls everything. As a result all things move
at slow speed; developments and changes take place at the speed of a snail. In Iceland policy is
deliberated for one evening and implemented the next day.. . .In the U.S. cabinet secretaries always
express the same view on issues of the day. Members of the government do not publicly disagree.
Great emphasis is placed upon this rule of conduct. Cabinet ministers have two choices; either
they speak on behalf of one common policy, the President’s policy, or they resign . . . The Icelandic
situation is very different; it is quite common for Cabinet ministers to publicly disagree on policy.
Such behaviour is considered acceptable so far as I can observe. (Morgunbladid, 9 Jan 1992)
The dominance of the minister over his/her ministry starts with his/her power to hire
civil servants. As Oddsson has put it: ‘The power of the individual minister in hiring
people is absolute.’ (Frjáls verslun 1996). Each minister also actively cultivates contacts out-
side the ministry, including ones with individuals, companies, and interest groups. All
ministers make themselves available for personal appointments. Anyone can request an
appointment to discuss something with a minister. Some people might be required to
wait a long time for such an appointment, but the waiting time is not as long as that for
seeing some medical specialists or getting an operation in a hospital! An individual might
even call a minister on the telephone, even at home. (He will find the number in the
phone book.)
Iceland 413
The argument is not that anyone can get a minister to do his/her bidding at any
time. What is important here is that ministers are highly accessible. Most people,
companies, and interest groups can get the minister’s attention. This access undercuts
the authority of the civil service hierarchy and decreases the autonomy of the civil
service. Ministers can oversee and control civil servants by cultivating the relevant
‘customers’ of public service.
Ministers have powerful incentives to nurture outside contacts. To start with, ministers
are by convention almost invariably also MPs and as such they are in constant danger
of not being re-elected. As noted above, the nomination stage is particularly hazardous
given the prevalence of open primaries in all but one of the major political parties. The
open primary is a personal vote. To survive under such circumstances, ambitious MPs,
such as the ministers, must emphasize constituency service.
The ministers run their departments and have full formal control of the civil
servants under their own jurisdictions. Civil servants, who are of only one type, that is,
career civil servants, are accountable direct to the minister. The greatest restrictions to
their rule come from outside the government departments. The unclear division of
power between the Althingi and the executive branch of government cuts both ways.
It enables ministers to play a crucial role in the preparation and interpretation of legis-
lation, but also allows for direct involvement of the Althingi in executive functions.
Sometimes the minister and the ministry work together in order to decrease the
scope of administrative discretion and strengthen bureaucratic autonomy. This is most
obvious in the case of decisions regarding the size and distribution of fishing quotas,
which are among the most important decisions made by government agencies at any
given time. Given their importance, the Department of Fisheries has issued detailed
and transparent rules to govern such decisions.
E X T E R NA L C O N S T R A I N T S
Parliamentary democracy became the basic constitutional principle of the Republic
established in Iceland and it remains dominant today. However, parliamentary government
has evolved in response to several factors external to it (or at least to its ideal-type).
From the beginning the parliamentary system was mixed with another constitutional
principle, that of semi-presidential government. The parliamentary system
established its identity partly in power struggles—first with the Regent and then with the
President. The political parties also contributed to the development of a coherent parlia-
mentary system with the Althingi as the pivotal institution. Nevertheless, the principles
of delegation and accountability were poorly developed within this system, because the
logic of Iceland’s patronage system embedded political relationships in a social system of
mutual exchange and favours. A number of important changes have occurred over time.
One change is that a new form of governance has emerged—that is, corporatism
without parties.
A very comprehensive corporatist agreement, termed The National Consensus, was reached
early in the year 1990. This consensus, which was initiated and formulated in great detail by the
414 Svanur Kristjánsson
employers’ association and the labour unions, included agreements on wage-raise policies. The
role of the cabinet was primarily to get the necessary legislation through parliament, pay the bill
from the budget, and keep the employees’ wage settlements in line with those included in the
National Consensus. (Kristjánsson 1998: 180)
This type of decision-making, which was repeated in 1998 when a new comprehensive
proposal on pension funds was approved in a similar procedure, pushes both the
Althingi and the executive branch to the sidelines of policy-making and allows interest
groups to dominate.
Parliamentary government is also affected by the constitutional provision of the
national referendums and by international factors such as NATO membership and the
country’s formal relationship with the EU, through the EEA. The Constitution auto-
matically calls for a national referendum under three sets of circumstances:
1. A resolution to unseat the President has been passed by a three-quarters majority in
the Parliament. The President is immediately stripped off his official power, but his
impeachment must be ratified in a national referendum. If the President wins the
referendum he reclaims full authority and the Althingi is immediately dissolved.
2. The President has refused to sign legislation passed by the Parliament. In this case
the legislation goes into effect, but must be ratified in a national referendum. This
law is rescinded if a majority of those participating in the referendum vote against it.
3. Parliament has passed a proposal changing the status of the national state
Lutheran church. This decision must be put to a vote among the whole electorate.
The Althingi alone can initiate national referendums other than in these three situ-
ations. It can simply attach a clause to any law that it passes stating that the law’s com-
ing into effect is contingent upon its acceptance in a national referendum.
In the history of the present Constitution, no national referendum has been held.
This should not, however, lead to the conclusion that national referendum is insignificant
in the Icelandic political system. First, the constitutional provisions with regard to the
holding of referendums clearly bypass the chain of delegation and accountability char-
acteristic of the ideal model of parliamentary government. Through national referen-
dums, the power to resolve fundamental conflicts between the Althingi and the
President is vested with the people. Second, from time to time demands are raised for
national referendums to settle controversial issues. Such demands have been voiced
inside and outside the Althingi. These demands often include references to the fact that
the national referendum has been used five times in Iceland. It was used twice (in 1918
and 1944) to decide Iceland’s relationship with Denmark and to approve the basic con-
stitution. It was used in 1908 and 1933 to settle the issue of whether or not to prohibit
alcohol. Finally, in 1916 a proposal for compulsory civic work for young men was
defeated in a national referendum. Given the rather large gap that now exists between
the formal Constitution and constitutional practice, it is possible that a national
referendum has the potential to shift the balance of power between the parliamentary
majority and the President.
Another issue that has affected the country’s democratic chain of delegation and
accountability is the NATO base on Iceland. Iceland occupied a strategically crucial
Iceland 415
position during the Cold War. The country’s political leaders quickly developed great
skill at turning this position into dollars and cents. The US and other NATO govern-
ments used economic assistance and favours to facilitate Iceland’s acceptance of the
stationing of US troops on Iceland in the face of strong domestic resistance. Iceland’s
political leaders also successfully pursued the goal of repeatedly extending Iceland’s
fishing limits, from 12 miles in the late 1950s to 200 miles in 1975. In this struggle
Iceland ‘defeated’ both Great Britain and West Germany, both of whom lost their
traditional rights to fish in Icelandic waters. The mobilizing of the Icelandic population
was important for securing this outcome, but so was the government’s masterful
manipulation of ‘the American connection’. The government hinted at a reconsideration
of its NATO membership and the defence agreement with the United States. The US
promptly used its political leverage to help settle the dispute.
More recently, Iceland’s membership in the EEA has created an important environ-
mental constraint on its political system. Iceland signed the treaty without changing its
Constitution and without holding a national referendum. The government argued that
the agreement did not curtail Iceland’s sovereignty or change the constitutional status
of its national institutions. Despite these claims, the position of the country’s political
institutions has indeed been altered insofar as Iceland must fulfil obligations established
by the agreement but has no real influence in EU policy-making. In fact, Icelandic
legislation must conform to the rules and regulations established by the EU in all areas
covered by the EEA agreement (see Kristjánsson and Kristjánsson 2000).
Prior to the adoption of the EEA agreement, in 1991, pressure from the European
Commission of Human Rights and the European Court of Human Rights played a very
important role in the decision of the Althingi to change the judicial system in order to
separate the executive and judicial role of the local magistrates. Environmental ‘pres-
sure’ thus helped to bring about the most fundamental revision of the Icelandic judicial
system since 1874. In connection with this, Icelanders increasingly have both the means
and the will to seek redress for their grievances against the domestic institutions with actors
and institutions outside of the nation state—often by emphasizing their rights as estab-
lished by Iceland’s membership in European and international organizations and insti-
tutions. The increasing concern for codified citizen rights is also evident in the 1995
constitutional change, adding fifteen new clauses covering the full range of human
rights. The judiciary, notably the Supreme Court, has in several cases in the 1990s ruled
against the acts of the Althingi and the Cabinet on Constitutional grounds.
C O N C LU S I O N
One can fruitfully place the parliamentary system of Iceland in two contexts: the
domestic level and the environment outside the boundaries of the nation state. On the
domestic level, clearer delegation has developed within the Icelandic parliamentary
system, in particular due to the sharper definition of the Althingi as the principal insti-
tution which, in turn, delegates power to the executive branch. At the same time, the
parliamentary system has lost much of the coherence previously provided by patronage-
oriented political parties. The political system has thus become more complex.
416 Svanur Kristjánsson
Organizationally weak political parties survive in the electoral arena, partly by
abnegating their collective responsibility for offering the voters a choice between clear
alternatives. Individual MPs, on the other hand, are now more directly accountable to
the voters.
The increasing accountability of individual MPs, and clearer definition of the
Althingi as separate from the executive, is more than matched by countervailing devel-
opments. Of particular importance in this regard is greater pluralism in society as a
whole as various actors—including firms, professionals, interest groups, government
institutions, mass media—have gained considerable autonomy from (party) politics.
Neither individual MPs nor political parties prepare or initiate the most important laws
passed by the Althingi. The government bureaucracy and specialists play a key role in
the political system. A clear example of this is the fact that the preparation and enact-
ment of the state budget is predominantly in the hands of the cabinet, particularly the
Ministry of Finance.
With regard to the environment beyond the nation state, the picture is rather clearer.
Political leaders in Iceland are subject to increasing constraints at a time when they have
fewer opportunities than before to ‘milk’ external relations for the country’s benefit as
well as their own. Other Icelanders—as individuals and members of groups or firms—
share these constraints to some degree, but they have also gained considerable leverage
on Icelandic politicians. Icelandic citizens now have better chances of redressing
grievances against their own government by external means.
Now that the Cold War has ended, the EEA agreement is Iceland’s most important
treaty. EEA membership involves costs and benefits for the country as a whole. From
the point of view of the average citizen, the opportunities clearly outweigh the new
constraints. For rulers, the picture is more mixed. Their opportunities to participation
in and exert influence on decision-making at the European level are extremely limited,
while at the same time the constraints they are subject to are considerable. They face
both direct constraints due to the EEA agreement and other international agreements
and indirect ones because of the new opportunities citizens have to make their own
government more accountable by making use of the external commitments binding
the Icelandic nation state.
REFERENCES
Bergman, Torbjörn (1995). Constitutional Rules and Party Goals in Coalition Formation. Department
of Political Science. Umeå University.
The Economist (1997). February 8–14.
Frjáls verslun (Icelandic journal) (1996): 1, 32–9.
Gallagher, Michael and Marsh, Michael (eds.) (1988). Candidate Selection in Comparative Perspective.
London: Sage.
King, Anthony (1976). ‘Modes of Executive-Legislative Relations: Great Britain, France and
Germany’. Legislative Studies Quarterly, 1: 37–65.
Kristinsson, Gunnar H. (1994). Embættismenn og stjórnmálamenn. Reykjavík: Heimskringla.
–––– (1996). ‘Parties, States and Patronage.’ West European Politcs, 19(3): 433–57.
Iceland 417
Kristjánsson, Svanur (1998). ‘Electoral Politics and Governance: Transformation of the Party
System in Iceland, 1970–96’, in Paul Pennings and Jane Erik Lane (eds.), Comparing Party System
Change. London: Routledge.
—— and Kristjánsson, Ragner (2000). ‘Delegation and Accountability in an Ambiguous System:
Iceland and the European Economic Area (EEA)’, in Torbjörn Bergman and Erik Damgaard
(eds.), Delegation and Accountability in European Integration: The Nordic Parliamentary Democracies
and the European Union. London: Frank Cass.
Morgunbladid (Icelandic daily newspaper) (1992): 9 January.
Strøm, Kaare (1986). ‘Deferred Gratification and Minority Governments in Scandinavia’.
Legislative Studies Quarterly, 11: 583–605.
Styrkársdóttir, Audur (1986). ‘From Social Movement to Political Party: The New Women’s
Movement in Iceland’, in Drude Dahlerup (ed.), The New Women’s Movement. London: Sage.
—— (1998). From Feminism to Class Politics. Department of Political Science. Umeå University.
13
I N T RO D U C T I O N
Given that the political histories of Britain and Ireland are inextricably linked it is
perhaps not surprising that the governing institutions of independent Ireland are an
adaptation of the Westminster system, and that both countries continue to have much
in common. On the other hand, the new Irish state was established by revolutionary
secession from the United Kingdom and the circumstances of its birth helped shape its
institutions and political parties.
Sinn Féin, which had won a resounding victory at the 1918 election to the Westminster
Parliament (taking 73 of the 105 seats on the island of Ireland), refused to attend that
Parliament and instead established in 1919 its own revolutionary assembly, Dáil Éireann. The
ensuing war of independence from Britain resulted in a military truce and subsequently
‘Treaty’ between the two sides during 1921. The terms of the Treaty, which ‘went much
further than conceding Home Rule, but stopped well short of permitting complete separa-
tion’ (Coakley 1999: 16) led to a bitter civil war (1922–3) that the anti-Treaty forces lost. This
division is of more than historical interest since the pro- and anti-Treaty sides constituted
the origins of the two parties that came to dominate independent Ireland, Fine Gael and
Fianna Fáil, respectively. Thus the new Irish state was created in 1922 by a partial secession
from the United Kingdom that resulted in a partition of Ireland into two political entities.
The new state was conceived as a parliamentary democracy with many features such
as the organization of Parliament, procedures for law-making, and the administrative
structure clearly inherited from Westminster. Nevertheless, Irish political institutions
were designed as a variation on rather than a carbon copy of the Westminster model.
The most important institutional divergence from Britain was the adoption of limited
government by a written Constitution (first in 1922 and superseded by the current 1937
Constitution) protected by a Supreme Court and developed by judicial interpretation
and popular referendum. The 1937 Constitution was enacted by the people in a refer-
endum as a demonstration of popular sovereignty.1 Other important institutional
1
Ireland, or rather southern Ireland, officially became a republic in 1948 but is henceforth referred to for
simplicity as ‘Ireland’.
Ireland 419
2
Note that the electoral system is protected by the Constitution and thus can only be changed by refer-
endum. The electoral system and presidency are discussed in sections below.
3
Any citizen can test the constitutionality of any law, treaty, or action of the Parliament and government
by bringing a case before the high court, or in an appeal against its decision, before the Supreme Court. The
citizen must establish, however, that she has locus standi—that she is in someway affected by the action or law
(Gallagher 1999a: 83). Gallagher (1999a: 83) comments that ‘in Ireland, judicial review has proved to be the
main method by which the constitution has been developed’. In addition, the President (Article 26) can elect
to refer a bill passed by the Oireachtas directly to the Supreme Court to test its constitutionality rather than
sign it into law. Of course while judicial interpretation often serves valuable functions the dilemma is that
judges of the Supreme Court are not accountable to anybody.
4
In 1999 the voters in a referendum approved of a constitutional amendment recognizing the existence
of local government and requiring local elections at least once every 5 years (Gallagher 1999a: 81).
420 Paul Mitchell
The bill is formally introduced in one of the houses which constitutes the first stage of
the legislative process. The second crucial stage is the debate on the general principle
of the bill. This is the main partisan event that largely determines whether the bill will
be passed. The third stage is a detailed examination by the relevant committee at
which amendments that do not conflict with the general principle of the bill may be
proposed.5 The fourth stage is the report stage at which further new amendments can
be made. The final fifth stage is the formal passing of the bill, which is then sent to the
other house where the same stages are repeated.6
Although most (non-money) bills can be introduced in either house, the Dáil and
Seanad are in no sense equal. While some powers are shared (e.g. the impeachment of
a President), in law-making the Dáil is pre-eminent. In essence the most that the Seanad
can impose on an ordinary bill is a delay of 90 days. If the Seanad fails to pass a bill that
has already passed through the Dáil or proposes amendments unacceptable to the
lower house, the Dáil prevails. Following the delay the bill is ‘deemed to have been
passed by both Houses of the Oireachtas’ (Article 23.1 of the Constitution). In the case
of money bills (the Chairman of the Dáil determines what constitutes a money bill) the
Seanad’s powers of delay are reduced to just 21 days. If the Seanad rejects a bill and the
Dáil overrules its objections the Seanad can invoke the ‘Article 27 procedure’ (Gallagher
1999b: 200), whereby a majority of the Seanad plus one-third of the Dáil can petition
the President not to sign the bill on the grounds that it ‘contains a proposal of such
national importance that the will of the people thereon ought to be ascertained’
(Article 27.1). If the President accedes to such a petition the bill cannot become law
until either it passes a referendum or the Dáil passes it again following a general
election (Article 27.5.1). However, no Article 27 petition has ever been presented to the
President, partly because the government usually controls a majority in the Senate.7
Given that the Senate is relatively powerless and rather bizarrely composed, some
parties, such as the Progressive Democrats, have called for its abolition or reform
(e.g. see Lynch 1996; Laver 1996b; Coakley and Manning 1999).
As in other parliamentary democracies the formal position that the legislature makes
laws and the government executes them, elides the reality that the government (at least
a majority government) largely controls the legislature. Most bills are government bills
5
In practice the relevant minister decides whether a proposed amendment is consistent with the principle
of the bill. Since committee assignments are proportionate to party strength, majority governments could in
any case out vote the opposition on any committee. In addition, committee decisions can be overturned by
the Dáil (Gallagher 1999b: 196).
6
Any member may demand a quorum which is twenty TDs. The decision rule is simple plurality (50 % + 1
of those voting). Actually in practice it is 50 % since in the event of a tie there is a strong
convention that the Ceann Comhairle (Speaker) will vote with the government.
7
Government control of the upper house is facilitated by the fact that an incoming Taoiseach (following
an election) nominates eleven of the Seanad’s sixty members. The importance of this prerogative was
demonstrated during the 1994–7 government which was the first time that one coalition replaced another
without an election. This meant that the new government formed in 1994 inherited the eleven ‘Taoiseach’s
nominees’ of the previous government, so that the combined opposition had a majority in the Seanad. ‘The
result was a major change in the manner in which Senate business was conducted . . . The government suf-
fered two defeats on legislation and avoided defeat on other occasions by either conceding on issues or by
postponing them altogehter’ (Coakley and Manning 1999: 200).
Ireland 421
and certainly only bills supported by the government have much prospect of being
enacted. In Ireland the government’s privileged position in the legislature has constitu-
tional foundation through Article 17.2 that states that no provision involving revenue or
spending can be passed by the Dáil unless it has been recommended by the government
in a message signed by the Taoiseach. This stipulation is ‘clearly motivated by the fear that
were it not in force, parliament might vote for the spending of money but against
government efforts to raise it’ (Gallagher 1999b: 187–8).
Apart from an ‘Article 27 referendum’ any proposal to amend the Constitution must
be initiated in the Dáil and passed (or deemed passed) by both houses before being sub-
mitted to the people for a decision in a referendum.8 Referendums are decided by a
majority of valid votes cast. The frequency of referendums has been increasing: while
only two occurred before 1968, twenty more took place between 1968 and 1998, aver-
aging about one a year in the 1990s (see Gallagher 1996a: 90). Most referendums have
fallen into three broad categories: institutional issues (for example electoral laws);
European integration (accession to EC, ratification of the Single European Act,
Maastricht and Amsterdam Treaties); and finally, moral issues (abortion, divorce).
Electoral behaviour during referendum campaigns also appears to be progressively less
determined by the positions of the political parties (Gallagher 1996a). Interest groups
and well-organized single issue movements have dominated some recent referendum
campaigns, with parties playing a relatively low-key role, sometimes because they
themselves are internally divided. However, on controversial, especially ‘moral’ issues
referendums may help insulate rather than undermine the party system (Sinnott 1995;
Gallagher 1996a). In sum, the referendum device is an important but occasional paral-
lel chain of delegation and accountability, which despite having some life outside of
parties is nevertheless structured by them since only Parliament can authorize a refer-
endum and the government determines the wording of the question. In all other areas
of parliamentary and executive behaviour, parties and especially their leaders play the
dominant role.
P O L I T I C A L PA RT I E S
Political parties are neither presumed nor protected by the Constitution. While various
offices and institutions are constitutionally mandated, including an Uachtarán
(President), Taoiseach (PM), Tánaiste (deputy PM), and Dáil, political parties are not
even mentioned. This fact and the relative lack of statutory regulation mean that Irish
political parties have been essentially private organizations, though there are some
signs that this is beginning to change due to new legislation designed to regulate party
finances. The state plays no role in regulating party membership, candidate selection,
the internal organization, or the election of party leaders. General election campaigns
8
Thus citizens cannot initiate referendums. Although Parliament (or in reality the government) initiates
and determines the wording of the proposed ammendment to the Constitution, ‘parliament does not vote
on the substance of the proposal, only on the question of whether the proposal should be put to the people’
(Gallagher 1996a: 88).
422 Paul Mitchell
must be a minimum of 3 weeks and a maximum of 4 weeks. As of 2003, there are
no restrictions on the publication of opinion polls during elections, though sev-
eral governments have considered restrictions. The state does play some role in
providing financial subsidies and access to election broadcasting. While Ireland intro-
duced a more explicit system for the state financing of political parties in the 1990s, it
has indirectly financed parties through subsidies to parliamentary groups (block grants)
and individual deputies (travel expenses, secretarial, and research assistants). By 1989
David Farrell (1994: 235) estimated that the total amount of state aid for deputies and
their parties (excluding broadcasting) was about £4 million. A series of corruption scan-
dals in the 1990s stimulated concern about the private financing of parties and led to
The Electoral Act 1997, a piece of legislation that Laver and Marsh (1999: 157) suggest is
‘destined to have a fundamental impact on party politics in modern Ireland’. Essentially,
much more public money will be given to parties and candidates (based on their past
and present electoral performances, but weighted in favour of opposition parties) in
order to reduce their reliance on anonymous donors. ‘The 1997 Act relies entirely upon
making political donations transparent, rather than upon regulating either the size or
the source of acceptable donations’ (Laver and Marsh 1999: 157). The new law also
places clear limits on how much individual candidates and parties can spend during
elections. In addition free political commercials on national TV during elections repre-
sent an important hidden public subsidy to parties. These are allocated to parties
proportionally based on votes in the previous election.9
Irish parties are generally very cohesive and disciplined and in most respects are
increasingly centralized under the direction of the party leaderships (Gallagher 1985;
Mair 1987; D. Farrell 1994; Laver and Marsh 1999). Candidate selection is a partial
exception since ordinary party members are reluctant to cede their only real power, but
even here the party head offices exert considerable influence by promoting certain
candidates or imposing an additional candidate on the constituency (more below). In
general though, most powers in the parties lie with the leadership and to a lesser extent
the parliamentary party. In all parties members of the parliamentary party can hope to
influence policy and in most cases they elect the party leader.
For most of its history Irish politics has been dominated by three parties: Fianna Fáil,
Fine Gael, and the Labour party. Much confusion results from the difficulty of assigning
Fianna Fáil and Fine Gael to the standard ideological families of European politics.
Unfortunately there is no easy answer. In comparative terms both parties are best
regarded as parties of the centre or centre-right. They are not, however, clearly
members of the two principal centre-right families in Europe, the secular conservatives
or the Christian Democrats (see Mair 1999: 129–31). Content analyses of party mani-
festos and expert surveys suggest that there has been considerable movement between
these two parties and they have even on occasion appeared to leapfrog each other in
9
In 1965 (the first televised election) free broadcasting time was allocated to parties with at least seven
TDs. In 1977 the allocation rule was changed to parties or groups with at least seven candidates, though the
amount of time is based on votes at the last election and the geographical distribution of candidates (Farrell
1992). Parties cannot purchase additional time.
Ireland 423
ideological space (see Mair 1986, 1987; Laver 1992, 1994, 1998; Laver and Hunt 1992;
Benoit and Laver 2003). However, most commentators would probably rank the parties
from Left to Right as follows: Sinn Féin, the Greens, Democratic Left,10 Labour, Fianna
Fáil, Fine Gael, and the Progressive Democrats as the most right-wing party.
Of these seven parties only five have ever been in government (exceptions are Sinn
Féin and the Greens) and three—Fianna Fáil, Fine Gael, and Labour—have dominated
government tenure.11 The history of the post-1945 party system is conveniently divisible
into three periods: before 1973, 1973–89, and after 1989 (Mitchell 2000b). Before 1973
Ireland had a multiparty system in which the typical election produced a single-party
(though quite often minority) government. Fianna Fáil governed for most of these
decades, including two separate periods of continuous sixteen-year rule (1932–48,
1957–73).
During the second period (1973–89) the logic of competition remained Fianna Fáil
versus ‘the rest’, but ‘the rest’ consisted of just two parties (Fine Gael and Labour).
Thus, during these sixteen years Fianna Fáil single-party governments alternated with
‘the coalition’. The third period began in 1989 when Fianna Fáil transformed the
bargaining environment by entering its first executive coalition, ceding two cabinet
seats to the Progressive Democrats. Since that time Ireland has been governed by six
successive coalitions, thus breaking the earlier pattern of alternating single-party and
coalition governments. Indeed, until 2002, no government since 1969 had been elected
to a second successive term so that alternation, and coalition reshuffles, have become
the norm in the 1990s. From 1948 to 1997 there have been eight single party govern-
ments and eight coalitions. While Ireland has some history of minority governments
(six in the same period accounting for 27 per cent of duration), the recent pattern of
coalitions replacing coalitions suggests that the incidence of minority administrations
will decline, since most coalitions build towards a majority.12
The primary constraint on all parties and leaders in the parliamentary and governing
arenas are the legislative resources at their disposal, that is, the Members of Parliament
that they can count on.
D E L E GAT I O N F RO M VOT E R S TO M P s ( T D s )
Ireland’s candidate-centred electoral system can be expected to influence the incentives
and loyalties of deputies (TDs) in a manner that may attenuate the agency losses we
normally expect to characterize the relationship between voters and their representat-
ives. Electoral institutions by definition structure the delegation relationship between
10
In early 1999 Democratic Left ceased to be a party and merged with the Labour Party. This consolidated
party has retained the name the Labour Party.
11
Indeed in the period 1973 – 97 the Labour party was in government longer than any other. The
figures for tenure in government in days (and years) are as follows: Labour, 4,810 days (13.4 years); Fianna
Fáil, 4,330 (12); Fine Gael, 4,155 (11.5), Progressive Democrats, 1,183 (3.3); Democratic Left, 892 (2.5). Note
the PDs were only founded in 1985 and DL in 1992.
12
Minority coalition governments do occur and were formed in 1948, 1981, and 1997. Nevertheless,
83 per cent of coalition duration (1948–97) has been in majority governments (Mitchell 1999b: 252).
424 Paul Mitchell
voters and Members of Parliament (Mitchell 2000a). In most systems of proportional
representation each voter’s principal decision is to choose between rival party lists.
Although many countries’ list systems have intraparty preference voting, electors
nevertheless usually vote first for the party, and their vote may help elect an individual
whom they oppose. The central feature of Ireland’s electoral system (proportional
representation by single transferable vote (STV)) is that the electorate votes directly for
individual candidates in multimember constituencies.
The adoption of STV in Ireland owes much to the electoral reform debate in Britain
in the nineteenth century. Reformers such as John Stuart Mill and Thomas Hare
opposed strong political parties, and for this reason ‘the single transferable vote had for
them a particular attraction in being a personal and not a party system of voting’
(Carstairs 1980: 194). Similar reservations about the role of political parties existed in
Ireland with the result that the party affiliation of candidates was not even listed on the
ballot paper until 1965 (Sinnott 1999: 100).13 Thus there was (and is) a normative prefer-
ence amongst the electorate for a pattern of (small district) constituency representation
similar to the Westminster plurality system, but without the latter’s pronounced
disproportionality at national level. This is not to suggest that the rules governing the
electoral system are uncontentious. Indeed, Fianna Fáil single-party governments have
twice attempted to replace STV with the plurality system. On both occasions (1959 and
1968) the proposed electoral reform was rejected in a popular referendum, the second
time quite decisively (60.8 per cent against, 39.2 per cent for). While there are still peri-
odic calls for electoral reform, there is little prospect of any fundamental change, espe-
cially after the Constitutional Review Group (Whitaker et al. 1996) did not recommend
changing the electoral system. Thus STV is likely to be retained.
Two principal aspects of STV are of importance to delegation and accountability
between the electorate and their representatives. First, it is a preferential electoral
system in which voters have the opportunity to rank individual candidates in consti-
tuencies with a very small district magnitude (since 1947 between three and five seats).14
This typically means that only two or three of a major party’s candidates have much
chance of being elected and that the voters alone decide which of the party’s candidates
are successful. A frequent result is intense intraparty competition at all levels of electoral
campaigning and candidate selection. From the perspective of individual candidates the
electoral threat from intraparty competition is tangible and severe. For the 1948–77
period Carty found that on average 30 per cent of all seat turnovers at general elections
were intraparty, with incumbents losing to party colleagues (Carty 1983: 115).15 This
13
Antipathy to political parties clearly was not the only or main reason for the largely uncontested adop-
tion of STV in Ireland. The British Proportional Representation Society proposed STV as a way of easing the
Home Rule crisis, arguing that it would guarantee the representation of minorities. Irish Nationalist leaders
were persuaded of the merits of STV over first-past-the-post and in any case tended to equate PR with STV
since there was a ‘profound ignorance’ concerning list system PR (Gallagher 1987: 27).
14
TDs are elected for 5 years. There are no term limits or mandatory retirement age, though in principle
a member could be expelled for corruption, electoral offences, or bankruptcy.
15
Gallagher has re-examined this threat at the 1987, 1989, and 1992 elections. Of the 33 Fianna Fáil incum-
bents that were defeated twenty were displaced by party colleagues. In Fine Gael ten of the thirty-three
Ireland 425
provides incentives towards constituency service since three or four candidates of one
party (with limited ideological differentiation feasible) compete for the two or three
seats that the party can realistically hope to win.16
The second relevant feature of STV is that preference voting is not limited to an
intraparty choice: voters can (and do) vote across party lines. As Gallagher (1987: 41)
says: ‘[u]nder list systems with preferential voting, deputies would need to concern
themselves only with the support of those who will, or might, vote for their party,
whereas under STV even the fifth or sixth preference of a supporter of another party
could be important, so that deputies need to be concerned about their reputation in every
voter’s eyes’. This creates incentives for candidates to try to attract lower preferences
from partisans of other parties, and clearly has important consequences in terms of
democratic accountability.
One of the sophisticated (and unique) aspects of STV is that votes cannot help an
individual candidate unless they explicitly express a preference for him or her (Gallagher
1988: 128). Thus the electorate can easily punish an incumbent deputy without having
to abandon the party.17 Prominent politicians can and do lose their seats. Leading politi-
cians from the smaller parties are particularly vulnerable: at the June 1997 election, one
cabinet member and three junior ministers lost their seats and one party leader only
survived on the final count. At election time there is nowhere to hide. There is nothing
equivalent to the ‘safe’ spots near the top of a party’s list that exist in non-preferential
electoral systems.
Candidate Selection
The individual accountability of incumbent deputies at elections, enhanced in Ireland
by the voters’ ability to desert an individual without abandoning the party, is one insti-
tutional control that should help contain agency losses. With rival party colleagues
active in their constituencies and after their seats, deputies cannot afford to stray too far
from the interests of their constituents, the ultimate principals.
However, the primary mechanism for containing agency losses in the delegation chain
from voters to representatives is clearly political parties and partisanship (Müller 2000). It
is difficult though not impossible to get elected without the support of an established
defeated TDs lost to running mates. Thus, the risk of being ousted by a party colleague (if you belong to
Fianna Fáil or Fine Gael) is real. This very rarely applies to other parties since they typically run single
candidates in each constituency. In the same elections no incumbents of other parties were defeated by
a running mate (Gallagher 1996b: 510).
16
The electoral incentive towards candidate differentiation does not of course necessitate that
constituency service be the method of competition. Candidates could compete as effective legislators,
ministers, or take up distinctive ideological positions if these were thought highly valued by voters.
17
Assuming that the party has nominated more than one candidate. An important (and often overlooked
point) is that the degree of choice in an STV electoral system crucially depends on nomination strategies.
For example, since the smaller parties typically nominate a single candidate in each constituency there is no
intraparty choice. A substantial degree of choice depends on parties nominating more candidates than the
seats they expect to win.
426 Paul Mitchell
party. Independents and minor party candidates are elected at every election—which is
unusual in Europe—though in fairly small numbers. For 1944–97 an average of 6.8 such
deputies were elected at each election. For the more recent period between 1981–97,
on average, 3.5 per cent of parliamentarians were independents or minor party candi-
dates. Thus 96 per cent of all members of the Dáil belonged to the established parties
(Fianna Fáil, Fine Gael, Labour, Democratic Left, and Progressive Democrats).
Hence, selection as a candidate by one of the main parties is the primary route to
a political career. The choice of candidates that the electorate have the opportunity to
reward or punish is largely predetermined by the political parties in the ‘secret garden’
of candidate selection (Gallagher and Marsh 1988). In practice candidate selection is
entirely a matter for the parties. There are no constitutional or legal provisions relating
to candidate selection for general elections.
The main Irish parties have broadly similar methods of candidate selection. First,
just like most parties in Europe candidate selection is a power retained by the parties
and their members: there are no primaries in which ordinary voters play a role. Second,
candidates are chosen at constituency level by selection conventions composed of dele-
gates representing the constituency party branches. Thus, a considerable number of
party members are involved in selection through their role as convention delegates, the
number depending on the size of the party. For example, Gallagher (1988: 121) reports
that the average Fianna Fáil convention has about 220 delegates (varying from 40 to 440),
whereas the Labour Party average has about fifty delegates (varying from 20 to about
200).18 Selection is by election, with most parties (though not Fianna Fáil) using STV.
However, in 1996 Fine Gael introduced an important reform whereby all members in
a constituency of at least 8 weeks’ standing (who belong to a branch that has been reg-
istered for at least 3 months) could participate and vote at candidate conventions. The
introduction of ‘one member, one vote’ by Fine Gael stimulated competitiveness,
increased attendance, and generally enlivened selection conventions for the 1997
election. For example, in the party leader’s constituency of Meath, eight candidates
sought the three available places on the party ticket, and 845 members (out of almost 950
eligible) actually voted (Galligan 1999: 63).
18
Differences of detail are as follows. In Fianna Fáil selection conventions are held at which each branch
can have three delegates. Election is by an elimination system for each place to be filled. The national leader-
ship can add or delete a candidate. From 1945 until 1996 Fine Gael also used a branch delegate system in
which each branch was entitled to a number of delegates proportionate to its size (see above). In the Labour
Party, selection conventions are held in which each branch is entitled to four delegates and voting is by STV
secret ballot. The Progressive Democrats’ generally use one-member-one-vote but their procedures are flexible
so that they occasionally use branch delegate conventions, as they did in 1997 in Limerick East (Galligan 1999:
61). Voting is by STV secret ballot. The Greens allow all ordinary members in a constituency to participate
in candidate selection by a postal vote.
Ireland 427
reinforcing aspects. First, the one personal quality that a prospective candidate must
have is strong local connections. Irish electoral behaviour is quite different from the
pattern in Britain where parachutage is widely practised.19 In Ireland, a Dublin candidate
standing in Kerry or Donegal would do well to save his or her deposit. No party would
select a candidate without a strong local base. Thus, it is not really possible to test
whether voters insist on strong local connections, because the parties hardly ever select
candidates without them (Gallagher 1988: 130). What is clear is the selectors’ strong
preference for local candidates, which is not entirely surprising. The selectors want to
reward members of the local organization, and the national party largely acquiesces
since it is in their interest to have local party workers campaigning enthusiastically for
the party’s candidates. Thus, most prospective parliamentary candidates have assidu-
ously endeavoured to build a local political machine, one aspect of which is holding
local office. Of deputies elected in 1997, 74 per cent were members of a local authority
before their initial election (O’Sullivan 1999: 191). Indeed, virtually the only way of
acquiring a local pedigree without building a strong local base is to ‘inherit’ the seat
from a relative, quite a frequent practice in Ireland (Gallagher 1988: 139).20
Second, in electoral systems featuring multimember constituencies, the selectorate
often attempt to ‘balance the ticket’ according to various criteria such as gender, race,
locality, or ideological persuasion. In Ireland, it is well known that selection conven-
tions attempt to balance the geographical dispersion of their candidates (Marsh
1981a,b; Gallagher 1988). If a party selects three candidates it is highly unlikely that it
will pick three individuals whose bases are all in the north-west corner of the con-
stituency—even if they are the ‘best’ candidates. Parties attempt to geographically dis-
tribute their candidates throughout the constituency both to maximize their vote, and
to contain negative aspects of intraparty competition stimulated by two or more party
candidates concentrating on the same part of the constituency.
While the candidates are mostly selected locally by the convention, all the national
party leaderships reserve the right to determine the number of candidates nominated
and the right to veto or add to the list of candidates selected locally. In practice the
parties rarely veto the local selection of candidates since this would lead to considerable
resentment. Increasingly, however, the central leadership will add a candidate to
those selected locally if it believes that doing so provides a competitive advantage. Since
incumbents’ seats are most vulnerable to party colleagues, some deputies attempt to
lessen the danger by having fewer and weaker candidates selected, a practice known as
‘quota-squatting’. Of course, such defensive tactics are not in the interests of party leaders.
Therefore, they increasingly add an additional candidate if they believe the party’s
constituency ticket to be too uncompetitive (or if one of their favoured candidates
19
Parachutage is the French term for the imposition of non-local candidates by the national party leader-
ships. Indeed it is Britain that is exceptional in that it is the only country in Europe where many candidates
were ‘outsiders’ when they were first selected.
20
Of the 166 deputies elected in 1992, 39 (23 %) were related to a present or former TD. More
significantly, in 26 cases (16 %) the relationship was judged sufficiently close to have made a difference (such
as a son or daughter taking over a seat from a parent) (Gallagher 1993c: 73). This pattern was repeated in 1997
(O’Sullivan 1999: 190).
428 Paul Mitchell
failed to get selected). Even then, however, party leaders attempt to do this sensitively,
and usually add a local candidate, often someone who was unsuccessful at the selection
convention. The nature of STV facilitates the leadership’s intervention. As Gallagher
(1988: 129) points out, ‘Under STV the national executive can add a name without
displacing anyone’ or pushing anyone further down a list. All of a party’s candidates are
formally equal.
The fact that candidate selection is largely controlled locally might suggest that
deputies are free to do as they please as long as they look after local interests, a delega-
tion relationship that is principally between voters and individual TDs. And certainly
there is an element of truth to this: some rebels can survive sanction by the national
leadership if they remain locally popular (see footnote 21). Nevertheless, Irish voters
are primarily party voters, and loyalty to a party usually outweighs local orientations
(as evidenced by transfer patterns; see Gallagher 1977; Marsh 1981b; Carty 1983;
Sinnott 1995). Even if a local TD has built up an elaborate personal election machine,
he or she will often find that loyalty is contingent on remaining within the party
(Gallagher 1988: 133).
Party discipline in the legislature is strong. Apart from deputies’ general commitment
to their parties (which should not be underestimated), they know that promotion will
be denied if they oppose the leadership and that if they repeatedly vote against the
party line in Parliament they will lose the party whip. Persistent rebellious behaviour
could lead to deselection or expulsion from the party, in which case the deputy loses his
primary electoral asset, the party label. Moreover, incumbents know that deselection will
most likely end their political careers. Mair (1987: 67) states that ‘a survey of the rela-
tively poor electoral results of candidates who leave established parties and attempt to
win seats as independents suggest that there is not much life outside party, and
that the prior success of these individuals derived mainly from their original party
affiliation’.21
In summary, localism, while important, usually does not override partisanship.
Loyalty to party results in disciplined party voting in Parliament and structures
electoral behaviour. Thus, the chain of delegation is mixed: while it primarily flows
between voters and parties, the nature of STV means that voters can punish individuals
without deserting party.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
While the legislatures of parliamentary democracies are officially sovereign, in practice
legislatures and executives are not even equal. In most countries majority governments
with disciplined parties are in an ascendant position to Parliament. In terms of its rela-
tionships with the executive branch, the role of the legislature can be reduced to two
21
Counterexamples are few and therefore relatively exceptional. In 1997 two prominent politicians from
the largest parties (Michael Lowry and Jackie Healy-Rae) stood as independents after they were denied
nominations. Both of them topped the polls in their respective constituencies. Thus, while this is unusual, the
election of independents in Ireland means that there may be some life after party.
Ireland 429
main sets of functions. The first are the institutional powers of the legislature in the
‘making and breaking’ of governments (Laver and Shepsle 1996). The defining charac-
teristic of parliamentary democracies is that the government is at least accountable to
the legislature. The cabinet is linked to the legislature by a series of ‘tests’, investiture
votes, confidence motions, and dissolution powers, by which a parliamentary majority
can select or terminate a government at will (assuming that Parliament is in session).
Of course in many countries the legislature has the important ex ante screening power
that cabinet members must be members of the legislature. In Ireland all cabinet
members must be members of one of the houses of Parliament and no more than two
can be from the Senate. In addition the Constitution stipulates that the Taoiseach,
Tanaiste (deputy PM), and Finance Minister must be members of the Dáil (Article 28.7.1).
In practice, all cabinet members are selected from the Dáil, the last exception occurring
in 1981. Of course these ‘hire and fire’ powers have an all or nothing quality that
provide little opportunity to influence ongoing policy-making. Thus the second role of
Parliament is to monitor, investigate, and generally exercise legislative oversight of the
executive powers of government.
22
The quorum is twenty and the vote is won by a candidate securing a majority of those voting.
430 Paul Mitchell
allow further time for negotiations between the parties. After 2 weeks (and one more
unsuccessful meeting of the Dáil) a government was finally elected, the first ever
Fianna Fáil-led coalition government.
It is generally expected that once formed a government will govern without much
‘interference’ from Parliament. Of course all governments must be able to win periodic
votes in Parliament. But while legislative defeats might be an irritation, only lost confid-
ence motions are terminal. In Ireland the Constitution (Article 28) specifies that the
Taoiseach must resign from office if he ceases to retain the support of a majority in the
Dáil (and if this happens all other members of the government are automatically
deemed to have resigned).
In practice no-confidence motions are fairly rare and some that do occur are a form of
symbolic protest by the opposition rather that a realistic attempt to topple the government.
Strong party discipline means that most governments can survive no-confidence motions
with impunity. Any TD can propose a no-confidence motion. Confidence motions are not
explicitly in the parliamentary rules of procedure. It seems that they must be governed by
the rules pertaining to all motions. In confidence votes the quorum is twenty and the
decision rule is a simple majority of those voting.23 Of course the strategic context changes
if the government lacks an overall majority, thereby empowering the legislature. Only two
Irish governments have ever been dismissed by the Dáil. The first was in November 1982
when the Fianna Fáil minority government was defeated on its own confidence motion
(proposed by the Taoiseach) after the Workers’ Party withdrew its external support
(Mitchell 1996: 123–4). Similarly in November 1992 another Fianna Fáil minority govern-
ment lost a confidence motion the day after its coalition partners, the Progressive
Democrats, had resigned (Mitchell 1993: 113).
Nevertheless, the fact that the Dáil has only directly brought down two governments
underestimates the confidence procedure as an institution framing parliamentary gov-
ernment. For example, Gallagher counts at least eight other occasions (August 1927,
1938, 1944, 1951, 1957, January 1982, 1987, and 1994) when governments opted for an
election rather than face almost certain defeat on a confidence motion (Gallagher
1999b: 183). Table 13.1 includes the resignation of the Haughey IV cabinet in this group.
While the President plays only a formal role with regard to the inauguration of
governments he or she does have one potentially important discretionary power in
relation to dissolutions. In normal circumstances, when a Taoiseach requests the President
to dissolve the Dáil and hold an election, the latter is required to do so (Article 13.2.1
of the Constitution). However, according to Article 13.2.2, the President ‘may in his
absolute discretion refuse to dissolve Dáil Éireann on the advice of a Taoiseach who has
ceased to retain the support of a majority in Dáil Éireann’ (see Gallagher 1988; Laver
1996a). Constitutional ambiguity arises because it is not clear (and is not specified in
the Constitution) what should count as evidence that the Taoiseach has lost the Dáil’s
23
The Dáil’s Standing Orders (SO) say ‘All motions to be put on the order paper for any day, shall be in writ-
ing, signed by a member, and shall reach the clerk not later than 11 a.m. on the fourth preceding day’ (SO 28),
though by permission of the Ceann Comhairle motions may be made on shorter notice. Dáil standing
orders also stipulate that: ‘No member shall re-open a discussion on a question already discussed during the
preceding six months’ (SO 51), which seems to limit the frequency of such motions.
Ireland 431
confidence. Losing an explicit confidence motion or suffering defeat on the Finance Bill
would certainly count but other situations are less clear-cut. For example, if one of the
coalition partners essential to the government’s majority resigns, could the President
use this as tangible evidence that the government has lost the confidence of the Dáil
and hence refuse to dissolve? While this has never been tested, to conclude otherwise
would undermine presidential discretion since a Taoiseach faced with certain defeat on
a confidence motion could always ask for a dissolution (Laver 1996a). On the other
hand, the President by using this power could encourage coalition renegotiation after
a government falls as an alternative to the traditional resort to an election. While no
President has yet exercised this power, it could become a source of controversy if
Ireland were to experience a period of coalition turbulence and government instability.
Table 13.1 summarizes the behavioural record with respect to cabinet investiture,
no-confidence and confidence votes, and parliamentary dissolution.
The tables shows that, while unsuccessful investiture votes have generally been rare
in Ireland, a total of six unsuccessful votes occurred both before the Haughey IV cabi-
net in 1989 and the Reynolds II cabinet in 1993. No data exists on the number of
no-confidence votes held during each cabinet, but two cabinets were forced to resign
because they lost no-confidence votes, and six cabinets resigned to pre-empt a no-confi-
dence vote. On sixteen occasions early elections were held after the Taoiseach dissolved
Parliament.
24
This has begun to change in the last few years. See section on ‘Executive Secrecy’, and Connolly and
O’Halpin (1999).
Table 13.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Ireland 1944–2000 a
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv)b Confidence Did the Who
formation votes (cv) cabinet end dissolved
No. of Voting results of successful iv No. of Cabinet Cabinet under specific with an early Parliament
unsuccessful (1: pro gov., 2: Against gov., ncve removed resigned institutional election? in this
investiture 3: abstentions, and 4: other) by ncv to mechanismc case?d
votes 1 2 3 4 pre-empt n.a.
ncv
will pass unamended. Third, most Irish committees (at least those set up before the
mid-1990s) tended not to directly shadow government departments but rather to cover
a bundle of jurisdictions that cut across departments (this was changed after the 1997
election so that now there is a committee monitoring each department).25 Fourth, com-
mittees have lacked powers to compel ministers and civil servants to attend and give
evidence, and more generally have lacked administrative resources. Finally, individual
incentives have militated against the development of very strong committees. When push
comes to shove, TDs prioritize loyalty to party over loyalty to committee for reasons dear
to their hearts: promotion and re-election. Since committees do not distribute ‘goodies’,
TDs do not expect electoral credit in their constituencies for good committee work.
Indeed, they anticipate that too much time devoted to committees could be electorally
counter-productive, or even suicidal. Nevertheless, committees do appear to be very
gradually gaining in prestige and influence and play an increasing role in developing
specialist knowledge and creating ‘information efficiencies’ (Krehbiel 1991).
D E L E GAT I O N W I T H I N T H E C A B I N E T
The Taoiseach is the boss both formally and in practice. He (to date all have been men)
has the constitutional right to hire or fire any member of the government. The
Taoiseach’s ministerial nominees will be appointed by the President providing that they
have been previously approved by Dáil Éireann (Article 13.1.2). For majority govern-
ments this is a formality. The Taoiseach’s power over ministerial selection is amended
during coalition governments, so that the leaders of the other governing parties pick
their own ministers.26 Similarly the Taoiseach can fire any member of the government
‘for reasons which to him seem sufficient’ (Article 28.9.4) though again in practice a
Taoiseach cannot fire a minister of another party without that party leader’s consent.
While ministers are not dismissed very frequently, the Taoiseach’s power is always
manifest. Thus, the delegation relationship is not so much from cabinet to minister as
from Taoiseach (and other governing party leaders) to individual ministers, since the
latter serve at the former’s pleasure.
Irish cabinets contain fifteen ministers (all governments since 1973 have filled this
constitutional maximum) who are both collectively responsible for the work of all gov-
ernment departments and individually responsible for virtually all the actions of their
own department. The style of cabinet decision-making is structured by Article 28.4.2 of
the Constitution: ‘The Government shall meet and act as a collective authority, and shall
be collectively responsible for the Departments of State administered by the members of
the Government’. The cabinet acts as a clearing house for most major decisions and the
doctrine of collective cabinet responsibility ‘normally denies ministers the right to record
25
The new committee system is also a source of patronage in that there are four paid positions (Gallagher
1999b).
26
Technically, the nominees are recommended to the Dáil as members of the government rather than to
particular departments, a point that was made clearly to the Dáil by Eamon de Valera in 1944 (Chubb 1983: 60).
This underlines the Taoiseach’s power to alter assignments or jurisdictions at will.
436 Paul Mitchell
private dissent, let alone public opposition, to cabinet decisions’ (Farrell 1993a: 174). This
collective responsibility of the cabinet is reinforced by a doctrine of executive secrecy.
Nevertheless, there is a difference between the cabinet being formally collectively
responsible for government decisions and extensive use of actual collective decision-making
(Laver and Shepsle 1994: 298, 1996). There is a far-reaching division of labour in
government and this gives decision-making a departmental structure in which each
cabinet minister has considerable agenda powers to determine which policies and
policy options are brought to cabinet for a formal decision. And cabinet ministers have
considerable discretion as the unequivocal heads of their departments. This role is
established by the Ministers and Secretaries Act 1924 (and subsequent amendments)
whereby the minister is legally a ‘corporation sole’, and can be sued in that capacity.
‘The effect of this is that the acts of a department are the acts of its minister, for which
he alone is responsible . . . in effect, the minister is the department’ (Dooney and
O’Toole 1992: 115).27 However, while energetic and determined ministers no doubt
have considerable scope to shape the policies of their departments, this autonomy
should not be exaggerated. Particularly during coalitions party leaders cannot afford to
allow full ministerial discretion since they would then lose all input in jurisdictions in
which they did not control the minister. Parties have strong incentives to underwrite
the credibility of their key policies—and attempt to commit their coalition partners to
them—by negotiating their inclusion in the coalition policy document and sub-
sequently policing these promises by a variety of devices designed to monitor progress
towards the party’s goals (see also Strøm 1998; Mitchell 1999a).
The Irish cabinet has not made extensive use of cabinet subcommittees (which are
established on an ad hoc basis) and as a collective body appears to at least ‘see’ an
extraordinary volume of papers. Farrell (1993a: 173) estimates that about 800 papers a
year are handled by the cabinet, as compared to about 60–70 in Britain, where much
more is delegated to cabinet committees. The Taoiseach controls the cabinet agenda
through the government secretariat (the Department of the Taoiseach), and the
cabinet usually meets once a week. There is no quorum, and in any case voting is not
common. In addition the Minister for Finance clearly has a special role in exercising
financial control over other departments. The Finance Minister is briefed on every
proposal that comes before the cabinet and prior to this all submissions to the government
‘must have been presented in advance to the Department of Finance so that that
department may include its comments and advice to the government’ (Dooney and
O’Toole 1992: 234). After the Taoiseach, Finance is clearly the most important ministry
and norms of mutual non-intervention do not apply to it.
In conclusion, the attempt to contain agency losses between individual ministers and
their party leader relies heavily on their assumed incentive compatibility through their
co-partisanship and their shared electoral goals. Leaders, ministers, and MPs of the same
party do usually sink or swim together. Thus, though some ministers are more forceful
or more able than others, they are generally agents of their parties and in practice (espe-
cially when in government) the party leader is the arbiter of the party interest. He or she
27
The ‘corporation sole’ principle was partially amended in 1997. See section titled Delegation from
Ministers to Civil Servants.
Ireland 437
can, after all, fire any of them. However, while some commonality of interest may
develop between coalition partners, they have different policy priorities and often suffer
diverging electoral fates. Agency losses are a continuous problem in that a party is likely
to be interested in the policy outputs (and electoral credit if available) of departments
controlled by cabinet ministers belonging to their governing partners. While interven-
tion in other ministries is a decidedly uphill task, a party cannot simply abandon a key
party policy simply because it does not control the relevant minister. The increased fre-
quency of coalition governments has led to the evolution of enforcement mechanisms
by which parties attempt to safeguard their key interests. These range from contract
design, in which parties negotiate the inclusion of brand identity items in the coalition
policy document, to the detailed tracking, reporting, and monitoring of all departments
by policy advisors and junior ministers in an attempt to ensure that party policies are
being implemented in all ministries (see section titled Delegation from Ministers to Civil
Servants). Smaller parties have been more enthusiastic about these enforcement mech-
anisms than larger ones. This might reflect their policy-driven profiles, but also indirectly
testifies to the considerable agenda powers inherent in controlling ministries. Quite sim-
ply, large parties control more of the government agenda by virtue of the fact that they
control more ministries, so that ancillary enforcement mechanisms that essentially
attempt to compensate for less direct control (fewer portfolios) are less crucial to the big
guns. In short, delegation relationships are more complex in coalition governments: they
are a tangled web of agents monitoring other agents.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
While civil servants are in a general sense servants of the public they are more directly
agents of the minister. Indeed, in Ireland ‘the minister in charge of each department is
legally the employer of all staff in that department’ (Dooney and O’Toole 1992: 109).
Most civil servants are recruited by means of competitions held by the Civil Service
Commission. The minister in each department appoints all civil servants from names
submitted by the Commission, subject to the posts being made available by the Finance
ministry. The one exception is the Secretary of each department (the top civil servant),
who is appointed by the cabinet on the recommendation of the relevant minister.
As noted, Irish ministers were traditionally regarded as ‘corporation soles’: the minister
is the department. Civil servants had no legal basis for issuing a decision in their own
names and the minister is responsible for virtually everything done by her department.
This position has been somewhat transformed by the Public Service Management Act
1997, ‘which for the first time identified the role and responsibilities of senior civil
servants and distinguished their responsibilities from those of the minister. The Act
gave a public, legal accountability to the Secretary General of each government
department’ (Connolly and O’Halpin 1999: 264).
Evidence concerning senior civil servant’s potential for independent influence on the
political process is mostly anecdotal. Brian Farrell (1994), one of the few to have
interviewed cabinet ministers about such matters, generally judges the civil service’s
influence as quite extensive. While some ministers claimed to be able to identify quite
438 Paul Mitchell
easily civil servants engaging in delaying and blocking tactics, many others were
frustrated by the difficulty of forcing their agents to do their bidding.
As coalition politics has evolved ministers have increasingly resorted to political
advisors—often people familiar with the civil service—whose chief characteristic is
their direct loyalty to the minister. Thus, unlike the civil service, these advisors should
not shirk their responsibility or engage in hidden actions, because, if carefully selected,
their incentive structure should closely correspond to that of their principal, the minister.
Their tenure depends entirely upon keeping the minister in power. These advisors are
of two types. One is the ministerial political advisor who acts (in the words of senior
civil servants) as ‘a replication of the minister in many respects, as an extension of his
political personality, as an extra pair of eyes and ears, doing for him what the minister
would do for himself if he had the time’ (Dooney and O’Toole 1992: 39).
Many civil servants were very unhappy, however, with a 1993 innovation whereby
the incoming Fianna Fáil–Labour government created a new set of appointments in
which each cabinet minister would appoint a ‘programme manager’. These were sen-
ior political appointments operating within the civil service, but accountable directly to
the minister rather than part of the normal line management of a department
(O’Halpin 1997: 81). Whereas political advisors look after the minister generally, the
programme manager’s central task was to submit policy implementation to detailed
tracking, to overcome bureaucratic and political obstacles to policy delivery, and gen-
erally to make sure that the party’s policies were actually implemented. Thus, apart
from overcoming civil service obstacles, a key function of the programme manager
was to enable the Labour Party to ‘maximise its strength in a coalition and to exercise
its influence across a range of issues, including those where Labour ministers had no
direct responsibility but on which the party itself had some view or definite policy posi-
tion’ (O’Halpin 1997: 80). Labour’s programme managers met weekly as a group with
party officials and TDs to monitor the progress of the party’s policy. The programme
managers are agents of their respective ministers: their job is to protect the minister
against agency losses vis-á-vis the permanent civil service. In effect they constitute an
attempt to compensate for the information asymmetry that favours the civil service in
increasingly complex policy areas.
These appointments were repeated by the next government in 1995–6 but scaled
down by the 1997 Fianna Fáil–Progressive Democrat coalition, partly because the PDs
in particular had criticized the creation of the new political appointees as wasteful.
However, the practice is likely to re-emerge since it was generally quite successful, and
more to the point popular with most ministers who welcomed the addition of a senior
operative unequivocally loyal to themselves.
‘ E X T E R NA L’ C O N S T R A I N T S
The Impact of ‘Foreign Affairs’
Two sets of ‘external’ events are of particular importance to representative politics in
Ireland: Northern Ireland and the EU. While the status and future of Northern Ireland
Ireland 439
has long been one of the biggest problems facing the Irish government, the special
relationship with Britain has improved dramatically in the 1980s and 1990s. After the
Anglo-Irish Agreement of 1985 and especially following the negotiating breakthroughs
in 1998, the Irish government has become increasingly involved in North–South
cooperation at all political and administrative levels (see Mitchell and Wilford 1999).
Nevertheless, despite its intrinsic importance, increased involvement in Northern
Ireland affairs imposes responsibilities rather than serious constraints on Irish policy
makers in Dublin.
The same cannot be said of the EU. Like other member countries, the Irish
Parliament’s ‘sole and exclusive’ legislative authority is seriously amended by member-
ship of the EU. Nevertheless, and unlike its larger neighbour, Ireland has generally been
an enthusiastic member of the European club. Yet despite Ireland’s active participation
in the development of European institutions, a domestic democratic deficit in terms of
a lack of accountability has occurred in relation to EU decision-making. In 1973 the
Oireachtas established the Joint Committee on the Secondary Legislation of the
European Communities, ostensibly to compensate for the pooling of sovereignty
implied by membership. In practice, however, although ‘its reports were often of high
quality. . . the Oireachtas generally showed little interest in them’ (Gallagher 1993b:
139). In addition, the government undertook to place two reports a year before
Parliament but ‘these reports generally arrive too late for parliament to give serious
consideration to the issues they raise’ (Keatinge and Laffan 1999: 339). However, a Joint
Committee on European Affairs was created in March 1995 with a deputy from the
governing coalition as chairperson. This committee like others in Ireland has no real
binding power on government. EU membership, however, has dramatically affected the
work of most departments and the civil service. One obvious effect has been to
increase the domestic stature and importance of the Foreign Affairs ministry, which
plays a key coordinating role.
Executive Secrecy
Ireland traditionally adhered strictly to doctrines of collective cabinet responsibility and
executive secrecy. While some privacy of cabinet discussion is usually considered
prudent, when applied rigorously these doctrines constitute an important constraint on
executive accountability. This became a problem of public concern in the 1990s as the
Irish political establishment was rocked by a series of scandals. Matters came to head
in 1992 during a Tribunal of Enquiry into the Beef Industry, which among other things
was investigating whether there had been improper links between some senior ministers
and business interests (see Collins and O’ Raghallaigh 1995; O’Toole 1995). The Supreme
Court (in 1992) declared that the Tribunal could not have access to cabinet discussions;
the Court thereby established a constitutional right to absolute cabinet confidentiality.
Even the cabinet could not waive this right. The Chief Justice declared that ‘confiden-
tiality is a constitutional right which, in my view, goes to the fundamental machinery
of government and is, therefore, not capable of being waived by any individual
member of the government’ (Farrell 1993a: 175, 1993b). A former Taoiseach, Garret
440 Paul Mitchell
FitzGerald, argued that the Supreme Court decision ‘turned a valuable political
convention into a dangerous constitutional ban of an absolutist character’ (The Irish
Times, 18 October 1997).
The Irish political class is still struggling with such issues of executive accountability. A
number of senior ministers have been forced to resign and former Taoiseach Charles
Haughey has admitted improperly accepting a ‘gift’ of over one million pounds. Further
scandals necessitated further tribunals of investigation, which found themselves ham-
strung by the confidentiality ruling and by their inability to compel witnesses to testify. In
an unsatisfactory attempt to deal with the first of these problems, a referendum on cabi-
net confidentiality was narrowly passed (53 per cent to 47) in November 1997. It
enshrined this doctrine in the Constitution while at the same time allowing the High
Court to authorize disclosure in inquiries by official tribunals.28 Related reforms are grad-
ually moving Ireland from a position of almost total executive secrecy to gradual and lim-
ited openness. New electoral laws now govern and limit campaign spending and require
financial disclosure. The establishment of a ‘Standards in Public Office Commission’ was
announced in June 1998 as a permanent investigative tribunal, replacing the previous pat-
tern of ad hoc individual enquiries. The head of the Commission is to have the powers of
a High Court judge and will investigate complaints against politicians and examine finan-
cial donations to deputies and parties (The Irish Times, 13 June 1998). A Freedom of
Information Act was passed in 1997, although with twelve significant exemptions.
Nevertheless, every department has now trained designated civil servants as freedom of
information officers and citizens have a new legal right to information, or failing
disclosure, to a justification. Connolly and O’Halpin (1999: 267) comment that the Act
‘represents a reversal of the presumption of secrecy that has underpinned Irish govern-
ment...It is, for example, in practice a far more powerful investigative instrument than a
parliamentary question...’ Legislation has also been introduced allowing committees to
compel testimony from senior public servants and grant immunity to certain witnesses.29
These reforms are overdue and essential if tribunals, courts, and parliamentary commit-
tees are to have the opportunity to exercise effective ex post scrutiny of politicians’
behaviour in government.
C O N C LU S I O N
In terms of the ‘bigger picture’ the newly independent state of Ireland established in the
1920s has been remarkably successful. The political forces initially engaged in a bitter
civil war subsequently managed to abandon violence in favour of peaceful alternation
in office and there have been no serious threats of democratic breakdown or regime
instability. Political institutions, parties, and norms of parliamentary behaviour were
quickly institutionalized in a manner considered legitimate and reasonably efficient.
28
Many who voted ‘no’ did so because they felt that the amendment was much too restrictive, not
because they were in favour of confidentiality.
29
The Committees of the Houses of the Oireachtas (Compellability, Privileges, and Immunitiies of
Witnesses) Act 1997.
Ireland 441
REFERENCES
Benoit, Kenneth and Michael Laver (2003). ‘Estimating Irish Party Policy Positions Using
Computer Wordscoring: The 2002 Elections’, Irish Political Studies, 18:1, forthcoming.
Carstairs, Andrew McLaren (1980). A Short History of Electoral Systems in Western Europe. London:
George Allen and Unwin.
Carty, R. K. (1983). Party and Parish Pump: Electoral Politics in Ireland. Brandon: Mercier
Press.
Chubb, Basil (1983). A Source Book of Irish Government. Dublin: The Institute of Public
Administration.
Coakley, John (1999). ‘The Foundations of Statehood’, in John Coakley and Michael Gallagher
(eds.), Politics in the Republic of Ireland, 3rd edn. London: Routledge and PSAI Press.
—— and Manning, Maurice (1999). ‘The Senate Elections’, in Michael Marsh and Paul Mitchell
(eds.), How Ireland Voted 1997. Boulder: Westview Press.
Collins, Neil and O’ Raghallaigh, Colm (1995). ‘Political Sleaze in the Republic of Ireland’.
Parliamentary Affairs, 48: 697–710.
Connolly, Eileen and O’Halpin, Eunan (1999). ‘The Government and Governmental System’, in
John Coakley and Michael Gallagher (eds.), Politics in the Republic of Ireland, 3rd edn. London:
Routledge and PSAI Press.
Dooney, Sean and O’Toole, John (1992). Irish Government Today. Dublin: Gill and Macmillan.
Farrell, Brian (1983). ‘Coalitions and Political Institutions: The Irish Experience’, in Vernon
Bogdanor (ed.), Coalition Government in Western Europe. London: Heinemann.
—— (1993a). ‘The Government’, in John Coakley and Michael Gallagher (eds.), Politics in the
Republic of Ireland. Dublin: Folens and PSAI Press.
—— (1993b). ‘ “Cagey and Secretive”: Collective Responsibility, Executive Confidentiality and the
Public Interest’, in Ronald J. Hill and Michael Marsh (eds.), Modern Irish Democracy. Dublin:
Irish Academic Press.
442 Paul Mitchell
—— (1994). ‘The Political Role of Cabinet Ministers in Ireland’, in Michael Laver and Kenneth
A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge
University Press.
Farrell, David (1992). ‘Ireland’, in Richard S. Katz and Peter Mair (eds.), Party Organizations:
A Data Handbook on Party Organizations in Western Democracies, 1960–90. London: Sage.
—— (1994). ‘Ireland: Centralization, Professionalization and Competitive Pressures’, in Richard
Katz and Peter Mair (eds.), How Parties Organize. London: Sage.
Gallagher, Michael (1977). ‘Party Solidarity, Exclusivity and Inter-Party Relationships in Ireland,
1922–77: the Evidence of Transfers’. Economic and Social Review, 10: 1–22.
—— (1985). Political Parties in the Republic of Ireland. Manchester: Manchester University Press.
—— (1987). ‘Does Ireland Need a New Electoral System?’ Irish Political Studies, 2: 27–48.
—— (1988). ‘Ireland: The Increasing role of the Centre’, in Michael Gallagher and Michael Marsh
(eds.), Candidate Selection and Comparative Perspective: The Secret Garden of Politics. London: Sage.
—— (1993a). ‘The Constitution’, in John Coakley and Michael Gallagher (eds.), Politics in the
Republic of Ireland. Dublin: Folens and PSAI Press.
—— (1993b). ‘Parliament’, in John Coakley and Michael Gallagher (eds.), Politics in the Republic of
Ireland. Dublin: Folens and PSAI Press.
—— (1993c). ‘The Election of the 27th Dáil’, in Michael Gallagher and Michael Laver (eds.), How
Ireland Voted 1992. Dublin: Folens and PSAI Press.
—— (1996a). ‘Ireland: the Referendum as a Conservative Device?’, in Michael Gallagher and Pier
Vincenzo Uleri (eds.), The Referendum Experience in Europe. London: Macmillan.
—— (1996b). ‘Electoral Systems’. Appendix 4 of Report of the Constituition Review Group. Dublin:
The Government of Ireland.
—— (1999a). ‘The Changing Constitution’, in John Coakley and Michael Gallagher (eds.), Politics
in the Republic of Ireland, 3rd edn. London: Routledge and PSAI Press.
—— (1999b). ‘Parliament’, in John Coakley and Michael Gallagher (eds.), Politics in the Republic of
Ireland, 3rd edn. London: Routledge and PSAI Press.
—— and Michael Marsh (eds.) (1988). Candidate Selection and Comparative Perspective. London:
Sage.
Galligan, Yvonne (1999). ‘Candidate Selection in 1997’, in Michael Marsh and Paul Mitchell (eds.),
How Ireland Voted 1997. Boulder: Westview Press.
Keatinge, Patrick and Laffan, Brigid (1999). ‘Ireland: A Small Open Polity’, in John Coakley and
Michael Gallagher (eds.), Politics in the Republic of Ireland, 3rd edn. London: Routledge and
PSAI Press.
Krehbiel, Keith (1991). Information and Legislative Organisation. Ann Arbor: University of
Michigan Press.
Laver, Michael (1992). ‘Coalition and Party Policy in Ireland’, in Michael Laver and Ian Budge
(eds.), Party Policy and Government Coalitions. London: Macmillan.
—— (1994). ‘Party Policy and Cabinet Portfolios in Ireland 1992: Results from an Expert Survey’.
Irish Political Studies, 9: 157–64.
—— (1996a). ‘The Government Formation Process in Ireland: Implications for the Constitutional
Role of the President, the Government and the Dáil’. Appendix 1 of Report of the Constitution
Review Group. Dublin: Government of Ireland.
—— (1996b). ‘Notes on a New Irish Senate’. Appendix 7 of Report of the Constitution Review Group.
Dublin: Government of Ireland.
—— (1998). ‘Party Policy in Ireland 1997: Results from an Expert Survey’. Irish Political Studies,
13: 159–71.
Ireland 443
Laver, Michael and Hunt, W. Ben (1992). Policy and Party Competition. New York: Routledge.
—— and Marsh, Michael (1999). ‘Parties and Voters’, in John Coakley and Michael Gallagher
(eds.), Politics in the Republic of Ireland, 3rd edn. London: Routledge and PSAI Press.
—— and Shepsle, Kenneth A. (1994). ‘Cabinet Members in Theoeretical Perspective’, in Michael
Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government.
Cambridge: Cambridge University Press.
—— —— (1996). Making and Breaking Governments: Cabinets and Legislatures in Parliamentary
Governments. Cambridge: Cambridge University Press.
Lynch, Kathleen (1996). ‘Seanad Éireann’. Appendix 6 of Report of the Constitution Review Group.
Dublin: Government of Ireland.
Mair, Peter (1986). ‘Locating Irish Parties on a Left–Right Dimension: An Empirical Enquiry’.
Political Studies, 34: 456–65.
—— (1987). The Changing Irish Party System: Organisation, Ideology and Electoral Competition.
London: Frances Pinter.
—— (1999). ‘Party Competition and the Changing Party System’, in John Coakley and Michael
Gallagher (eds.), Politics in the Republic of Ireland, 3rd edn. London: Routledge and PSAI Press.
Marsh, Michael (1981a). ‘Localism, Candidate Selection and Electoral Preferences in Ireland: the
General Election of 1977’. Economic and Social Review, 12: 267–86.
—— (1981b). ‘Electoral Preferences in Irish Recruitment: The 1977 Election’. European Journal of
Political Research, 9: 61–74.
Mitchell, Paul (1993). ‘The 1992 General Election In The Republic of Ireland’. Irish Political
Studies, 8: 111–17.
—— (1996). The Life and Times of Coalition Governments: Coalition Maintenance by Event
Management. Ph.D. Dissertation. Florence: European University Institute.
—— (1999a). ‘Coalition Discipline, Enforcement Mechanisms and Intra-Party Politics’, in Shaun
Bowler, David Farrell, and Richard S. Katz (eds.), Party Cohesion, Party Discipline and the
Organization of Parliaments. Ohio: Ohio State University Press.
—— (1999b). ‘Government Formation: A Tale of Two Coalitions’, in Michael Marsh and Paul
Mitchell (eds.), How Ireland Voted 1997. Boulder: Westview Press.
—— (2000a). ‘Voters and Their Representatives: Electoral Institutions and Delegation in
Parliamentary Democracies’. European Journal of Political Research, 37: 335–51.
—— (2000b). ‘Ireland: From Single-Party to Coalition Rule’, in Wolfgang C. Müller and Kaare
Strøm (eds.), Coalition Governments in Western Europe. Oxford: Oxford University Press.
—— and Wilford, Rick (eds.) (1999). Politics in Northern Ireland. Boulder: Westview Press and
PSAI Press.
Müller, Wolfgang C. (2000). ‘Political Parties in Parliamentary Democracies: Making Delegation
and Accountability Work’. European Journal of Political Research, 37: 309–33.
O’Halpin, Eunan (1993). ‘Policy Making’, in John Coakley and Michael Gallagher (eds.), Politics
in the Republic of Ireland. Dublin: Folens and PSAI Press.
—— (1997). ‘Partnership Programme Managers in the Reynolds/Spring Coalition, 1993–4:
An Assessment’. Irish Political Studies, 12: 78–91.
O’Sullivan, Mary-Claire (1999). ‘The Social and Political Characteristics of the Twenty-Eighth
Dáil’, in Michael Marsh and Paul Mitchell (eds.), How Ireland Voted 1997. Boulder: Westview
Press.
O’Toole, Fintan (1995). Meanwhile Back at the Ranch: The Politics of Irish Beef. London:
Vintage.
444 Paul Mitchell
Sinnott, Richard (1999). ‘The Electoral System’, in John Coakley and Michael Gallagher (eds.),
Politics in the Republic of Ireland, 3rd edn. London: Routledge and PSAI Press.
—— (1995). Irish Voters Decide: Voting Behaviour in Elections and Referendums Since 1918.
Manchester: Manchester University Press.
Strøm, Kaare (1998). ‘Institutions and Strategy in Parliamentary Democracy: A Review Article’.
Legislative Studies Quarterly, 23: 127–43.
Whitaker, T. K. et al. (1996). Report of the Constitution Review Group. Dublin: Government of
Ireland.
14
I N T RO D U C T I O N
The Italian political system has usually been described as a stable parliamentary demo-
cracy with proportional representation rules and a fragmented party system. About
10 years ago, this image was undermined by a deep crisis whose consequences, after sev-
eral years of political turmoil, are still not clear. In this Chapter I discuss the constitu-
tional and historical context of parliamentary democracy in Italy, starting from the
‘classic’ model set out in the republican Constitution of 1948. Of course, parliamentary
democracy began much earlier, under the so-called Liberal Italy (1861–1924). The old
Constitution marked the birth of an elective Parliament without changing the ‘private’
nature of the monarchic government1 and without introducing broad political parti-
cipation.2 After 20 years of Fascism (1924–43), the entire political landscape was funda-
mentally altered—including the party system, the political class, and the predominant
political cultures.
The constituent assembly elected in 1946 represented a complicated constellation of
parties. Nevertheless, the new carta was approved by a large majority (around 90 per cent)
of assembly members. Many strands of thought influenced the contents of the
1948 Constitution, which turned out to be both inflexible and long. Left parties and the
Christian Democrats, despite the great differences between them, supported the intro-
duction of social rights and welfare principles. In addition, the Constitution included
basic liberal principles. Institutionally, after long debate, a parliamentary model pre-
vailed. This choice signalled the dominance of mass parties, which would set the con-
ditions for Italy’s consensus democracy. The parties (or party factions) occupied all
institutional levels, thus serving as the gatekeepers of democratic consolidation
(Morlino 1991).
1
The Statuto albertino was issued by the King of Piedmont in 1848 and then extended to the unified state.
Article 5 of the Statuto Albertino stated that ‘the executive power belongs to the King’. Article 65 added that
‘the King appoints and dismisses his ministers’.
2
Less than 5 % of the population had the right to vote before the first extention of the suffrage in 1882.
Until the reform of 1912, which introduced universal male suffrage, only about 10 % of the population had
the right to vote.
446 Luca Verzichelli
The 1948 Constitution provides the two Chambers with symmetrical legislative
powers. Both chambers have to approve all legislative items. A quorum of 50 per cent ⫹1
of the MPs is required, even though it is not checked by default. Both chambers feature
similar procedural options: an abbreviated procedure by committee (Commissione in sede
legislativa) can be applied, with the exception of constitutional amendments, electoral
laws, delegations to cabinet, ratification of international treaties, and Budget laws.3
The President of the Republic has the power of promulgation (Article 73),4 thus he
is formally involved in the legislative process, while the cabinet has only limited
powers of initiative and the authority to issue decrees on specific urgent issues (Article 77).
A decree loses its legislative force unless Parliament approves it within 60 days. However,
cabinets have traditionally reiterated the same decrees, challenging the majority to
accept its initiatives (De Micheli 1997).
Constitutional amendments and budgetary policy (after 1978) do not follow this
legislative process. The budget consists of a number of documents written by the cabinet
(the Treasury and the Prime Minister (PM) ) and approved by the Parliament during its
autumn budgetary session. One of them is the formal budget law, which includes the
final balance for the current year and the estimate for the next. All authorizations to
change the existing expenditures and revenues have to be recorded in the Financial Bill
(Legge finanziaria). A decade after the 1978 budget reform, the unsatisfactory results
(and rising deficit) convinced Parliament to reinforce the cabinet’s power, shorten the
budgetary session, and constrain parliamentary amendments (Hine 1993: 180–7).
According to Article 138 of the Constitution, constitutional amendments have to pass
two votes in each chamber, with the second vote held at least 3 months after the first.
If an amendment is approved by a majority smaller than two-thirds of all MPs, then
a referendum on the amendment can be requested by one-fifth of MPs or by five
(of twenty) Regional Councils.
The legislative procedures, which do not follow the ordinary process, also include
the referendum provided by the Constitution (Article 75), which allows citizens to abro-
gate ordinary laws. There have been few changes in the legislative process since 1
January 1948. The most important transformations have been due to non-constitu-
tional reforms. For example, the anonymous vote (1988) has replaced the dominance of
secret voting in Parliament. On the other hand, the rise of strong party government
and the cleavage between ‘Atlanticists’ and Left parties rapidly changed the conditions
for the implementation of some provisions of the Constitution. Government parties,
in an effort to increase the freedom of action of their fragile majorities, tried to reduce
the Constitution’s power-dispersing system of checks and balances (Hine 1993: 157)
3
The abbreviated procedure has been heavily used, especially in the past, typically to produce an output
called little-laws (leggine), micro-sectional pieces of legislation decided by variable majorities inside the com-
mittees (Di Palma 1977; Capano and Giuliani 2002b).
4
Article 73 of the Constitution states that the President must sign new laws within 1 month of their
approval, less if the Chambers declare the legislation urgent. Article 74 also states that the President may
request a reconsideration by means of a message to both Houses, but if the Houses reapprove a law then it
must be promulgated.
Italy 447
and delayed the actual creation of important institutions contained in the Constitution
(the Constitutional Court, until 1958, the referendum, until 1970, and above all the
regional autonomies, also until 1970). Paradoxically, Left parties, after favouring a cen-
tralized political system, pushed for ‘regionalization’ to strengthen their local and
administrative power. The delay in the implementation of the Constitution distorted
some of the intentions behind its provisions, thus making institutional mechanisms
more flexible than one would expect given the formal Constitutional framework.
Political scientists and constitutional lawyers have paid particular attention to the
problem of party government (see further). In contrast, institutional evolution has been
largely unexplored. The usual image of Italian democracy is therefore relatively static:
formally, the framework of parliamentary democracy has not changed, despite strong
public demand for institutional and administrative reforms over at least the past
fifteen years. However, in reality, the institutional structure has always been flexible
(Calandra 1996). In this writer’s view, this flexibility is connected to cycles of party gov-
ernment (Pasquino 1987; Verzichelli and Cotta 2000). In a broad sense, the success or
dominance of different governance styles has been linked to different stages of Italian
party government and, specifically, to the political formulas that supported the fifty-one
cabinets of the 1945–98 period. In particular, we can identify a number of small changes
related to parliamentary delegation and ministerial accountability.
The literature illustrates this flexibility in the practice of governing Italy. Maurizio
Cotta (1994) shed light on the dynamics of the interaction between Parliament and cab-
inet. Using a revised Lijphart model (Lijphart 1984), he discovered the alternation
between ‘majoritarian’ phases (during the 1950s and the 1980s) and ‘proportional’ ones.
Majoritarian phases have been characterized by a clear distinction between cabinet and
parliamentary opposition, a relatively strong PM, and greater cabinet autonomy. In the
proportional phases, a progressive integration of the (Left) opposition has been visible
both in the legislative process and in appointments, and Parliament has prevailed over
a constrained cabinet. Pasquino (1987) and Vassallo (1994) confirm the existence of
a historical evolution of party government. The latter, in particular, observes changes
in the ruling party’s ability to control the political agenda and appointment processes.
Four phases of party government have been manifest in postwar Italy. Interestingly
enough, Vassallo raises doubts about Italian institutional stability, asserting that some
constitutional prerogatives (for instance the role of the PM during government formation)
have been implemented in quite different ways at different times.
Differently from the publications quoted above, I will discuss whether the empirical
record makes it possible to consider Italy as a good approximation of the ideal type of
parliamentary delegation and accountability. Hence, I review the Italian delegation
process as established in law and discuss whether or not significant modifications
actually have occurred. If we want to understand ‘delegation and accountability’ in
Italy, we must first assess to what extent the republican experience fits parliamentary
democracy’s ideal type. In the conclusion I return to the theoretical framework to
discuss whether this model of delegation and accountability is applicable to the Italian
case, not least in the light of recent changes.
448 Luca Verzichelli
P O L I T I C A L PA RT I E S : T H E T R A D I T I O NA L GAT E K E E P E R S
Italian politics after Second World War was deeply influenced by a fragmented multi-
party system and by the powerful role of those parties. A strong polarization between
two permanent oppositions and the governmental centre was the basic characteristic of
a polarized, multiparty system (Sartori 1976). The Left was dominated by the largest
Communist party in Western Europe (PCI), while the small but stable neo-
fascist party (MSI) was the main actor on the Right. The DC (Christian Democrats)
dominated that part of the political spectrum from which governments came. It was
flanked by the so-called secular parties ( partiti laici): the Liberals (PLI) on the Centre-
right, and the Social Democrats (PSDI) and the Republicans (PRI) on the Centre-left.
Finally, the Socialist party (PSI) was always a middle-sized party, acting as a second
power in both the Left opposition (until 1963, when it distanced itself from PCI) or in
the governmental arena. A broad literature has analysed the impact of parties in the
recruitment of parliamentary elites (Cotta 1979), the distribution of cabinet positions
(Calise and Mannheimer 1982), or policy-making (Pasquino 1987). But how (and to
what extent) have political parties influenced democratic delegation in Italy?
Traditional interpretations argue that a strong party government rapidly transformed
parliamentary democracy in Italy by filling all political institutions with politicians from
governing parties, a ceto politico selected and formed by traditional mass parties. Under
this view, political parties worked mainly as office-seeking and/or patronage agencies
(Vassallo 1994; Cotta and Verzichelli 1996). However, it is important to keep in mind
(Panebianco 1982) that Italy’s traditional parties have produced different types of
politicians and exercised different degrees of control over their voters (Cotta 1979). In
general terms, DC and PCI represent two alternative models. In the first case, a
factionalized party (with discipline progressively enforced at the elite level) could
control the selection of politicians on the basis of a checks-and-balances logic between
different ‘electoral forces’ (territorial patrons and national factions) and a powerful
central elite. Parliamentary seniority was the basic requisite for a Christian Democratic
leader both to enter government (Calise and Mannheimer 1982) and to influence the
party. On the other hand, the ‘apparatus model’ of recruitment of the Communist
Party implies centralized candidate selection. Parliamentary turnover (i.e. regular
replacement of party MPs) and circulation between representative institutions and
positions within the party organization were the dominant methods used to secure this
control, and served to concentrate recruitment and selection decisions in the hands
of a small group of leaders. The Communist model implied a minor role for voters.
Party elites were responsible for selection and elected politicians could be indirectly
sanctioned by that same elite. In the other model, different party actors (individuals,
local elites, national elites) and voters were all involved in a bargaining process in the
first step of delegation. The DC’s procedure is more interesting in the present context
because, for a long time, it was the most important ‘constraint’ on the linkages we are
studying here (i.e. the delegation between electorate and governmental MPs and
between governmental MPs and cabinet). Of course, this procedure is particularly
relevant to relationships among DC voters and politicians, but even other parties in
government had similar selection methods.
Italy 449
D E L E GAT I O N F RO M VOT E R S TO M P S : T H E W E A K L I N K
The Electoral Systems
In keeping with the consensus democracy model, the electoral connection between
citizens and MPs in Italy was long governed by a PR system. Until 1992 two different
PR systems enabled many parties to get at least a few seats in Parliament. The electoral
system for the lower chamber was pure PR, based on multimember constituencies and
a national distribution of the remainders (corrected quota system). Three or four ‘mega-
constituencies’ in different areas of the country increased proportionality. Elections to
fill the 315 Senate seats were based on a single-member district system at the first stage,
but the majority required to win a seat was so high (65 per cent of the constituency
vote) that the normal outcome was a second stage, in which all seats were proportion-
ally distributed (d’Hondt method) on a regional basis. In elections to the lower chamber
voters chose from among parties: symbols for each party list, not candidates, were
printed on the ballot. Voters could exercise a preference vote by writing in a maximum
450 Luca Verzichelli
of four names from the party lists. Although preference voting was not mandatory, it
was widely used by voters (especially DC voters) to determine the internal ranking on
each list. On the other hand, the single-member competition for the Senate gave only
the ‘illusion’ of personal voting. Parties were not encouraged to form electoral cartels
(because of the impossibly high threshold) and commonly imposed their own candid-
ates in each constituency. Therefore, in reality, there were no significant differences—in
terms of selection process, electoral campaign, or party results—between the two elec-
toral systems.
In the early 1990s a strong pro-majoritarian movement, supported by local and sec-
ond rank politicians from different parties, promoted a referendum that abolished pre-
ference voting in lower chamber elections. Though the referendum had few practical
consequences, it was rich in symbolic meaning. The 1992 elections were technically dif-
ferent from the previous ones, but still ‘proportional’. However, major change was on
its way. A real reform was introduced in 1993 (after another referendum had definitively
killed the old electoral system). In the elections of 1994 and 1996, 25 per cent of
deputies and senators were elected proportionally on regional bases and 75 per cent
through a single-member plurality system.
‘public party financing programme’.5 The funds are distributed proportionally based on
national electoral support (Bardi and Morlino 1994). At the beginning of the 1990s
public contributions to parties reached 100 billion lira per year, representing 40–90 per cent
of the overall income of each party. A 1993 referendum abrogated the 1974 party
finance law and introduced a new system of public financing based on the principle of
‘electoral reimbursement’ (of campaign costs). The distribution of funds was still
proportional and the parliamentary parties were still the beneficiaries. A subsequent
law (no. 157 of 1999) confirmed the same principle and quadrupled the annual volume
of subsidies (Pujas 2000). In general, after 1974, two important features continue to
characterize the public financing of political parties: First, a party needs to enter the
Parliament in order to receive public funds. Second, money is always distributed
proportionally.
In parliamentary recruitment, the common pattern for most parties was ‘local involve-
ment’ in the selection of nominees, but strong ‘central control’ (Wertman 1988). During
the period popularly known as the first republic (1948–1994), candidate selection was
entirely controlled by parties, which could easily prevent any genuine, direct participation
by voters—and to a large extent, by party members as well. Consequently only a very
small percentage of voters were involved in candidate selection, which was largely a
process of bargaining between local executives and national elites,6 and intraparty com-
petition was very limited. A significant exception was competition among Christian
Democratic list of candidates. However, this competition did not take place according to
transparent mechanisms, but was governed by informal agreements among strong can-
didates (the so-called cordate). The Socialist party occasionally also manifested a high
degree of factionalism, but internal competition was controlled by national leaders. In
fact, most party statutes reserved a ‘quota’ of parliamentary slots for party executives
(Bardi and Morlino 1994).
The recent transformation of the party system and the shake-up of the old parlia-
mentary establishment have changed the political professionalism of MPs (Verzichelli
1998, 2002), but recruitment continues to be centralized. In no party are voters or
members directly involved in determining parliamentary nominees, and very few have
added rules about term limits or retirement age. The relative weakness of the principal
(the voter) and the hidden information controlled by the agent (the MP) thus seem to
characterize this delegation process. Already before the 1990s, certain external con-
straints had progressively changed the opportunities for political parties to limit the
5
Law no. 195/1974 introduced an annual subsidy to all parties that won at least one seat in Parliament or
2% of the valid votes nationwide. A 1978 referendum to abrogate this law was supported by only about 40 %
of voters. The law was amended several times (in 1978, 1980, and 1981), though none of these were signifi-
cant changes. Other funds subsidized parties running for European and local elections (Law 422/1980).
6
Interparty differences did, of course, exist. PCI members had little opportunity to control candidate
selection since the involvement of the national executive began very early in the process. DC local elites had
some power, which was reflected in the structure of the traditional candidate list: a top group of candidates
(maybe with some national names), a second group of ‘local’ and ‘strong’ incumbents, and finally the rest of
the nominees in alphabetical order. For a more detailed description see Wertmann (1988).
452 Luca Verzichelli
prerogatives of the voters and thus protect the privileges enjoyed by their elites.7 But
the real change happened only with the electoral revolution of 1994. Since this election,
a remarkable instability of parliamentary parties and a high degree of party indiscipline
have characterized parliamentary life, thus impeding efforts to enhance citizen control
(Verzichelli 2003). A decisive step towards the consolidation of the new party system
occurred with the 2001 election, but this is not a sufficient condition to argue that
constituency control and individual accountability of MPs have emerged. Yet, a rather
unusual lack of incumbency power persists after three applications of a first past the post
electoral system (Verzichelli 2002).
At the end, while the old partisan model of political recruitment is still at work, a
clear decay of information (Strøm 2000), due to a deep and persistent crisis of the party
system, seems to endanger the tools of ex ante voter control of politicians.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T:
A S LOW C O N S O L I DAT I O N
Constitutional Features
Formal mechanisms of government formation have remained unaltered since the
establishment of the Italian Republic. After consultations involving parties and institu-
tional personalities, the President of Republic appoints the PM and, on advice of the
PM, the other members of the cabinet. In reality, parties have been the key negotiators,
reducing the role of institutional actors and imposing their own nominees (Verzichelli
and Cotta 2000). According to Article 92 of the Constitution, Parliament is not formally
consulted during the first phase of the process. Nonetheless, during preliminary consul-
tations, the President of Republic typically invites the speakers of the chambers and the
former Presidents of the Republic to discuss the political situation. Party leaders and
parliamentary whips (capi-gruppo parlamentari) also participate in the consultations. This
practice can be seen as a screening procedure designed to find a stable parliamentary
majority, though it does not seem to be terribly important in the appointment process.
Italy features positive parliamentarism (Bergman 1993) insofar as the Constitution
requires that governments win a mandatory vote of confidence in both Chambers of
Parliament within 10 days of the appointment of ministers (Article 94). The investiture
vote was debated for a long time by those who drew up the Constitution and remains
a complicated matter for scholars. The Constitution aims to give the cabinet a strong
parliamentary legitimization. The two-Chamber investiture vote requirement was
intended to force the Government to establish a stable majority in each chamber. In
reality, the symmetrical political configuration of the two chambers, a consequence of
the PR voting systems, made the second vote (in the Senate) unnecessary and
7
A good indicator can be the consolidation of the abrogative referendum, which is discussed below.
Another progressive sign of change is the emergence of a number of parliamentary actors with different
form of ‘anti-party’ message: the Radicals (1976), the Greens (1983), other minor formations between 1987
and 1992, and a more conspicuous number of politicians elected by the new parties after 1992 (Verzichelli
2000a).
Italy 453
predictable.8 The situation is a bit different since the electoral reform of 1993.9 This is
partly due to a minor rule difference between the two chambers with regard to the dis-
tribution of the seats under the proportional quota (Katz 1996), but it is also caused by
uncertainty in the party system in the wake of the breakdown of the old partitocracy.
In the past, the roll-call votes of confidence were characterized by a high degree of
party discipline. During the chaotic phase of the transition that started in 1992, how-
ever, individual MPs changed their partisanship, parties changed their coalition strate-
gies, and some parliamentary groups split (Verzichelli 2003).
In addition to the inauguration vote, there are two types of confidence votes in Italy:
the no-confidence motion (mozione di sfiducia), which can be introduced by MPs, and
the confidence motion (mozione di fiducia), which is an instrument of the government.
The former, which can be directed against a single minister or the whole cabinet, must
be signed by at least one-tenth of the members of either house and can be voted on
3 days after its submission. The quorum requirement is the simple majority (50 per cent
⫹ 1 of MPs), while the majority requested to censure a cabinet member or the entire
cabinet is 50 per cent ⫹ 1 of the MPs present and voting.10 The same rules apply to the
confidence motion, which can be tied to a single piece of legislation like bills or reso-
lutions (questione di fiducia).11 The cabinet has to request a confidence vote at least
twenty-four hours before the scheduled vote on the relevant legislation.
8
The order in which the chambers hold the vote of confidence alternates: the Chamber goes first after
one election, then the next time the Senate votes first.
9
In 1994 Berlusconi had a stable majority in the lower chamber (366 yes and 245 no in the inauguration
vote), but was forced to gain support from some independents in the Senate (and even the vote of one MP
elected from the Popular party) to obtain a very slight majority (159 votes to 153). Prodi’s difficulties (1996)
arose in the Chamber, where he won the vote of confidence thanks to the external support of the neo-
communist party (322 votes to 299). In the Senate, MPs from the coalitional parties were sufficient. After the
2001 elections, a solid majority supports the Berlusconi II cabinet in both the chambers. However, an impor-
tant difference is that at the Senate the votes of the Northern League can be necessary to pass the cabinet’s
initiatives.
10
In the Chamber of Deputies, this means the majority of MPs who actually vote, while in the Senate the
rule is the positive vote of the majority of the MPs who are present at the moment of the vote. Thus, in the
Senate abstention is in effect counted as a vote against the government.
11
The regulations of the lower Chamber forbid the use of this motion in cases of parliamentary investiga-
tion, personal questions, or the modification of internal regulations. The latter eruption (internal regula-
tions) applies to the Senate as well.
12
This did not usually occur on confidence motions, but the franchi tiratori often acted in the first secret
vote held after the confirmation of confidence. The abrogation of the secret vote in 1988 (with the exception
of voting on decisions about individuals) eliminated the phenomenon.
454 Luca Verzichelli
hand, used to verify their support in different arenas, especially in summits with party
leaders from the majority parties. Cabinet resignations very often occurred before a
vote on a parliamentary resolution, thus pre-empting a no-confidence vote. Hence the
prevalence of the so-called extra-parliamentary crisis: a resignation directly announced to
the President of Republic, before a decisive parliamentary vote. Even when a cabinet
was sent back to Parliament to verify its majority or when there was a ‘threat’ of pos-
sible defeat, the PM preferred to resign in anticipation—perhaps in order to obtain a
new appointment.
In contrast, the cabinet often uses the confidence motion to support its legislative ini-
tiatives. As already mentioned, this special procedure can be applied to normal
legislation (with few exceptions). A vote on a piece of legislation ‘marked’ by a request
of confidence is conducted according to the normal voting procedure (under existing
rules, a non-secret vote without a roll-call). Cabinets have employed the confidence
motion more than 230 times in either or both Chambers between 1948 and 2000 and
the practice increased in the 1990s. For example, the government’s annual financial bill
and proposed economic reforms are often amended in Parliament. The cabinet has
begun turning passage of these pieces of legislation into confidence votes to pressure its
parliamentary majority to accept the original proposals.
The lack of balance in the use of confidence votes (i.e. rarely by MPs, often by
governments) shows the gap between the actual delegation process and the one
formally envisaged in legislation, including the Constitution. The repeated use of
confidence motions, for instance, indicates the ‘ambitiousness’ of a government seeking
to control parliamentary behaviour in particular policy areas, even when Parliament
has not delegated authority in those policy areas to the government.
Table 14.1 summarizes the behavioural record with respect to cabinet investiture, no
confidence and confidence votes, and parliamentary dissolution. The table shows that
unsuccessful investiture votes have been rare; only five have occurred in the postwar era,
the most recent before the formation of the Goria cabinet in 1987, when the five-party
coalition was living a period of crisis due to the conflict between the DC and PSI.
Numerous cabinets have faced votes of no confidence, but no cabinet has been removed
through such a vote, though several have resigned to pre-empt one. The number of con-
fidence votes, as noted above, has been large, although only the Prodi government in 1998
was removed by a failed confidence vote. Seven cabinets ended in early elections, after
having been dissolved by a parliamentary majority.
A final constitutional tool to delegate legislative powers is the legislative delegated
decree (decreto delegato legislativo). It enables the Parliament to give the cabinet the power
to legislate in a given matter. A clear ‘ex ante’ control is provided by Article 76 of the
Constitution: ‘the exercise of the legislative function may be delegated to the government
only after having established principles and criteria, for a limited period of time and for
specific objects’. In practice, there are very few delegated decrees, though in recent years
there has been an increasing tendency to resort to them. For example, a 1992 delegation
of authority allowed the cabinet to produce an important decree on public finance. In
addition, both the technocratic cabinets (Ciampi 1993; Dini 1995) and the political ones
(Prodi 1996; Belusconi 2001) implemented important delegated decrees. According to
most observers, this is clear evidence of growing cabinet power (Calise 1994).
Table 14.1 Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Italy 1945–2000a
No. Cabinetb Date of Investiture votes (iv) No-confidence Confidence votes Did the Who
formation votes (ncv) (cv)c cabinet dissolved
end with Parliament
No. of Voting results of successful ive No. of Cabinet Cabinet No. of Cabinet an early in this
unsuccessful (1: pro gov., 2 :against gov., ncv removed resigned cv under removed election? case?d
investiture 3: abstentions, and 4: other) by ncv to spec. by failed
votes 1 2 3 4 preempt inst. cv
ncv f mech.
1 Parri 45/06/21 — — — — — — — — —
2 De Gasperi I 45/12/09 — — — — — — — — —
3 De Gasperi II 46/07/13 0 398 53 7 n.a. 0 —
4 De Gasperi III 47/02/02 0 292 53 1 n.a. 0 —
5 De Gasperi IV 47/05/31 0 274 231 4 n.a. 0 —
6 De Gasperi V 48/05/23 0 346 167 0 n.a. 0 1 No
7 De Gasperi VI 50/01/21 0 314 89 12 n.a. 0 3 No
8 De Gasperi VII 51/07/26 0 291 175 42 n.a. 0 3 No
9 De Gasperi VIII 53/07/16 0 263 282 37 n.a. 0 0
10 Pella 53/08/17 1 315 215 44 n.a. 1 No Yes 0
11 Fanfani I 54/01/18 0 260 303 12 n.a. 0 0
12 Scelba 54/02/10 1 300 283 1 n.a. 1 No Yes 8 No
13 Segni I 55/07/22 0 293 265 12 n.a. 1 No 7 No
14 Zoli 57/05/19 0 305 255 11 n.a. 0 1 No
15 Fanfani II 58/07/01 0 295 287 9 n.a. 0 1 No
16 Segni II 59/02/15 0 333 248 1 n.a. 0 0
17 Tambroni 60/03/25 0 300 293 0 n.a. 1 No 0
18 Fanfani III 60/07/26 0 310 156 96 n.a. 1 No 2 No
19 Fanfani IV 62/02/21 0 295 195 83 n.a. 2 No Yes 2 No
20 Leone I 63/06/21 0 255 225 119 n.a. 0 0
21 Moro I 63/12/04 0 350 233 4 n.a. 1 No 25 No
Table 14.1 Continued
b
No. Cabinet Date of Investiture votes (iv) No-confidence Confidence votes Did the Who
formation votes (ncv) (cv)c cabinet dissolved
end with Parliament
No. of Voting results of successful ive No. of Cabinet Cabinet No. of Cabinet an early in this
unsuccessful (1: pro gov., 2 : against gov., ncv removed resigned cv under removed election? case?d
investiture 3: abstentions, and 4: other) by ncv to spec. by failed
votes 1 2 3 4 preempt inst. cv
ncv f mech.
13
Every week there is at least one question period in each Chamber. Answers must be delivered within
two (for oral questions) or three (for written questions) weeks.
Italy 459
‘technocratic’ and political cabinets of the 1990s were able to conclude important
pieces of labour legislation and reforms of the welfare system. The cabinet also increas-
ingly appears to be in control of Italy’s European policies (including security and public
order, regulatory policies, and communications). The ‘new dualism’ created by cabinet
and Parliament at the sunset of the partitocratic age (Pitruzzella 1997) implies
a stronger cabinet role in ordinary policy-making and a clear dominance for the
Parliament with regard to institutional reform.
Another feature emerged recently, corroborating the new image of executive–
legislative relations depicted by the analysts (Capano and Giuliani 2002): the growing
screening role of legislative committees, more and more involved in the (cooperative)
work of ‘improving’ the quality of the cabinet proposals, but less and less engaged as
‘alternative sources of legislation’ through the use of direct legislative power and the
consequent production of leggine. This feature, which has to be interpreted in the direction
of a more rational application of the ‘delegation and accountability rule’, was already
noticed at the end of the first republic, especially observing the work of so-called
‘horizontal committees’14 Today, up-to-date information about legislative behaviour
and outcomes confirm this trend (De Micheli 1997; Capano and Giuliani 2003).
Another important issue is the role of extraparliamentary actors. With the
delegitimation of national and partisan political elites, regions and the President of the
Republic have in recent years become increasingly active. In particular, regions, which
used to have very little legislative influence, have begun to exercise their constitutional
prerogatives in national legislation—for example, by proposing legislative initiatives or
promoting abrogative referendums. Furthermore, during the 1990s, two active
Presidents sought to influence agenda-setting in a number of ways (see the section on
‘External Constraints’ below).
Legislatively, the Italian cabinet has always been very active, interpreting broadly its
constitutional responsibility to promote ‘urgent and necessary’ decrees with the force
of law (Article 77). The use of this decree authority, which increased progressively until
the early 1990s (De Micheli 1997; Capano and Giuliani 2002),15 best enables the
cabinet to impose its political programme when faced with a fragmented Parliament.
On the other hand, Parliament can respond with strong counter-force. Each decree has
to be put to a vote no later than 60 days after its publication. This form of parliament-
ary control has been used in very different ways. The parliamentary response to cab-
inet decrees has varied with the subject matter of the decree and the strength of the
cabinet that proposed it. Recently, Parliament has tended to delay the adoption of cab-
inet decrees and, occasionally, to reject them (Della Sala and Kreppel 1998).
In sum, the cabinet is a strong institutional agent by virtue of its constitutional
authority. This authority is counterbalanced by the parliamentary instruments
discussed earlier, which give the chambers the power to enforce parliamentary control.
14
Namely, they are the Budget committee and the Institutional Affairs committees. Thy are called ‘hori-
zontal’ because their main task is to monitor the reliability of all the proposals on the parliamentary agenda
with the existing norm concerning budgetary and constitutional constraints.
15
A recent judgement of the Constitutional Court limited to the use of the reiterated decrees by the cabinet.
460 Luca Verzichelli
These countervailing forces have continued to evolve historical perspective: the institu-
tionalization of the typical measures of cabinet accountability, for instance, required a
certain time (Cotta 1994; Furlong 1996). The same has been true for the increasing
delegation of legislative power to the cabinet. Potential agency losses have emerged in
the decline of parliamentary screening and selection mechanisms during cabinet for-
mation (especially under the recent transitional phase) and in the absence of effective
institutional checks on the cabinet. This has to do with the flexibility of the legislative–
executive relationship in the long cycle of Italian republican experience (Cotta and
Verzichelli 1996). Generally speaking, the cabinet remains relatively constrained.
However, the effectiveness of the principal (i.e. the Parliament) depends upon the locus
of interparty bargaining and the influence of ‘third powers’ like the President.
D E L E GAT I O N W I T H I N C A B I N E T: S T I L L A N
U N C E RTA I N R E S P O N S I B I L I T Y
Cabinet Structure
The complexity of the executive institutions is created by the Constitution, which
solemnly states that: ‘The government of the Republic consists of the President of the
Council, and of the Ministers, who together constitute the Council of Ministers’. Three
institutions are made into one. This institution has three different types of responsibilit-
ies as well as three actors all seeking to control the others in a complicated game of
checks and balances.
According to the Constitution, the PM sets the guidelines of the cabinet agenda.
Article 95 states that the President of Council of Ministers conducts and is responsible
for the general policy of the government. He is to ensure the unity and consistency of
the political and administrative programme ‘by promoting and coordinating the activ-
ity of ministers’. The same article states that all the ministers ‘are collectively respons-
ible for the decisions of cabinet, and individually for their own ministries’ acts’. This
constitutional provision links each minister both to an individual political agenda setter
(the PM) and to a collective actor (their respective ministries). Therefore, this cabinet
model is a mixed one: it cannot tout court be a collective government, because each
minister has a clear sector of responsibility. It is not a pure collegial government either,
because the PM is not absolutely a primus inter pares (Andeweg 1993). The progressive
use of the PM’s constitutional prerogatives to set policy guidelines and, above all, the
changes provoked by the reform of this office in 1988 (see below), particularly the
creation of a policy staff and the strengthening of the PM’s agenda coordination powers,
have implied a significant change in the delegation process, without fully resolving the
nature of the Italian PM.
From about twenty ministers in the early postwar period (roughly fifteen portfolios
plus a variable number of ministers without portfolio), the size of the cabinet has
progressively increased to more than thirty. If one counts junior ministers—who are
appointed on the basis of party (and faction) proportionality—then more than
Italy 461
100 political appointments were made when the last Andreotti government came to
power in 1991. It is easy to explain the increasing cabinet size with reference to the
enlargement of the governing coalition and the balance among governing parties
(Verzichelli and Cotta 2000). The presence of a Vice-PM is a clear example: this posi-
tion exists in some multiparty cabinets in order to balance the appointment of a
‘strong’ leader from the largest party as PM. However, after 1992 the size of the gov-
ernment has declined. The number of ministries and the number of ministers have
both been reduced. The number of junior ministers has also been reduced, particularly
in the technocratic cabinets, and the total number of political appointments is now
down to about seventy. Nonetheless, the 1997 reform, providing a remarkable reduction
of ministerial structure and a more functional link between executive units and
legislative committees, has been implemented in a rather ‘shy’ mode.
The PM-designate has the formal right to propose ministers to the President of
Republic, that is, the right to submit a name for each ministry. Much more complicated
is the question of who has the right to dismiss ministers. The Constitution is silent on
this point, and since there is no example of a minister being removed at the request of
the PM, many observers argue that the PM does not have this power. Technically, the
only mechanism to remove a minister is the individual no confidence motion in
Parliament, but in practice other instruments have always been used: the cabinet
reshuffle, the minister’s resignation with no substitution (in this case, the PM generally
takes on the portfolio responsibility ad interim), and in some cases a full cabinet crisis
followed by the birth of a very similar cabinet.
All ministers take part in the weekly meetings of the Council, but only portfolio
ministers have a full right to vote on all items. The golden rule among cabinet members
is unanimity. Nonetheless, reforms in the 1980s gave the PM some limited opportunity
to get his proposals passed by an absolute majority vote.
These characteristics push the Italian cabinet in a collective direction. At the same
time, however, the Constitution provides an opportunity for government by com-
mittee by permitting the creation of collegial bodies inside the cabinet. Despite this,
together with the cabinet councils established in the 1980s, only a few interministerial
committees have been active.
The reduction of this network into a distinction between the ‘collegial’ responsibil-
ity of the cabinet and the ‘autonomy’ of each minister—the traditional legalistic
approach—is of very limited use. Empirical analyses describe a system in which every
minister has a different set of resources and confronts a different set of opposing
powers depending on the nature of the department and his or her political ‘mandate’.
Some ministers are asked to make policy decisions (particularly the former ‘troika’ of
economic ministers, who were forced to work together). Others, the more generalist
politicians, control important areas of policy implementation, for example, agriculture,
communications, or industrial relations, in addition to operating as ‘watch-dogs’ with
regard to the PM’s political programme. A third type of cabinet figure is the ‘specialized’
minister who works in a specific policy area with limited political autonomy. This kind
of minister seems to have become more common, not only in the ‘technocratic
462 Luca Verzichelli
governments’ but also in the political cabinets where, some observers argue, a new
configuration characterized by a stronger PM is emerging (Fabbrini 2001).
Independently of the type of ministerial appointment, party control was tradition-
ally a defining cabinet feature. Until the end of the 1980s, parliamentary seniority and
party background were the natural prerequisites for a cabinet appointment. The pro-
portional representation of parties (and the DC and PSI factions) was the most impor-
tant criterion, which killed all initiatives aimed at specializing the ministers. As a result,
the typical government had a unifying element (the partyness of personnel) and a vari-
able one—that is, the experience and competence (and interest) of each minister.
Therefore, ministers with quite different individual roles were characteristic of cabinets
under the first republic. In contrast, a certain continuity seems to characterize the polit-
ical execut-ives of the period from 1994 (the centre-right Berlusconi cabinet in 1994, the
centre-left cabinets between 1996 and 2001, and again a Berlusconi cabinet in 2001).
Nevertheless, there are some indications of a return to party government: for instance,
the use of majority summits among coalition parties (including those who provide
external support) and the creation of an informal directorium of leaders of the coalition
parties (Criscitiello 1996). Another related sign is the re-emergence of intraparty nego-
tiations (e.g. among heads of parliamentary groups, party experts, and individual min-
isters) on specific policy issues before the meetings of the cabinet. All these actions
traditionally promoted the cabinet as a collective body and reduced the powers of indi-
vidual ministers (Cassese 1981). At the same time, however, they limited the PM’s ability
to control and coordinate the cabinet. On the other hand, ‘majoritarian reforms’ have had
some obvious effects. The PM is now ‘indirectly chosen’ by the voters as leader of his elec-
toral cartel and his role of coordinating of ministers seems to be slowly increasing. These
limited changes can be interpreted as a transformation of the government decision-mak-
ing process in a more hierarchical direction (Andeweg 1993): the PM seems to have
strengthened his political authority, while other ministers have become more
accountable with regard to the management of their specific competencies (Fabbrini
2001).
council (consiglio di gabinetto).16 This was followed by a new law (400/1988) which
re-enforced the PM’s control over the cabinet agenda and provided him with political
instruments to coordinate the actions of ministers. The same reform allowed the
PM to give instructions to implement cabinet directives and to suspend temporarily
a ministerial initiative in order to bring it up in a subsequent cabinet meeting.
Law 400/1988 authorized the creation a prime ministerial staff (Hine and Finocchi
1991) in order to support the PM’s constitutional coordination role. Yet, it did not clarify
the PM’s real impact as a policy coordinator. As David Hine argues (1993: 197), this
touches on a central issue in Italian parliamentary democracy. The strengthening of the
Prime Minister should imply not only new ‘positive’ power for the PM to direct the
ministerial team, but also ‘negative’ power to limit ministers’ own actions. The actual
situation with regard to the PM’s power is currently quite unclear. He is more than
a primus inter pares but is still unable to hold ministers accountable on a day-to-day basis.
The evolution of the role of the PM clearly has important consequences for policy
formulation and implementation. The PM can ask his personal staff to raise questions
and to provide independent data that can be compared with information from
ministers. He can review the operation of departments and coordinate intracabinet
disputes. On the other hand, in particular policy areas, such as administrative reform,
the prerogatives of the PM appear not to have been effectively enforced and have not
significantly enhanced his control over the administrative actions of the ministers.
Recent practices do not have much in common with the classic vision of collegial
government, insofar as they establish a new hierarchy inside the cabinet. On the other
hand, the governmental process has retained a collegial structure, something that
seems appropriate given Italy’s fragmented polity. In particular, the PM’s enforcement
authority makes individual ministers more accountable (and not purely administratively)
when they are delegated authority to draft a legislative act or conduct a preliminary
investigation. At the same time, some super-ministers tend to have the same role of
coordinator over specific issues—for instance, the Treasury Minister on economic issues—
as the PM has over government policy in general.
The traditional fragmentation of the Italian government remains the most evident
obstacle to delegation based on specific departmental functions and prime ministerial
steering and coordination. Recent experience suggests a paradox. While technical and
‘temporary’ governments were strongly directed by their respective PMs (with the
evident support of the Head of State), PMs from the ‘political’ cabinets (Berlusconi,
Prodi, and even D’Alema) seem, for different reasons, to have been weaker. The con-
stitutionally based logic of cabinet collegiality, developed during the first decades of the
Republic (Criscitiello 1994), is still present. In sum delegation within the cabinet is
fluid. Despite the formal division of labour provided by constitutional and ordinary
norms, different circumstances make the real game quite uncertain. Some of these
16
The size of consiglio di gabinetto was flexible. The practice was to include all the financial and economic
ministers, the foreign office, and the defence ministers. Another informal rule was to include at least one rep-
resentative from each coalition party.
464 Luca Verzichelli
circumstances are directly connected with the fragmented nature of the Italian multi-
party coalitions, which increases the number of veto points within the government.
Others are due to the different degree of accountability by individual ministers whose
competencies, duties, and policy autonomy are rather flexible over the time. Again, the
idea of coalition cycles (where each cycle includes caretaker, weak, consolidated, and
critical cabinets) can help explain these variations (Verzichelli and Cotta 2000). But one
can also discern a slow process of change, based on attempts to rationalize and enforce
delegation within the executive. An excellent example is provided by the Treasury
Minister, who has greatly increased (specially during the 1990s) his autonomy from
spending ministers and also become much more accountable to the PM.
Yet, the reforms of the past and those currently in progress cannot remove the risk
of agency losses. Particularly weak links remain in the definition of the ministerial
competencies (on the side of ex ante delegation) and in the real impact of the reports
to the PM (on the ex post side of accountability).17
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S :
C O M M U N I C AT I O N P RO B L E M S
A wide variety of departments and agencies, under direct or indirect cabinet control,
have emerged in Italy since the postwar reconstruction. The main reason was the
dominance of the administrative system (Hine 1993), which was made possible by the
consolidation of a huge class of professional bureaucrats capable of translating and
interpreting the difficult ‘language’ of law. A second reason is the continuity of admin-
istrative elites. The new political class of the late 1940s had to make compromises with
the top bureaucrats who had survived the regime change. From that moment on,
control over bureaucratic recruitment became an important source of political struggle.
In principle, all positions in public administration are filled in open, public examination.
In reality, this method is used at low and middle levels, while the small, central top
administration (dirigenza generale) is directly appointed by the cabinet. Other basic
principles are that officials are to have generalist backgrounds and that administrative
uniformity is to apply to the whole civil service. Administrative unity means that all
posts other than the central top administration formally belong to a single class, even if
different levels of skill and pay exist.18 With the passing of time these general rules have
been weakened. Specialized training was introduced in some administrative branches
as early as the 1970s in the form of mandatory enrolment at a school for the higher
civil service (Scuola superiore di pubblica amministrazione). Later reforms introduced
17
As mentioned earlier, the implementation of the reform of the cabinet structure has not been very easy.
Other measures to reduce the public administration bill, both using the contracting-out system and creating
of specialized agencies, are still under discussion.
18
In the past there were eight functional divisions. Doormen or workers, recruited without a particular
educational title, made up the first four divisions. The middle ones were composed of office workers, who
generally held a high school diploma. Heads of sections, who held advanced degrees, were found in the two
highest divisions. Recent reforms reduced this fragmentation to three basic levels corresponding to educa-
tional requirements.
Italy 465
E X T E R NA L C O N S T R A I N T S
I already mentioned the influences of some external actors on the chain of delegation
and accountability. Here I will discuss the weight of these actors. The President of
Republic, who under Article 87 of the Constitution represents national unity, may send
messages to both houses of Parliament, promulgate laws, and issue decrees. The same
article lists other presidential prerogatives as commander of the armed forces, head of
the High Council of Judiciary, etc.—but these are not directly related to the parliament-
ary chain of delegation. The President is not formally invested with strong powers
related to the crucial link between cabinet and Parliament. On the other hand, because
he formally appoints ministers and dissolves the houses of Parliament, the Constitution
gives him a limited opportunity to exert influence.
The decade 1992–2001 has been referred to as a time of ‘semi-presidential manage-
ment of the Italian transition’ (Calandra 1996: 473; Pitruzzella 1997: 244), that is to
say a period bringing no real change in the rules of the game21 but, in the meantime,
19
The same law adopted an open system of evaluation for senior staff via the introduction of an evaluation
team (nuclei di valutazione) in each administrative unit.
20
Important steps have been taken in this direction, in particular the adoption of a law about transparency
in public administration (Law no. 241/1990) and the introduction of measures to simplify administrative
procedures (1993). Nonetheless, implementation will probably take a long time.
21
A proposal passed by a bicamerale committee in June 1997 endorsed a ‘weak’ semi-presidentialism:
a directly elected President with a 6-year term, who has responsibilities in matters of foreign and defence
Italy 467
discovering the possible practical evolution of the Presidential role. There is evidence
that supports this characterization: the active participation of the President in bringing
about institutional reforms, his active role in government formation (i.e. he proposes
some candidates for cabinet positions), and his increasing participation in setting the
policy agenda—by means of messages to Parliament, summits with members of gov-
ernment, and even explicit requests for policy initiatives. It is true that the President
was not irrelevant in the past, either.22 The difference between the old style of presi-
dency and the ‘transitional’ one probably lies in the degree of engagement. At one time
the President, who was never fully autonomous, occasionally acted as a ‘tool’ to pro-
mote the policy goals of some political group (e.g. a DC faction). Today the President
has become an initiator of important change, by taking advantage of the disappearance
of parties as ‘engines’ of political life (Cotta 1996; Pasquino 1997).
A similar development has occurred with regard to the role of the central bank. For
a decade after its ‘divorce’ from the Treasury in the early 1980s, Bankitalia played
a subordinate role in economic policy-making. Its leader (the governatore and his advisers)
were ‘guardians’ of economic soundness, but their suggestions were often rejected by
the executive. When the old ruling class was destroyed by political crisis, the central
bank had the opportunity to assume a significant role in determining the guidelines of
economic policy, and some of its important figures have even assumed powerful
government roles.23
While the changed roles of actors like the President and the central bank are
evidence of important differences between the first republic and the thereafter, transition,
the popular referendum has undergone much slower, incremental change. Since 1970,
when an ordinary law established the referendum,24 Italy has witnessed an increasing
number of requests to abrogate legislation. The Constitutional Court, which has to
verify that these requests conform with legislation, has interpreted constitutional
norms quite rigidly, and has therefore approved very few referendums. Over time, how-
ever, the Court has become more flexible, approving, for instance, referendums about
the electoral systems which, by abrogating existing rules, could create dangerous gaps
in legislation. Since the end of the 1980s voters have approved some referendums (and
thus abrogated laws). It is interesting to consider which actors use the referendums.
The referendum was first used in 1974 when Catholics (who are not directly
policy. As envisaged in the proposal, the President could be dismissed by a qualified majority in Parliament.
This proposal did not become law because the government and the centre-right opposition could not
reach an agreement, and all decisions concerning the new form of government were postponed to the next
legislature (the one begun in 2001).
22
President Gronchi, for instance, had a significant influence during the rise of a centre-left coalition in
the early 1960s, while President Pertini tried several times to balance the relationship between the DC and its
‘secular’ allies by appointing a PM who was not a Christian Democrat.
23
Carlo Azelio Ciampi and Lamberto Dini, technical Prime Ministers during the 1990s, were both top leaders
of Banca d’Italia. Interestingly, both decided to remain in politics: Ciampi as ‘expert’ economic minister (and then
as President of the Republic) and Dini (who started a new party in 1996) as Foreign Minister in the Prodi cabinet.
24
Law 352/1970 introduced the referendum. Article 75 of the Constitution forbids the use of referenda
on fiscal and budgetary legislation, amnesties, and the ratification of international treaties.
468 Luca Verzichelli
represented by the Christian Democrats) were mobilized to oppose the legalization of
divorce. Later, small and unconventional minorities (e.g. radical parties, the new Left,
the Greens) used the referendum to oppose the ‘consociative’ legislation of ‘national
solidarity’. During the 1990s, abrogative referendums have been called for by groups
representing ‘societal forces’ (meaning industrialists, associations) as well as by individ-
ual politicians such as Mario Segni, who was able to transform himself from a second-
tier DC politician to the leader of the pro-majoritarian movement. In 1995, regional
councils also began exercising their constitutional right to request abrogative referen-
dums on some pieces of legislation.
Between 1974 and 1987 nine proposals of abrogation were rejected by referendum.
Among them were a proposal to abolish divorce (introduced in 1970), an attempt
to rescind the law on abortion, and a proposal to cancel a reduction of wage-
indexation decided in 1984. The first two proposals were supported by the Christian
Democrats, the last one by the Communists. In 1987 five referendums passed. Three of
them called for a reduction in the use of nuclear energy. In 1991 and 1993 it was the
proportional electoral system to be changed by the popular vote. All the other seven
referendums held in 1993 were successful, abrogating various pieces of legislation. The
most significant ‘anti-party’ outcome was the abrogation of a large part of the 1974 law
on party financing. Moreover, three ministries were abolished (among them, the
famous Ministero delle partecipazioni statali, which had been a crucial agency for party
control over economic resources) and important reforms introduced in the health sys-
tem (giving more competencies to the regional governments) and drug detention.
In 1995, five out of twelve referendums passed. With the reconstruction of a bipolar
political spectrum, a clearer pattern of alliances between societal and political forces
emerged, but with mixed results. Some ‘neo-liberist’ proposals by the radicals and by
the industrialists (and strongly supported by the centre-right leader Silvio Berlusconi)
cancelled old privileges of the national trade unions (namely, the automatic union affilia-
tion and the special quota reserved for these organizations). The Left and the unions
avoided an even worst result, as some other proposals were rejected. On the other
hand, the same actors (plus a number of other political and non-political promoters)
failed to gain a majority for the so-called ‘anti-Berlusconi’ referendum: two proposals
to abrogate legislation from the 1980s on television network concessions and the use of
paid television advertisements for political purposes.
In the last three rounds of referendums, in 1997 (7), 1999 (1), and 2000 (7), a lack of
voter quorum (50 per cent ⫹1 of the registered voters) has resulted in three consecu-
tive failures. Even the proposal to abrogate the party ballot in lower chamber elections
did not reach the quorum (neither in 1999 nor in 2000), despite the (formal) support of
the most important parties. This appears to be a clear sign of decay for an important
‘external’ democratic tool which has been crucial in reshaping the democratic chain of
delegation.
In addition to its role in screening referendum proposals, the Constitutional
Court decides on controversies about the constitutionality of laws, acts emanating
from the regions, and also on conflicts between constitutional powers (Article 134).
Italy 469
C O N C LU S I O N
Parliamentary democracy in Italy has always shown much more flexibility than one
would suspect given its rigid constitutional framework, as it has continuously adapted
to changing circumstances. Traditionally, a partisan parliamentary class representing
different political traditions controlled the individual actors who collectively formed
the government. Bargaining among (and inside) parties reduced the opportunities for
well-defined delegation relationships. Another evident deviation from the ideal-type of
a parliamentary democracy was visible in the last link in the chain, where the ability of
the cabinet to coordinate the bureaucracy has always been limited.
During the last decade, the crisis of old parties and the delayed consolidation of a
new party system has changed the political framework. In the wake of the decline of
the established political class, voters and other principals—for example, the PM vis-à-
vis his ministers—have regained some of their powers to hold their agents accountable.
The previously powerful intervening actor—the political parties—have lost some of
their prerogatives. However, a new set of external forces has begun to assert itself. The
final outcome is, at the moment, unclear, but we can argue that, despite the extraord-
inary expectations raised between 1992 and 1994, another incremental change, rather
than a revolutionary one, has occurred. Thus, the recent transitional phase is another
example of the adaptability of the ‘old’ constitutional order.
In theory, the legal framework provided by the Constitution established a set of
mechanisms which formed a chain of delegation. With the exception of the link
between voters and MPs, the mechanisms of control in the hands of principals were
well defined. In reality, however, the structure worked in very different ways. As has
been explained above, the major limitations on interpreting Italian parliamentary
democracy as a chain of delegation and accountability have to do with two critical
links. The first is delegation from the cabinet to individual ministers who sometimes act
autonomously and competitively. The second problem is the lack of accountability of
many civil servants.
In general, the whole chain has suffered as a result of the weak institutionalization
of delegation mechanisms, as the story of executive–parliamentary relations illustrates.
The reasons for this weakness are to be found in the historical circumstances of Italy’s
democratic experience, in particular a highly fragmented, polarized polity and a society
with deep cleavages. Throughout the history of the Republic voters have not been
active and effective principals, and the political class—strongly linked to parties—has
not been held accountable. On the other hand, the distance between political and civil
society also explains the lack of ex post control mechanisms in the hands of cabinet
ministers vis-à-vis civil servants. That is to say, ministers do not enjoy the benefit of
attentive social interests (‘fire alarms’) who continuously monitor the performance of
government agencies.
Several environmental factors have constrained democratic delegation and account-
ability. The first is the legacy of party rule born during a war of liberation and result-
ing in a high level of partitocracy. The second factor is the continued existence of an
administrative and legal system born in a proto-democratic era and consolidated under
Italy 471
REFERENCES
Andeweg, Rudy (1993). ‘A Model of the Cabinet System: The Dimension of Cabinet Decision
Making Processes’, in Jean Blondel and Ferdinand Müller-Rommel (eds.), Governing Together:
The Extent and Limits of Joint Decision-Making in Western European Cabinets. London: Macmillan.
Bardi, L. and Morlino L. (1994). ‘Italy: Tracing the Roots of the Great Transformation’, in
R. S. Katz and P. Mair (eds.), How Party Organise. London: Sage.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
Bindi, Federiga M. (1996). ‘Italy: In Need of More EU Democracy’, in Svein S. Andersen and Kjell A.
Eliassen (eds.), The European Union: How Democrtatic Is It? London: Sage.
472 Luca Verzichelli
Bufacchi, V. and Simon Burgess (2001). Italy after 1989. Events and Interpretations. London: Palgrave.
Calandra, Piero (1996). I Governi della Repubblica. Vicende, Formule, Regole. Bologna: Il Mulino.
Calise, Mauro (1994). Dopo la Partitocrazia. Torino: Einaudi.
—— and Mannheimer, Renato (1982). Governanti in Italia. Un Trentennio Repubblicano. Bologna:
Il Mulino.
Capano Giliberto and Giuliani, Marco (2002). ‘Governing Without Surviving? An Italian Paradox:
law-making in Italy 1987–2001’. Journal of Legislative Studies, 7(4): 13–36.
—— and —— (2003). ‘The Italian Parliament: in Search of a New Role?’ Journal of Legislative
Studies, 9(2): 8–34.
Cassese, Sabino (1981). Esiste un Governo in Italia? Roma: Officina.
—— and Franchini, Carlo (1994). L’amministrazione Pubblica Italiana. Un Profilo. Bologna: Il Mulino.
Cotta, Maurizio (1979). Classe Politica e Parlamento in Italia. Bologna: Il Mulino.
—— (1990). ‘The Centrality of Parliament in a Protracted Democratic Consolidation: The Italian
Case’, in Ulrike Liebert and Maurizio Cotta (eds.), Parliament Consolidation in Southern Europe.
London: Pinter.
——(1994). ‘The Rise and Fall of The Centrality of the Italian Parliament’, in G. W. Copeland and
S. C. Patterson (eds.), Parliaments in the Modern World. Ann Arbor: The University of Michigan
Press.
—— (1996). ‘La Crisi del Governo di Partito all’Italiana’, in Maurizio Cotta and Pierangelo Isernia
(eds.), Il Gigante dai Piedi di Argilla. La Crisi del Regime Partitocratico in Italia. Bologna: Il Mulino.
—— and Verzichelli, Luca (1996). ‘Italy: Sunset of a Partitocracy’, in Jean Blondel and Maurizio
Cotta (eds.), Parties and Government. London: Macmillan.
Criscitiello, Annarita (1994). ‘The Political Role of Cabinet Ministers in Italy’, in Michael Laver
and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge:
Cambridge University Press.
—— (1996). ‘Alla Ricerca della Collegialità di Governo. I Vertici di Maggioranza dal 1970 al 1996.’
Rivista Italiana di Scienza Politica, 26: 365–89.
Della Sala, Vincent and Kreppel, Amie (1998). ‘Dancing Without a Lead: Legilsative Decrees in
Italy’, in John M. Shugart and Matthew S. Shugart (eds.), Executive Decree Authority. Cambridge:
Cambridge University Press.
De Micheli, Chiara (1997). ‘L’attività Legislativa dei Governi al Tramonto della Prima
Repubblica’. Rivista Italiana di Scienza Politica, 27: 151–88.
Dente, Bruno (1990). ‘Introduzione’, in Bruno Dente (ed.), Le Politiche Pubbliche in Italia. Bologna:
Il Mulino.
Di Palma, Giuseppe (1977). Surviving Without Governing. Berkeley: University of California Press.
Fabbrini, Sergio (2001). Tra Pressioni e veti. Il Mutamento Politico in Italia. Roma Bari: Laterza.
Furlong, Paul (1996). ‘The Italian Parliament and European Integration. Responsibilities, Failures
and Successes’, in Philip Norton (ed.), National Parliaments and the European Union. London:
Frank Cass.
Fusaro, Carlo (1994). ‘Media, Sondaggi e Spese Elettorali’. Rivista Italiana di Scienza Politica,
24: 427–64.
Hine, David (1993). Governing Italy. The Politics of Bargained Pluralism. Oxford: Oxford University
Press.
—— and Finocchi, Renato (1991). ‘The Italian Prime Minister’. West European Politics, 14(2):
79–96.
Katz, Richard S. (1996). ‘Electoral Reforms and the Transformations of Party Politics in Italy’.
Party Politics, 2: 31–53.
Italy 473
I N T RO D U C T I O N
The Grand Duchy of Luxembourg was created by the Great Powers at Vienna in 1815
and became a unified constitutional parliamentary monarchy under its first liberal
constitution in 1848. The Constitution created a unicameral Parliament of deputies
directly elected in a two-round majoritarian system with restricted voting rights.
Democratization was temporarily reversed when the reactionary Constitution of 1856
created a non-elected Council of State with an important legislative role.
During the First World War,1 the Grand-Duchess Marie-Adelaïde exercised the
monarch’s prerogatives to dissolve the Parliament and keep the Catholic Party (Party of
the Right) in power without a parliamentary majority. Following the turmoil provoked
by this decision, a 1919 referendum settled the future form of the State in favour of the
monarchy. The Constitution of 1919 was a decisive democratic improvement. It stipul-
ated that sovereignty rests with the nation, introduced universal suffrage for men and
women, and created a new electoral system—list proportional representation (PR).
Environmental Constraints
Ever since its creation, Luxembourg’s small size has heavily influenced its choices. First,
it accounts for the unitary Constitution. Second, militarily weak and dependent on
foreign markets, Luxembourg has always participated in (or even initiated) international
political and economic agreements. From the German Confederation (1815–66) and the
Zollverein (1842–1918) to the 1921 Union économique belgo-luxembourgeoise (UEBL) and the
1944 Benelux Treaty on the Customs Union, the country has always been associated
with international cooperation and economic integration. Luxembourg’s participation
We thank former Prime Minister Pierre Werner for his generous comments on a draft chapter. We also wish
to thank Mrs Maryse Baustert, Mr Ben Fayot, and Mr Etienne Schneider.
1
Despite its neutrality, Luxembourg was occupied during Second World War. Yet, domestic civil author-
ities continued to function, as German military authorities worked along with them. The Grand Duchy
adopted a neutral policy to avoid making the occupation worse, but the behaviour of Grand Duchess
Marie-Adelaïde (and the Catholic government) was widely seen as pro-German.
Luxembourg 475
Legislative Process
Since 1848, the Constitution stipulates that both the unicameral Parliament (the
Chamber of Deputies) and the monarch (in practice, his government)2 have the right
to initiate legislation and that the assent of the Parliament is required. The Grand Duke
signs and promulgates all bills, but a member of the cabinet responsible for imple-
menting them must countersign. The Grand Duke has not used his formal veto power
after the Second World War. The Council of State, designed in 1856 as the legal advi-
sor to the government, has a right of initiative, as well as a much broader role in the
legislative process. First, it gives compulsory advice on each bill or amendment.
Though the advice is non-binding, voting in Parliament cannot take place until it is
received. Second, it can decide whether or not legislation must be put to a second vote
(Article 59). The Constitution stipulates, in effect, that all bills must pass a second vote
in the Parliament no earlier than 3 months after the first vote. This second vote can
only be avoided if plenary sessions of the Parliament and Council of State decide that
it is unwarranted. The Council of State thus has a suspensive veto, which in practice is
rarely used (Als and Philippart 1994).
Until 1990, the parliamentary Standing Orders required a vote to authorize the
consideration of a legislative proposal before the proposal could be sent to the Council
of State. Since 1990, the Speaker of the House decides on behalf of the Parliament, and
then submits the proposal to the relevant committee as soon as the advice of the
Council of State is received. Legislative proposals are considered in one or several
permanent specialized committee(s) (introduced by the 1965 revision of the Standing
Orders),3 or in a special ad hoc committee. The committee(s) in charge draft(s) a report
that is (are) then discussed in a plenary session. Under the urgency procedure, a bill can
go through committee and plenary deliberation even if the advice of the Council of
State has not yet been received, but the Parliament will nevertheless have to wait for
this advice before the final vote. More than half of the MPs must be present for voting
to be valid. Voting is done article by article. If any amendments are accepted (by the
successive voting procedure), then they must be sent to the Council of State. When the
Parliament receives the Council’s advice on the amendments accepted, a new ‘statutory
vote’ on the articles and amendments is held, followed by a vote on the entire bill. As
with the first vote, a quorum of half of all MPs plus one is required, and a bill is passed
2
For simplicity and to avoid awkwardness, we use the male pronoun when we refer generically to the
monarch, a minister, or a legislator. During the period of interest, males were indisputably overrepresented
in these positions.
3
Before 1965 the Parliament was divided into three central sections, chosen by lot. The sections were not
specialized, and within a particular section the opposition might form a majority. Bills were examined simul-
taneously by the three sections, then by a group formed by the sections’ rapporteurs, before plenary delibera-
tion. This system was gradually replaced by specialized committees, which in some fields (e.g. agriculture,
finance) have been in use since before 1945.
476 Patrick Dumont and Lieven De Winter
if 50 per cent plus one of all valid votes are cast in favour of the bill (blank votes are not
counted). If the Council of State demands the second, so-called ‘constitutional vote’,
the Parliament has to wait at least 3 months before definitively adopting the bill (with
the same quorum and decision rule). On government-initiated legislative proposals,
including budget bills, the avant-projet de loi is examined first by the Council of State.
The latter may introduce modifications or even a counter-proposal (by urging the
government to legislate or to modify the existing legislation, the Council of State even
has a right of initiative). If it agrees with the Council of State, the government may ask
the Council to draft the amendments itself. The definitive proposal is then submitted
to the Grand Duke for authorization to introduce it in Parliament. The rest of the
procedure is the same as for parliamentary initiatives.
A 1924 law established other legislative actors—the elected professional chambers
(see below). While they are not enshrined in the Constitution, and therefore failing to
consult them would not constitutionally nullify the law, in practice the Parliament usu-
ally waits for their (non-binding) advice before holding a final vote. Hence, they play an
important role on bills relating to particular professional interests, including appropria-
tions in the national budget. In contrast to laws, governmental decrees (‘Grand Ducal
decrees’ and ‘regulations’, see below) passed without the advice of the professional
chambers can be rendered null and void.
Since 1966, the government may also consult the Economic and Social Council (see
below) on broader economic, financial, and social legislation. Although this body also
has the right to make proposals to the government based on the reports it drafts at the
government’s or its own initiative, it plays no part in any compulsory phase in the
legislative process. Its main function is rather to help reconcile divergent interests
(e.g. conflicting advice from the professional chambers). A 1977 government bill created
the ‘tripartite committee’ (see below) for crisis management. It issues advice when one
of three well-defined unemployment thresholds is reached. Not only is this process
compulsory, but in practice its recommendations are also binding. The Parliament does
not take part in this interchange between labour and government representatives. In
principle, it remains sovereign and may accept, refuse, or modify the deals made by this
committee but in practice it is required to ratify agreements and to formalize them into
law. This new legislative procedure is therefore a more serious deviation from pure
parliamentary democracy than the others are.
The constitutional amendment process has two stages. First, a declaration of revi-
sion enumerates all articles to be amended, without specifying the nature of the
change. Parliament is dissolved immediately after passing this declaration. The normal
quorum and decision rules apply. Once a new Parliament has been elected, the consti-
tutional articles mentioned in the declaration of revision can be amended through
a parliamentary vote in which three-quarters of all MPs are present, and two-thirds of
the votes (blank votes excluded) are cast in support of the amendments. International
treaties negotiated and signed by the government (officially the Grand Duke) must be
ratified by the Parliament. Since 1956, international agreements that involve transfers
of authority to supranational organizations are subject to the same quorum and
decision rule as constitutional amendments.
Luxembourg 477
Otherwise, the powers of the Grand Duke (in casu his government) are constitu-
tionally limited to the making of decrees and regulations necessary to the execution of
laws. He cannot suspend a law or dispense with its execution, nor initiate autonomous
regulations on matters that have not been legislated. ‘Grand Ducal decrees’ are
governmental decisions aimed at implementing an act of Parliament, whilst ‘Grand
Ducal (or public administration) regulations’ are decrees limited to administrative or
policing matters. Both of these instruments are adopted in the Council of Ministers,
signed by the Grand Duke and countersigned by a minister, and finally published in the
official journal. For either type of governmental decree to be promulgated, the advice
of the corresponding professional chamber(s) must have been requested and received.
Administrative and policing regulations must also receive the advice of the Council of
State, except if the Grand Duke invokes an urgency provision. Finally, if it is stipulated
in the law, instruments issued in the execution of the law may be submitted to the
Parliament’s Commission de travail (working committee, Conférence des présidents since
2000) for advice. ‘Ministerial decrees (or regulations)’ are supplementary rules for the
execution of a law issued by a single minister. The nonconformity of all these decrees
and regulations to a law can be sanctioned by the courts. Other specific executive
decrees are used for the implementation of European directives (see below). Finally, by
adopting a bill called loi habilitante, Parliament delegates to the government its compe-
tencies in specific policy areas for a fixed time. The government can then take impor-
tant measures in these fields through Grand Ducal regulations (see below).
Featuring an indirect chain of delegation and a reasonable correspondence to the
singularity principle, the Grand Duchy of Luxembourg thus presents a number of
characteristics that approximate the ideal type of parliamentary democracy. The coun-
try is a unitary parliamentary monarchy with a unicameral Parliament, and it has never
used a referendum in the postwar period. Yet, a number of domestic institutions and
policy-making procedures do deviate from this ideal–typical picture. Another con-
straint has been the country’s involvement in international organizations and arrange-
ments that continuously reduce its sovereignty and thus the significance of the national
chain of delegation and accountability.
P O L I T I C A L PA RT I E S
Because political parties are not legally recognized, they receive no governmental
subsidies for central party offices, which are very poor in both infrastructure and per-
sonnel. Parliamentary groups (which must include at least five MPs) were recognized
in the Parliament’s Internal Rules in 1965 and have since received state support to main-
tain offices and staff.4
A four-party system existed from the 1920s to the early 1980s. Socialists, Liberals, and
Catholics became structured parties between 1902 and 1914, while the Communist Party
was founded in 1921. The system was rooted in a two-dimensional cleavage structure.
4
A ‘technical group’ was created in 1990 for parties that did not have enough MPs to form a parliamentary
group.
478 Patrick Dumont and Lieven De Winter
The socio-economic class cleavage was superimposed upon the older clerical/anti-clerical
one, which as early as 1848 pitted Catholics against Liberals (Hearl 1987). Language has
never been a conflict issue, although the country lies at the crossroads of the Roman and
German worlds.
The CSV (Christian Democrats) has always been the largest party in Parliament, usu-
ally followed by the Socialists, Liberals, and Communists (in that order). Between 1968
and 1989, when new parties first performed better than the KPL (Communists), the
four traditional parties’ combined share of the vote dropped from 99.6 per cent to 80.2
per cent. The Greens appeared in 1984, fought the 1989 elections on two separate tick-
ets, but reunited in 1993. The Aktiounskomitee fir Demokratie a Rentegerechtegkeet (ADR)
was created in 1987 as a pressure group for increased private sector pensions. It has sub-
sequently become a single-issue, populist party. The Communist Party lost
parliamentary representation in 1994, but re-entered Parliament in 1999 (one seat)
under the name of Déi Lénk (DL), an alliance of the KPL and a number of extreme-left
organizations.
From 1945 onwards, Luxembourg’s party system has shared all the features of
Sartori’s ‘moderate pluralism’ (Sartori 1976: 173). All governments have been coali-
tions, and centripetal competition has favoured the most centrist party (CSV), which
typically has governed in coalition with Socialists or Liberals. However, the small
ideological distance between relevant parties has enabled all possible two-party
coalitions (CSV–LSAP, CSV–DP, but also DP–LSAP from 1974 to 1979) to emerge. The
Communist Party, the only anti-system party, was excluded from all coalition negotia-
tions after 1947. Furthermore, Luxembourg has displayed extraordinary cabinet and
prime-ministerial stability. The five most recent governments have all served full terms,
and there have been only nine Prime Ministers since 1919.
Parties influence the chain of democratic delegation at various stages and to various
degrees. Candidate selection takes into account both the segmented nature of
Luxembourg’s society and the specific features of the electoral system. Each of the
three traditional parties is part of and seeks to represent an extensive network, a more
or less integrated ‘pillar’ of social organizations. Parties also try to recruit local nota-
bles and media figures not strongly associated with the ‘pillars’ to benefit from
panachage (see below). Candidate selection is a decentralized process in which the
initiative rests with the constituency party and lists are adopted nationally. All
constituency party members,5 or delegates of the lowest party institution, ratify the
selection. There are no retirement age limits, but during its stint in opposition (1974–9),
the CSV decided to refresh not only its ideology but also its personnel by introducing
a party statute stipulating that not more than two-thirds of the candidates could be
more than 40 years old.
5
Closed primaries are used only by the Greens. Within the KPL, the CSV (a quarter of whose lists are
drafted by the national party officials), the LSAP, and the DP, final candidate selection decisions are made
nationally, though national organs rarely reject the constituencies’ candidates. The major exception occurred
in 1974 when DP President Thorn drafted the party’s lists. In 1999 the Socialists began to use closed primar-
ies to choose their national Spitzenkandidat and the Spitzenkandidaten of the four constituencies.
Luxembourg 479
The electoral system (see below) reduces the importance of partisan candidate selec-
tion and even implies that, once elected, MPs are accountable to their constituency
voters. This gives MPs an important incentive to pay attention to local interests (to bring
home ‘pork’, to attend social constituency events, etc.). MPs who deviate from their
party’s national policy positions to protect their constituency do not usually have to
worry about losing their seat in the next election. Candidates of traditional parties (espe-
cially the CSV and the LSAP) who are officially supported by pressure groups may occa-
sionally deviate from party positions considered harmful to the group. However,
rebellion against the party as a whole would provoke harsh reactions in the powerful
partisan press and rejection by the ‘pillar’. Hence, party splits in parliamentary voting are
rare.
Permanent monitoring and reporting enhance party cohesion in legislation. Private
members’ bills are subject to the approval of the parliamentary group, which also
decides on party voting discipline. Hence, an MP’s voting discretion is usually very
narrow: either he complies with the group’s decision, or he declares in advance in
a group meeting that he wants to abstain. Parliamentary party meetings are attended
by several different party organs: national MPs and their parliamentary staff, MEPs, the
party’s government members, and leaders of the party organization. In the traditional
parties, members of the Council of State, representatives of newspapers that support
the party, as well as representatives of sub-organizations (e.g. youth and women’s
organizations) are also invited. These meetings are not public. In coalition government
parties, the meetings also serve as preparatory sessions before the ‘inter-fractional’ (or
coalition parliamentary group) meetings (see below), and thus help to assure coalition
loyalty in Parliament.
Political parties more effectively structure the link between Parliament and cabinet
than the one between voters and MPs. Although cabinet ministers have frequent meet-
ings with their respective party presidents, the Prime Minister (PM) or the vice-PM is
the real leader of each party’s ministerial team. A party president customarily resigns
this position when he becomes PM or vice-PM. During the late 1960s and early 1970s,
party organizations were more influential and managed to solve coalition conflicts
through ‘inter-fractionals’. Since the 1980s, when almost all coalition party heavy-
weights have had cabinet positions and parliamentary groups have received generous
financial aid, the national party leadership outside the cabinet has become less influen-
tial than the party’s cabinet and parliamentary team. Finally, political parties affect the
link between cabinet and civil servants through political nominations and promotions
for top civil service appointments (see below).
D E L E GAT I O N F RO M VOT E R S TO M P S
Voting in Luxembourg is compulsory for all citizens aged eighteen or older.6 The coun-
try is divided into four electoral districts, and MPs are elected for a five-year term (the
6
The minimal voting age was lowered from twenty-one to eighteen in 1972.
480 Patrick Dumont and Lieven De Winter
number of terms they can serve is not limited).7 Until 1954, the term was 6 years, with
half of the Parliament—two of the four constituencies—up for re-election every
3 years. Elections were held in the southern and eastern constituencies at the same time,
with the central and the northern district constituting the other pair. Before 1989, the
number of seats in each constituency was linked to total population figures. Today
the Constitution stipulates a fixed number of parliamentary seats, 60, which can only
be changed through a constitutional amendment. A law passed by a constitutional
majority determines the allocation of seats to the four constituencies. Since 1989, the
south has elected twenty-three MPs, the centre twenty-one, the north nine, and the east
seven. Proportionality is determined within the four districts, not nationally. Articles
136, 137, and 138 of the Electoral Code establish a first allocation of seats as follows:
the total number of valid constituency votes is divided by the total number of seats plus
one and rounded up in order to produce an electoral quota (the Hagenbach–Bischoff
quota). After each list has received as many seats as the number of times its total vote
meets the quota, the remaining seats are filled through a largest average method. This
formula discourages newcomers in the two smallest constituencies.
As Hearl (1987) puts it, ‘Luxembourg possesses a unique and ingenious electoral sys-
tem, which not only ensures party proportionality but also makes every successful can-
didate’s seat dependent upon his personal vote rather than upon party preference’. As
a convention, parties do not rank their candidates, but list them alphabetically.
Moreover, if the voter casts a straight party vote list by ticking the appropriate cell, each
candidate is awarded one vote, regardless of party rankings. The voter can also choose
specific candidates, either from the same party or from a variety of parties (panachage).
Voters who opt for preferential voting, whether intraparty or interparty, may cast
a maximum of two votes for each candidate (though the maximum number of prefer-
ence votes equals the total number of constituency seats). Since the 1970s preferential
voting has increased, from 31.3 per cent of the votes cast in 1979 to 41 per cent in 1999,
with the proportion of interparty panachage jumping from 18 per cent of the total vote
in 1979 to 29 per cent in 1999. Traditionally, those parties (DP and CSV) that presented
well-known candidates benefitted most from panachage. Parties based on a well-defined
ideology or/and local structures rather than specific candidates (KPL and LSAP) had more
disciplined and ‘exclusive’ electorates. The 1994 and 1999 election figures, however,
7
Apart from death, voluntary retirement, and the end of term or dissolution, there are various conditions
under which an MP can lose his seat. First, there are six positions constitutionally incompatible with the man-
date of MP (Article 54 of the 1848 Constitution): minister, member of the judiciary, member of the Chambre
des comptes, district commissioner, finance civil servant, and active army member. A seventh one (member of
Council of State) was added in 1948. According to a 1968 amendment to the 1924 electoral law, an MP who
accepts an incompatible position is to be removed from his seat (Article 101). Article 102 of the electoral law
stipulates that MPs cannot be the parents, close relatives, or spouses of other MPs. Since 1985, Article 100 enu-
merates more extensively than the Constitution the incompatible public positions (including the Railways
Company). Since 1984, the European MP mandate is also incompatible with the national one. Absentees can
also lose their seat in Parliament if they in two consecutive parliamentary years fail to attend at least half of
the meetings. The 1911 law on parliamentary enquiries stipulates that an MP can be imprisoned and lose his
voting and eligibility rights for 5–10 years after the Chamber has lifted his immunity. MPs can also be tem-
porarily expelled from Parliament for a maximum of six sessions if they prove to be trouble-makers.
Luxembourg 481
reflect a new trend: the LSAP has now become the party that relies most on inter- and
intraparty panachage votes (45.2 per cent), and the DP won the 1999 elections after
15 years in the opposition by mobilizing a proportion of list votes (above 60 per cent)
it had never reached before. The newer parties, the Greens and the ADR, constantly
had a figure of party votes higher than 60 per cent, but whilst for the former this pro-
portion decreased in 1999, the latter recorded a high of 73.9 per cent such votes (Centre
de Recherche et d’Information Socio-Politiques 1995; Centre de Recherche Public,
Gabriel Lippmann 2000).
Since 1974, the Liberals have chosen a Spitzenkandidat (top candidate) to lead their
campaign, win support for the party list, help people identify their candidates for PM
and ministerial portfolios, and attract a maximum of voters through panachage. The
remainder of the party list is presented in alphabetical order. Other parties sub-
sequently adopted this strategy—CSV in 1979, LSAP in 1984, followed by the non-
traditional parties in 1994. As expected, the Spitzenkandidaten score better than other
candidates, and their electoral fate is seen as a genuine popularity contest. Since pre-
ferential votes determine who gets elected, the voters decide which candidates become
MPs. Hence, parties have less influence over the composition of their parliamentary
group than they do in most countries with list systems.
Party organizations do not appear in any legal document and are not considered to
be legal entities. It is therefore impossible for the state to finance them or to control
their resources or campaign expenses. Only parliamentary groups, and therefore only
parties that are already represented in Parliament, get financial support to hire personnel.
Since 1999, political parties represented in Parliament receive public subsidies for
electoral campaigns, but there is no legal cap on campaign spending. As campaigning
has become increasingly costly, in both financial and manpower terms, richer candid-
ates (or parties) backed by extra-parliamentary organizations stand a better chance of
winning. Moreover, the fact that the main newspapers are either owned by or friendly
to a particular party benefits these parties.8 The publication of opinion polls is forbid-
den in the month preceding the elections. Electoral campaigns last between two
(parties must introduce their lists at least sixty days before the elections) and three
months (the maximum time between a parliamentary dissolution and the elections).
Several fairness rules govern broadcast time on national radio and television. Parties
that are not represented in Parliament are given time according to the number of
lists and candidates they are presenting (the minimum is a two-min spot), while the oth-
ers are granted time in proportion to their parliamentary strength. However,
governing parties tend to generate publicity by sponsoring cultural or sports events.
Since broadcast time cannot be freely purchased, this kind of publicity usually
advertises a specific candidate rather than the party.
Luxembourg’s electoral system not only influences the link between voters and MPs:
the Grand Duke usually asks the leader of the largest party to form a new government,
8
Until 1999, it was impossible for a party to buy space for political advertisement in a newspaper outside
its pillar. In 1999, the leading newspaper (the Catholic Luxemburger Wort) first opened its pages to parties other
than the CSV. The other newspapers decided to follow the same policy.
482 Patrick Dumont and Lieven De Winter
but the junior partner is almost always the traditional party that has made gains rela-
tive to the previous election. Moreover, coalition parties usually select ministers on the
basis of their personal electoral success. This direct link between electoral success and
office, for parties as well as ministrables, rests on convention rather than law. It repre-
sents a deviation from the simple and indirect chain of delegation common in parlia-
mentary democracy. Together with the growth of panachage voting, these conventions
make cabinet ministers accountable not only to the parliamentary majority, but also to
the voters. Accordingly, cabinet ministers try to promote their constituency interests,
sometimes against the will of their formal principal, the parliamentary majority.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
According to the Constitution, the Grand Duke nominates and can dismiss ministers. In
practice, his power has been very weak since 1945, and he only nominates the formateur.
The selection of ministers is left to the coalition parties, although the formateur can pro-
pose names for his own party. By drafting comprehensive coalition agreements before
the cabinet takes office, parties strongly condition the policy link between Parliament
and government. Although the coalition composition generally respects the will of the
voting majority, government formation is controlled entirely by political parties.
Since 1848, ministers may not simultaneously hold seats in Parliament, although they
have the right to attend parliamentary meetings. The principle of internal selection of
ministers is widely respected, and almost all cabinet members over the past 50 years have
come from the Parliament. Symbolic of ministerial accountability, the Parliament has the
right to require ministers to appear at any time. The first democratic Constitution intro-
duced the notion of ministerial responsibility. This very general provision (in so far as it
fails to spell out the authority to which the ministers are accountable or the nature of
their responsibility) reflects the fact that the Grand Duke is non-responsible (Article 45
requires that a minister must countersign all his acts). However, two other articles stipul-
ate that only the Parliament has the right to accuse ministers, thus implying their
accountability to the Parliament alone. Although the matters for which a minister can be
held accountable have never been spelled out in law, the convention is that political
accountability to the Parliament exists, and that certain types of votes imply dismissal.
occasion must positively endorse this agreement. A veto (or even the anticipation of
one) by any party congress means that negotiations have failed. In practice, however,
the party congress has always ratified the agreement, and mostly without a hitch
(Dumont and De Winter 2000). Once approved, coalition agreements bind all coalition
parties and all decision-making organs within each party.
By convention, a parliamentary investiture vote takes place after the reading of the
governmental declaration and the subsequent debate. Luxembourg uses a positive
formation rule (Bergman 1995: 40–55) in which a relative majority must actively
support the government (i.e. 50 per cent of the MPs plus one must be present, and
50 per cent plus one of the votes cast, blank votes excluded, must support the govern-
ment). Because party discipline is rigorously enforced in investiture votes, dissident
government MPs tend to express their discontent by verbally opposing certain provisions
of the coalition agreement, rather than by voting against the agreement as a whole.
Investiture votes can be seen as a screening procedure in the delegation from
Parliament to cabinet. However, the investiture motion introduced by majority MPs is
purely a formality. In the postwar period, no government has lost such a vote of confid-
ence, and there have been very few abstentions.
Motions of confidence and no confidence are not institutionalized or regulated. In
practice, when Parliament refuses to pass a governmental bill, it can be considered
a personal defeat for the competent minister. If he decides to resign, then cabinet
solidarity—among ministers of the same party or in the form of collegial accountability
(see below)—brings down the government. Parliamentary action has brought down
only two cabinets: one after a majority turned a vote on a specific motion into a ques-
tion of confidence (1958), and another when a cabinet resigned to pre-empt such a vote
(1966). In the first case an information mission unsuccessfully tried to avoid an early
parliamentary dissolution. In the second case, the coalition partners (CSV and LSAP)
managed to re-form a cabinet without early elections. Under this new cabinet,
Parliament was nevertheless dissolved one year before the normal end of its term, the
second premature dissolution in Luxembourg’s postwar political history. Table 15.1
summarizes the behavioural record with respect to cabinet investiture, no-confidence
and confidence votes, and parliamentary dissolution.
The Parliament’s vote on the government’s annual budget bill is considered to be
a vote of confidence, because under the Constitution the Parliament is sovereign with
regard to state finances. The budget vote thus serves as another ex ante control over the
government’s annual program. However, the influence of the Parliament over the
budget seems even lesser than with regard to the ‘investiture vote’. This is due in part
to the practice of ‘debudgetizing’—taking expenditures off-budget—which deprives
Parliament of its power to oversee large parts of state spending by providing subsidies
to semi-public bodies like the Social Security agency, but also by creating special invest-
ment funds for telecommunications, roads, etc. Parliament’s budget authority is also
undermined by the fact that it does not have the requisite resources for a thorough
scrutiny of the annual budget bill (1,000 –1,500 pages of highly technical articles). In
short, the Parliament cannot in this case effectively control its agent. In addition, the
government might deliberately hide information from the Parliament through, for
example, obscurely labelled budget items.
Table 15.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Luxembourg 1945–2000 a
No. Cabinet Date of Investiture votes (iv) No-confidence Confidence Did the cabinet Who
formation votes (ncv) votes (cv) end with an dissolved
under specific early election? Parliament in
No. of Voting results of successful iv No. of ncvd Cabinet Cabinet institutional this case?b
unsuccessful (1: pro gov., 2: against gov., removed resigned mechanism
investiture 3: abstentions, and 4: other)c by ncv to (n.a.)
votes pre-empt
1 2 3 4 ncv
9
The Speakership has with one exception gone to the party that holds the most seats in Parliament (the
CSV since 1945).
486 Patrick Dumont and Lieven De Winter
Council of Ministers) attends all the meetings of the Commission de travail to ensure that
the plenary agenda is drafted in full compliance with government interests. Thus,
although government bills are not formally given agenda priority, the government can
in practice always assure that its interests come first.
On permanent parliamentary committees, parties are represented in proportion to
their strength in the Parliament. Committee chairmen and vice-chairmen are elected
by committee members at the beginning of each parliamentary year, while committee
composition normally remains the same throughout the legislative term. Committee
competences generally correspond to the main ministerial departments, and almost all
chairmanships go to majority MPs (usually from the same party as the minister who
heads the corresponding department).10 Since committees have no deadline for their
reports to the plenary (unless the Speaker proposes one), majority members can block
opposition bills by voting against the relevant report, setting the committee agenda, or
failing to attend meetings so that no quorum exists. For these reasons, private member
bills often ‘die’ at the committee stage. Provisions in these bills, which majority parties
find interesting, are sometimes reintroduced in governmental bills.
Finally, inter-fractional meetings (meetings of coalition parliamentary groups) fur-
ther promote coalition cohesion on government-sponsored legislation. Although not
institutionalized in the parliamentary Standing Orders, these meetings serve to verify
that all government projects comply with the coalition agreement and to eliminate any
parliamentary initiative that violates it. Even though there are no official barriers
prohibiting majority MPs from exercising their parliamentary initiative, coalition gov-
ernance rules in practice limit its effectiveness. Since voting discipline in Parliament
depends on the results of inter-fractional meetings, the initiatives of dissident MPs are
never voted upon.11 ‘Inter-fractionals’ were often used in the 1970s, when cabinets did
not have comfortable parliamentary majorities. Such meetings became more struc-
tured (with regard to frequency, thematic agenda, etc.) in 1984–5. Although they might
seem to increase MP involvement in the legislative process, inter-fractionals actually
remove the decision-making process (at least on delicate political issues) from
Parliament. This is because only coalition MPs are informed and because their behav-
iour is constrained by strict voting discipline.
The low success rate of private members’ bills is also due to the fact that the Council
of State must give mandatory advice on every bill or amendment. Due to limited time
and resources, this politicized body gives priority to government bills, and thus ‘kills’
private members’ bills through delays. Moreover, by invoking urgency (with parlia-
mentary approval), the government can speed up the legislative process. In matters
deemed to be urgent, the Council of State is not asked to give mandatory advice before
the final vote (see below).
10
The petitions committee, which has no corresponding ministerial department, is normally chaired by
an opposition representative.
11
‘Extraordinary’ inter-fractional meetings can be held at the request of the PM. Nicknamed ‘the fire
brigade’, they were convened most often during the crisis years of the late 1960s, but have since been rare.
They have traditionally dealt with issues such as the interpretation of the coalition agreement. They have also
been held when the Council of Ministers has been unable to resolve a conflict.
Luxembourg 487
Parliament may, through a bill supported by a simple majority, grant the cabinet
‘special powers’ for a limited time (though government must report to the Parliament).
The government is only allowed to legislate on a restricted number of subjects and in
accordance with certain fundamental guarantees. This provision was first used in 1915 to
preserve the country’s economic interests in wartime. It was later used in response
to the Great Depression of the 1930s, again in 1939 to promote the country’s interests
in wartime, and then in 1946 to facilitate reconstruction. Over time this practice of
transferring legislative authority has become a ‘regular’ (rather than ‘crisis’) technique
of governance. In fact, while until the late 1950s Parliament tended to grant special
powers to the government unanimously, since then only majority MPs vote in favour.
While the original intent of special powers was at times ignored until the 1960s, since
then it has been abused every year, even though in truth Luxembourg’s socio-economic
problems have not been so severe as to justify the granting of special powers year in and
year out.
12
In the most recent legislative term (1994–9), the rate of reply was 94 %.
13
In contrast to other types of questions, the number of these questions that can be asked during each
legislative session is limited. Each parliamentary group can ask twice as many questions as it has members.
Luxembourg 489
A final type of parliamentary debate introduced in the early 1990s is a public debate on
the government’s foreign policy.
Accountability mechanisms also exist in parliamentary committees. Ministers and civil
servants may be called upon by a parliamentary committee to answer questions in con-
nection with the committee’s scrutiny of a government bill. Moreover, Article 64 of the
Constitution gives Parliament investigative powers, for which it may seek expert reports
and/or call witnesses to testify. A 1911 law gives the parliamentary committee of inves-
tigation the same powers as a criminal law judge. After almost half a century of disuse,
three parliamentary investigations have been held in the postwar era. In 1989, following
a parliamentary investigation, a minister was found guilty by the Parliament and forced
to resign. A third instrument is described in Article 24 of the Standing Orders, which stipu-
lates that a committee can, in order to prepare its legislative work, by majority vote seek
information and advice from persons or institutions not belonging to Parliament.
Because parliamentary committees already benefit from the advice of one or several
professional neo-corporatist chambers, such hearings are not very frequent. Hearings
with a greater scope may be requested by parliamentary committees and organized in
plenary meetings. They do not have to be linked to a particular bill, but rather are aimed
at gathering general information from a wide range of pressure groups and citizens’ asso-
ciations. The Parliament decides whether or not the hearing will be public. Because such
hearings are time-consuming, only four were held before the 1990s.
D E L E GAT I O N W I T H I N T H E C A B I N E T
Cabinet Internal Organization
The Grand Ducal decree of 1857 (since amended seven times) about the organization of
the government stipulates that all decisions that need the consent of the Grand Duke are
to be discussed and decided collectively by the cabinet (Majerus 1990: 174). The PM
chairs the Council of Ministers (cabinet meeting), sets the agenda, and coordinates
ministerial departments. Yet, the PM is not legally a hierarchical chief over the other
ministers (Bodry 1996). Although it is not a formal rule, either the PM or the vice-PM
acting on the PM’s behalf may nevertheless call for a cabinet discussion of any depart-
mental issue. This authority follows from the PM’s responsibility scrupulously to verify
that all ministers’ proposals conform to the coalition agreement (see below). Thus, a
minister’s individual responsibility is residual in the sense that the management of his
department’s competencies are subject to collective deliberation. Even if consensus is
always sought in the Council of Ministers, it is not required. Decisions can be made by
majority rule. In recent years, this has only occurred with regard to non-political
questions (i.e. appointments). Ministers who do not vote in favour of a particular
decision can excuse themselves from collective responsibility by recording their opposition
in the written report of the cabinet meeting. In this event, dissident ministers cannot be
held legally or politically accountable for the cabinet decision. If the dissident minister is
the minister responsible for implementing the decision, he is only responsible for executive
acts following the decision.
490 Patrick Dumont and Lieven De Winter
Since the cabinet is small, there is no institutionalized inner cabinet (though informal
contacts between the two coalition leaders—the PM and the vice-PM—take place very
frequently). Specialized ministerial committees are sometimes created, but only on
very pragmatic grounds. Recent examples include European affairs, employment, and
information policy. Even when there is a special committee, however, all issues are dis-
cussed and decided in the cabinet.
The position of junior minister was created in the 1950s to give ministers in very
large or highly compartmentalized departments an efficient, day-to-day assistant.
These Secretaries of State never belong to a party other than the minister’s.
Ministerial portfolios are concentrated in a few hands, all of whom have clear par-
tisan identities. Even PMs commonly head one or two departments beyond the
Ministry of State. Parties are very important in this regard, as they seek ministerial
competencies that are particularly important to their own electorate or departments
that are highly visible and have large budgets.
Government Agreements
Compliance with the coalition agreement is essential to cabinet stability. Apart from
regular elections, the main cause of cabinet resignation in Luxembourg is intra-
coalition conflict (though this has not occurred since 1968). The cabinet’s great respect
for the agreement has several explanations. First, ministers are all very familiar with the
agreement because most of them helped negotiate the government themselves. While
they know in advance that they are almost certain to get a ministerial post, however,
they do not know for sure which portfolio they will receive. Hence, most negotiators
participate in all plenary bargaining sessions and understand the spirit of policy inten-
tions on any topic. Second, and perhaps most importantly, ministers are constantly
monitored by the PM and the vice-PM, who are the guardians of the execution of the
coalition pact.
Changing the PM does not trigger the drafting of a new coalition agreement, nor
does it automatically lead to a new governmental declaration or investiture vote.
Partners may take the opportunity to amend the agreement, but extensive revisions are
rare, since the equilibrium reached during the formation of the preceding government
is difficult to maintain in renegotiations.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
Civil service positions are awarded on the basis of competitive exams. The candidates
with the best results may choose their department from among those that have vacan-
cies. Politicization in the middle and lower career levels is very low. However, top civil
service appointments and promotions are highly politicized, despite the formalism of
the process (collective cabinet decisions signed by the Grand Duke for appointment of
the five administrateurs généraux, the top civil servants who collectively cover all min-
istries). Cooperation between government and the civil service is assured via the
appointment of about fifty government advisers organized in a four-level hierarchy,
Luxembourg 491
who become the ministers’ personal staffs. Most are drawn from the top civil service to
ensure that they have the requisite experience. They keep their civil servant status and
often continue their careers after they leave their political posts. Other advisers are
private sector experts who become ‘civil serviced’ without having taken an exam.
A General Secretary of the Council of Ministers is chosen from among the more than
400 civil servants in the Cabinet Office. Because most of these ‘promotions’ are made
along party lines, those who are assigned to a department other than the one in which
they were previously working are likely to be reassigned to much less attractive positions
if their parties are not part of the subsequent government. Screening and selection
(ex ante) procedures thus also intervene in the relationship between ministers and civil
servants to ensure the former that their agents will not go against their preferences.
Ministers head one or more departments, while the Minister of State heads the entire
government administration.14 The accountability of civil servants working in ministries
is thus doubly indirect: They are accountable for their doings neither directly nor via
their departmental minister. Rather, citizens complain to the Minister of State, not to
the relevant minister. Individual ministers are nevertheless politically accountable to the
Parliament for the actions of their civil servants, and as such may be asked to respond to
parliamentary enquiries. On the other hand, MPs are only provided information from
the civil service through the government as intermediary. The problem of hidden informa-
tion, specifically the problem of obtaining information for the purpose of holding the
government accountable, is a constant concern of parliamentary committee members.
Orders to withhold (or delay) information come from cabinet ministers, and the politi-
cization of the top civil service means that ministers are obeyed, despite the fact that
hearings and the consultative advice of professional chambers (third-party testimonies)
are designed to mitigate this potential information problem.
We noted earlier that the government’s legislative predominance and the parlia-
ment’s lack of resources reflect the fact that law-making requires increasingly diversi-
fied competencies. In Luxembourg, the civil service and through it the executive
branch have the necessary human and financial resources to perform the functions
assigned to them. As a result, the government’s legislative initiatives and responses to
parliamentary questions and interpellations are well thought-out and professionally
drafted by civil servants. Thus, successful delegation from ministers to civil servants
tends to reinforce agency loss between the legislature and the executive.
E X T E R NA L C O N S T R A I N T S
There are a number of external actors who intervene in the chain of parliamentary
delegation and accountability in Luxembourg. The first three discussed below have
a potential impact, though in practice they do not constrain the delegation process. The
other three lead to serious deviation from the ideal-type of parliamentary democracy.
Keep in mind that the effects of the creation of a Constitutional Court in 1997 cannot
yet be judged.
14
The Ministry of State has always been held by the PM, except for the 1995–9 period ( Juncker I).
492 Patrick Dumont and Lieven De Winter
Head of State
The Grand Ducal decree of 1857 stipulates that all decisions that need the consent of
the Grand Duke are taken collectively by the cabinet and countersigned by a minister.
These decisions include all bills passed by the Parliament, Grand-Ducal decrees, and
high-level appointments. Yet, the powers of the Head of State are purely formal, and in
practice he cannot act autonomously in public life. Yet, the Grand Duke retains the
right to be consulted by his ministers. He occasionally influences the country’s political
choices, most effectively in times of political crisis. In general, political leaders appreci-
ate his mediation.
The Referendum
Direct democracy is enshrined in the Constitution (Article 51), but in practice no ref-
erendums have been held since 1937. Referendums are formally held only at the request
of the Parliament. Only three questions have ever been subject to a referendum, two in
1919 and one in 1937. Even though past referendums can be interpreted as non-bind-
ing, the question of their decisive or consultative character is not answered in any legal
document. The uncertainties associated with the referendum helps explain why parlia-
mentary majorities refrain from using it.
15
On the other hand, in 1982 the Martens government devalued the Belgian franc without even warning
the Luxembourg government.
Luxembourg 493
of State deem it unnecessary. This second constitutional vote is intended to give the
Parliament time to reflect (and thus compensates for the absence of a second chamber).
In fact it has rarely been used during the postwar period. The Council of State always
justifies its decision that a second vote is not necessary with a reference to the urgency
of the matter at hand. However, when the opinion of the Council of State fundament-
ally diverges from the Parliament’s—on average, almost once each year—the Council
refuses to grant an exemption.
The twenty-one members of the Council of State are appointed by the Grand Duke
(i.e. in casu the government), either on his own or on the basis of a list of three candi-
dates proposed either by the Chamber of Deputies or by the Council of State itself. As
a result, the Council of State is highly politicized and in effect colonized by the three
traditional parties. The Council meets behind closed doors.
(Neo-)Corporatism
Since 1924, there have been five professional chambers in Luxembourg: agriculture,
labour, commerce, private employees, and craftsmanship. In 1964, a sixth chamber was
added for civil servants. These bodies were created to represent all economic forces. Each
chamber is directly elected by those who work in the corresponding field. All wage earn-
ers and independent workers are obliged to pay dues to their respective chamber, and all
who pay have the right to stand as candidates for election. The chamber(s) give advice on
bills and governmental decrees concerning their particular professional interests. The
cabinet must have received this advice before it can approve a decree or a budgetary bill,
although it does not have to follow it. Nonetheless, the Parliament never neglects the
chambers’ views on matters that significantly affect them.
A 1966 law created a Economic and Social Council of thirty-five members: twenty-
eight appointed by the Council of Ministers on the basis of proposals from professional
organizations in twelve sectors (e.g. small companies, steel industry, banks, transport,
insurance, agriculture), and seven independents—four of whom are selected by the
twenty-eight, three by the cabinet. All members serve four-year terms. The Council
studies economic, social, and financial issues and can initiate studies on its own or at
the request of the government. During the first trimester of each year, the government
asks the Council’s advice on the government’s report on the economic, social, and
financial evolution of the country and on its related policy programme. The govern-
ment can ask for advice on matters of general interest and on questions of principle on
which the professional chambers have expressed divergent views. In the latter case, the
Council must provide a ‘single and coordinated’ opinion.
The Luxembourg model of neo-corporatism is based on mixed committees both at
the national and at the individual company level (‘cogestion’ was introduced in 1974 for
large companies) and has evolved with the unfolding of the steel industry crisis. A 1977
law triggered the formation of the ‘Tripartite Co-ordination Committee’, whose task is
to give advice on unemployment measures. Consisting of four members of the gov-
ernment, four employers’ representatives, and four union delegates, and presided over
by the PM, it meets whenever one of three unemployment thresholds is reached. Since
494 Patrick Dumont and Lieven De Winter
1977, it has given advice on other politically delicate matters and has helped social part-
ners reach consensus. In theory, the Chamber of Deputies is free to ignore agreements
reached in the Tripartite Committee, but in general it considers itself bound to ratify
them. As a result, this tripartite structure gives economic interests (‘social partners’) the
power to block the government’s socio-economic proposals. The golden age of the model
was between 1977 and 1983. Its importance has since declined (Govaert 1997; Hoffman
1996), but it remains an important feature of Luxembourg’s consensus democracy.
both national MPs and MEPs, does not seem to have much impact on European policy.
Although there are few persons in Luxembourg with competence in EU matters, coor-
dination in EU bodies mostly occurs along party lines.
Through 1995, over 90 per cent of EU directives had been transposed into
Luxembourg’s legal order through four types of executive decrees that do not require
parliamentary participation (Carlier 1995). Legislation is only used when a European
directive introduces a new obligation, or if the subject matter is one that the
Constitution vests exclusively in the legislature. Apart from these objective criteria, the
scope of economic and political changes included in the directive might be taken into
account. Luxembourg’s cumbersome legislative procedure, which has difficulty oper-
ating within EU deadlines, justifies the use of governmental decrees and thus further
enhances cabinet predominance vis-à-vis the Parliament.
C O N C LU S I O N
Luxembourg seems to fit rather well the simple and indirect chain of delegation and
accountability of parliamentary democracies. Strøm’s proposition that this chain might
be most applicable to small, homogeneous, and stable societies in which principals have
greater faith in their screening procedures is thus confirmed in this case. Deviations from
the ideal-type do exist, but almost all the links in the chain display features of
delegated power to an agent and accountability mechanisms aimed at controlling the
agent’s actions. Political parties play an important role, but they serve to channel the
delegation process from one actor in the classical chain to another, rather than distort it.
External actors, on the other hand, create some deviations from the singularity principle.
While a direct link between voters and ministers is discernible, this is not only a result
of an evolution found in most parliamentary democracies (Strøm, Chapter 3), but also
due to structural and conventional arrangements in Luxembourg.
Between voters and MPs, the electoral system assures a representation that is relat-
ively immune to party discretion, and that approximates the ideal-type of parliament-
ary delegation. However, long-standing conventions aimed at creating a high level of
responsiveness (e.g. the largest party is almost always in government, its junior partner
is usually the traditional party that had the best election performance, and ministers are
usually those persons who got the most preference votes) and increasing panachage vot-
ing have tended to establish a direct link between voters and cabinet ministers. This
presence of multiple principals could increase the likelihood of failed delegation.
Majority parties structure the link between Parliament and cabinet. Because parties
dominate all parliamentary bodies, draft comprehensive coalition agreements, and set
up very strict ministerial accountability mechanisms towards their party and the coali-
tion as a whole, the executive is free from parliamentary interference in the implemen-
tation of coalition policies. Even though accountability tools at the Parliament’s
disposal have increased, they are not sufficient to counter the downgrading of
Parliament, even with respect to legislative initiative. The power delegated by the
principal is not balanced by effective accountability mechanisms to the Parliament as a
whole, but rather the executive is formally responsible to the parliamentary majority
496 Patrick Dumont and Lieven De Winter
(majority parties)—which is consistent with the model of parliamentary democracy.
These features account for the extraordinary stability of cabinets in Luxembourg.
External constraints such as the Council of State, neo-corporatism, and EU member-
ship further reduce the role of Parliament, but also create deviations from the singu-
larity principle of delegation and accountability.
The link between cabinet and individual ministers is probably the one that deviates
most from the ideal model presented in this volume. Collective decision-making
reduces the scope of matters that are effectively delegated. Ex ante mechanisms—the
comprehensive coalition agreement, the screening and selection procedure during nego-
tiations—and ex post mechanisms—permanent monitoring of the PM and vice-PM on
behalf of the cabinet and as party leaders—serve to make the process different from the
ideal-type, as delegation is limited, while accountability requirements are high.
Finally, the delegation between ministers and civil servants is in keeping with the
ideal model. Top civil servants are important both in implementation and in policy-
making. This delegation of power is necessary in a country in which all ministers hold
multiple portfolios and cannot be specialized in all fields. Accountability is ensured by
the politicization of nominations and promotions of top civil servants, and by the
coordination tasks performed by ministerial advisers.
REFERENCES
Als, Nicolas and Philippart, Robert L. (1994). La Chambre des Députés. Histoire et lieux de travail.
Luxembourg: Chambre des Députés du Grand-Duche de Luxembourg.
Bergman, Törbjorn (1995). Constitutional Rules and Party Goals in Coalition Formation. Umea:
Umea University, Department of Political Science.
Bodry, Alex (1996). ‘La fonction de Premier Ministre dans le système politique du Luxembourg’.
Bulletin du Cercle François Laurent, 1: 3–27.
Carlier, Katty (1995). Autonomie et pouvoir d’appréciation de la Chambre luxembourgeoise dans l’appli-
cation du droit communautaire. Dissertation. Départment des Sciences Politiques et Sociales,
Université Catholique de Louvain-la-Neuve.
Centre de Recherche et d’Information Socio-Politiques (CRISP) (1995). Les élections au Grand-Duché
de Luxembourg. Données sur les scrutins de 1974, 1979, 1984, 1989 et 1994. Résultats et comportements.
Etude réalisée pour la Chambre des Députés du Grand-Duché de Luxembourg. Bruxelles: CRISP.
Centre de Recherche Public– Gabriel Lippmann (CRPGL) (2000). Les élections au Grand-Duché de
Luxembourg: Rapport sur les élections législatives du 13 juin 1999. Etude réalisée pour la Chambre
des Députés sous la direction de F. Fehlen and Ph Poirier. Luxembourg: CRPGL.
Dumont, Patrick and De Winter, Lieven (2000). ‘Luxembourg: Stable Coalitions in a Pivotal
Party System’, in Wolfgang C. Müller and Kaare Strøm (eds.), Government Coalitions in Western
Europe. Oxford: Oxford University Press.
Eurobarometer 48 (Autumn 1997). European Commission.
Govaert, Serge (1997). ‘Le Grand-Duché de Luxembourg: Une stabilité trompeuse?’ Revue
Internationale de Politique Comparée, 4: 585–99.
Hearl, Derek J. (1987). ‘Luxembourg 1945–82. Dimensions and Strategies’, in Ian Budge, David
Robertson, and Derek Hearl (eds.), Ideology, Strategy and Party Change. Cambridge: Cambridge
University Press.
Luxembourg 497
I N T RO D U C T I O N : T H E C O N S T I T U T I O N A N D B E YO N D
The Constitutional Framework
The government and the two Houses of Parliament together make up the constitutional
legislative power of The Netherlands. The two Houses of Parliament are not equal in
this respect, however. Only the populary elected Tweede Kamer (Second Chamber, or
Lower House) is granted the constitutional power to initiate or amend legislation.
Ministers and individual MPs can propose bills, which must be approved by a simple
majority (votes are public), with a quorum rule of 50 per cent plus one. The Eerste
Kamer (First Chamber, or Upper House), which is elected by an electoral college
composed of all provincial councillors, lacks the right to initiate or amend legislation,
which underlines the fact that the constitutional engineers intended it to be a chambre
de reflexion. Once legislation is approved in the Lower House, the Eerste Kamer is mainly
a potential veto player, as it can only approve or reject bills at large. If approved, the bill
has to be signed by the monarch and countersigned by a minister to become law.
For bills fixing departmental budgets and for constitutional revisions, a special legislative
procedure is required. Bills containing proposals for the annual budget can only be sub-
mitted by government ministers, which is done formally on the third Tuesday of
September. Before that date, ministers draft their respective budgetary proposals according
to an internal schedule drawn up by the cabinet. Changes of the Constitution can be pro-
posed by ministers and members of the Second Chamber. As with ordinary legislation,
both Houses of Parliament must approve proposals for a revision of the Constitution by
simple majority. This is followed by a second reading, after at least the Second Chamber
has been dissolved and newly elected. In this second reading a two-thirds majority is
required in both Houses. As in other cases, the quorum requirement is 50 per cent plus one
vote. In practice, the second reading is usually held after regular elections in order to avoid
cutting short the government’s term of office by calling special elections.
The Dutch Constitution is based on the principle of representative democracy. This
does not mean, however, that Parliament alone is responsible for public policy-making.
The Netherlands 499
The most obvious other policy-making body is the government. The Constitution
mentions both Parliament and the government as the legislative powers, and in prac-
tice the government usually takes the initiative. In addition to formal legislation, the
government produces executive decrees, which flesh out formal legislation. This kind
of ‘pseudo legislation’ has become more important over time, and already in the early
1970s one observer took this as a reversal of the doctrine of Trias Politica: the execut-
ive (government) has become the main legislative power (Koopmans 1970). With only
few exceptions, Parliament is not involved directly in the making of these decrees,
and it takes highly alert MPs to bring these decrees within the scope of effective
parliamentary control.
The Netherlands has been described both as a consociational democracy and an
example of neo-corporatism. Both concepts are closely related. Catholic, Protestant,
and Socialist groups have emerged and formed networks of subcultural organizations.
This feature of Dutch political life is widely known as pillarization. Although the social
pillars have eroded to a large extent, their leaders still share responsibility for governing
the country in a culture of compromise and cooperation. In specific policy areas such
as socio-economic policy and education policy, the leaders of the relevant subcultural
networks also work together, often in special institutional arrangements such as tripartite
advisory councils. The Socio-Economic Council, in which representatives of employers
organizations, trade unions, and government appointed experts meet, is a typical
example.
P O L I T I C A L PA RT I E S A N D C O N S T I T U T I O NA L
P R I N C I PA L S A N D AG E N T S
Political parties have their own chains of delegation and accountability that go beyond
the Constitution and may also cut across the formal constitutional relationship
between Parliament and cabinet. Yet, parties are not mentioned in the Constitution,
there is no law on parties, and the term ‘party’ does not even figure in the electoral law.
In the Dutch system of political minorities, policy-making usually involves interparty
bargaining. In this context, constitutional bodies such as government and Parliament
are arenas as much as they are principals and agents. Parties condition the chain of del-
egation through their strategies and through mostly unwritten rules that have emerged
within and between them.
The four main parties with stakes in these arenas are the Social Democrats (PvdA),
the Christian Democrats (CDA), the Liberal Conservatives (VVD), and the Liberal
Democrats (D66) (see Timmermans and Andeweg 2000). With the exception of D66,
the social and ideological roots of these parties date to the early years of the twentieth
century, much like similar parties in other European parliamentary democracies. In
their present form, however, all main parties are postwar organizations. The PvdA and
the VVD were formed in the immediate postwar years, as modern versions of the pre-
existing Social Democratic and Liberal parties. D66 began as a democratic reform party
in 1966, while the CDA was created through a merger of two smaller Protestant and
one Catholic party in 1980.
These relatively centralized and disciplined parties also dominate the governmental
arena. All postwar Dutch cabinets have been composed of at least two but more often
three or more parties. In these cabinets, the Christian Democrats (before 1980 mainly
the Catholic party) long took a pivotal position, having been in office continuously
from 1917 to 1994. Dutch government coalitions often control more than a simple
majority in Parliament, although the number of minimum winning coalitions has
increased since the late 1960s.
D E L E GAT I O N F RO M VOT E R S TO M E M B E R S O F PA R L I A M E N T
Electoral Engineering in a Quasi-Multiple District System
The formal election procedure in the Dutch system of proportional representation
(PR) is that the voter expresses a preference for an individual candidate from a party list
in one of nineteen electoral districts. In reality, however, the electoral system does not
produce a real relationship of delegation between voters in a particular district and an
individual MP. For the distribution of the 150 seats in the Second Chamber, all nineteen
districts are combined into a single nationwide district, and district votes are aggregated
to one national party vote. The districts mainly play a role in the organization of the
elections (supervising of polling stations, counting votes, etc.), and the party vote in a
particular district determines what share of the seats that the party has won nationwide
will be filled by candidates from the district list. Until the 1960s, the electoral districts
had some practical relevance, as the larger political parties presented different lists (and
list leaders), if not in each electoral district, then at least in different regions of the
The Netherlands 501
country. Probably under the impact of the electronic media, the scale of electoral cam-
paigns has become national, and the names topping parties’ lists are now the same in
all nineteen districts. If the remaining names on the lists differ from one district to
another, this is to allow a party to nominate a larger number of candidates nationwide
than the maximum (thirty for new parties; twice the number of seats currently held
with a maximum of eighty for parties already represented) that it is permitted in an
individual district. Most candidates are nominated in several districts, though their
ranking on the lists varies across districts.
20 per cent of votes cast. Research has revealed that the main reason for preference
voting is to choose a female candidate (the list leaders of most parties are male).
Regional representation is the second most important motivation (Hessing 1985).
Occasionally, preference voting may indicate voter disagreement with the ordering
of the candidates on the party list. Sometimes this disagreement takes the form of
ex post accountability. In 1986, for example, about one-third of the VVD electorate cast
a preference vote, primarily to support incumbent MP Theo Joekes, who had fallen out
with the party leadership over his role in a Parliamentary Inquiry. He had criticized one
of his party’s cabinet ministers. When he was relegated to an unelectable position on
the party list, a campaign started to get him elected in defiance of party leaders, with
the result that he retained his seat.
However, it is not always disagreement with the party’s list of candidates that creates
an incentive for preference voting. It may be used also to signal discontent with party
policy more generally, as in 1952, when a political weekly launched an appeal to vote for
the seventh candidate on the party list as a sign of opposition to the country’s defence
policy. Less often, preference voting might be meant to show support for the party’s
choice of candidates. For example, in 1986 a significant number of PvdA voters cast their
votes for Wim Kok, the intended successor of long-time party leader Joop den Uyl.
The real impact of preference voting on who wins a seat in Parliament is quite small.
Since 1945, only a few candidates have obtained a seat due to preference voting. The
threshold for obtaining a seat via this process has only gradually been reduced from
50 per cent of the electoral quota within one of the administrative electoral districts to
the current 25 per cent of the electoral quota nationwide. Moreover, parties do not
always allow their candidates to campaign for preference votes. In some cases parties
do not even allow their candidates to accept election on the basis of preference voting
without the party’s consent—for example, candidates are required to sign undated letters
of resignation. Increasing the impact of preference votes is controversial. In 1989, the
government rejected a proposal to bring the threshold down from 50 to 25 per cent, on
the grounds that this would transfer too much influence from the parties to voters
(Proceedings of the Second Chamber 1987–8, 20264 no. 3: 59). In 1993, an advisory
committee rejected the abolishment of a threshold because it would create more intra-
party competition, decrease job security of incumbent MPs, and hinder the balance of
expertise within parliamentary parties (De Koning Commission 1993: 28).
The absence of a personal relationship between voters and MPs is generally felt to
be one of the system’s shortcomings. In the 1960s and early 1970s, proposals to intro-
duce PR within multimember districts were rejected because doing so would make
nationwide party representation less proportional. In the 1990s, advocates of electoral
reform have sought to circumvent this objection by advocating variations on the
German electoral system, which combines nationwide PR with the election of at least
some candidates in single or multimember districts. These proposals have failed
because they have been deemed to be too complicated or to have other undesirable
side-effects (Andeweg 1997a: 239–44).
As a result, there is little or no delegation directly from voters to MPs. There are no
provisions for recall elections, and especially when an incumbent MP is placed relatively
504 Arco Timmermans and Rudy B. Andeweg
high on the party list in subsequent elections, there is no way in which voters can hold
the individual MP accountable. Voters tend to vote for a party rather than an individual
candidate, and even in the rare event of an MP being elected through preference
voting, the MP does not know who his/her voters are. For the members future political
career, it is probably wiser to try to impress those within the party who have power over
the party lists than to rely on preference voting. The chain of delegation is therefore
indirect: voters delegate to parties, and parties are, or contain, the selectors to whom
MPs are responsive. It is this interposition of the party in the chain of delegation
between voter and MP that has led to rejections of American models of representation
(Thomassen 1994; Andeweg 1997b).
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
The constitutional link between Parliament and the government most clearly distin-
guishes parliamentary democracies from other kinds of systems. As in most parlia-
mentary systems with a written Constitution, the Dutch Constitution delegates
general policy-making competence to the government, and it grants control powers to
Parliament. In this section we discuss the specific forms that these governing and con-
trol powers take as instruments of delegation and accountability, as well as their relev-
ance in practice. Before the coexistence of the two bodies is considered, however, we
consider the way in which the government, the non-elected body, is formed.
1
In contrast to countries such as Belgium and France, there are few patronage appointments in the Dutch
government formation process.
The Netherlands 505
2
Ironically, when this party prominent was in the government, a conflict between a party colleague in the
cabinet and the party’s parliamentary group led to the premature resignation of the government in 1989.
506 Arco Timmermans and Rudy B. Andeweg
(people who, almost without exception, are MPs), but ministers are also recruited from
outside Parliament. Article 57 of the Constitution prohibits ministers from simultane-
ously holding seats in Parliament.3
After an often protracted process of government formation, the inauguration of gov-
ernments illustrates the complex interplay of parties and constitutional bodies. A party’s
involvement in the formation of a government activates an internal chain of delegation
and accountability within the party. Drafts of the coalition agreement, for example, are
scrutinized and often amended by the parliamentary groups. The overall outcome of
the negotiations (coalition agreement and portfolio allocation) will be voted on at a spe-
cial party congress (PvdA, D66), or scrutinized by party councils (CDA and VVD)
and/or the parliamentary group (VVD). Within the constitutional arena, the govern-
ment formation is formally concluded when the Head of State signs the Royal Decrees
appointing the ministers, but the new Prime Minister countersigns these decrees (includ-
ing the one pertaining to his own appointment) (Article 48), and in that way the forma-
tion of the government falls under the general responsibility of the government to
Parliament (Article 42). There is no formal investiture vote. The Prime Minister reads
the ‘Government Declaration’ (usually a summary of the coalition agreement) in
Parliament, followed by a parliamentary debate. At the end of the debate, opposition
parties may introduce a motion of no confidence in the new government. As long as
such a motion is not approved by a majority, the government is assumed to have
Parliament’s confidence (i.e. negative parliamentarism; see Bergman 1993).
3
The exception is the period between elections and the inauguration of the new cabinet, when a minis-
ter remains a member of the outgoing government even if he or she has been elected to parliament.
The Netherlands 507
4
This is the constitutional rule for the Tweede Kamer. Members of the Eerste Kamer do not face such a
support threshold when they submit motions to a plenary vote. This is remarkable because in most other
respects the Eerste Kamer has fewer competences.
508 Arco Timmermans and Rudy B. Andeweg
consider motions of no confidence are those that express ‘disapproval’ (which is the
commonly used terminology) of the behaviour of a minister or (more rarely) a whole
government. Most no-confidence motions proposed between 1981 and 1997 were
rejected (see Table 16.1). Although we have no systematic information about the
number of no-confidence motions in the period before 1981, we do know from parlia-
mentary history that few were approved and/or led to the resignation of a minister or
of the government. Only twice between 1945 and 1981, in 1960 and 1966, did the gov-
ernment resign after a motion approved by a parliamentary majority was interpreted
as a vote of no confidence.5 In 1951 and more recently in 1989, the government
resigned after a motion was tabled but before the Parliament had voted. In both cases,
the motion criticizing the government had been introduced by one of the coalition
partners, indicating that the coalition had broken down.
Whether or not parliamentary sanctions are used to bring down governments, the
Constitution limits a government’s term to four years, or more precisely, it states that
elections must be held every four years (governments remain in office after an election
until a new government is formed, which usually takes several months). While the
Constitution does not explicitly give Parliament the power to end the term of a gov-
ernment (i.e. it does not contain a motion of no confidence), the government can dis-
solve Parliament (Article 64). However, the early dissolution of Parliament is rarely
brought about by conflict between the government and Parliament as an institution.
When the Parliament is dissolved prematurely, it is more often a consequence of the
escalation of a conflict within the governing coalition. Table 16.1 summarizes the
behavioural record with respect to confidence votes and parliamentary dissolution. As
shown in the table, a total of seven cabinet have ended in parliamentary dissolution and
early elections. In October 2002, the Balkenende cabinet (formed after the 2002
elections) also ended prematurely, and early elections were held in January 2003.
What other working rules exist in these two arenas of coalition politics? There is a
rich literature on the tension in coalitions between the need to cooperate and maintain
coalition cohesion, and each party’s desire to profile itself. Most of this literature
emphasizes the strategic aspects of interparty relationships, although some recent
work has begun to consider institutional constraints (Strøm, Budge, and Laver 1994;
Müller and Strøm 2000). Moving beyond formal constraints such as the majority
requirement and the procedures for fixing the budget and engineering constitutional
revisions, rules pertaining to party interaction during a government’s term have
evolved. These rules are endogenous to the political system and concern mainly the
‘peaceful coexistence’ of coalition parties. An example of such an unwritten rule is that
outgoing governments or caretaker governments should leave politically controversial
issues to the new government formed after parliamentary elections.
One of the rules relating to interparty relationships that has evolved since the 1960s
is that parties must observe the coalition agreement.6 We previously referred to the
5
In the case of 1960, however, cabinet resignation was revoked, while in 1966 the MP who tabled the
motion was reported to have said that he had not intended to bring down the government.
6
The rule has evolved, though it was written down in the first published coalition agreements in 1963 and
1965. Later the rule was implicit (Timmermans 2003).
Table 16.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: The Netherlands 1945–2000 a
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the cabinet Who
formation votes (iv) votes (cv) end with an dissolved
(n.a.) No. of Cabinet Cabinet under specific early election? Parliament in
ncvc removed resigned to institutional this case?b
by ncv pre-empt ncv mechanism
(n.a.)
1 Schermerhorn 45/06/24 0
2 Beel I 46/07/03 0 Yes Const.
3 Drees I 48/08/07 1 Yesd Yes Cabinet
4 Drees II 51/03/15 0
5 Drees III 52/09/02 0
6 Drees IV 56/10/13 0
7 Beel IIe 58/12/22 0 Yes Cabinet
8 De Quay 59/05/19 1 Nof
9 Marijnen 63/07/24 0
10 Cals 65/04/14 1 Yes
11 Zijlstrag 66/11/22 0 Yes Cabinet
12 De Jong 67/04/05 0
13 Biesheuvel I 71/07/06 0
14 Biesheuvel IIh 72/08/09 0 Yes Cabinet
15 Den Uyl 73/05/11 0
16 Van Agt I 77/12/19 1 No
17 Van Agt II 81/09/11 0
18 Van Agt IIIi 82/05/29 1 No Yes Cabinet
19 Lubbers I 82/11/04 4 No
Table 16.1. (contd.)
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the cabinet Who
formation votes (iv) votes (cv) end with an dissolved
(n.a.) No. of Cabinet Cabinet under specific early election? Parliament in
ncvc removed resigned to institutional this case?b
by ncv pre-empt ncv mechanism
(n.a.)
D E L E GAT I O N W I T H I N T H E C A B I N E T
Rules of Delegation and the Role of the Prime Minister
It is only since 1983 that the Dutch Constitution has recognized the existence of a
Council of Ministers, which ‘shall consider and decide upon overall government policy
and shall promote the coherence thereof ’ (Article 45), and of a Prime Minister who
‘chairs the Council of Ministers’ (Article 45). Although the Constitution allows for min-
isters without portfolio, these are exceptions. Practically all cabinet ministers head a
department. The Constitution does not regulate the size of the cabinet, but it has never
included more than sixteen members.
To the extent that individual ministers have delegated powers, these come from the
Council of Ministers (cabinet) rather than from the Prime Minister. As the formateur
of the new government, the Prime Minister formally nominates ministers for appoint-
ment by the Queen, but he has real influence only over his own party’s nominations (in
his capacity as party leader). The Prime Minister cannot change the departmental juris-
dictions of ministers, dismiss ministers, or unilaterally reshuffle the cabinet, nor can
he issue policy directives to individual ministers. In the case of a tie vote in the cabinet,
the Prime Minister casts the decisive vote; however, this is of limited importance since
the cabinet rarely takes decisions by voting. The Prime Minister heads the Department
of General Affairs, but most of its officials work in the Government Press Office or the
semi-autonomous Scientific Council for Government Policy. He has only a dozen
policy advisers, all career civil servants, each of whom focuses on a few policy areas,
but who lack formal monitoring competences.
The influence of the Prime Minister is sometimes much greater than that which
comes from his limited formal powers or small staff, and some commentators argue
that there has been a trend towards a stronger premiership in recent years (cf. Andeweg
1991). This is based partly on the increased external role of the Prime Minister (EU
summitry, press conferences), although greater visibility connected to external respons-
ibilities does not automatically imply greater political power. A more important
reason for this trend is that, as the need for coordination grows with the complexity
of modern government, so to does the acceptance of a more powerful Prime Minister
to serve as coordinator. Nonetheless, even if this has led to a more influential
The Netherlands 513
Prime Minister, his role is still mostly reactive, more like an arbitrator than a true
policy activist (Blondel 1988). Moreover, the norms and rules relating to this more
influential role are still emerging.
The Dutch government has permanent cabinet committees, but they have no delegated
powers and they do not delegate power to ministers. With a few insignificant exceptions,
these committees are only authorized to prepare items for discussion within the cabinet.
Only a few of these committees meet frequently. All ministers are welcome to attend com-
mittee meetings in which they are interested. Civil servants also serve as members of
cabinet committees. The Dutch cabinet also has a number of informal committees, some
of which are more important than the formal ones. While most of these are ad hoc, some
are more or less permanent. The group of ministers with responsibility for socio-economic
policy (called the ‘triangle’ or the ‘pentagon’ depending on the number of ministers
involved) has even been accused of forming a de facto inner cabinet. Of course, even if this
informal committee does reach agreement, decisions are not binding on other ministers
unless they have been channeled through the cabinet.
Formally speaking, delegation from the cabinet to individual ministers is accompa-
nied by detailed supervision. The Standing Orders of the Council of Ministers spell out
which ministerial decisions are to be submitted to the cabinet for prior approval. These
include all proposals for legislation and regulation, all important appointments, the cre-
ation of committees or advisory councils, and in general any decision that might
impact on other ministers’ policies or might have major financial or political con-
sequences. When in doubt, a minister is expected to consult the Prime Minister. Since
ministers’ decisions eventually become public, it is difficult to ignore this rule.
Moreover, the rule is not merely pro forma. The Dutch Council of Ministers not only
meets frequently, it actually debates and amends proposals, even against the wishes of
the sponsoring ministers.
The importance of this for cabinet outcomes has increased since the 1960s. Perhaps less
obvious is the fact that ministers quite often ‘go native’ and become departmental ambas-
sadors to the cabinet. Especially new ministers are often targets of such attempts by high-
ranking department officials who seek to co-opt them. This brings us to the question of
the relationship between ministers and their departmental organizations.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
Ministerial Responsibility, Civil Service Loyalty, and their Limits
The relationship between ministers and the administrators in their departments is
probably the purest case of delegation in the whole chain from voters to policy imple-
mentation. Ministers are responsible for all public-policy related activities of their civil
servants’, including decisions involving the use of discretionary power. Civil servants
are supposed to be loyal to their minister. This rule of loyalty is considered to be so self-
evident that explicit arrangements for the public accountability of civil servants to their
ministers do not exist.
The civil service is organized along functional departmental lines. Formal internal
departmental structures are hierarchical in accordance with the classical Weberian model.
Departments are not staffed by a minister’s political affiliates, and in general recruitment
of department officials is based on expertise in the department’s policy area or in more
general management practices. There have been efforts to rotate top civil servants
among departments increasingly in order to avoid over-compartmentalization in the
administrative management of departments.
Some civil servants have powers granted to them directly by law. Tax inspectors, for
example, can determine an individual citizen’s tax liability, just as university professors
determine students’ grades without ministerial interference. Much more important,
however, are the powers delegated to civil servants by their department minister. For
example, every year Justice Department officials, acting on behalf of the Minister of
Justice, make some 30,000 decisions regarding the naturalization of immigrants. Over
600,000 decrees setting student scholarships are issued on behalf of the Minister of
Education (Scheltema Commission 1993: 36). These ministers may provide civil servants
with decision-making guidelines, but they are rarely cognizant of specific decisions,
despite the fact that they are fully responsible for each and every one of them.
Dutch ministers have few control instruments, despite the fact that they must, of
necessity, delegate considerable authority. It is physically impossible for a minister to
actually monitor the behaviour of his/her numerous departmental civil servants.
There is considerably more continuity in the civil service than in cabinet, even though
Dutch cabinets are relatively stable (on average they remain in office for about three years),
and ministerial careers may continue beyond a particular cabinet (the average ministerial
career lasts almost four years; see Bakema 1991: 93). Furthermore, departmental decisions
often require a level of technical expertise that most ministers do not have. None of this is
necessarily peculiar to the Netherlands, but it seems to be more pronounced here
than in most other countries. Civil servants are highly specialized (and are recruited on
516 Arco Timmermans and Rudy B. Andeweg
this basis), and in most cases they spend their entire career within a single department.
The asymmetry in expertise between ministers and their departmental staffs might be
limited by the fact that Dutch ministers are also recruited in part on the basis of policy-
specific expertise (recall that ministers can be recruited from outside Parliament).
Another factor that may limit the expertise asymmetry is that Dutch ministers are often
single post ministers (within Europe only Austrian ministers are less mobile across
departments; see Bakema 1991: 90). In practice, however, a minister’s relevant expertise
seems to be less useful in helping him/her control the department, and more useful
to top civil servants seeking to co-opt the minister. That is, it seems to increase the
likelihood that the minister will become an ambassador for the department. This leads
to a paradox about the ‘weakness’ and ‘strength’ of ministers. Even if departmental
civil servants want a minister who is susceptible to their policy beliefs and can be influ-
enced, at the same time they prefer a ‘strong’ ambassador who can get things done in
cabinet.
Departments are completely staffed by career civil servants. There are no party polit-
ical appointments or ministerial cabinets. When top civil service positions are vacant,
party allegiance does matter. However, the pervasive rule of proportionality among the
main political parties prevents a minister from exclusively appointing party friends. As
of 2003, only Deputy Prime Ministers have a small personal cabinet. However, it is sup-
posed to focus on external activities (to help coordinate their party’s ministers), rather
than to monitor the department to ensure that it is faithfully carrying out ministerial
policy. Ordinary ministers are allowed to hire a political adviser, but if they do so the
adviser’s job is primarily to serve as a liaison with the party or with Parliament. In short,
ministers—including the Prime Minister—have no source of regular, policy-specific
advice and support other than their departmental officials.
All of what has been reported here about the nature of decisions, required level of
policy expertise, broad ministerial responsibility, and limited instruments of control,
is fuel for the argument that ministerial responsibility is a fiction. In fact, there have
been proposals to limit ministerial responsibility by introducing the criterion of
reproachability. This means that Parliament should only hold a minister responsible
(and censure him/her) when she/he is actually at fault, that is, when a policy failure
is due to the minister’s own policy guidelines, or to his neglect of departmental man-
agement. Opponents of such a ‘narrow’ interpretation of ministerial responsibility
have pointed out that this would create situations in which civil servants are account-
able to no one.
The practical ‘solution’ for reducing the gap between the fiction of full ministerial
responsibility and the reality of civil servants’ discretion was to transfer responsibility
for some public tasks from departments to other (semi-public) organizations. The pro-
posal intended to create so-called ‘core departments’ of a size more amenable to min-
isterial supervision (Wiegel Commission 1993). These core departments were to be
formed by decentralizing tasks (transferring from central to provincial and municipal
governments), by privatization, and by splitting off parts of departments to create new,
quasi non-governmental organizations (quangos). While there are examples of all three
strategies, their overall importance should not be overestimated.
The Netherlands 517
E X T E R NA L C O N S T R A I N T S
The institutional factors affecting the democratic chain of delegation and accountability
that have been discussed so far are internal to the delegation process, such as political
parties or the structure of the civil service. There are also ‘constraints’ or institutional
factors relevant to delegation, which are external to the formal chain of delegation.
While there is a whole range of possible and more or less stable external influences on
delegation and accountability, we will limit this discussion to those influences mentioned
in the introductory chapter which have immediate relevance in The Netherlands.
The Judiciary
A fundamental principle of the Rechtsstaat is that the judiciary remains outside the
formal chain of delegation and accountability, or at least outside that part of the chain
in which relationships are political. Looked at this way, the courts are a constraint by
virtue of their authority to judge the legality and constitutionality of government
decrees and actions in court cases. Legislation approved by Parliament is not subject to
judicial review, not even if there have been procedural errors in the legislative process.
The absence of judicial review does not remove the judiciary as a constraint on the del-
egation process, as so much of government policy is laid down in decrees rather than
laws (see Van Koppen and Van Kate 1994). In addition to the ‘ordinary’ judiciary, the
Council of State acts as the highest administrative court. In practice, the administrative
courts deal primarily with the implementation of government policy, which is often the
518 Arco Timmermans and Rudy B. Andeweg
task of subnational governments. They are therefore less important as a constraint on
delegation and accountability at the national level.
C O N C LU S I O N
To the extent that the ideal-type chain of delegation is a normative model, it should
perhaps be thought of as a cycle, in which the policy output of the system of repres-
entative government is evaluated by voters. If the ideal-type model starts with voters,
it should thus also end with voters. In the idea of democratic delegation and account-
ability, elections are prominent features. It is through elections that politicians are held
accountable by the voters on the basis of their judgements about the extent to which
their policy preferences have been followed, misunderstood, or ignored since the last
election. To the extent that many voters consider the government to have misunder-
stood or ignored their preferences, voters can be said to have experienced agency loss.
The Netherlands 519
Although this can and often does occur within the administrative sphere as well, in the
Dutch case it is the politicians who are held accountable, at least once every four years.
We conclude our contribution by highlighting the most important deviations from this
ideal–type chain or cycle of delegation and accountability in the Netherlands.
The most striking deviation concerns the interposition of political parties as key
actors in the chain. The various roles of parties cut across the formal chain at several
points (voters–MPs, MPs–cabinet, cabinet–ministers, sometimes ministers–civil ser-
vants as well). Parties may even be seen as political actors in broader policy coalitions
that extend beyond the formal political bodies in parliamentary democracy. All these
actual roles and linkages involving parties are sometimes seen as harmful to the ideal
type, as for example, in Lord Bryce’s cri de coeur about the ‘Decline of Parliament’.
However, a good argument can be made that parties have not usurped power, but
rather have saved the parliamentary system of government. In institutional terms, it is
not true that parties constrain the operation of the chain of delegation; in fact, they
facilitate the operation of the system of representative democracy. Indeed, they have
been key actors in the creation of the system, for example, by pressing for universal
electoral suffrage, which was granted in 1919. The ideal-type relationship between
voters and individual MPs probably would not have survived the shift to a mass
electorate, and for this reason the responsible party model is perhaps more appropriate
as a point of departure for studying modern parliamentary democracy.
Second, the empirical picture not only differs from the norm; it is also considerably
more complex than the ideal type. This complexity creates its own delegation problems
and forms of agency loss. Arrows in the empirical chain or cycle do not point unam-
biguously in one direction, and the scheme contains numerous loops reaching beyond
the constitutional system. The main reason for this deviation from the ideal type is that,
in practice, the various elements within the chain are not only institutions but also
arenas in which struggles for power and policy take place. This is most clearly the case
with regard to Parliament and cabinet, where incumbents are constrained not only by
these bodies’ institutional rules, but also by the institutional rules of other organiza-
tions, such as parties and departments, which also see these incumbents as agents. For
example, ministers not only represent the cabinet in Parliament, they also represent
their departments and their parties within the cabinet.
This complexity is not only a matter of level of analysis and conceptual confusion; it
is a real problem for parliamentary democracy insofar as delegation ‘down’ the chain
(starting with voters) has no monopoly. Signals from the opposite direction may jam sig-
nals from the citizens as initial principals. Parliamentary democracy seems to allow only
for input into the system, not for ‘withinput’. Even the validity of inputs from citizens is
problematic in the Dutch case. The crucial normative feedback loop in which policies
and politicians are subject to the judgement of voters is often obfuscated. The out-
standing example of this is the weak relationship between electoral shifts and the subse-
quent results of the government formation process, in which parties winning
parliamentary seats often fail to obtain a seat at the cabinet table.
Coalition politics also creates ambiguity in the relationship between Parliament as
principal and the cabinet as the agent, and this manifests itself through the practice of
520 Arco Timmermans and Rudy B. Andeweg
formulating and enforcing coalition agreements. Ostensibly, Dutch cabinets are bound by
the comprehensive and detailed coalition agreements drafted and adopted by the parlia-
mentary groups. In this case, the governing majority in the Second Chamber, and not the
Second Chamber as a body, acts as the principal. In fact, the informal character of policy
negotiations during cabinet formations seriously limits the potential of Parliament to func-
tion as a genuine principal with regard to major policy issues. But even if parliamentary
groups are very important actors in coalition formation negotiations, the commitments
they make also tie their own hands, or at the very least limit the possibility that they can
take initiatives on politically sensitive issues once the cabinet is in office. Indeed, on such
matters the rules of the game related to coalition agreements prevent individual coalition
parties in Parliament from forming ad hoc legislative coalitions with non-coalition parties.
At the same time, coalition agreements often leave appreciable discretion to ministers on
matters below the sphere of high coalition politics. Particularly with regard to depart-
mental matters, information asymmetries between cabinet and Parliament (or between
individual ministers and MPs) exist and hidden action occurs. On such matters, parties in
Parliament (perhaps with the exception of small and impotent opposition parties) are not
inclined to complicate the lives of ministers by making hidden actions public.
All this led to demands for more direct democracy in the 1960s and 1970s. Today, this
same problem has led to renewed demands for procedures that bypass the chain of
delegation and establish direct links between citizens and policies. This is to be accom-
plished by introducing judicial review, through the introduction of a referendum (see
the previous section on External Constraints), and by creating new accountability
arrangements that give people a voice in the performance of public organizations
(citizens’ charters, etc.). Alongside such reform suggestions, public agencies increasingly
take recourse to non-legislative policy instruments, for example, covenants with private
actors in specific policy areas. The possible result of these developments is not the
reform of the chain of delegation running through Parliament, but a hollowing out of
the parliamentary system of government.
REFERENCES
Andeweg, Rudy B. (1990). ‘Tweeërlei Ministerraad; besluitvorming in Nederlandse kabinetten’,
in R.B. Andeweg (ed.), Ministers en Ministerraad. The Hague: SDU.
—— (1991). ‘The Dutch Prime Minister: Not Just Chairman, Not Yet Chief ?’ West European
Politics, 14(2): 116–32.
—— (1997a). ‘Institutional Reform in Dutch Politics: Elected PM, Personalized PR, and Popular
Veto in Comparative Perspective’. Acta Politica, 32: 227–57.
—— (1997b). ‘Role Specialisation or Role Switching? Dutch MPs between Electorate and
Executive’. Journal of Legislative Studies, 3(3): 110–27.
—— and Bakema, Wilma E. (1994). ‘The Netherlands: Ministers and Cabinet Policy’, in Michael
Laver and Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government,
Cambridge: Cambridge University Press.
Bakema, Wilma E. (1991). ‘The Ministerial Career’, in Jean Blondel and Jean-Louis Thiébault
(eds.), The Profession of Government Minister in Western Europe. London: Macmillan.
The Netherlands 521
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments.’ European Journal of
Political Research, 23: 55–66.
Blondel, Jean (1988). ‘Cabinet Structures and Decision-Making Processes in Western Europe.’
European Journal of Political Research, 16: 115–23.
Bovend’Eert, P.P.T. (1988). Regeeraccoorden en Regeringsprograms. ‘S-Gravenhage: SDU.
De Koning Commission (1993). ‘Het Bestel Bijgesteld.’ Proceedings of the Second Chamber 1992–3,
21427, no. 36–7.
Hessing, R.C. (1985). ‘Bij Voorkeur: een Onderzoek naar het gebruik van voorkeurstemmen.’
Acta Politica, 20: 157–76.
Hillebrand, Ron (1992). De Antichambre van het Parlement; kandidaatstelling in Nederlandse politieke
partijen (with English language summary). Leiden: DSWO Press.
Koole, Ruud A. and Leijenaar, Monique H. (1988). ‘The Netherlands: The Predominance of
Regionalism’, in Michael Gallagher and Michael Marsh Candidate (eds.), Selection in
Comparative Perspective. London: Sage.
Koopmans, T. (1970). ‘De rol van de wetgever’, in Honderd jaar rechtsleven. Nederlandse
Juristenvereniging. Zwolle: Tjeenk Willink.
Laver, Michael and Shepsle, Kenneth A. (1996). Making and Breaking Governments. Cambridge:
Cambridge University Press.
Luebbert, Gregory M. (1986). Comparative Democracy. Policymaking and Governing Coalitions in
Europe and Israel. New York: Columbia University Press.
March, James G. and Olsen, Johan P. (1995). Democratic Governance. New York: Free Press.
Müller, Wolfgang C. and Strøm, Kaare (eds.) (2000). Coalition Governments in Western Europe.
Oxford: Oxford University Press.
Nijpels, E. T. M. (1984). ‘Over monisme als strategie’. Liberaal Reveil, 15: 4–7.
Peterson, Robert L., De Ridder, Martine M., Hobbs, J. D., and McClellan, U. F. (1983). ‘Government
Formation and Policy Formulation in Belgium and the Netherlands’. Res Publica, 18: 49–82.
Richardson, Jeremy (1996). ‘Eroding EU policies: Implementation Gaps, Cheating and Re-steering’,
in Jeremy Richardson (ed.), European Union: Power and Policy-Making. London: Routledge.
Scheltema Commission (1993). ‘Steekhoudend Ministerschap; betekenis en toepassing van de
ministeriele verantwoordelijkheid’. Proceedings of the Second Chamber 1992–3, 21427, no. 40–1.
Shepsle, Kenneth A. (1996). ‘Political Deals in Institutional Settings’, in Robert A. Goodin (ed.),
The Theory of Institutional Dessign. Cambridge: Cambridge University Press.
Strøm, Kaare, Budge, Ian, and Laver, Michael, J. (1994). ‘Constraints on Cabinet Formation in
Parliamentary Democracies’. American Journal of Political Science, 38: 303–35.
Thomassen, Jacques J. A. (1994). ‘Empirical Research into Political Representation: Failing
Democracy or Failing Models?’, in M. K. Jennings and T. E. Mann (eds.), Elections at Home and
Abroad. Essays in Honor of Warren E. Miller. Ann Arbor: University of Michigan Press.
Timmermans, Arco (1991). ‘Königreich der Niederlande’, in Winfried Steffani (ed.),
Regierungsmehrheit und Opposition in den Staaten der EG. Opladen: Leske⫹Budrich.
—— (1998). ‘Policy Conflicts, Agreements, and Coalition Governance’. Acta Politica, 33: 409–32.
—— (2003). High Politics in the Low Countries. Aldershot: Ashgate.
—— and Andeweg, Rudy, B. (2000). ‘The Netherlands: Still the Politics of Accomodation?’, in
Wolfgang C. Müller and Kaare Strøm (eds.), Coalition Governments in Western Europe. Oxford:
Oxford University Press.
Toirkens, José (1988). Schijn en werkelijkheid van het bezuinigingsbeleid 1975–86. Alphen aan den
Rijn: Tjeenk Willink.
Van Koppen, P. J. and Van Kate, J. (1994). ‘Judicialisation of Politics in the Netherlands. Towards
a Form of Judicial Review’. International Political Science Review, 15: 143–53.
522 Arco Timmermans and Rudy B. Andeweg
Van Raalte, E. (1977). Het Nederlandse Parlement. ‘S-Gravenhage: SDU’.
Van Schendelen, M. P. C. M. (1995). ‘From Founding Father to Mounding Baby’. Journal of
Legislative Studies, 1: 60–74.
Wiegel Commission (1993). ‘Naar Kerndepartementen; kiezen voor een hoogwaardige en flexibele
rijksdienst’. Proceedings of the Second Chamber 1992–3, 21427, no. 51–2.
17
I N T RO D U C T I O N
Norway is an old and stable democracy, but not originally a parliamentary one. The
Norwegian Constitution of 1814 is currently the oldest living codified constitution in
Europe and indeed second only to that of the United States in the democratic world.
Contrary to Denmark and Sweden, Norway has not in recent decades fundamentally
overhauled and updated its Constitution. Consequently, it is silent on many aspects of
modern governance (such as the role of political parties), and on other issues there are
major discrepancies between formal constitutional provisions and contemporary
practice. The 1814 Constitution was a separation-of-powers construction, under which
considerable powers were vested in the monarch. Contrary to the intentions of the
founding fathers, however, the Head of State in practice became occupied by the dual
monarchy of Sweden and Norway. Hence from the 1870s on, liberals became increasingly
vociferous in their demands for cabinet accountability to the Storting (the Parliament).
In 1884, they through impeachment forced this reform upon a recalcitrant cabinet of
conservatives loyal to the king. Under such duress, the conservatives gradually
accepted the realities of parliamentary government. Thus, although Norwegian
democracy has in practice been parliamentary since 1884 or shortly thereafter,1 the
written Constitution gives scant recognition to this practice. It is important to note,
however, that long-standing practice has given parliamentary government the status of
a constitutional convention (sedvanerett). The most controversial political reform of the
nineteenth century thus became a non-issue in the twentieth.
Although parliamentarism is not well entrenched in the formal Constitution,
Norway does conform to most key features of parliamentary government. During the
struggle over parliamentary government, the battle cry of Johan Sverdrup and his
Liberal Party was ‘all power in this assembly’, meaning the Storting. Although the
Liberals may not have realized their goal in full, they have come far in terms of the
We are grateful to Bjørn Erik Rasch, Eivind Smith, Lars Svåsand, Henry Valen, and the co-editors for valuable
information and comments.
1
Nordby (2000) claims that the Conservatives did not fully accept the principles of parliamentary
government until 1905.
524 Kaare Strøm and Hanne Marthe Narud
formal distribution of authority in Norway, which places few and relatively weak con-
straints on the parliamentary majority.
Under the Norwegian Constitution, all legislative authority is exercised by the people
through Parliament (Article 49). The same constitutional article specifies that the
Storting consists of two divisions: the Odelsting, with three-quarters of the total num-
ber of members, and the Lagting, with one-quarter. These two divisions of the Storting
are not truly independent chambers, since they are jointly elected. The assignment of
individual representatives to the Odelsting or Lagting occurs when the Parliament first
convenes after an election, on the first weekday in October. The members of the
Lagting are then elected by the parliamentarians at large, and the remaining members
assigned to the Odelsting. In reality, party leaders negotiate the membership of the two
divisions, and since the advent of parliamentary government interparty proportional-
ity has prevailed. Thus, the parties’ relative strength in the two divisions is effectively
identical, and the majorities in the two divisions very rarely diverge. The division of the
Storting into two quasi-chambers thus has virtually no consequence for legislation.
Note also that the Storting is so divided only when it deliberates on non-financial
bills.
Bills can be introduced by any member of the Odelsting, and there is no requirement
of a second sponsor. In reality, however, government bills, known as propositions
(proposisjoner), dominate the legislative agenda. Non-financial bills must be approved
by the Odelsting and the Lagting separately by a simple majority of those present
and voting. If the Lagting twice fails to approve the bill adopted by the Odelsting,
the bill goes to a plenary session, which must pass the bill by a two-thirds majority
(Article 76). After legislation has successfully passed the Storting, it is signed by
the King and Prime Minister in a formal cabinet meeting (Article 78). The
Constitution thus technically gives the King and the Prime Minister ex post veto power.
However, parliamentarism has made this power a pure formality. The volume of
Storting legislation has gradually increased. At the end of the nineteenth century, it
passed an average of twenty laws per year. This has after Second World War grown to
about ninety. Between 1989 and 1994, it ranged from 76 in 1989–90 to 155 in 1992–3
(Østbø 1995: 88).
Budgetary bills are decided by the Storting in plenary sessions. The budget proposal
(Stortingsproposisjon No. 1) is always the first bill introduced by the government when
the annual Storting convenes in October. Since 1961, the budget year begins in January
and thus coincides with the calendar year, and budget deliberations therefore dominate
the fall session of the Storting.
Constitutional amendments must be introduced during one of the first three years
of a legislative term and then be adopted in one of the first three years of the subse-
quent parliamentary term. Thus, the voters will always be consulted between the time
that a constitutional amendment is proposed and the final parliamentary vote.
Ordinary constitutional amendments are considered by a plenary session of the
Storting and require a two-thirds majority (Article 112). Constitutional amendments
that involve a transfer of sovereignty to an international organization face an even more
Norway 525
rigorous requirement. According to Article 93, passed in 1962 in the wake of Norway’s
first application for European Union (EU) (then EEC) membership, such amendments
must command a three-quarters majority.
Outside the Parliament, there are few domestic constraints on the legislative policy
process. A Constitution that provides for centralized government and singular institu-
tions has proven to be conducive to parliamentary governance. Norway is a unitary
state, and the constitutional status of subnational government is weak. Members of
Parliament are the only national agents elected by the people. And although some con-
straints exist on delegation and accountability through the parliamentary chain,
Norway is in practice a fairly pure case of parliamentary democracy.
Although the formal Constitution vests a number of powers in the cabinet (formally
the King), much of this authority has been eroded by convention. Yet, as we will dis-
cuss below, the cabinet retains significant legislative and appointment powers.
More importantly, it has strong legislative agenda powers. Moreover, in practice
Parliament delegates much of its policy-making authority to executive agencies
( fullmaktslovgivning).
At the level of the central government, two additional constraints on legislative
authority need to be mentioned, although neither is particularly strong by international
standards. One is the referendum mechanism. Since the Norwegian Constitution does
not mention referendums, they can only be consultative. Only six national referendums
have in fact been held, the most recent being the 1972 and 1994 EU membership refer-
endums. Second, the Supreme Court has established a doctrine of judicial review, but
the courts are in practice very reluctant to challenge the constitutionality of parlia-
mentary decisions. A less constitutionally entrenched, but in practice more significant,
constraint is the practice of remiss (høring) in legislative matters. This practice, which
is embedded in ordinary legislation, provides a right for affected interests, such as gov-
ernment agencies, institutions, and interest organizations, to be consulted on legislative
initiatives of significant concern to them.
P O L I T I C A L PA RT I E S
Though political parties were not recognized by Norwegian law until well into the
twentieth century, formal party organizations have existed since the 1880s, and looser
parliamentary clubs began forming a couple of decades before that. The two original
parties, the Liberals (Venstre) and the Conservatives (Høyre) date back to 1883, when
they represented the two sides in the battle over parliamentary government. The
Labour Party (Arbeiderpartiet) followed in 1887, reflecting the beginning of Norwegian
industrialization. All other parties that currently enjoy parliamentary representation date
from the twentieth century. The Centre Party (Senterpartiet) and the Christian People’s
Party (Kristelig Folkeparti) both broke away from the Liberals before the Second World
War. The Socialist Left Party (Sosialistisk Venstreparti) and the Progress Party
(Fremskrittspartiet) both formed in the wake of the destabilizing 1972 European
526 Kaare Strøm and Hanne Marthe Narud
Community (EC) referendum, although the former built on a precursor, the Socialist
People’s Party (Sosialistisk Folkeparti), that was founded in 1961.
The party system has been defined around six dimensions of conflict, determined by
economic, geographical, and cultural circumstances (see Rokkan 1967, 1970; Valen and
Rokkan 1974). The class cleavage and the urban–rural divide were determined by eco-
nomic conflicts in the labour market and the commodity market, respectively. A terri-
torial cleavage between centre and periphery partly overlaps with three cultural
cleavages: a socio-cultural conflict between two different versions of the Norwegian
language, a moral conflict articulated by the teetotalist movement, and a religious con-
flict originally pitting the liberal leadership of the Lutheran state church against a
revivalist lay movement. A number of cleavages have thus influenced Norwegian party
politics. Although the labour market (Left–Right) conflict has in the twentieth century
dominated and generated a bloc division between Socialists and non-socialists, the
cleavages cross-cut in a complex and dynamic fashion.
Through the 1960s Norway had one of the most stable party systems in Western
Europe. With the first EC referendum in 1972, however, Norway began to experience
more party system fragmentation and volatility than at any time since the 1920s. After
the early 1970s about one out of three voters changed partisanship from one election to
the next, as compared to the 1960s when the corresponding figure was about one out of
four (Valen and Urwin 1985). In the 1993 Storting election, individual volatility reached
almost 44 per cent, a proportion that was virtually unchanged in 1997 and 2001.
Norwegian parties, both those on the Left and those on the right, traditionally had
strong mass membership organizations. For many years about fifteen per cent of the
electorate were dues-paying members, according to reports from party headquarters
(Svåsand 1985: 49–53). In some smaller parties, particularly the Centre Party, the ratio
of members to voters has at times been as high as one in three. However, national elec-
tion surveys indicate that during the 1990s party membership declined to around 10–
11 per cent (see also Svåsand, Strøm, and Rasch 1997). In such parties as the Labour
Party and the Liberals, it had the late 1990s come closer to 5 per cent than to double
digits (Aftenposten, 16 July 1998). According to the parties’ own records, the Labour
Party and the Conservatives (the two largest membership organizations) both lost
about one-third of their gross membership figures between 1993 and 1999, whereas the
smaller parties all suffered somewhat smaller losses.
The parties are well provided for financially. Public financing of political parties was
first introduced in 1970, and has since increased greatly in volume. The largest part of
these subventions goes to the national party organizations. The total volume of such
aid was 8 million NOK in 1970, increasing to 21 million in 1980, and almost 59 million
in 1990. By 1997, it had passed 100 million, and in 1999 it stood at close to 119 million.
Only registered parties that gain at least 2.5 per cent of the national vote are favoured
in the funds distributed nationally. There are also province-level and local subsidies, in
total almost 100 million in 1999. The distribution of all these funds is largely propor-
tional to the electoral support of the party. Public subsidies make up an ever-increasing
share of party finances. The Labour Party, for example, in 1997 received more than 41
million NOK from the government, whereas the unions contributed around 10 million
Norway 527
NOK, and individual members only just over one million. Similarly, public funds
accounted for 71 per cent of the Centre Party’s revenue and 70 per cent of that of the
Conservatives (Aftenposten, 16 July 1998).
Norwegian parties are critical players at most stages of the policy process. As the section
on Delegation from Voters to Parliament will show, they control the selection of candi-
dates for parliamentary office to a greater extent than in any other Nordic country. In the
legislative arena, party cohesion is very high, and the same is the case in the cabinet. In
coalition governments, the most serious disagreements have tended to be handled by party
leaders, rather than through formal cabinet channels. It is only in the civil service that party
control is significantly less prominent. Though there are frequent tales of partisan promo-
tions at high levels of the civil service, the administration is by and large professional and
non-partisan. Partisan quotas or favouritism would be considered illegitimate. During the
period of social democratic dominance, Labour Party members and sympathizers gained
a disproportionate share of top-level appointments in state enterprises including the
Norwegian Broadcasting Corporation (NRK). With Labour’s vote share declining and pri-
vatization gaining support, however, the importance of such partisan preferences seems to
be decreasing. All in all, Norwegian voters have become less enamoured with their parties,
but there is as of yet little evidence that the parties have suffered significant decline in their
control of the democratic chain of delegation.
D E L E GAT I O N F RO M VOT E R S TO PA R L I A M E N T
The Electoral System
Delegation from voters, the ultimate principals, to their representatives in Parliament
is surely one of the most exhaustively studied and debated phenomena in politics, and
for good reason. Norway is no exception, as electoral reform was on the political
agenda for most of the twentieth century. The rules by which Members of Parliament
(the Storting) are selected and held accountable are embodied in legislation of different
status. Some of these rules are contained in the Constitution; others are spelled out in
ordinary legislation, the most important of which is the Election Law of 1 March 1985
(Overå and Dalbakk 1987; Aardal 2002).2 Unlike most parliamentary systems, legislative
terms are constitutionally fixed at four years, and there is no provision for early parlia-
mentary dissolution. Nor does the Constitution provide any mechanism for the recall
of individual representatives, or specify conditions under which a representative would
lose his or her seat. Parliament does not have the right to expel its own members. There
are no term limits or any mandatory retirement age, even though parties do not com-
monly nominate candidates who will turn seventy during the parliamentary term.
Thus, the independence of the elected representatives is very strongly protected.
Three principles established by the constitutional assembly in 1814 have had a lasting
effect upon the electoral system (Valen 1981, 1985; Matthews and Valen 1999). First,
basic principles, such as the definition of constituencies and the apportionment of seats,
2
Prior to that date, the relevant legislation was contained in separate laws concerning parliamentary elec-
tions and nominations, both of which were adopted in 1920 (Valen 1985).
528 Kaare Strøm and Hanne Marthe Narud
were incorporated into the Constitution. Since constitutional amendments require a
two-thirds parliamentary majority, this decision made reapportionment or electoral
reform very difficult. Second, the constitutional assembly decided that peripheral
regions should be overrepresented compared to central areas. And third, the assembly
decided that urban and rural areas should vote separately, with a seat allocation in the
ratio of 1:2. This latter provision, which originally favoured urban areas and gradually
came to overrepresent rural ones, was abolished as late as 1952. From then on, each of
the 19 provinces, including both urban and rural areas, has been a separate constituency.
Norway has since 1920 employed proportional representation (PR). Until 1952, a
d’Hondt formula was used in a single-tier system with relatively small-magnitude dis-
tricts. Since 1952, Norway has used a modified St. Laguë system with somewhat greater
district magnitude. The total number of seats has been increased incrementally, from 150
to 155 in 1972 (effective with the 1973 elections), and to 157 in 1985. The main purpose
of these increases has been to ease the under-representation of such areas of population
growth as the Oslo region. Only in 1988 (effective with the 1989 elections), however, was
complex districting introduced, when eight adjustment seats were created. Since then,
there have been 157 first-tier seats and a pool of eight national second-tier seats.3
The 2002 report of an Electoral Reform Commission provided the impetus for sig-
nificant institutional reform. The total number of seats was increased to 169, and the
number of adjustment seats raised from eight to nineteen. The geographical appor-
tionment of seats was removed from the Constitution and subjected to periodic adjust-
ments based on census figures. Yet, the deliberate overrepresentation of rural areas was
retained. The Commission’s recommendation of personal preference voting was, how-
ever, rejected by the Storting. The reforms are effective as of the 2005 election.
As in the case of adjustment seats, the demand for greater partisan proportionality
(‘fairness’) was the main argument when the Storting reintroduced apparentement
(listeforbund) in 1985, after it had been abolished in 1949 (Matthews and Valen 1999). This
reform greatly benefited the non-socialist coalition parties. Apparentement was removed
again in 1988, and this time prohibited by constitutional amendment. Yet, demands for
strict geographical proportionality in representation (‘one person, one vote, one value’)
were never seriously considered until the Electoral Reform Commission of 2002.
3
Supplementary seats, allocated to parties which are under-represented on the basis of constituency repre-
sentation, are given to provinces in which the respective parties have the highest remainders (in raw numbers)
of ‘unused’ votes. The highest remainders tend to be found in large and under-represented constituencies.
Norway 529
have only recently become an issue. The debate on this topic was triggered by a 1997
election campaign incident in which the Progress Party attempted to purchase such time
on Norway’s premier commercial television channel, TV2. The consumers’ ombuds-
man intervened and stopped the advertising, but the Market Council later reversed this
decision, creating a situation of legal ambiguity. Hence, in July 1999 the Storting passed
a new law permitting political advertising on radio but not on television. Previously, the
political parties were not allowed to advertise in any of the broadcast media, the main
argument being that it would favour the most resourceful parties and groups. Those in
favour of political advertising, on the other hand, pointed to the fact that such advertis-
ing was already allowed in the print media as well as on the internet.
Another issue that has only recently been regulated is campaign polling. Until 1994,
the only restrictions on polling and the publication of poll results were unofficial
gentlemen’s agreements between the parties and the polling organizations. Before the
1994 EU membership referendum, however, Parliament banned the publication before
the polls were closed of surveys conducted on election day. In addition, restrictions
were introduced on exit polling, as no voter surveys or interviews were permitted
inside the polling station or in its immediate surroundings. From 1995 on, these restrictions
have been applied to ordinary local and national elections as well.
4
In order to be registered, a party must submit a petition signed by at least 3,000 eligible voters, who need
not, however, be party members. A party’s registration lapses if it does not present candidates in any con-
stituency in two successive elections (Overå and Dalbakk 1987).
530 Kaare Strøm and Hanne Marthe Narud
The law gave all parties monetary incentives to select their candidates in each con-
stituency (province) by conventions of delegates elected by the parties’ dues-paying
members in the province’s subdivisions. These nomination procedures were not
mandatory, but if the party used them, the national government would fund the
provincial conventions. In general, the parties tended to follow the law, except in Oslo,
where distances are so small that expenses for nomination meetings are negligible
(Valen 1988: 212; Valen et al. 2002).
The organizational procedures for candidate selection do not differ much from one
party to the next. In each locality the delegates are elected by majority vote. These
meetings normally take place 6–9 months before the election. In all parties the typical
nomination committee asks incumbents whether they are available for renomination,
though in recent years it has become more common for interested incumbents to be
turned down for renomination (see below). Normally the first draft of the nomination
committee’s list is sent out to the local branches for comments. All parties report that
local organizations carry a great deal of weight when the committees draw up their
lists (Valen 1988: 213; Valen et al. 2002). Although it is difficult to identify the precise
party arena in which the key decisions are made, it is clear that the candidate selection
process in Norway is decentralized. As Valen (1988: 228; Valen et al. 2002) points out,
the process is not one of oligarchic selection directed from the top of the political par-
ties. Rather, the ultimate decision lies with the constituency party organizations.
5
The independents are Anders Aune and Steinar Bastesen. Aune was in 1989 elected as a representative
of a local non-partisan slate (Folkeaksjonen Fremtid for Finnmark) in the province of Finnmark. He did not run
for re-election in 1993. Bastesen, a controversial and outspoken whaling captain, was elected from Nordland
on the slate of the Coastal People’s Party ( Kystpartiet) in 1997 and re-elected in 2001. Interestingly, both inde-
pendents have been elected from northern provinces, a fact that may reflect a peripheral protest vote.
Norway 531
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
More than a hundred years ago, Bagehot already identified executive selection as the
most important function of the British Parliament (Bagehot 1963 [1867]; see also Cox
1987). This stage of delegation may be either internal or external. In a ‘pure’ form of
parliamentarism, the delegation from Parliament to cabinet is internal, in that only
Members of Parliament may be appointed to the cabinet. In many contemporary par-
liamentary systems including Norway, however, this process is external, as the mem-
bers of the cabinet need not, and indeed may not, simultaneously hold parliamentary
seats (for a cross-national survey, see Andeweg and Nijzink 1995).
Cabinet Formation
Formally, the Norwegian Constitution gives the King wide discretion to appoint the
members of the cabinet, which is still formally known as the King’s Council. In prac-
tice, however, the King has exerted no influence on the composition of any cabinet
since 1928, and it is questionable whether he could constitutionally exercise any such
authority today.6 In reality, when he formally calls upon someone to form a new govern-
ment, the King always follows the advice of the leaders of the parliamentary parties. In
practice, the choice of a Prime Minister-designate has rarely been difficult. The use of
informateurs has no codified place in the Norwegian Constitution, and the practice has
been rare indeed. In the postwar period, there is only one notable case.7 If the recent
trend toward party system fragmentation and coalition fluidity continues, however, it
is not inconceivable that a stronger tradition of informateurship may develop.
6
In 1928, King Haakon VII called upon the leader of the Labour Party, the largest party in Parliament, to
form a new government. The King’s decision was in accordance with established procedure, but contrary to
the advice of the outgoing Prime Minister and the President of the Storting (Nordby 2000: 105). The result
was the formation of the first Socialist government in Norwegian history (see Björnberg 1939). The govern-
ment proved short-lived but the King’s behaviour did much to solidify his support among Norwegian Social
Democrats.
7
In the difficult cabinet crisis of 1971, when a bourgeois majority coalition had just broken down over the
EC issue and the Prime Minister’s conduct in this area, the King formally gave the President of the Storting,
Conservative Bernt Ingvaldsen, the mandate of investigating the opportunities for another non-socialist coalition.
It was obvious that the role of Ingvaldsen, a senior, right-wing, and somewhat formal member of his party,
was purely that of an informateur. At any rate, his efforts failed for no fault of his own, and the informateur
institution has never again been used.
Norway 533
No Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the cabinet
formation votes (iv) votes (cv) end with an
n.a. No. of ncv Cabinet removed Cabinet resigned under specific early election?
by ncv to pre-empt ncv institutional n.a.
mechanism
n.a.
1 Gerhardsen II 45/11/05 5 No
2 Gerhardsen III 49/10/10 0
3 Torp I 51/11/19 1 No
4 Torp II 53/10/12 4 No
5 Gerhardsen IV 55/01/22 1 No
6 Gerhardsen V 57/10/07 3 No
7 Gerhardsen VI 61/09/11 4 Yes
8 Lyng 63/08/28 0
9 Gerhardsen VII 63/09/25 0
10 Borten I 65/10/12 1 No
11 Borten II 69/09/07 1 No
12 Bratteli I 71/03/13 0
13 Korvald 72/10/18 1 No
14 Bratteli II 73/10/16 7 No
15 Nordli I 76/01/15 1 No
16 Nordli II 77/09/11 3 No
17 Brundtland I 81/02/04 0
18 Willoch I 81/10/14 3 No
19 Willoch II 83/06/08 1 No
20 Willoch III 85/09/08 1 No
21 Brundtland II 86/05/09 6 No
22 Syse 89/10/16 1 No
23 Brundtland III 90/11/03 2 No
24 Brundtland IV 93/09/13 5 No
25 Jagland 96/10/25 0
26 Bondevik 97/10/17 —
a
n.d.⫽no data (missing data); n.a. or ‘—’ ⫽ not applicable in this case.
538 Kaare Strøm and Hanne Marthe Narud
same committee throughout the four-year parliamentary term,8 and re-elected repre-
sentatives often retain their committee assignments. In general terms, the jurisdictions
of parliamentary committees tend to mirror those of the ministerial departments,
though exceptions to this rule have become increasingly common (Rommetvedt 1996).
Each of the standing committees engages in various forms of oversight of the execu-
tive agencies under its jurisdiction. For most of the post-1945 period, the Storting has had
one standing committee with more general and oversight functions. Until 1972, this was
the Protocol Committee, which was a committee established as early as 1814 and author-
ized to read and audit cabinet minutes. Under the early separation-of-powers
Constitution, this was a powerful committee, but under parliamentarism it gradually
faded into obscurity. By the 1950s, members complained publicly that service on the com-
mittee gave them stomach aches (Sejersted 2000: 176). In 1972, this dreaded committee
assignment was abolished and its oversight functions divided between the various other
permanent committees. This reform proved to impede efforts at coordination, which in
1981 led to the establishment of another centralized oversight committee, the Control
Committee, which also took over responsibility for constitutional affairs (Hansen and Mo
1994: 28–30). Yet, the fact that the members had their primary committee assignment
elsewhere weakened this new committee, and in 1992 it was reorganized once again and
put on a par with other committee assignments. This latest reorganization ushered in a
renaissance for the Control Committee during the 1993–7 parliamentary term, when the
committee vigorously scrutinized a string of executive ‘scandals’ (Sejersted 2000).
Parliamentary questions have long played an important role in Norwegian parlia-
mentarism, and their importance has tended to increase over time. As Rasch notes,
‘parliamentary questions are in fact used quite extensively as a means of controlling the
executive. This is not because the legislators necessarily intend to control. Rather, con-
trol is realized as a by-product of behaviour motivated primarily by an attempt to reach
other (less collective) aims related to—or derived from—the electoral arena’ (Rasch
1994: 247–8; see also Kuhnle and Svåsand 1984). Some form of parliamentary questions
have been asked in the Storting as far back as 1885 (the year after the introduction
of parliamentary government), and interpellations found their way into the Rules
of Procedure in 1908. A regular question time, held every Wednesday beginning at
11 a.m., was established in 1949 (Rasch 1994: 254–5).
Several other types of parliamentary questions have since evolved. For a long period of
time, ‘long’ questions (which allowed the questioner and responding minister five minutes
each) coexisted with short ones, but in 1989 the former were abolished. An October 1996
reform instituted two additional forms of parliamentary questions: a spontaneous ques-
tion time (up to 1hour in duration) and written questions to be answered in writing (Rasch
1998). The former of these allows for the most immediate and spontaneous debate
between cabinet members and MPs, as the questions are not known in advance. Finally,
8
Members of the Storting who are appointed to the cabinet must relinquish their seats to their respective
deputies as long as they serve in the cabinet. If they return to the Storting before the end of the term,
however, they are not guaranteed reassignment to the same committees on which they served before their
cabinet appointments.
Norway 539
the Parliamentary standing orders permit representatives to raise urgent questions during
the daily adjournment debate. Known as ‘questions at the conclusion of the meeting’, such
inquiries are rare. Ordinary questions remain by far the most numerous (Rasch 1998: 9).
Outside their own ranks, members of the Storting possess more specialized vehicles
of oversight, such as the Auditor General and the various parliamentary ombudsmen.
The Auditor General’s Office (Riksrevisjonen) traditionally engaged in more technical
review of the government’s expenditures and financial dispositions. During the 1980s,
however, these audits became more comprehensive and often critical. From 1993 on,
these more incisive audits were increasingly seized upon by a reformed and embold-
ened parliamentary Control Committee. At the same time, the Auditor General’s
Office began issuing separate review of particularly important and controversial gov-
ernment agencies. Thus, there have in recent years been strong synergies between the
Storting’s Control Committee and the Auditor General’s Office, and their simultaneous
emergence as important fora of ex post scrutiny of the executive branch is no mere
coincidence (Sejersted 2000).
The Parliamentary Ombudsman for Civilian Affairs is an institution established in
1962 to investigate administrative abuses against ordinary citizens. The Ombudsman is
a quasi-judicial institution that reports to Parliament. The Ombudsman cannot assess
penalties, but does report on inappropriate practices and may propose reforms and var-
ious forms of redress. This institution has since its inception virtually displaced
Parliament itself as a forum for petitions and individual casework concerning govern-
ment agencies. The EOS Committee, which was established in 1995 and reports to
Parliament, is a forum for the scrutiny of secret (intelligence) services. Finally, the
Storting may appoint independent commissions of inquiry. Until recent years, most
investigative commissions were in fact appointed by the cabinet, though often after
consultation with Parliament. Since 1985, however, there have been at least four impor-
tant commissions of inquiry established by the Storting itself (Sejersted 2000: 178).
Thus, there is no doubt that parliamentary scrutiny of the Norwegian executive has
become more intense since the 1990s, and that it has found an increasing number of effect-
ive institutional vehicles. Compared to Denmark, Sejersted (2000: 179) argues that the
Norwegian Parliament has focused less strictly and stringently on legal criteria. Instead, the
Norwegian Parliament has more and more jealously defended its right to be properly
informed by executive agencies, and critical scrutiny of individual ministers has often
focused on whether they have neglected to inform Parliament properly about executive
actions under their jurisdiction. Through this growing activism, the Norwegian
Parliament has also increasingly focused on ex post, rather than ex ante, control of the exec-
utive branch. The main factor that limits the effectiveness of such parliamentary scrutiny
is the bluntness (and hence lack of credibility) of the key sanctions that can be imposed:
no-confidence votes and impeachment (see below).
D E L E GAT I O N F RO M C A B I N E T TO I N D I V I D UA L M I N I S T E R S
The Norwegian cabinet is in the 1814 Constitution formally recognized as the Council
of State. Article 12 of the Constitution prescribes that the cabinet consist of
540 Kaare Strøm and Hanne Marthe Narud
a Prime Minister and at least seven other members. In practice the number of cabinet
members in Norway has since 1945 varied between thirteen and nineteen, with a secular
trend toward larger size. Junior ministers, known as secretaries of state (statssekretœr) and
personal advisers (personlig rådgiver), were first appointed in 1949. As of September 2000,
there were forty-six such appointees altogether. The cabinet normally meets three times
per week, with all formal decisions (Royal Resolutions) being made on Fridays in the
Council of State, in which the King is present. Council of State meeting are very largely
ritualistic, typically lasting only 30–45minutes and containing no real debate on the issues
under consideration. Decisions of the Council of State require the signatures of both the
King and the Prime Minister. More substantive discussions take place in the King’s
absence in the Monday and Thursday cabinet meetings, which last 2–3h, as well as
during the informal cabinet luncheons on Fridays.
Cabinet Authority
The cabinet (Council of State) enjoys extensive powers under the Norwegian
Constitution. It is the supreme, collective leadership of the central administration. In
the form of bills and white papers, it initiates and prepares most legislation that is sub-
sequently adopted by the Parliament. It can issue decrees with the force of law when
Parliament is not in session. It is routinely granted broad implementation powers
through regular legislation. It is in charge of a large system of public enterprises. It is
an institution to which citizens can appeal under certain circumstances. And it is a
branch of government with certain other constitutional prerogatives, such as appoint-
ment powers. Within certain limits, the cabinet can delegate authority, most commonly
to the individual ministries. Yet some decisions, such as certain appointments, are con-
stitutionally required to take place in the Council of State.
(Olsen 1983; Eriksen 1988b). Only Willoch appears to have appropriated some veto
powers in this respect.
The Prime Minister is in charge of the Prime Minister’s office. This used to be a very
small operation, as the Prime Minister had virtually no administrative help in the imme-
diate postwar period. When Einar Gerhardsen formed his first cabinet in 1945, his staff
consisted of two civil servants (one lawyer and one economist), three clerical employees,
and a chauffeur (Berggrav 1997: 15). In 1956, the office was reorganized and professional-
ized, and by the mid-1980s it had grown to a size of approximately thirty employees
(Berggrav 1985). When Bondevik took office in 1997, he employed six political appointees
(four secretaries of states and two political advisers), thirty-five civil servants, and eight
chauffeurs, for a total of forty-nine employees (Berggrav 1997: 41). The staff is divided into
an administrative office, an economic office, and an international office. The Prime
Minister’s office has an important function in preparing the cabinet agenda. It also per-
forms coordination functions, such as meetings to which both political appointees and
civil servants may be summoned, as well as ceremonial liaison functions vis-à-vis the King.
Cabinet Members
The general rule, spelled out in Article 2 of the standing orders of the cabinet
(Regjeringsinstruksen), is for each cabinet member to be the administrative head of some
ministry. (The Prime Minister heads his or her own office, but after the Second World
War no Prime Minister has been in charge of any regular ministry.) Occasionally,
though, ministers without portfolio have been appointed. This was most notably the
case in the years of reconstruction immediately following the Second World War, when
as many as three ministers without portfolio served simultaneously. It is even more rare
for any department to be represented by more than one cabinet member, although this
again has occasionally happened in periods of administrative reorganization. Thus,
Norway neatly fits one favourable condition for ministerial autonomy: being the head
of a ministry is to all intents and purposes a necessary and sufficient condition for
membership in the cabinet.
Though cabinet members may have close ties to Parliament, they cannot serve as rep-
resentatives while they hold cabinet office. Nor is it expected that cabinet members have
prior parliamentary experience. Between 1945 and 1978, 52 per cent of all Norwegian
cabinet ministers had no prior parliamentary experience (Olsen 1983: 93). This propor-
tion was especially high in the early postwar Labour governments under Gerhardsen.
Particularly in ministries with distinctive and well-organized clienteles (e.g. agriculture,
fisheries), it is considered much more desirable for cabinet members to have good
interest group ties than to have parliamentary experience. Prime Ministers and Finance
Ministers, however, tend to have extensive previous parliamentary experience.9
9
Recent Labour cabinets have occasionally violated this expectation. When Gro Harlem Brundtland first
became Labour leader and Prime Minister in 1981, one of the most serious criticisms leveled against her was
precisely her relative lack of parliamentary experience. Similar criticisms were directed against Prime Minister
Jagland, who had limited parliamentary experience and no cabinet experience when he took over in 1996.
542 Kaare Strøm and Hanne Marthe Narud
Ministerial Accountability
In Norway, ministers bear two kinds of responsibilities for their actions and those of
their subordinates. One is a constitutional responsibility for any illegal or unconstitu-
tional behaviour on the part of the cabinet or the minister individually, including any
non-compliance with acts of Parliament or any withholding of information to which
Parliament is entitled. A minister may also be constitutionally responsible for the
actions of a subordinate, such as a civil servant, if negligence is shown in the instruc-
tion or oversight of this subordinate. A formal protocol is kept of Council of State deci-
sions, and individual ministers can request that their dissent from cabinet policy be
registered in this document. This is the only way they can escape legal responsibility for
cabinet decisions they consider ‘unconstitutional or injurious to the realm’. However,
except for parliamentary scrutiny, the cabinet protocol is confidential, and a registered
dissent is no political excuse for public criticism of cabinet decisions.10
Violations of the constitutional responsibilities of ministers are punishable by
impeachment, but since the dramatic and unsuccessful impeachment proceedings
against Prime Minister Abraham Berge in 1926–7, this extreme instrument has fallen
into disuse. In recent years, ministers guilty of major violations of the law have tended
to resign and face regular criminal prosecution. Thus, during the 1980s two ministers
(one Conservative and one Labour) resigned under suspicion of embezzlement. Both
were subsequently convicted.
A more pressing concern for most cabinet ministers is their political responsibility to
the Storting. This is a form of collective responsibility shared by all members of the
cabinet and all other political appointees in the ministries. Whereas ministers can
escape some forms of constitutional responsibility by registering their dissent from
10
Not all cabinet decisions are formal and binding. Technically, cabinet ministers and their subordinates
are bound only by the formal decisions of the Council of State and not by the more informal agreements
reached in other cabinet meetings. Yet this distinction must be considered more formal than real. Individual
cabinet members would be ill advised to ignore or sabotage any cabinet decision, regardless of whether it has
been ratified in the Council of State.
544 Kaare Strøm and Hanne Marthe Narud
Council of State decisions, there is no way for an individual minister to exempt himself
from his political responsibility to the parliamentary majority.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
The Norwegian civil service was in its initial form an inheritance from Denmark. As one
of the few social organizations with professional expertise, it played a major role in the
political and economic development of Norway in the nineteenth century. Throughout
the development of the Norwegian civil service, there has been an ongoing debate over
the merits of a centralized and hierarchical, versus a decentralized and professional,
administrative structure. In organizational terms, this has been a debate over whether
the civil service should be organized around a smaller set of hierarchical departments
under strong political control (traditionally known as the ‘Danish model’, although
Denmark itself has in recent decades deviated somewhat from these principles) or in the
form of a larger set of more autonomous and professional agencies under less direct
political control (the ‘Swedish model’). In practice, the Norwegian administration has
become a hybrid of these two models. The former principle is embodied in a structure
of departments, the latter in a set of directorates (direktorater). The Danish model
dominated initially, but from about 1850 on an increasing number of directorates were
Norway 545
established. This process was reversed after the introduction of parliamentary govern-
ment in 1884, when the parliamentary majority wished to bring the civil service under
stricter political control. Since the Second World War, there has been no consensus on
either of these two basic principles, and a pragmatic balance has evolved that places
Norway somewhere in between Denmark and Sweden in this respect. As of 2001, there
were seventeen departments and seventy-five directorates in the Norwegian central
administration (Christensen et al. 2002).
Since the early 1980s, the total number of civil servants in the Norwegian central
administration has hovered between 3,000 and 4,000, numbers that had increased
slightly over the previous decade. Over a longer period of time, most of the growth in
the central administration has occurred in the directorates. Thus, between 1945 and
1991, the number of departmental positions increased by 70 per cent, whereas the cor-
responding growth in the directorates amounted to no less than 210 per cent
(Christensen and Egeberg 1994). There are two classes of Norwegian civil servants:
higher civil servants (embetsmenn) and ordinary civil servants (tjenestemenn). According
to Article 28 of the Constitution, higher civil servants must be appointed by the King
in a formal cabinet meeting. Their tenure is also strictly protected, in that they can only
be dismissed if found guilty of malfeisance in a court of law. Ordinary civil servants
have lesser protections, but in practice Norwegian civil servants of both categories have
always tended to be professional and non-partisan.
Under professional civil services systems such as the Norwegian one, politicians have
very little ex post control over the civil servants to whom they delegate. Certainly, civil
servants cannot be hired or fired at will. Accountability takes place in two ways: (1)
Civil servants have explicitly designed contracts that prevent them from taking various
types of hidden or arbitrary action; and (2) all decisions taken in the executive branch
explicitly carry the authorization of the head of department. A great deal of formal
authority is therefore placed in this minister’s hands. Ex ante controls, on the other
hand, are quite strong, in that the credentials that are required of civil servants tend to
be quite substantial.
Since the late 1960s, ex post control has been facilitated by legislation designed to open
the administration up to third-party oversight. The Administrative Procedures Act
(Forvaltningsloven) of 1967 established a right for affected interests to be consulted before
the administration adopts a general rule that significantly impacts upon them. The
Freedom of Information Act (Lov om offentlighet i forvaltningen) of 1970 (implemented in
1971) gave citizens a general right of access to administrative documents.
E X T E R NA L C O N S T R A I N T S
As mentioned in the introductory section, Norway is in practice a fairly pure parlia-
mentary democracy. Although some constraints exist on delegation and accountability
through the parliamentary chain, these are modest compared to many other European
nations. Below we discuss four sources of domestic constraint on the national chain of
delegation, namely subnational governments, direct democracy, judicial review, and
corporatism. Roughly speaking, these are listed in increasing order of importance. Yet,
546 Kaare Strøm and Hanne Marthe Narud
it is important to note that none of these constraints is of major importance in cross-
national perspective, and that most of them exist by virtue of delegation from
Parliament.
Norway is a unitary state. Although local government was established as early as
1837 and enjoys broad support and participation, it has weak constitutional support.
Local governments have weak taxation powers, and much of their revenue comes in
the form of transfers from the national government. Although some of their responsi-
bilities have remained fairly constant since the early nineteenth century, others have
changed, and the central (national) government has frequently changed these terms.
Provincial governments, of which there are nineteen, are even weaker. Following a
reform in the 1970s, each province now has an elected government, but their jurisdic-
tions are narrow (mainly parts of higher education), and several parties have in recent
years called for reductions in their authority or the abolition of provincial governments
altogether. Thus, subnational government is largely dependent on authority and
funds delegated by the national government and poses no significant constraint on the
latter.
Nor does direct democracy. Since Norwegian independence in 1905, only six
national referendums have been held: two in 1905 on national independence and on the
form of new Norwegian state (monarchy versus republic), two in the interwar period
over prohibition, and most recently the 1972 and 1994 EU membership referendums.
Because the Norwegian Constitution does not mention national referendums, they can
only be consultative, although in practice Parliament has never failed to follow the will
of the majority.11 Yet, there is no constitutional requirement that referendums be held
on any particular issue (including constitutional amendments), nor any mechanism by
which any institution except Parliament can call such an event. Thus, Parliament is con-
strained by direct democracy only to the extent that it decides to submit itself to this
mechanism.
Judicial review, however, is a less optional constraint on the national government. More
strongly than either Denmark or Sweden, Norway has established a doctrine of judicial
review. In fact, the Norwegian tradition of judicial review postdates only that of the
United States, and in developing this practice Norwegian legal authorities relied heavily on
American doctrines (Smith 1993: 32). Judicial review was not clearly and explicitly estab-
lished in the 1814 Constitution, but evolved gradually through a series of Supreme Court
decisions. Smith (1993: 158–9) argues that the practice can be traced back to the 1840s, that
it is clearly evident from 1866 on, and that it is first explicitly articulated in a 1890 Supreme
Court decision. The latter decision inaugurated the ‘golden age’ of judicial activism in
Norway. From about 1930 until the mid-1970s, however, this activism subsided, and
the courts became extremely reluctant to challenge the constitutionality of government
11
Before the EU membership referendum in 1994, representatives of several anti-EU parties did suggest
that they would not feel obligated by a narrow pro-EU majority. However, the popular response to this stance
was largely negative, so that it is unclear whether the anti-EU representatives would in fact have followed
through on their threat to disregard the popular vote.
Norway 547
C O N C LU S I O N
Norway is not the epitome of parliamentary government. In fact, the 1814 Constitution
was drawn up to establish a very different kind of regime, a separation-of-powers sys-
tem. Only gradually, informally, and incompletely has the Constitution been trans-
formed into a parliamentary one. It still lacks some of the typical features of
parliamentarism, such as parliamentary dissolution powers. Nonetheless, delegation of
authority follows the parliamentary pattern fairly closely. This, however, is largely due to
two interrelated factors: the strength of party government and a strong reliance on
ex ante controls over agents. Ex post controls, on the other hand, though increasingly
important, are much less well institutionalized. Nonetheless, the performance of
Norwegian democracy has traditionally kept most of its citizens happy.
Whether this happy outcome can be maintained is an entirely different matter. Both
of these features of Norwegian parliamentarism, political party government and
ex ante controls, have been eroded by recent developments in the Norwegian polity.
Though political parties so far maintain effective control of candidate selection as well
as policy-making in the Storting, they are rapidly declining as vehicles of mass parti-
cipation, and their previously high level of popular support is slipping (Strøm and
Svåsand 1997). Given these developments and the clear trend toward more open
candidate selection processes in other Nordic countries, it is an open question how long
Norwegian parties can maintain their strict control of parliamentary politics. The
weakening of ex ante controls has occurred simultaneously with the decline of parties,
Norway 549
and the two phenomena are in many ways interrelated. It is also in part the reflection
of the rise of new parties, such as the Progress Party and the Socialist Left, that reject
many traditional forms of social control. But the increasing difficulty of relying on
ex ante mechanisms is also driven in part by the emergence of a more diverse and com-
petitive, and less transparent, society. In some form, Norwegian politicians seem to rec-
ognize these developments and to respond to them by strengthening mechanisms of
ex post control (e.g. committee hearings and more rigorous audits), particularly in the
parliamentary arena. Their success or failure in these endeavours could have a major
effect on the future trajectory of Norwegian democracy.
REFERENCES
Aardal, Bernt (2002). ‘Electoral Systems in Norway’, in Bernard Grofmen and Arend Lijphart
(eds.), The Evolution of Electoral Systems and Party Systems in the Nordic Countries. New York:
Agethon Press: 167–224.
——, in cooperation with Valen, Henry, Narud, Hanne Marthe, and Berglund, Frode (1999).
Velgere i 1990-årene. Oslo: NKS-Forlaget.
—— and Valen, Henry (1989). Velgere, Partier og Politisk Avstand. Oslo: Central Bureau of Statistics.
—— —— (1997). ‘The Storting Elections of 1989 and 1993: Norwegian Politics in Perspective’,
in Kaare Strøm and Lars Svåsand (eds.), Challenges to Political Parties: The Case of Norway. Ann
Arbor: University of Michigan Press.
Andeweg, Rudy B. and Nijzink, Lia (1995). ‘Beyond the Two-Body-Image: Relations Between
Ministers and MPs’, in Herbert Döring (ed.), Parliaments and Majority Rule in Western Europe.
New York: St. Martin’s Press.
Bagehot, Walter (1963 [1867]). The English Constitution. London: Fontana/Collins.
Berggrav, Dag (1985). ‘Regjeringen’, in Trond Nordby (ed.), Storting og regjering 1945–1985:
Institusjoner—rekruttering. Oslo: Kunnskapsforlaget.
—— (1994). Slik styres Norge: Kongen, regjeringen og Stortinget i norsk statsliv. Oslo: Schibsted.
—— (1997). ‘Statsministerens kontors historie siden 1945’, in Dag Berggrav (ed.), Maktens høyborg.
Oslo: Grøndahl Dreyer.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
Björnberg, Arne (1939). Parliamentarismens utreckling i Norge efter 1905. Uppsala: Almqvist & Wiksell.
Christensen, Dag Arne (1997). ‘Europautvala i Danmark, Sverige og Noreg’. Nordisk
Administrativt Tidsskrift, 78: 143–62.
Christensen, Tom, Egeberg, Morten., (1994). ‘Sentraladministrasjonen—en oversikt over trekk
ved departement og direktorat’, in Tom Christensen and Morten Egeberg (eds.),
Forvaltningskunnskap, 2nd edn. Oslo: TANO.
Christensen, Tom, Egeberg, Morten., Larsen, Helge O., Lægreid, Per, and Roness, Paul G. (2002).
Fovalting og politikk. Oslo: Universitetsforlaget.
Cox, Gary W. (1987). The Efficient Secret. Cambridge: Cambridge University Press.
Eckhoff, Torstein and Smith, Eivind (1997). Forvaltningsrett, 6th edn. Oslo: TANO.
Eliassen, Kjell A. (1985). ‘Rekrutteringen til Stortinget og Regjeringen 1945–1985’, in Trond Nordby
(ed.), Storting og regjering 1945–1985: Institusjoner—rekruttering. Oslo: Kunnskapsforlaget.
—— and Pedersen, Mogens (1978). ‘Professionalization of Legislatures: Long-Term Change in
Political Recruitment in Denmark and Norway’. Comparative Studies in Society and History, 20:
286–318.
550 Kaare Strøm and Hanne Marthe Narud
Eriksen, Svein A. (1988a). Herskap og tjenere. Oslo: TANO.
—— (1988b). ‘Norway’, in Jean Blondel and Ferdinand Müler-Rommel (eds.), Cabinets in Western
Europe. London: Macmillan.
Hansen, Guttorm and Mo, Erik (1994). Om Stortingets arbeidsordning. Oslo: Stortinget.
Karmly, Dag (1975). Kings Bay-saken. Oslo: Gyldendal.
Kuhnle, Stein and Svåsand, Lars (1984). ‘Spørreordningene og politiske profiler i Stortinget
1977–1981’, in Ole Berg and Arild Underdal (eds.), Fra valg til vedtak. Oslo: Aschehoug.
Lijphart, Arend (1999). Patterns of Democracy. New Haven: Yale University Press.
Matthews, Donald and Valen, Henry (1999). Parliamentary Representation. The Case of the
Norwegian Storting. Columbus: Ohio State University Press.
Miller, Arthur H. and Listhaug, Ola (1990). ‘Political Parties and Confidence in Government:
A Comparison of Norway, Sweden, and the United States’. British Journal of Political Science,
29: 357–86.
Myhre-Jensen, Kjell and Fløistad, Brit (1997). ‘The Storting and the EU/EEA’, in Matti Wiberg (ed.),
Trying to Make Democracy Work. Stockholm: The Bank of Sweden Tercenary Foundation and
Gidlunds Förlag.
Narud, Hanne Marthe (1998). ‘Norwegen: Professionalisierung zwischen Parteien- und
Wahlkreisorientierung’, in Jens Borchert (ed.), Politik als Beruf. Die politische Klasse in westlichen
Demokratien. Opladen: Leske & Budrich.
—— and Valen, Henry (2000). ‘Does Social Background Matter?’, in Peter Esaiasson and Knut Heidar
(eds.), Beyond Westminister and Congress. The Nordic Experience. Columbus: Ohio State
University Press.
Nordby, Trond (1994). Korporatisme på norsk 1920–1990. Oslo: Universitetsforlaget.
—— (2000). I politikkens sentrum: Variasjoner i Stortingets makt 1814–2000. Oslo: Universitetsforlaget.
Olsen, Johan P. (1983). Organized Democracy. Oslo: Universitetsforlaget.
Overå, Oddvar and Dalbakk, Steinar (1987). Den Norske Valgordningen. Oslo: Sem & Stenersen.
Raaum, Nina Cecilie (1995). ‘The Political Representation of Women: A Bird’s Eye View’, in
Lauri Karvonen and Per Selle (eds.), Women in Nordic Politics: Closing the Gap. Aldershot: Dartmouth.
Ranney, Austin (1981). ‘Candidate Selection’, in David Butler, Howard R. Penniman, and
Austin Ranney (eds.), Democracy at the Polls. Washington, DC: American Enterprise Institute.
Rasch, Bjørn Erik (1994). ‘Question Time in the Norwegian Storting—Theoretical and Empirical
Considerations’, in Matti Wiberg (ed.), Parliamentary Control in the Nordic Countries: Forms of
Questioning and Behavioural Trends. Jyväskylä: The Finnish Political Science Association.
—— (1998). ‘Electoral Incentives to Control Government Ministers’. Paper prepared for presen-
tation at the International Conference on the Significance of the Individual Parliamentary
Member in Parliamentary Politics, Budapest, July 1–5.
Rokkan, Stein (1967). ‘Geography, Religion and Social Class: Crosscutting Cleavages in
Norwegian Politics’, in Seymour Martin Lipset and Stein Rokkan (eds.), Party Systems and Voter
Alignments. New York: Free Press.
—— (1970). Citizens, Elections, Parties. New York: David McKay.
Rommetvedt, Hilmar (1984). Borgerlig samarbeid. Stavanger: Universitetsforlaget.
—— (1996). ‘Norwegian Parliamentary Committees: Performance, Structural Change and
External Relations’. Paper presented at the International Conference on the Changing Roles
of Parliamentary Committees, Budapest, June 20–22.
Sejersted, Fredrik (1996). ‘The Norwegian Parliament and European Integration—Reflections
from Medium Speed Europe’, in Eivind Smith (ed.), National Parliaments as Cornerstones of
European Integration. London: Kluwer.
—— (2000). ‘Stortingets kontrollfunksjon’. Nytt Norsk Tidsskrift, 17: 173–85.
Norway 551
Skjeie, Hege (1997). ‘A Tale of Two Decades: The End of a Male Political Hegemony’, in
Kaare Strøm and Lars Svåsand (eds.), Challenges to Political Parties: The Case of Norway. Ann
Arbor: University of Michigan Press.
Smith, Eivind (1993). Høyesterett og folkestyret. Oslo: Universitetsforlaget.
Stavang, Per (1968). Parlamentarisme og maktbalanse. Oslo: Universitetsforlaget.
—— (1971). ‘Negativt fleirtal i norsk parlamentarisme’. Lov og rett, 145–66.
Strøm, Kaare (1990). Minority Government and Majority Rule. Cambridge: Cambridge University
Press.
—— and Lars Svåsand (eds.) (1997). Challenges to Political Parties: The Case of Norway. Ann Arbor:
University of Michigan Press.
Svåsand, Lars (1985). Politiske partier. Oslo: Tiden.
——, Strøm, Kaare, and Rasch, Bjørn E. (1997). ‘Party Organization’, in Kaare Strøm and
Lars Svåsand (eds.), Challenges to Political Parties: The Case of Norway. Ann Arbor: University of
Michigan Press.
Valen, Henry (1981). Valg og Politikk. Oslo: NKS-forlaget.
—— (1985). ‘Valgsystemet’, in Trond Nordby (ed.), Storting og regjering 1945–1985: Institusjoner—
rekruttering. Oslo: Kunnskapsforlaget.
—— (1988). ‘Norway: Decentralization and Group Representation’, in Michael Gallagher and
Michael Marsh (eds.), Candidate Selection in Comparative Perspective. London: Sage.
——, Narud, Hanne Marthe, and Skare, Audun (2002). ‘Norway: Party Dominance and
Decentralized Decisionmaking’, in Hanne Marthe Narud, Mogens Pedersen, and Henry Valen
(eds.), Party Sovereignty and Citizen Control. Odense: Odense University Press.
—— and Rokkan, Stein (1974). ‘Norway: Conflict Structure and Mass Politics in a European
Periphery’, in Richard Rose (ed.), Electoral Behavior: A Comparative Handbook. New York: Free
Press.
—— and Urwin, Derek W. (1985). ‘De Politiske Partiene’, in Trond Nordby (ed.), Storting og
Regjering 1945–1985, Vol. 2. Oslo: Kunnskapsforlaget.
Østbø, Ivar Buch (1995). Stortinget: All Makt i denne sal. Oslo: Schibsted.
18
I N T RO D U C T I O N
An elected Parliament has long been a part of Portugal’s political landscape. It was first
established after the Liberal revolution of 1820. Throughout the nineteenth century,
however, parliamentary life in Portugal was markedly oligarchic and clientelistic. In
1910, with the foundation of the Republic, Parliament had for the first time an oppor-
tunity to assert itself as an institution responsive to popular demands. But the so-called
First Republic fell prey to radicalism, instability, and coup attempts, and soon
foundered. Its final collapse in 1926 led to the emergence of a corporatist authoritarian
regime—known as the Estado Novo (New State)—under the personalistic leadership of
António de Oliveira Salazar. This non-democratic regime held power for decades.
However, it was badly hurt by its participation in colonial wars in the 1960s, and was
finally toppled in April 1974 in a revolution led by the armed forces. The April 1974
Revolution, also dubbed the Revolution of the Carnations, paved the way for a quick,
albeit rocky, transition to democracy. Elections for a national constituent assembly
were held in 1975. In 1976 a democratic Constitution was promulgated, and free and
fair presidential and parliamentary elections were held. Thus, 1976 is the first year of
actual parliamentary democracy in Portugal’s history.
Any analysis of the chain of delegation and accountability in Portugal’s democracy
must consider the political role played by the Chief of State. The 1976 Constitution
established a popularly elected President of the Republic, and endowed this office with
extensive powers over government formation and termination, the survival of the
legislative assembly, and policy-making. Constitutional reforms enacted in 1982
The author acknowledges the support of a University of Umeå grant that financed the fieldwork in Portugal
on which this study is based. The author is grateful to Mário Bacalhau, Cristina Leston-Bandeira, Arend
Lijphart, Pedro Coutinho Magalhães, Antonio Ortiz-Mena, Matthew S. Shugart, and the book editors for
their useful comments and suggestions, and to Manuel Braga da Cruz, Margarida Guadalpi, and José M.
Magone for kindly sharing their data on Portugal’s political institutions.
Portugal 553
reduced some of the powers of the President and enhanced those of the Parliament
and the cabinet. Despite this, the Chief of State has retained an important role in
policy-making.
The law-making process in Portugal displays the following features. Individual MPs
and the cabinet can initiate bills in the unicameral Assembleia da República (Assembly of
the Republic). In the Portuguese constitutional jargon, the bills sponsored by MPs are
called Projetos de Lei (Projects of Law), and those presented by the cabinet are denomi-
nated as Propostas de Lei (Proposals of Law). The approval of any bill requires a quorum
of 50 per cent plus 1 of all MPs and the support of a majority of those voting. The
budget process is identical to that of law-making. As for constitutional amendments,
they can only be proposed by MPs, and require a two-thirds majority of all MPs to be
enacted.
Additionally, in Portugal the Constitution endows the cabinet with the power to
issue decrees, which enables the latter to unilaterally legislate in policy areas that are
not constitutionally under the jurisdiction of the Assembly.1
A unique feature of the 1976 Portuguese Constitution relative to its Western
European counterparts was that it granted the President the power to veto both bills
approved by the Assembly of the Republic and decrees of the cabinet (Articles 139, 277,
and 278). A President’s veto could be overridden by an absolute majority of the members
of the Assembly of the Republic (Article 139: 2). However, paragraph 3 of the same
Article stated that a two-thirds majority of those present was required to override
a president’s veto on the following matters: (1) boundaries between the public, private,
and cooperative sectors; (2) foreign affairs; and (3) rules governing elections. Moreover,
the President had an absolute veto over administrative decrees issued by the cabinet
(Articles 138: 2, 278). Also, the president had the power to exert what the Constitution
calls pre-emptive control of constitutionality. After consulting with the Council of the
Revolution2 on the constitutionality of a legal document, the President could veto
international treaties, cabinet decrees, and laws passed by the Assembly (Article 277). If
the Council of the Revolution ruled that a legal document was unconstitutional, then
the President was required to veto it (Article 278).
The 1982 constitutional revision actually enhanced the President’s prerogatives with
regard to his or her veto right. After the revision a two-thirds majority is required to
override a President’s veto in a larger number of policy areas than was the case under
the 1976 Constitution. The new areas were state of siege and state of emergency,
1
The Constitution, Article 164, defines a wide set of policy areas over which Parliament has absolute
legislative jurisdiction, such as basic education, the referendum regime, organization of the Constitutional
Tribunal, organization of national defence, state of siege and state of emergency regimes, organization
of political parties and associations, the electoral system of local legislative assemblies, and appointment of
members to European Union (EU) organs. It should be noted that the government has an exclusive right to
introduce legislation on its internal organization and functioning.
2
The Council of the Revolution institutionalized the role of the military as guardians of the democrati-
zation process. This role was enshrined in the 1976 Constitution, which, in Article 142, determined that the
Council have the functions of advising the President and ensuring the regular functioning of the democratic
institutions faithfully in the spirit of the 1974 Revolution. The Council was presided over by the President,
and was composed of the PM and high-ranking military officers.
554 Octavio Amorim Neto
organization of national defence, organization of the armed forces, and organization of
the Constitutional Tribunal (Article 139: 3).3 The revision also strengthened the
President’s prerogatives with regard to the pre-emptive control of constitutionality.
With the elimination of the Council of the Revolution, from 1983 on the pre-emptive
control has been performed by the Constitutional Tribunal upon a request by the
President (Article 278).
In acknowledgement of the fact that Portugal deviates from a typical parliamentary
democracy because of the role played by the President, Portugal experts and compar-
ativists have aptly classified the country’s system of government as semi-presidential
(Duverger 1980; Cruz 1994; Frain 1995; Sartori 1997) or premier-presidential (Moreira
1989; Shugart and Carey 1992).4 This system has helped Portugal consolidate its young
democracy because it proved to be flexible enough to respond quickly to challenges
posed by an unstable and ideologically polarized multiparty system in the early phase
of democratization. When the parties finally achieved a stable pattern of interaction
among themselves and with voters in the mid 1980s, Portugal’s democracy managed
not only to become institutionalized but also to promote high rates of economic
growth and an unprecedented process of social change and modernization. Such
achievements are evidence that Portugal’s parliamentary chain of delegation, despite
all the complexity created by an elected President, seems to be responding well to the
demands of voters. This is what will be argued below.
P O L I T I C A L PA RT I E S
Despite the fact that political parties did not play any part in the governing of Portugal
during three-quarters of the twentieth century, since the April 1974 Revolution
Portuguese elites have managed to successfully build a stable system of national polit-
ical parties. According to Bruneau (1997: 1), ‘In no other country, … , have political
parties played anywhere near as important a role in the democratization [process] as in
Portugal’. Such a role is enshrined in the Constitution. Article 153 grants political
parties monopoly over the presentation of candidates for parliamentary elections, and
Article 51 prohibits the organization of parties with regional scope and goals.
Since 1976 four parties have emerged as the key contenders for the popular vote:
the Partido Social Democrata (PSD—Social Democratic Party), the Partido Socialista
(PS—Socialist Party), Centro Democrático e Social (CDS—Social Democratic Centre
Party), and the Partido Comunista Português (PCP—Portuguese Communist Party). Four
parties have been members of at least one cabinet: the PSD, PS, CDS, and the tiny
Partido Popular Monárquico (PPM—Popular Monarchy Party).
The PSD was founded right after the April 1974 revolution as the representative of
the Right in the new regime. The PSD drew support from a variety of social groups
3
The role and composition of the Constitutional Tribunal will be explained in the section on External
Constraints.
4
A semi-presidential or premier-presidential system has the following attributes: (1) a president elected by
popular vote, (2) endowed with considerable powers, and (3) a PM and a cabinet accountable to Parliament
and invested with executive functions.
Portugal 555
and political forces formerly associated with the Estado Novo, including liberals, univer-
sity student movements, masons, and Catholic associations (Frain 1997: 78). In order to
survive in the radical post-revolutionary environment, its leaders were forced to shift
the party’s programme to the left of the median position of party members and sym-
pathisers (Frain 1997: 83). The contradiction between the conservative political lean-
ings of the party’s core constituencies and the more leftist tone its leaders have adopted
in order to increase the party’s electoral appeal has left its mark on the PSD. According
to Frain (1997: 83), the PSD is not ideologically unified, and therefore it is highly fac-
tionalized: ‘…liberals, technocrats, populists, social democrats all find a place in the
PSD’. Factionalism within the party has been further worsened by the strong person-
alities of its leaders—a characteristic trait of Portuguese political culture—and the
clashes stemming from this.
Despite the many internal conflicts that resulted from factionalism, PSD leaders
were wise enough to devise a decentralized and democratic organizational structure
that could accommodate the diversity of its elites and at the same time ensure strong
party identification at the mass level (Frain 1997: 85). Moreover, the fact that the PSD
was in the opposition in the first years of the new democratic regime further helped to
hold its factions together (the PSD entered the cabinet for the first time in 1980). In
addition, when the PSD formed a minority single-party government in 1985 the desire
to hold on to power helped to constrain and appease its factions. The party’s victories
in the 1987 and 1991 elections—in which it won absolute majorities—had a similar
positive effect.
The PSD joined seven of the the fifteen cabinets formed between 1976 and 2000—
once as a coalition partner, three times as the member of an electoral alliance, and
three times as the leader of single-party minority or majority governments. It entered the
cabinet for the first time in January 1980, as the major partner in a winning electoral
alliance with the CDS and the PPM—the so-called Aliança Democrática (Democratic
Alliance). This alliance formed two additional cabinets during the 1980–3 legislature
before being dissolved after the 1983 elections. It was replaced by a government coali-
tion, the Bloco Central (Central Block), composed of the two largest parties, the PS and
PSD. In the 1985 elections the PSD won 35.2 per cent of the seats in the Assembly of
the Republic, which made it the country’s largest parliamentary party. The PSD then
formed a single-party minority government. In the 1987 election the PSD achieved its
greatest electoral success, winning an absolute majority of 59.2 per cent of the seats in
parliament. This allowed it to form the first single-party majority government in
Portugal’s democratic history. In 1991 the PSD repeated its strong electoral perform-
ance, managing to retain an absolute majority of seats in the Assembly (58.7 per cent).
It therefore continued to single-handedly run the government. In the 1995 election the
PSD finally lost its majority, and had to give up government power to the PS.
The PS was founded in 1973 by a group of exiled Portuguese Socialists in Germany.
Its first programme called for a classless society and for the nationalization of key
industries (Sablosky 1997: 59). This ideological programme made the PS the pivotal
party in post-revolutionary Portugal and during the democratization period. Like the
PSD, factionalism and clashes among its leaders have also plagued the PS. However,
unlike the PSD, the PS has a vertical and centralized organizational structure. The party
556 Octavio Amorim Neto
is controlled from the top, which makes it difficult to achieve efficient coordination of
its leaders, factions, and rank-and-file. According to Sablosky (1997: 62), over time the
complicated organizational structure proved a weakness as the PS failed to develop a
strong cadre of middle level managers.’
A number of factors led the PS to moderate its initial ideological radicalism
and move towards the centre of the ideological spectrum. These include the party’s
participation in the first democratic governments, its decision to ‘Europeanize’ in the
late 1970s, and electoral defeat from the mid-1980s to the early 1990s. Its ideological
transformation, combined with an economic downturn in the mid-1990s, contributed
to the PS’s increasing popularity and made it possible for the party to win both the 1995
and 1999 parliamentary elections and the 1996 presidential race.
The PS was the dominant party during the first years of the new democratic regime.
It formed a single-party minority cabinet in 1976 and a coalition government with the
CDS in 1978. This coalition soon collapsed. However, because of the highly polarized
atmosphere of the late 1970s the parliamentary parties were unable to agree on the
composition of a new government. The parties’ failure to agree led to a period of
sixteen months of undisguised presidential rule, during which President Ramalho Eanes
appointed three non-partisan cabinets. The period of presidential rule was followed by
the liberal governments of the Democratic Alliance. The PS returned to power in 1983
in a government coalition with the PSD. The latter’s rise to dominance in the mid-1980s
deprived the PS of government power for a period of 10 years. The party finally
regained power after the 1995 elections, this time leading a single-party near majority
government. After the 1999 elections, the Socialists were returned to power, again
forming a single-party near majority cabinet.
The CDS5 was founded in 1974 by individuals closely linked to some of the elites of
the old regime, particularly elites in the universities, public administration, and business.
Its leaders placed the party to the right of the PSD (Robinson 1996). During its first
years the CDS was basically a party dominated by notables. Over time, however, it was
able to develop a decentralized party structure (Frain 1997: 86). The CDS has been
Portugal’s fourth largest party and participated in one coalition government (in 1978
with the PS), and in the three Democratic Alliance cabinets between 1980 and 1982.
D E L E GAT I O N F RO M VOT E R S TO M P S
As mandated by Article 149 of the Constitution, MPs are elected in twenty multi-
member districts on the basis of proportional representation and the D’Hondt method
of seat distribution for a four-year term.6 Article 151 states that candidates are put forth
by parties, and Article 152 that MPs represent the country and not their districts. The
1979 Electoral Law, Article 15, mandates that parties are to present their candidates on
closed lists.
5
In January 1993 the CDS changed its label to CDS–PP (Popular Party).
6
Eighteen districts are located in continental Portugal. The Madeira and Açores Islands have one district
each, and there are two districts for Portuguese emigrants, one for those in Europe, and the other for those
outside the continent.
Portugal 557
According to the comparative literature (Katz 1980; Carey and Shugart 1995), closed
list-proportional representation electoral races are mainly won by party organizations,
not candidates. This is because the fate of individual candidates depends on the
electoral success of their party label and their place on the party list. It is presumed that
the better the reputation the candidate has with the party leadership, the higher on the
list he or she will be placed. It is further believed that candidates’ dependence on the
electoral performance of their party lists will both strengthen the party leadership and
generate strong incentives for individual candidates to cultivate a party vote.
The Portuguese case largely corroborates the comparative literature. By all
accounts, the national party leadership is the most important actor in the selection of
candidates for parliamentary elections (Cruz 1995: 175–219; Frain 1997; Freire 1997;
Sablosky 1997). There is no institutionalized form of involvement of either voters or
ordinary party members in candidate selection. Moreover, there are no term limits nor
a mandatory retirement age for MPs. These aspects further enhance the party elite’s
control over party activities. However, the ability of the national leaderships to shape
electoral lists varies across parties. This is partially explained by the different organiza-
tional structures of the parties, as indicated in the previous section. Because left parties
are more centralized, their national leaders are stronger than their centre and right
counterparts. Compelling evidence of such variation is provided by Freire (1997: 42). In
the fourth Parliament elected after the establishment of democracy (1985–7), 23.9 per
cent of MPs elected from the PS and 17.1 per cent from the PCP occupied posts on the
national committee of their parties. Comparative figures for the PSD and CDS were 6.9
per cent and 12.2 per cent, respectively.
Voter surveys seem to confirm the hypothesis that closed list-proportional represen-
tation fosters a party vote. Opinion polls show that Portuguese voters have an increas-
ing attachment to parties. In a 1979 national survey, 32.8 per cent of respondents said
that parties were very necessary. This same question was put to voters in 1993, and this
time an impressive 66.8 per cent of respondents said that parties were very necessary
(Bacalhau 1997: 121). In 1993 voters were also asked why they voted for a party.
The most frequent response, given by 81.4 per cent of respondents, was ‘I vote for the
party that is closest to my ideas’—which indicates a strong ideological commitment to
parties. After reviewing extensive data from four public opinion polls over fifteen years,
Bacalhau (1997: 136) concludes that Portuguese parties ‘…have emerged as the key
actors in politics.’
In addition to candidate selection mechanisms and electoral incentives, institutional
aspects concerning electoral competition, party finance, and legislative organization
also contribute to the strength of parties in Portugal. Legislative campaigns tend to be
rather short. They lasted twenty days in the elections held between 1976 and 1991. In
the election in 1995 the campaign period was only thirteen days. It is safe to say that the
shorter the campaign period, the heavier the candidates’ reliance on their party labels to
win electoral races. Moreover, parliamentary parties are endowed with substantial pub-
lic resources. They have free broadcasting time allocated in proportion to their seat
shares, whereas non-parliamentary parties receive a minimum quota of time. Since 1988
parliamentary parties have started to receive public financing. Each receives annually
558 Octavio Amorim Neto
an amount equivalent to 1/225 of the national monthly minimum salary for each vote
obtained in the last election for the Assembly of the Republic. For legislative campaigns
an amount equivalent to 2,500 national monthly salaries are given to parties which have
fielded candidates for at least 51 per cent of the seats, and have won at least 2 per cent
thereof. Twenty per cent of that amount are equally distributed, and the remaining
80 per cent are divided according to vote share. According to Katz and Mair (1995), by
entrenching the parliamentary parties in the structure of the state, the grant of such
public resources to the former helps to cartelize the party system. Portugal can be
safely listed as another evidence of such a phenomenon.
Finally, as argued by Leston-Bandeira (1998: 146), individual MPs are endowed with
meager resources in the Assembly of the Republic to do their legislative work. An elo-
quent indicator of this provided by this author is that since 1991 a typical MP has to
share an office with four other colleagues. Consequently, as MPs are deprived of the
means to independently operate in Parliament in an efficient manner, the role of party
leaderships is considerably strengthened in the organization of legislative activities.
Hence, party discipline is very tight in the Assembly. Significantly, all parties have
enshrined party discipline in voting decisions in their statutes (Cruz 1995: 193).
In short, political parties are the central agents of Portuguese voters in the electoral
and legislative arenas. However, as stressed by some authors (Bruneau and Macleod 1986:
146–64; Cruz 1995: 269–96; Leston-Bandeira 1998), the flip-side of the strength and
centrality of nationally-oriented political parties is that MPs have weak links and scarce
contact with voters. As MPs are elected through constituencies, such pattern of
relationship generates agency loss to voters. Yet, given that Portugal is a small and
homogeneous country, this loss can be said to be rather small.
D E L E GAT I O N F RO M P R E S I D E N T TO C A B I N E T A N D
F RO M PA R L I A M E N T TO C A B I N E T
In Portugal’s system of government the national electorate designates two agents to
whom the cabinet is accountable: the President and the Assembly of the Republic.
Because the chain of delegation and accountability in this regime is more complex than
that of a typical parliamentary system, delegation from the President to the cabinet will
be described separately from that of the Parliament to the cabinet. In addition, the insti-
tutional rules regulating the relations among the Chief of State, the Parliament, and
the cabinet have undergone important changes since the promulgation of the 1976
Constitution, particularly after the 1982 constitutional reforms. This also speaks in
favour of separate subsections for these two delegation–accountability chains.
Prime Minister (PM) was appointed by the President after consultations with the
Council of the Revolution (which the President himself presided over) and the parlia-
mentary parties, and bearing in mind the electoral results. Cabinet members were
appointed by the President at the recommendation of the PM. Articles 193 and 194
established that the cabinet and the PM were politically responsible to the President.
Moreover, the President could dissolve the Assembly of the Republic if the Council of
the Revolution supported such an action. However, there were restrictions on these pow-
ers to dissolve: the Assembly could not be dissolved under a state of siege or a state of
emergency, and during an exceptional extension of the President’s term of office.
There were also automatic constitutional provisions on parliamentary dissolution in
case Parliament voted down a confidence motion or cast a no-confidence vote on the
third government formed in a single legislature.
Endowing a popularly elected Chief of State with such an array of formal powers in
the context of the instability that characterized party politics in Portugal in the first
years of the democratic regime had a straightforward consequence: the President
became a major player in domestic politics. Portugal’s first democratic President was
General Ramalho Eanes, the hero of a counter-coup staged in 1975 against a commu-
nist uprising. This counter-coup decisively smoothed the path from revolution to
democracy. A forceful character, General Eanes assumed the presidency in 1976 intent
on making himself heard in the governance of the country.
Additionally, as stressed by Sousa (1983: 14), Article 190—which stated that the
President should consult with the Council of the Revolution and parties in the selec-
tion of a government—was not binding, which gave Eanes greater discretion in gov-
ernment formation. In fact, he had so much discretion that after the failure of two
PS-led governments and disagreements among the parties over the composition of
a new government, between 1978 and 1979 he appointed three non-partisan cabinets,
staffed by independents. This show of presidential autonomy was obviously opposed
by the parties, and gave rise to the Democratic Alliance. The Democratic Alliance gov-
ernment managed to restrain Eanes somewhat. However, in order to place strict limits
on Eanes’s actions and amend the proto-socialist economic articles of the Constitution
with which he was identified, a constitutional revision was necessary (Bruneau and
Macleod 1986: 120).
On 30 September 1982 a revised Constitution was promulgated. The revision abolished
the heavily militarized Council of the Revolution and replaced it with a ‘civilianized’
Council of State. Since this constitutional amendment the President’s power has been
reduced in many important respects. Now the President can only dismiss the
government when such a measure is essential to ensure the ‘regular functioning of the
democratic institutions’, and after consulting with the Council of the State. Furthermore,
the Assembly of the Republic cannot be dissolved during the first six months after the
election of a President nor during the last six months of a President’s term. The President
is explicitly prohibited from exercising pocket vetoes, that is to say, he or she cannot refuse
to sign the promulgation of a bill approved by the Assembly (Article 139: 1,4). This means
that if the Chief of State dislikes a bill, he or she must bear the cost of explicitly issuing
a veto. In addition, presidential elections are not to be held within ninety days of a
560 Octavio Amorim Neto
parliamentary election (Article 128). The original Constitution required at least sixty days
between presidential and parliamentary elections.
It was noted in the introduction of this chapter that the president’s power was actu-
ally enhanced by the 1982 revision with regard to his veto right. Since then the veto has
been the touchstone of presidential power in Portugal, and should be understood as
a kind of ex post mechanism to control the cabinet and the Parliament. No other pop-
ularly elected President in Western Europe has this kind of reactive power, not even the
powerful French Chief of State. Furthermore, Portuguese Presidents, unlike their
Austrian and Icelandic counterparts, have not abdicated their powers. For example,
President Mário Soares (1985–95), a Socialist, vetoed key policy initiatives of the cen-
trist cabinets led by PM Cavaco Silva, who Soares ‘cohabited’ with during his two terms
in office (Cruz 1994: 255–6; Frain 1995: 668; Magone 1997: 42). By using his veto
powers—which he did particularly frequently during his second term—Soares was able
to push Cavaco Silva’s policy initiatives closer to the preferences of the median voter in
the presidential election, which leaned to the left. Thus, Portuguese Presidents must be
seen as more than figureheads. They actually affect government policy by exercising
their formal powers and political clout. During situations of ‘cohabitation’, Portuguese
Presidents are stronger than their French counterparts by virtue of their veto power.
Mário Soares was certainly more influential in domestic policy-making in 1985–95 than
was Mitterrand in 1986–8 and 1993–5 and than Chirac has been since 1997. On the
other hand, when the President and PM come from the same party, French Presidents
are definitely stronger than their Portuguese colleagues. The current Portuguese
President Jorge Sampaio and former Prime Minister Antonio Guterres are both
Socialists, and because of this Sampaio kept a lower profile than his predecessor, Mário
Soares. In sum, despite the 1982 Constitutional revision, Presidents have managed to
retain an important policy-making role, albeit a negative one, mainly due to their
ability to block the governmental agenda through the use of the veto.
Assembly vote against it. The cabinet also has the right to request a vote on a motion
of confidence, and an absolute majority of all MPs is needed for approval. With the
exception of the first and third governments led by Mário Soares and the first Guterres
cabinet, the opposition has always proposed the rejection of the programme of incoming
governments. Yet it has managed to reject only one, that of PM Nobre da Costa, which
led to his dismissal by President Eanes. And the only cabinet that fell because of a failed
motion of confidence was precisely the first Mário Soares cabinet.8
The practice of requesting a vote to reject the government’s programme after the
appointment of a PM can be considered as a strong convention in Portugal. Moreover,
as the rejection of a programme can lead to the dismissal of the cabinet, Bergman
(1993: 58) defines such convention as a kind of informal negative investiture vote. For
this author, ‘according to the negative rules, the onus is not on the government to prove
that it is supported by the parliament. Rather it is left to the parliament to prove that
the government is not tolerated’ (Bergman 1993: 57). This is precisely the nature of
Portugal’s vote to reject the government’s programme.
In addition, students of Portuguese politics agree that those negative rules facilitated
the formation of minority and presidential cabinets in the 1976–87 period (Sousa 1983;
Moreira 1989; Cruz and Antunes 1990). However, these same cabinets had difficulty
passing legislation because, obviously, parliamentary majorities could not be put
together without the cooperation of opposition parties. In addition, the factionalism
that plagued the largest parties in the early years of the new democratic regime affected
governments’ ability to make policy. Of course, with the emergence of majority
(or near majority) single-party governments in 1987 these difficulties disappeared.
As regards the Parliament’s accountability mechanisms vis-à-vis the cabinet, two
periods can be identified, between 1976–87 and 1987–95. In the first period—which was
characterized by fragile, majority-coalition cabinets, minority cabinets, and presidential
governments—parliamentary actors held the executive accountable by means of an
ex ante non-institutional mechanism: that is, coalition bargaining. In other words, given
that during this period no cabinet could count on a stable legislative majority, parlia-
mentary party leaders and MPs had plenty of freedom to trade support for the cabinet
for cabinet support of opposition policy preferences.
In the second phase—in which single-party majority cabinets have dominated—
there has been an impressive growth in the use of some ex post, formal, institutional
mechanisms to hold the government accountable. The works of Cruz (1995: 175–219)
and Cruz and Antunes (1990), and Leston-Bandeira (1995, 1996, 1998), all show
a marked increase in the use of interpellations by the opposition. In the era of single-
party majority cabinets the bargaining power of opposition parties simply evaporated.
Their natural response to this new strategic situation was to activate the institutional
mechanisms at their disposal. The Constitution stipulates that each parliamentary
party may request two interpellations per legislative session. In the 1987–91 and
8
For comprehensive data on the record of confidence and no-confidence motions, and proposals for rejec-
tion of governmental programmes, see Matos (1992).
562 Octavio Amorim Neto
1991–95 legislatures, both under majority single-party cabinets, the average number of
interpellations per legislative session was 5.5 and 6.3, respectively, as compared to 1.3,
3.0, and 3.5 in the first three democractic legislatures.
Another institutional mechanism that has been intensively used by the opposition in
the era of single-party majority cabinets is the committees of inquiry. According to
Leston-Bandeira,
the committees of inquiry have become the most visible and popular means of scrutinising gov-
ernment, in particular the discussion on the proposal to establish a committee. In fact, as António
Vitorino has pointed out, the proposal for an inquiry often constitutes a “formal charge against
the government, likely to put forward conclusions”, transforming its discussion into one of the
“most lively and polemic parliamentary debates”. (Leston-Bandeira 1998: 156)
In the 1991–5 legislature the average number of proposals for committees of inquiry
reached its peak, 8.3. It should also be noted that the lowest averages correspond to the
legislatures during most of which minority or presidential cabinets were in office:
0.5 in 1976–80, and 4.0 in 1985–7.
Yet, as also observed by Leston-Bandeira (1998: 155), the use of a classical parlia-
mentary control instrument—written questions—has declined with the rise of single-
party majority cabinets. Under the 1985–7 legislature, the average number of written
questions per legislative session peaked at 2,388.5, but decreased steadily in the next leg-
islatures. Leston-Bandeira contends that this trend is due not only to single-party
majorities but also to a rather low rate of answers.
A key problem concerning Portugal’s chain of delegation concerns how majority
parliamentary parties control their own single-party cabinets. As mentioned in the
introduction, the Constitution endows the cabinet with the power to issue decrees. For
obvious reasons, the cabinet has been prone to use this instrument of legislative
decision-making authority. Between 1977 and 1993, the total number of laws passed by
the Assembly of the Republic was 1,249 while the number of decrees issued by the
executive was 8,451 (Magalhães 1994: 19). According to Antunes (1988: 82), up to 1987
the Assembly frequently modified the decrees issued by the executive. However, from
1987 to 1991, that is, during the first single-party majority government of Cavaco Silva,
the proportion of executive decrees modified by Parliament plummeted to a low of
4 per cent, while no decree was rejected (Magalhães 1995: 98). This author interprets this
as indicative of a diminishing role for the Parliament in legislative decision-making.9 In
the language of the principal–agent model, Magalhães can be said to be arguing that the
Parliament has abdicated some of its law-making powers to the executive. However, this
is not necessarily true. It is only correct if those who hold cabinet posts are entirely at
odds with their party’s MPs in terms of policy preferences. If ministers and MPs have
similar policy preferences, which is very likely in a system in which the executive is
dependent on the confidence of Parliament, then the fact that few executive decrees are
modified or rejected can be seen as evidence that there is an efficient delegation
mechanism from the majority to the cabinet under single-party majority governments.
9
Leston-Bandeira (1998: 152–4) makes a similar point in her analysis of cabinet decrees in Portugal.
Portugal 563
Further evidence that such a mechanism exists is that under single-party majority
cabinets the Assembly of the Republic passed more laws delegating legislative author-
ity for the executive to speedily deal with relevant matters constitutionally falling under
the legislative jurisdiction of the Parliament than it did during the years of majority
coalition government and single-party minority governments (Magalhães 1995: 107;
Leston-Bandeira 1998: 152). According to the first author, during the 1987–91 legislature
and the first half of the 1991–5 one, one-third and one-half, respectively, of the relevant
laws enacted by the Assembly were Leis de Autorização Legislativa (authorization of pow-
ers bills). However, in passing such laws Parliament does not hand a carte blanche
to the cabinet. On the contrary, the Assembly defines the scope and time limit for
government action, thus establishing ex ante mechanisms of control. The efficiency of
legislative delegations is suggested by the following indicator: under the 1987–91 legis-
lature the passing of authorization of powers bills took on average eighty-three days,
while the average length of time required to pass a standard executive-initiated bills was
four months (Magalhães 1995: 107).
Finally, the cabinet is held accountable by means of moções de censura (no
confidence votes). Article 194 of the Constitution stipulates that at least one-quarter of
the membership of the Assembly and any parliamentary party can initiate a no confi-
dence vote on the cabinet’s execution of the governmental programme or on any mat-
ter of national interest. The approval of a no-confidence vote by an absolute majority
of all members of the Parliament leads to the dismissal of the cabinet (Article 195). The
first government led by Cavaco Silva (1985–7), a single-party minority one, was termi-
nated by a vote of no confidence initiated by the Partido Democrático Renovador (PRD—
Democratic Renewal Party), a centre-left party. The PS, which had been cooperating
with the government in Parliament, decided to join forces with the PRD in order to
avoid running the risk of losing voters to it. The no-confidence motion passed by a vote
of 134 votes to 108. Additionally, the Mota Pinto cabinet resigned in order to pre-empt
a no-confidence vote in Parliament.
Table 18.1 reports the behavioural record with regard to investiture votes, confi-
dence and no confidence votes, and early dissolution of Parliament.
Other than the motion that led to the resignation of the first Cavaco Silva cabinet,
no other motions of no confidence have been voted on in Parliament. A total of five
confidence votes have occurred since 1976, though only the first Soares cabinet has
been removed by a failed vote of confidence. Four cabinets have ended in early elec-
tions following the dissolution of Parliament, most recently in 1987, as of 2003.
D E L E GAT I O N W I T H I N C A B I N E T
The executive branch is the least studied organ of the Portuguese state during the
democratic period. There are only two academic works on this subject—an article by
Portas and Valente (1990) that examines the relationship between the PM and cabinet
ministers on the basis of interviews with twenty-three former ministers, and the first
volume of Amaral’s Curso de Direito Administrativo (A Course in Administrative Law),
Table 18.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Portugal 1976–2000 a
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv) Confidence votes (cv) Did this Who
formation cabinet end dissolved
No. of Voting results of successful ivc No. Cabinet Cabinet No. of cv Cabinet with an Parliament in
unsuccessful (1:pro gov., 2:against gov., of removed resigned under removed early this case?b
investiture 3:abstentions, and 4:other) ncv by ncv to spec. inst. by failed election?
votes 1 2 3 4 pre-empt mech. cv
10
The Gabinete da Presidência do Conselho de Ministros (office of the presidency of the cabinet) provides the
PM with support services for the coordination of interdepartmental affairs, and is the umbrella of many State
Secretariats that deal with specific policy issues.
Portugal 567
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
All civilian ministries in Portugal have the same structure, which tends to be rigid and
highly hierarchical. Each ministry has a general director, who is the head of the
ministry’s civil service and is jointly appointed by the PM and the relevant minister.
General directors appoint service directors and are their immediate superiors. Below
this are the chiefs of divisions, who are appointed by their immediate superiors, the
service directors. Finally, the chiefs of divisions appoint the bureau chiefs.
General directors are the highest-ranking civil servants, and are accountable to the
cabinet via the minister in charge of their respective ministry. Service directors, chiefs
of division, and chiefs of bureaus do not report directly to either the cabinet or the
minister. They are accountable to their immediate superiors. That is to say, a service
director is only accountable to the general director; likewise, a chief of division is only
accountable to the service director; and a chief of bureau to the chief of division. Thus,
on paper, Portugal has a completely vertical chain of command within bureaucracies.
According to Portas and Valente (1990), the relationship between ministers and
general directors tends to be rather conflictive. In interviews, some former ministers
report that general directors are a ‘major obstacle’ or a ‘conservative force’. They also
complain about the poor quality of bureaucrats’ work and the great amount of time
general directors wasted on various labour problems. Although Portas and Valente
only interviewed a relatively small number of former ministers, their findings largely
corroborate the conventional wisdom—of scholars but also of ‘the man on the
street’—about the limited competence of the Portugal civil service. This is no surprise
given the powerful legacy of the clientelistic and patrimonial practices prevalent in
the old authoritarian regime and the fact that the Portuguese civil service in general is
not highly skilled (Magone 1997: 128).
Although there have been great improvements in public administration as a result of
modernization prompted by high economic growth and Portugal’s entry into the
European Union (EU), the country’s bureaucracy still suffers from the impact of
decades of unrestrained patrimonialism. From the point view of the government, pat-
rimonial practices and legacies are a major source of agency loss because they mean
that bureaucrats respond to the demands of people other than their constitutional
principals.
E X T E R NA L C O N S T R A I N T S
The key external constraint on Portugal’s parliamentary chain of delegation and
accountability is the EU. The process of democratic consolidation in Portugal was inti-
mately linked to the process of integrating the country with the rest of Europe. In the
words of Magone (1996: 152), ‘The accession to the EC was seen as the only way to
consolidate the very young Portuguese democracy.’ Portugal applied for European
Community membership in March 1978, and finally became a member in 1986. The
decision to join the EU began to have a big impact on domestic policy-making in the
late 1980s. Thus, not long after a democratic Parliament had begun to take root in
568 Octavio Amorim Neto
Portuguese political life, it was faced with the challenge of dealing with a supranational
institution. Magone (1996: 157) reports that by 1990 Portugal was already the most
successful EU country in terms of enacting national legislation to complete the Single
European Market.
Unsurprisingly, the Assembly of the Republic did not take the lead in the process of
European integration. Given how busy it was with the domestic side of the democratic
consolidation agenda, the Parliament was content to let the cabinet run the show in the
early phase of the integration process. Only in 1987 did it pass a law on parliamentary
monitoring of Portugal’s EU policies. This law was replaced by a new one in 1988, that
was, in turn, superseded by yet another in 1994. These laws all aimed at establishing
legal mechanisms of parliamentary control over the government’s EU policies. The
1988 law established a Committee on European Affairs, which was granted some
prerogatives to participate in decisions on EU matters. The 1994 version mandates that the
government must submit an annual report on EU matters to the Assembly, an ex post
mechanism of control. Overall, the three laws have helped to improve the Assembly’s
information-gathering capacity and enhance its influence over EU affairs. ‘Nonetheless,
the largest share of EU-related laws has been enacted by the cabinet via executive
decrees’ (Magone 1996: 155). Yet this should not be seen as evidence that Parliament
has abdicated its role in EU affairs. Rather, as argued in the section on Delegation from
Parliament to Cabinet, the use of executive decrees can be considered as an efficient
delegation mechanism from the majority to the cabinet.
The other significant external constraint on Portugal’s parliamentary chain of dele-
gation and accountability is the Constitutional Tribunal. It was created in 1982, and was
charged with the pre-emptive control of constitutionality. Upon a President’s request,
the Tribunal must, within twenty days, rule whether a law approved by the Assembly
or a cabinet decree is constitutional. Up to 1997 the Constitution stipulated that this
tribunal be composed of thirteen judges, ten elected by the Parliament and three
appointed by the ten Parliament-elected judges for a renewable six-year term of office.
In 1997 the term was increased to nine years but it is no longer renewable. Candidates
for the Parliament-elected judgeships must be sponsored by at least twenty-five MPs
and at most fifty.
The Constitutional Tribunal has been active in Portugal’s political life. Between April
1983 and May 1993 it ruled over twenty-seven requests made by Presidents, and twenty
were deemed unconstitutional (Magone 1997: 52). According to this author, the
Tribunal has reviewed important pieces of legislation approved by sizeable parliamentary
majorities, such as the anti-corruption law, the law on incompatibilities, and the Official
Secrets Act, all passed in 1993, that is, under the second majority single-party cabinet
led by Cavaco Silva. It should be noted though that judicial review in Portugal is not
based solely on judges’ technical concerns with the legality of parliamentary acts. After
extensively surveying all the votes cast by the Constitutional Tribunal members on the
forty-seven rulings passed by the court between April 1983 and March 1998, Magalhães
and Araújo (1998) conclude that there is a ‘party connection’ underlying those
votes. Their statistical analysis clearly shows that in their voting decisions the judges
are politically responsive to the agendas of the parties that have supported their
Portugal 569
appointment to the Tribunal. At any rate, it is safe to affirm that the Constitutional
Tribunal is a relevant check on parliamentary majorities in Portugal.
Finally, referendums have increasingly become a new formal external constraint in
Portugal. In 1991 a constitutional amendment introduced the referendum as a new pol-
icy-making procedure. According to the ordinary law that regulated it, referendums
could be proposed by either the cabinet or the Assembly on ‘relevant issues of national
interest’, except for constitutional amendments, budgetary, fiscal, and monetary poli-
cies, and policy areas over which Parliament had absolute legislative initiative (see foot-
note 2). Additionally, the law established that (1) no electoral quorum was required for
a referendum to be valid; (2) referendums were to be subject to judicial review before
they were voted on; and (3) referendum results were binding on the cabinet and
Parliament. However, between 1991 and 1997 no referendum was held.
In 1997 a constitutional amendment allowed citizens to propose a referendum
which, however, has to be approved by Parliament before it can be held. And in 1998 a
new law regulating the referendum regime was passed. This law was very similar to its
1991 version. The only major change it brought about was to stipulate that a turnout
of 50 per cent⫹1 of registered voters was required for a referendum decision to be
valid. In 1998 two referendums on very relevant issues were held: one on abortion
rights, proposed by the PSD, and the other on decentralization, sponsored by the PS.
Both were defeated. Thus, so far the referendum can be considered only a potential
external constraint on Portugal’s chain of delegation and accountability.
C O N C LU S I O N
Portugal’s blend of semi-presidentialism or premier-presidentialism with multipartism
has created a regime capable of generating a wide variety of governing solutions:
presidential, single-party minority, majority coalition, near majority single-party, and
single-party majority cabinets. On the one hand, this flexibility has helped consolidate
one of Western Europe’s youngest democracies. On the other, it is indicative of how
complex the country’s delegation process is. The 1982 constitutional revision simplified
the delegation process because, for all practical purposes, it eliminated the possibility
of presidential government. However, despite the reduction in presidential prerogative
brought about in 1982, Presidents have remained important in policy-making, particularly
due to their use of the presidential veto, which is a unique characteristic of Portugal’s
institutional design vis-à-vis its Western European partners.
The second cohabitation period between President Mário Soares and Prime Minister
Cavaco Silva showed how important the veto can be. Soares used it as an ex post mech-
anism to influence Cavaco Silva’s policy initiatives. This feature of Portuguese political
life means that the country’s system of government deviates considerably from the
unity principle of delegation and accountability in the ideal-type chain. The national
electorate selects two agents—the President of the Republic and the Assembly of the
Republic—both of whom have the institutional means necessary to respond to their
principal’s demands, although to differing degrees.
570 Octavio Amorim Neto
Has this characteristic of Portuguese democracy created agency problems for the
voters? It certainly had in the early 1980s, when at a certain point President Ramalho
Eanes proved to be too meddlesome in parliamentary politics. Not surprisingly, some
of the powers of the presidency were curtailed in the 1982 constitutional revision.
However, since then there have not been significant calls to further reduce presidential
power. Moreover, the presidency has remained a prestigious and coveted office. In this
sense, it is plausible to say that Portuguese voters seem to be now satisfied with their
two competing but unequal agents.
Finally, it should also be emphasized that the President has an absolute veto over
cabinet decrees and the right to request that the Constitutional Tribunal review bills
approved by Parliament. As for cabinet decrees, they make the Portuguese executive
very strong in comparison with most of its Western European counterparts by virtue
of its monopoly over legislative initiative in many policy areas. Some observers argue
that the legislative prerogatives of the cabinet have rendered the Assembly of the
Republic almost voiceless in key aspects of national policy-making. In this chapter it
was argued that this is not necessarily the case, and that parliamentary majorities have
been able to devise efficient ex ante and ex post mechanisms to hold the executive
accountable. Yet it should be conceded that this topic needs further theoretical and
empirical study. What is indisputable, though, is that the President’s absolute veto over
cabinet decrees constitutes an important constraint on the powerful Portuguese cabinet.
It should also be viewed as a creative institutional solution to a recurrent problem in
parliamentary regimes, namely, unchecked executives created by the fusion of
legislative and executive power in the cabinet. Likewise, the President-activated
Constitutional Tribunal can be considered as a sensible constraint on potentially
perverse consequences of single-party parliamentary majorities.
REFERENCES
Amaral, Diogo Freitas do (1992). Curso de Direito Administrativo, Vol. I. Coimbra: Livraria
Almedina.
Antunes, Miguel Lobo (1988). ‘A Assembleia da República e a Consolidação da Democracia em
Portugal’. Análise Social, 24: 77–95.
Bacalhau, Mário (1997). ‘The Political Party System in Portugal: Public Opinion Surveys and
Election Results’, in Thomas C. Bruneau (ed.), Political Parties and Democracy in Portugal:
Organizations, Elections, and Public Opinion. Boulder: Westview Press.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
Bruneau, Thomas C. (1997). ‘Introduction’, in Thomas C. Bruneau (ed.), Political Parties and
Democracy in Portugal: Organizations, Elections, and Public Opinion. Boulder: Westview Press.
—— and MacLeod, Alex (1986). Politics in Contemporary Portugal: Parties and the Consolidation of
Democracy. Boulder: Lynne Rienner Publishers.
Carey, John M. and Shugart, Matthew S. (1995). ‘Incentives to Cultivate a Personal Vote: A Rank
Ordering of Electoral Formulas’. Electoral Studies, 14: 417–39.
Cruz, Manuel Braga da (1994). ‘O Presidente da República na Génese e Evolução do Sistema de
Governo Português’. Análise Social, 29: 237–65.
Portugal 571
Cruz, Manuel Braga da (1995). Instituições Políticas e Processos Sociais. Lisbon: Bertrand Editora.
—— and Antunes, Miguel Lobo (1990). ‘Revolutionary Transition and Problems of
Parliamentary Institutionalization’, in Ulrike Liebert and Maurizio Cotta (eds.), Parliament and
Democratic Consolidation in Southern Europe: Greece, Italy, Portugal, Spain and Turkey. London:
Pinter.
Duverger, Maurice (1980). ‘A New Political System Model: Semi-Presidential Government’.
European Journal of Political Research, 8: 165–87.
Frain, Maritheresa (1995). ‘Relações entre o Presidente e o Primeiro-Ministro em Portugal:
1985–1995’. Análise Social, 30: 653–78.
—— (1997). ‘The Right in Portugal: The PSD and CDS/PP’, in Thomas C. Bruneau (ed.), Political
Parties and Democracy in Portugal: Organizations, Elections, and Public Opinion. Boulder:
Westview Press.
Freire, André (1997). ‘Lógicas de Recrutamento Político: Caracterização Sócio-Política dos
Parlamentares Eleitos entre a Constituinte de 1975 e as Legislativas de 1995’. Secretariado
Técnico dos Assuntos para o Processo Eleitoral, working paper.
Katz, Richard S. (1980). A Theory of Parties and Electoral Systems. Baltimore: Johns Hopkins
University Press.
—— and Mair, Peter (1995). ‘Changing Models of Party Organization and Party Democracy: The
Emergence of the Cartel Party’. Party Politics, 1: 5–28.
Leston-Bandeira, Cristina (1995). ‘Controlo Parlamentar na Assembleia da República: A Translação
de Poder da IV para a V Legislatura’. Legislação: Cadernos de Ciência da Legislação, 12: 121–51.
—— (1996). ‘O Impacto das Maiorias Absolutas na Actividade e na Imagem do Parlamento
Português’. Análise Social, 31: 151–81.
—— (1998). ‘Relationship between Parliament and Government in Portugal: An Expression of
the Maturation of the Political System’, in Philip Norton (ed.), Parliaments and Governments in
Western Europe. London: Frank Cass.
Magalhães, José (1994). ‘A Constituição e as suas Revisões, a Lei e a Justiça’, in António Reis (ed.),
Portugal: 20 anos de Democracia. Lisbon: Círculo de Leitores.
Magalhães, Pedro Coutinho (1995). ‘A Actividade Legislativa da Assembleia da República e o seu
Papel no Sistema Político’. Legislação: Cadernos de Ciência da Legislação, 12: 87–119.
—— and Araújo, António de (1998). ‘A Justiça entre o Direito e a Política: O Comportamento
Judicial no Tribunal Constitucional Português’. Análise Social, 33: 7–53.
Magone, José (1996). ‘The Portuguese Assembleia da República: Discovering Europe’, in Philip
Norton (ed.), National Parliaments and the European Union. London: Frank Cass.
—— (1997). European Portugal: The Difficult Road to Sustainable Democracy. New York: St. Martin’s
Press.
Matos, Luís Salgado (1992). ‘O Sistema Político Português e a Comunidade Européia’. Análise
Social, 27: 773–87.
Moreira, Adriano (1989). ‘O Regime: Presidencialismo do Primeiro-Ministro’, in Mário Baptista
Coelho (ed.), Portugal: O Sistema Político e Constitucional 1974–1987. Lisbon: Instituto de
Ciências Sociais da Universidade de Lisboa.
Portas, Paulo and Valente, Vasco Pulido (1990). ‘O Primeiro-Ministro: Estudo sobre o Poder
Executivo em Portugal’. Análise Social, 25: 333–49.
Robinson, Richard A. H. (1996). ‘Do CDS ao CDS-PP: O Partido do Centro Democrático Social
e o seu Papel na Política Portuguesa’. Análise Social, 31: 951–73.
Sablosky, Juliet Antunes (1997). ‘The Portuguese Socialist Party’, in Thomas C. Bruneau (ed.),
Political Parties and Democracy in Portugal: Organizations, Elections, and Public Opinion. Boulder:
Westview Press.
572 Octavio Amorim Neto
Sartori, Giovanni (1997). Comparative Constitutional Engineering: An Inquiry into Structures,
Incentives, and Outcomes, Rev. edn. New York: New York University Press.
Shugart, Matthew S. and Carey, John M. (1992). Presidents and Assemblies: Constitutional Design and
Electoral Dynamics. Cambridge: Cambridge University Press.
Sousa, Marcelo Rebelo de (1983). O Sistema de Governo Português Antes e Depois da Revisão
Constitucional. Lisbon: Cognitio.
19
I N T RO D U C T I O N : T H E BA S I C C O N S T I T U T I O NA L
F E AT U R E S O F T H E S PA N I S H P O L I T I C A L S YS T E M
I N A H I S TO R I C A L P E R S P E C T I V E
Article 1.3 of the 1978 Spanish Constitution proclaims that ‘The political form of the
Spanish State is the Parliamentary Monarchy’. Despite the fact that both the Spanish
monarchy and Parliament are multisecular institutions—the first dating back to the
times of the Visigothic settlers, the second to the Middle Ages—no such statement ever
appeared in any previous Spanish Constitution. This, despite the fact that Spain has
gone through almost a dozen Constitutions in the last two centuries, almost all of
which have attempted to accommodate both institutions.
The Constitution’s proclamation arguably puts Spain ‘at the forefront of other
monarchic States’ (Álvarez Conde 1990: 51), since it ‘clearly reveals a design to rationalize
a parliamentary regime, expressing through written law what in [other] European
monarchies has been the outcome of an evolutionary and consuetudinary process’
(Sánchez Agesta 1980: 211). Although Article 1.3 of the Constitution has undeniably
greatly clarified the role and relationship between the Crown and the Parliament (see Bar
Cendón 1979; Torres del Moral 1983), the novelty of the Constitution on this particular
point should not lead us to forget the long process of evolution that preceded it.
A Parliamentary form of government was introduced early in the process of modern
constitutional evolution of Spain. To overcome the tensions, disputes, and discontinuities
that existed under the 1812 Constitution, the 1834 Royal Statute introduced formulas for
collaboration between Parliament, cabinet, and the Crown. These were subsequently
developed by ordinary legislation and unsanctioned constitutional conventions. Thus,
after 1834 the Council of Ministers was formally recognized, the holding of both a cabi-
net appointment and a parliamentary seat became accepted practice, collaboration
between Parliament and the Crown developed, the principle that the cabinet cannot stay
in office without the Parliament’s confidence was accepted, and the vote of confidence
and the motion of censure were introduced (Tomás Villarroya 1982: 35). These develop-
ments were gradually codified in subsequent constitutional texts, as well as comple-
mented by other practices—the Speech of the Crown, the use of interpellations, the
574 Carlos Flores Juberías
selection of ministers from the majority party—that are also characteristic of
parliamentary government.
However, through the first decades of the twentieth century, Spanish parliament-
arism was to a large extent distorted by royal meddling, military influence, the absence
of numerous basic political liberties, weak party structures heavily influenced by local
party bosses or ‘caciques’, and by periodic and widely accepted rigging of parliamentary
elections. Hence, it was not until the short-lived second Republic that an authentic par-
liamentarism emerged, one based on a broad recognition of civil liberties, a well devel-
oped party system, and free and competitive elections. However, the 1931 Constitution,
which featured a strong Prime Minister (PM) alongside a powerful President and
allowed for a deeply divided Assembly, proved ‘unable to channel the aggregated impact
of an explosive political participation, a deteriorated economic situation, very strong
social tensions and the appeal to violence by significant minorities’ (Martínez Sospedra
1988: 51). It collapsed after only five years. During the long parenthesis of Francoist
authoritarianism, the Head of State enjoyed sweeping powers (Zafra Valverde 1973:
200). Hence, a true parliamentary form of government was delayed until 1978.
The introduction of a true parliamentary government was not the only innovation
of the 1978 Constitution. The proclamation of one of the broadest sets of individual,
political, and social rights in Western constitutionalism, the reintroduction of the
Crown and the Senate, a new attempt to create an efficient judicial review, and above
all devolution, became the hallmarks of the new Spanish democracy. The creation of
‘el Estado de las Autonomías’ (see Flores Juberías 1998), a heterogeneous set of nineteen
territorial entities with different degrees of self-government, albeit very similar institu-
tions, was a huge task which took a decade and a half. In fact, although the process of
community creation and institution building is now completed, the task of transferring
competence remains unfinished. This process has substantially changed the way in
which the country is governed. The regional chain of delegation and accountability
now runs parallel to the national one.
EU membership has also dramatically changed the Spanish political system (see
Muñoz Machado 1993). Since Spain joined the European Community in 1986, Spanish
policies have had to accommodate EU requirements in a wide range of fields, national
law has come under the jurisdiction of the European Court of Justice, and decision-
making on many issues has shifted from Madrid to Brussels. Difficulties involved in
scrutinizing decisions adopted in the many EU forums and in holding EU bureaucrats
accountable—the so-called ‘democratic deficit’—has provoked scepticism and distrust.
Many citizens feel that the chain of delegation goes well beyond their national institutions,
while accountability does not.
P O L I T I C A L PA RT I E S : T H E R A P I D C O N S O L I DAT I O N
O F VOT E R S ’ P R E F E R E N C E
The 1978 Constitution was the first in Spain’s history to include an explicit reference to
political parties. This was not unexpected since parties played a key role in the transition,
and other European constitutions had already paved the way for such a recognition
(see García Cotarelo 1985: 142–88). Hence, Article 6, a part of the almost unamendable
Spain 575
Preliminary Title (see Rodríguez Díaz 1989), proclaims that ‘Political parties express
democratic pluralism, assist in the formulation and manifestation of the popular
will, and are a basic instrument for political participation’. The article also states that
the creation and activities of political parties should be free so long as they obey the
Constitution and the law, and commands that ‘their internal structure and operation
must be democratic’.
The key role of parties throughout the transition and the idea—widely accepted—
that strong parties are a precondition for a stable democracy quickly transformed Spain
into a ‘party democracy’. It has even been argued that Spain has become a ‘party state’,
‘the structure, performance, and real disposition of which is conditioned by the party
system in a relatively autonomous way in relation to its formal juridical configuration’
(García-Pelayo 1986: 90).
Perhaps the most telling evidence regarding the preponderance of parties in Spain
lies in the fact that their legal regulation was, until the law 6/2002 on political parties
was passed, limited to the six articles that comprised the 4 December 1978 Law, a text
that left most basic aspects of party activity virtually unregulated and provided no
effective guarantee for the constitutional requirement of internal democracy (Cascajo
Castro 1992: 192–3).
The configuration of the Spanish party system took place at the 1977 elections, and
since then the 1982 and 1996 parliamentary elections have been, by far, the most decisive
events in its evolution. The June 1977 elections—the first competitive ones since 1936—
reduced the extravaganza of parties, fronts, alliances, and coalitions that emerged after
General Franco’s death (see Equipo de Estudios 1976) to more manageable propor-
tions. Adolfo Suárez’s Union of the Democratic Centre (UCD) won the election, while
the Spanish Socialist Worker’s Party (PSOE) led by Felipe González came a close sec-
ond. To the right, some prominent francoist ministers embraced the reformist strategy
and rallied behind Manuel Fraga inside the Popular Alliance (AP). Most of the radical
Left converged in the Spanish Communist Party (PCE). These two forces polled slightly
less than 10 per cent of the vote each.
The situation created by the 1977 elections was largely reproduced two years later in
the first post-constitutional elections, but changed dramatically in 1982 (see Penniman
and Mujal-León 1985) when the UCD disbanded after suffering a huge loss to the
PSOE. The UCD, a loose conglomerate of former francoist cadres and local bosses,
with a few moderate opposition figures, was deeply marred by factions. These factions
only accepted the leadership of Adolfo Suárez as a matter of convenience (Huneeus
1985). Hence, when the party was forced to transform its election platform into
policies on several key issues, its weak cohesion quickly eroded (see Attard Alonso 1983).
In contrast to this, the PSOE, which developed strong organizational links while in
opposition, was able to carry out an ambitious set of reforms once it took office. This
ability to act was made possible by the undisputed leadership of Felipe González, the
party’s comfortable majority in both houses of Parliament and its dominance in most
local and regional governments, and a favourable European environment (see García
Cotarelo 1992).
Popular support for González began to decline in the early 1990s when economic
performance began to falter and several corruption scandals erupted. In the meantime,
576 Carlos Flores Juberías
the former AP—now renamed the Popular Party (PP)—had transformed itself into a
more centrist group, dismissed most of the ‘Old Guard’, and reinforced its internal organ-
ization under the leadership of José María Aznar. The new PP forged comprom-ises
with some minor regionalist parties, courted the stronger nationalist forces in
Catalonia and the Basque Country, and increased pressure on the Socialist Party as part
of its strategy for driving González out of government.
After mustering only a plurality of seats in Congress in the 1993 elections (see Martínez
Cuadrado 1996: 149–72), the PSOE was forced to form a minority government with the
external support of the Catalan nationalists. In the regional and local elections of 1995,
the Socialists were badly beaten by the PP in most important cities and regional
Governments (see Martínez Cuadrado 1996: 295–328). Discredited by corruption, and
affected by Gonzalez’s announced retirement, the party lost the 1996 elections to the
PP, which took over the government.
However, throughout its first term in office the new PP government was able to con-
trol only a plurality of seats and therefore needed the support of the Catalonian
Convergence and Union (CiU) and (to a lesser extent) the Basque Nationalist Party
(PNV) and both parties are ideologically right of centre and defend the historical rights
of their respective territories. Both have been in power in their respective autonomous
governments since they were first established. It was not until the sweeping victory in
the 2000 elections that José María Aznar was able to govern with the comfortable back-
ing of an absolute majority of both houses of parliament.
Generally speaking, the Spanish party system can be characterized in terms of five
features (Martínez Sospedra 1996: 250–6):
1. Two basic cleavages of class and centre/periphery, but no longer any serious
religious division.
2. An authoritarian political culture which explains low party membership ratios,
negative views of politics—especially at the national level—and a preference for
a strong, paternalistic State.
3. A rather stable distribution of political preferences since 1977, with more than
four-fifths of the voters defining themselves as either centre-right, centre, or
centre-left, the latter usually being the largest segment (Montero 1995: 105).
4. Parties that abandoned their responsibilities for political socialization once democ-
racy appeared sufficiently consolidated and mass mobilization no longer necessary.
5. And, therefore, parties characterized by low levels of social penetration and party
identification, a party membership with more interest in paid jobs in public
administration than grassroots activity, and the development of elitist and vote-
seeking electoral strategies (Share 1999).
D E L E GAT I O N F RO M VOT E R S TO M P S : H OW A R E
T H E C O RT E S E L E C T E D, A N D H OW D O T H E Y P E R F O R M ?
Considering the huge concentration of power in a single person that characterized
Franco’s regime, creating a strong, popularly-elected executive capable of ruling even
against the wishes of Parliament was unthinkable during the transition. The 1978
Spain 577
their parties (Punset 1992: 125–6). In this interaction, the position of the representatives
in the Spanish Congress vis-à-vis their party leadership is rather weak. While deputies
cannot legally be forced to vote against their will or compelled to give up their seats if
they dissent (Santaolalla 1992), parties have nevertheless developed a series of practices
that greatly limit this independence of mind and put recall essentially in the hands of
their party leaders. These practices include requiring their candidates to fill in a resig-
nation document with a blank date and imposing fines on deputies who fail to vote
along party lines.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T:
‘ R AT I O NA L I Z E D PA R L I A M E N TA R I S M ’
Despite the fact that the Spanish political system is constitutionally defined as a
parliamentary monarchy, the Constitution had created a strong executive, one quite
able to become the central pole of the state (Alzaga 1978: 429). Álvarez Conde (1990:
487) even characterized the Prime Minister (or ‘Presidente del Gobierno’) as ‘the cen-
tral axis of our political regime’.
This latter descriptive assertion is justifiably made about the Prime Minister, and
not about the cabinet as a whole, since the position of the former is constitutionally
reinforced vis-à-vis that of the other cabinet members in three main ways. First, the PM
is the only member of the executive directly elected by Congress. He drafts the
government’s agenda and presents it to the legislature, and he plays the key role in the
formation and operation of the cabinet. The PM proposes the appointment and dis-
missal of ministers, directs the government’s action, and coordinates the functions of
its other members. He is the only member of the executive who is directly accountable
to Congress, and the only one authorized to dissolve the legislature. In addition, he is
usually the undisputed leader of the largest party in parliament, which means that in
addition to his constitutional powers he has a significant degree of political authority.
It could be well said that the chain of delegation links not Parliament to cabinet, but
Congress to PM in the first instance, and then PM to the cabinet.
The prominence of the PM can be traced, as mentioned above, to the process of
cabinet formation itself (see Revenga Sánchez 1988). The government as a whole is
constitutionally obliged to resign after general elections, whenever the PM loses a
confidence vote in Congress, or upon the resignation or death of the PM (Articles 101.1
and 114.1). It is the King who proposes a candidate for PM. However, the King must
consult representatives designated by the political groups represented in parliament, and
must make his proposal through the President of the Congress of Deputies (Article 99.1).
At a second stage, the proposed candidate is required to submit to the Congress of
Deputies the political program of the government he intends to form and to obtain the
confidence of an absolute majority of its members in a first vote, or a plurality in a
second vote held forty-eight hours later (Article 99.2-3) and with a majority of members in
attendance. If the two voting rounds fail to invest the proposed candidate, the King pro-
poses new candidates until one succeeds. If no candidate has obtained the confidence of
Congress within two months of the first investiture vote, the King is to dissolve both
580 Carlos Flores Juberías
houses and call new elections (Article 99.4-5). This is the only exception to Article 115.3
of the Constitution, which prohibits a second dissolution of Parliament before a year has
passed since the previous one. The constitutional provision is designed to provide stability
for newly constituted governments, not to allow endless post-election negotiations.
It has been a subject of occasional debate whether the King has any discretion in
proposing a candidate for PM. The matter certainly depends on the election outcome,
and in particular on whether there is clear—or at least an unmatched—support for
a particular candidate over all others. Since this has always been the case, there is no
evidence about what might happen in a situation in which politically divergent
alternatives appeared equally feasible. Anyway, in Spain ‘the existence of governments
lacking the support of the majority of the House is perfectly possible, since this has
been consciously sought by the makers of the constitution, in their efforts to promote
political stability’ (Bar Cendón 1985a: 182).
Congress can withdraw its confidence in the PM at any time, thus forcing his imme-
diate resignation. This can happen either on a motion from the House or on a PM’s
motion, because the Spanish Constitution authorizes both votes of confidence and
votes of no confidence, that is, motions of censure.
The basic characteristics of the vote of confidence are set out in Article 112, which
states that ‘The President of the Government, after deliberation by the Council of
Ministers, may pose before the Congress of Deputies the question of confidence on his
program or on a declaration of general policy. Confidence shall be taken as granted
when a plurality of the deputies vote for it’. Article 114.1 adds that if Congress denies
its confidence to the government, then the government is to submit its resignation
to the King, and a new PM will be designated pursuant to the provisions of Article 99.
As in other parliamentary systems, the vote of confidence is aimed at allowing the
government to confirm the confidence expressed in the investiture vote. In addition,
it also compels parliamentary parties to define their positions, especially in the case of
minority governments or cabinets supported by a loose coalition of parties.
Several aspects of this procedure deserve consideration. To begin with, the initiation
of the vote of confidence depends entirely on the PM, since the cabinet only has the
right to deliberate upon proposals from the PM, but not to veto them. This is obviously
derived from the fact that only the PM—and not the cabinet—receive Congress confid-
ence upon investiture. As regards the object of the vote, the Spanish procedure is dif-
ferent from the French and German models, both of which allow a vote of confidence
to be linked to approval of a government bill. Confidence is, in such cases, granted
or withdrawn on the basis of level of support for a programme or general policy
declaration submitted by the government. However, the programme or policies in
question do not automatically become law upon a positive vote, thus ensuring that the
confidence vote cannot become an extraordinary law-making procedure. Third, the
government only needs the support of a plurality of deputies to stay in office—not
an absolute majority. This feature is in keeping with the investiture vote, but probably
transforms the institution into ‘a mere formality, devoid of any interest’ (Torres
del Moral 1992: 231). Finally, the Spanish vote of confidence also contrasts with
those of other European states in that it does not provide grounds for parliamentary
Spain 581
dissolution. If the PM loses the confidence of Congress, he has no choice but immediate
resignation.
The Spanish regulation of the no-confidence vote closely follows the precedent of
the German Grundgesetz, although the consequences of failure are not exactly the
same. The German experience was familiar to Spanish constitution-makers, but
imitation is not the only explanation for its adoption. Vírgala Foruria (1988) has argued
that the 1977 party system made single party majority cabinets unlikely. Given this
expectation, the two largest parties (UCD and PSOE) preferred constitutional
rules including no-confidence vote procedures, favourable to the formation and
maintenance of single-party minority governments.
The vote of no-confidence must be initiated by at least one-tenth of the deputies. It
must be motivated and propose an alternative candidate for Prime Minister. The pro-
posal cannot be put to a vote until five days after its presentation, and alternative motions
may be presented during the first two days of this period. Adoption of a vote of no con-
fidence requires the support of an absolute majority of the deputies (Article 113.1-3) and
forces the entire government to resign. Since the alternative candidate proposed in the
vote of no-confidence is understood to have the confidence of the house, the King must
appoint him PM (Article 114.2). If the proposal is unsuccessful, its signatories cannot
present another one during the same parliamentary session (Article 113.4).
One of the most criticized feature of the constitutional regulation of the no-
confidence vote (e.g. see Vírgala Foruria 1988: 242–5) is the possibility of introducing
alternative motions—that is, motions proposing alternative candidates—immediately
after a first motion has been proposed. Such a plurality of candidates might complicate
deliberations and make it more difficult for the opposition to agree upon a joint
strategy. In reality, alternative motions are unlikely. It is true that the number of
deputies whose signature is required to propose such a motion is not very high in
comparison to the number required for other regulations. Yet, given the configuration
of the Spanish party system, only the major opposition parties—PSOE in 1980, the
Popular Alliace in 1987—have actually been able to use this mechanism.
The most controversial regulation governing the no-confidence vote is found not in
the Constitution, but rather in Congress’s Standing Orders. It deals with the debate that
precedes the vote on a no-confidence motion. The regulation stipulates that the debate
is opened by one of the signatories of the motion, which may speak as long as he/she
likes. The alternative candidate speaks second, again without any time limit. After
a pause, representatives of those parliamentary groups willing to intervene are given
a chance to speak. These procedural rules mean that the sitting PM plays an entirely
secondary role in the parliamentary process to the alternative candidate.
Table 19.1 reports the behavioural record with respect to investiture votes, votes of
confidence and no-confidence, and parliamentary dissolution.
The only case of an unsuccessful investiture vote occurred before the Calvo-Sotelo
cabinet in 1981. Parliament twice voted on motions of no confidence, and twice voted
on motions of confidence, but in no instance was a cabinet removed by such a vote. Six
cabinets ended in early elections following the dissolution of parliament by the PM, for
example the González IV Cabinet in 1996.
Table 19.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: Spain 1977–2000 a
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv) Confidence votes (cv) Did the Who dissolved
formation cabinet end Parliament in
No. of Voting results of No. Cabinet Cabinet No. of cv Cabinet with an this caseb
unsuccessful successful iv of removed resigned under removed early
investiture (1: pro gov., 2: against ncv by ncv to spec. by failed election?
votes gov., 3: abstentions, and pre-empt inst. cv
4: other) ncv mech.
1 2 3 4
D E L E GAT I O N F RO M C A B I N E T TO
I N D I V I D UA L M I N I S T E R S : K A N Z L E R D E M O K R AT I E
The pre-eminence of the Spanish PM over his ministers derives from his privileged
position as the only member of the government directly elected by and directly
accountable to Congress. This pre-eminence is also revealed in the way the cabinet is
584 Carlos Flores Juberías
formed and operates. The PM alone is responsible for the appointment and dismissal
of ministers, and he directs the work of the government and coordinates the actions of
its other members.
The freedom with which the PM appoints ministers is almost unlimited. Congress
has no constitutional right to scrutinize the PM’s proposals, and the King’s role is a
mere formality. In addition, since the PM is usually the undisputed leader of the majority
party, ministerial appointments do not have to reflect to the views of parliamentary
groups or party notables. They usually reflect the PM’s personal inclinations more than
anything else. In fact, only during the first UCD Governments, when Adolfo Suárez had
to deal with a highly fragmented party, did factions become major players in the power
struggle, asserting control over specific policy areas via ministerial appointments. More
commonly, ministerial nominations are not revealed until after the investiture vote, and
they are made only after the PM himself has decided whom he wants to appoint. This
practice is supported by the Law 50/1997, which repeats the laconic wording of Article 100
of the Constitution, stating that ministers ‘shall be appointed and dismissed by the King
at the proposal of its President’.
A closely related issue, which illustrates the ever-increasing powers of the Spanish
PM, has to do with the determination of ministries. Since 1957 variations in the number,
denomination, and competencies of ministerial departments have needed to be estab-
lished by law. During the turbulent Transition, Royal Decree-Law 18/1976 allowed such
variations to take place via successive decrees passed by the Council of Ministers. Once
the Constitution entered into force, Royal Decree-Law 22/1982 and Law 10/1983
returned to the practice of requiring legislative authorization. Despite these prece-
dents, the new Law 6/1997 has gone so far as to allow the PM—not the government
as a whole—to determine the number, denomination, and competencies of ministerial
departments and Secretaries of State. He does so by issuing Royal Decrees in his own
name. Hence, the freedom the PM enjoys with respect to selecting and removing
ministers extends to the configuration of his cabinet and the basic inner structure of
government departments.
Regarding the PM’s leading role within the cabinet, Articles 2 and 3 of Law 50/1997
add several additional powers to the wide range of authority already granted by the
Constitution. The law makes the PM’s broad constitutional mandate of ‘direct[ing]
the Government’s action and coordinat[ing] the functions of its other members’ more
concrete by spelling out a list of responsibilities. These include representing the
government; drafting its political agenda; determining both national and foreign policies
and verifying their implementation; directing defence policy and commanding the
armed forces; calling, presiding over, and setting the agenda for meetings of the
Council of Ministers; solving conflicts of attribution that arise between ministerial
departments; providing instructions for the ministers; and entrusting the Deputy PM
(when such positions have been created) with the appropriate tasks.
Considering this sweeping enumeration of prime ministerial power, it may be hard
to appreciate the so-called ‘principle of collegiality’ and ‘departmental principle’—both
of which are articulated alongside the ‘prime ministerial leadership principle’ in
Spain 585
the exposition of the motives of Law 50/1997. The former stems in part from the
constitutional provision contained in Article 108, which states that ‘[T]he Government
in its political conduct is collectively accountable before the Congress of Deputies’. It
is codified in Law 50/1997, which spells out the functions of the Council of Ministers.
Among them, Article 5 lists drafting government bills and the state budget; passing
Royal Decree-Laws and Royal Legislative Decrees; negotiating and concluding interna-
tional treaties, and forwarding them to parliament when necessary; declaring states
of alarm and emergency, and proposing to Congress that it declare a state of siege;
issuing public debt and borrowing money as authorized by law; passing regulations for
the implementation of laws; creating, reforming, or suppressing executive organs
within ministerial departments; and adopting programmes, plans, and directives for all
state administrative agencies.
The way in which the Council of Ministers exercises all these competencies is
strongly determined by three elements. One is the primacy of the PM, who calls,
presides over, and determines the agenda for Council meetings. Another one is confid-
entiality. Cabinet members are legally required to keep their deliberations secret—to the
extent that they must swear to do so when they take their oath of allegiance. Minutes of
meetings only record cabinet decisions, not the preceding debates. Finally, the lack of
formality is also remarkable. Many issues clearly regulated in other institutions—for
example, quorum requirements, voting mechanisms, guarantees for minorities, debate
procedures—are simply unregulated. As López Calvo (1996: 206) has argued, this has
much to do with the fact that Spanish democracy has no experience with coalition
governments. It is bound to change once this possibility materializes. Until then, the
Council of Ministers appears elusive to both legal regulation and scholarly research.
The so-called ‘departmental principle’ aims at giving each minister autonomy
and responsibility within the scope of his jurisdiction. The Constitution itself, in
Article 98.2, requires that the leadership and coordinating functions of the PM
should be performed ‘without prejudice to [the ministers’] competence and direct
responsibility’ for the management of their respective departments. However, the
scope of these responsibilities is not clarified by Article 4 of Law 50/1997, which
basically grants the ministers the power to issue regulations and to carry out the
government’s business in their respective areas of competence, but always ‘in
conformity with the decisions adopted at the Council of Ministers and the guidelines
from the President of the Government’.
The corollary of this devaluation of ministerial autonomy is, obviously, a parallel
devaluation of ministers’ individual political responsibility, at least in its most relevant
dimension—political censure by Congress. Some scholars (e.g. see Vírgala Foruria
1988: 289–308) have argued that since ministers’ responsibility before Congress is not
expressly ruled out by the Constitution, it could very well derive from Article 98.2.
It is seen as potentially useful in cases in which individual ministers or their
subordinates commit important political mistakes in managing their departments with
regard to discretionary matters. Nevertheless, the Standing Orders of Congress have
never provided any specific procedure for censoring an individual minister, probably on
586 Carlos Flores Juberías
the assumption that if the minister’s conduct is supported by the PM the appropriate
mechanism is a vote of no confidence. If, on the other hand, the minister does not have
the PM’s support, then it is the PM who should dismiss him. Thus, the only way
Congress can criticize an individual minister appears to be by passing a non-binding,
resolution on the minister’s policies (see Fernández Segado 1985: 239–91). Motions
of this kind have been proposed and voted on several times, but since none has ever
passed it is impossible to tell whether such a resolution might force the PM to remove
the minister.
In conclusion, while one might expect the operation of Spanish governments to
exhibit a balance between the principles of prime ministerial leadership, collegiality,
and departmental autonomy, it does not do so. López Calvo (1996: 195) notes that
‘while the Council of Ministers has a collegial “appearance” it is in fact an organ in
which this principle yields to the monocratic principle’. In short, the Spanish cabinet is
a clear-cut example of Kanzlerdemokratie (see also Bar Cendón 1985b: 48).
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L
S E RVA N T S : T H E C H I M E R A O F A P O L I T I C A L LY
N E U T R A L A D M I N I S T R AT I O N, A N D A N
A D M I N I S T R AT I V E LY N E U T R A L G OV E R N M E N T
Perhaps the only scenario in which the relationship between government and adminis-
tration does not involve a potential for agency problems is when the spoils system has
penetrated the state structures so deeply that there are no administrative posts awarded
to neutral and professional civil servants. However, since it is easy to anticipate the end-
less problems that such a situation might create, it is more useful to consider the prob-
lematic agency relationship between government and administration.
Article 97 of the 1978 Constitution puts the administration—civilian and military—
under the direction of the government. However, this appearance of total and uncon-
ditional submission to the executive partially disappears as a result of language found
in Article 103.1, which proclaims that ‘the Public administration serves the general
interest with objectivity’, establishes the principles of ‘efficacy, hierarchy, decentraliza-
tion, deconcentration, and coordination’, and orders the administration to remain
subject to the rule of law.
The idea that the government leads and the administration simply implements poli-
cies is a gross oversimplification. To begin with, the border between government and
administration is not easy to draw. Many high-ranking officials in a variety of adminis-
trative categories are required to possess a technical expertise, but also need to retain
the minister’s—or even the cabinet’s—confidence. The line separating politicians and
political appointees from civil servants is rather permeable. In addition, the public
administration has its own set of values and interests. These are often quite different
from those that the government promotes and seeks to implement. Finally, the
administration sometimes has significant autonomy protected by the law.
For those reasons, prudent separation and mutual respect between government and
administration seems highly advisable. This is what Garrido Falla’s (1979) formula
Spain 587
E X T E R NA L C O N S T R A I N T S : D I R E C T D E M O C R AC Y,
J U D I C I A L I N D E P E N D E N C E , R E G I O NA L
S E L F - G OV E R N M E N T, A N D E U RO P E A N I N T E G R AT I O N
The Spanish political system includes a number of highly important institutions which
operate outside the specific process of delegation and accountability that has been
described thus far—that is, the core of which is the chain linking people, Congress, PM,
cabinet, and administration. The political relevance of these institutions and the extent
to which they influence the parliamentary system varies significantly, but should not be
underestimated.
The fact that the Franco regime frequently used direct democracy, namely the
referendum, to popularly ratify key political decisions made direct democracy suspect
to many of the framers of the 1978 Constitution. This led to a regulation that Torres
de Moral (1992: i. 106) calls ‘cautious and stingy’. The Constitution does provide for
several kinds of referendums, namely those mentioned in Articles 92 (of advisory
nature, about important political decisions), 151.1 (in order to ratify the decision to
become an autonomous community), 151.2.3 (in order to ratify a Statute of
Autonomy), 152.2 (in order to amend a Statute of Autonomy), 167.3 (facultative, in
order to ratify a constitutional amendment of secondary importance), and 168.3 (com-
pulsory, in order to bring about a complete revision of the Constitution or to make a
more specific but still significant constitutional amendment).
But this enumeration of referendums is deceptive. This is evidenced by the fact that
in more than twenty years, only one referendum has been called—leaving aside those
referendums strictly necessary to set in motion a new Autonomic State. During these
years, decisions as important as joining the European Union and NATO, ratifying
the Maastricht Treaty and participating in the EMU, the legalization of divorce and
abortion, the professionalization of the armed forces, and even a constitutional amend-
ment have all been implemented. However, not once did the government invoke the
constitutional provision that states that ‘political decisions of special importance may
be submitted for an advisory referendum of all the citizens’.
Lack of political will among parliamentary parties and the absence of popular pressure
largely explain why direct democracy has become almost a mirage in the new Spanish
democracy. However, the Constitution itself sets the tone for this marginalization. The
authority to call for a referendum rests entirely with the state. Autonomous
communities cannot hold regional referendums. Advisory ones can take place only if the
PM proposes them and he is supported by an absolute majority in Congress, an institu-
tion that can be expected to be reluctant to support such experiments. Last but not least,
advisory referendums merely ‘advise’ Parliament and the government, who might choose
to ignore it, though doing so might be somewhat politically risky (see Pérez Sola 1994).
‘Justice emanates from the people and is administered in the name of the King by
Judges and Magistrates who are members of the judicial power and are independent,
irremovable, responsible, and subject only to the rule of the law’. This statement,
found in Article 117.1 of the Constitution, clearly implies that members of the
judiciary should be kept outside the chain of delegation and accountability, and that the
Spain 589
C O N C LU S I O N : LOT S O F D E L E GAT I O N,
L E S S AC C O U N TA B I L I T Y, A N D V I C E V E R S A
If we were to underline a feature of the chain of delegation and accountability able to
characterize Spanish parliamentarism, our attention ought to perhaps focus on the
imbalance between these two mechanisms. Regarding the chain that links voters with
Members of Parliament, and Members of Parliament with the Prime Minister, the
impression is that there is a lot of delegation and very little accountability. Voters
simply cannot influence the actions of their representatives, much less remove them
from office before the expiration of their term. Though the Parliament—or at least the
Congress of Deputies—has mechanisms designed to hold the PM and his cabinet polit-
ically accountable, the more imposing these devices are on paper, the less effective they
appear to be in reality. This applies not only to ex post control mechanisms, but also to
those that operate ex ante. It is only necessary to remember that prime ministerial
candidates do not have to disclose their cabinet appointments before the investiture
vote or to recall that voter influence on candidate nomination is virtually non-existent.
When we look at the other half of the chain—the part that links the PM, cabinet
ministers, high-ranking officials, and civil servants of all ranks—the observation has to
be the opposite: delegation shrinks, while accountability soars. The powers of the PM
over his cabinet are indeed sweeping. Its structure, composition, agenda, and even
survival depend entirely on him. Much the same could be said about each minister with
regard to most domestic aspects of his department, and even about lower administrative
units and their heads. Here, delegation appears closely related with accountability,
ready to disappear as soon as accountability renders negative results.
In the end, the overall impression is an extreme concentration of power and authority
in the hands of the PM. He has become ‘the central axis of our political regime’—at
least as long as he is capable of remaining party leader and his party is able to maintain
a sufficiently large majority.
REFERENCES
Álvarez Conde, Enrique (1990). El régimen Político Español., 4th edn. Madrid: Tecnos.
Alzaga, Óscar (1978). La Constitución española de 1978. Comentario sistemático. Madrid: Ediciones
del Foro.
Arceo Vacas, J. L. (ed. ) (1993). Campañas Electorales y Publicidad Política en España. Barcelona: PPU.
Attard Alonso, Emilio (1983). Vida y muerte de UCD. Barcelona: Planeta.
Bar Cendón, Antonio (1979). ‘La monarquía parlamentaria como forma política del Estado
español’, in Estudios sobre la Constitución Española de 1978. Zaragoza: Universidad de Zaragoza.
—— (1985a). ‘Artículo 99. Nombramiento del Presidente del Gobierno’, in Óscar Alzaga (ed.),
Comentarios a las leyes Políticas. Madrid: EDERSA. Vol. VIII, 135–89.
—— (1985b). ‘La estructura y funcionamiento del Gobierno en España: una aproximación
análitica’, in Antonio Bar Cendón (ed.), El Gobierno en la Constitución Española y en los Estatutos
de Autonomía. Barcelona: Diputació de Barcelona.
Barrat Esteve, Jordi (1995). ‘Los procesos de selección de candidatos en los partidos políticos’.
Cuadernos Constitucionales de la Cátedra Fadrique Furió Ceriol, 13: 95–143.
Blasco Esteve, Avelino (1981). La responsabilidad de la Administración por los actos administrativos.
Madrid: Civitas.
592 Carlos Flores Juberías
Cascajo Castro, José Luis (1992). ‘Controles sobre los partidos políticos’, in José Juan González
Encinar (ed.), Derecho de partidos. Madrid: Espasa-Calpe.
Cazorla Prieto, Luis María (1986). Comentarios a la Ley Orgánica de Régimen Electoral General.
Madrid: Civitas.
Del Castillo, Pilar (1992). ‘Financiación de los partidos políticos: La reforma necesaria’, in José
Juan González Encinar (ed.), Derecho de partidos. Madrid: Espasa-Calpe.
Embid Irujo, A. (1987). Los parlamentos territoriales. Madrid: Tecnos.
Equipo de Estudios (1976). Lucha política por el poder. Grupos políticos en la actualidad. Madrid: Elías
Querejeta.
Fernández Segado, Francisco (1985). ‘Las mociones de reprobación y la responsabilidad política
individual (un paradigma de la dinamicidad del Derecho Parlamentario)’, in I Jornadas de
Derecho Parlamentario, Vol. I, Madrid: Congreso de los Diputados: 239–91.
—— (1986). Aproximación a la nueva normativa electoral. Madrid: Dykinson.
Flores Juberías, Carlos (1998). ‘Regionalization and Autonomy in Spain: The Making of the
“Estado de las Autonomías” ’, in Markku Suksi (ed.), Autonomy: Applications and Implications.
The Hague: Kluwer.
—— (1999). ‘A Chamber in Search of a Role: The Senate of Spain’, in Samuel C. Patterson and
Anthony Mughan (eds.), Senates: Bicameralism in the Contemporary World. Columbus: Ohio
State University Press.
García Cotarelo, Ramón (1985). Los partidos políticos. Madrid: Editorial Sistema.
—— (1992). La década socialista. Madrid: Espasa-Calpe.
García-Pelayo, Manuel (1986). El Estado de partidos. Madrid: Alianza Editorial.
Garrido Falla, Fernando (1979). ‘Constitución y Administración’. Revista Española de Derecho
Administrativo, 20: 5–11.
Huneeus, Carlos (1985). La Unión de Centro Democrático y la transición a la democracia en España.
Madrid: CIS-Siglo XXI.
López Calvo, José (1996). Organización y funcionamiento del Gobierno. Madrid: Tecnos.
Martínez Cuadrado, Miguel (1996). La democracia en la España de los años noventa. Barcelona: Ariel.
Martínez Sospedra, Manuel (1983). Aproximación al Derecho constitucional español. La Constitución
de 1978. Valencia: Fernando Torres Editor.
—— (1996). Introducción a los partidos políticos. Barcelona: Ariel.
Montero, José Ramón (1995). ‘Sobre las preferencias electorales en España: fragmentación y
polarización (1977–1983)’, in Pilar del Castillo (ed.), Comportamiento político y electoral.
Madrid: CIS.
Muñoz Machado, Santiago (1993). La Unión Europea y las mutaciones del Estado. Madrid: Alianza
Editorial.
Parejde, Luciano (1988). ‘El Gobierno de la Nación y los Gobiernos Je las Autonamías territori-
tales. Una problema de articulacion’. Documentación Administrativa, 215: 137–71.
Penniman, Howard and Eusebio Mujal-León (eds.) (1985). Spain at the Polls, 1977, 1979 and 1982:
A Study of National Elections. Durham: Duke University Press.
Pérez Sola, Nicolás (1994). La regulación constitucional del referéndum. Jaén: Universidad de Jaén.
Porras Nadales, Antonio (1987). ‘El Consejo General del Poder Judicial según la STC 108/1986,
de 29 de julio, sobre la Ley Orgánica del Poder Judicial’. Revista Española de Derecho
Constitucional, 19: 225–44.
Punset, Ramón (1992). ‘Prohibición de mandato imperativo y pertenencia a partidos políticos’,
in José Juan González Encinar (ed.), Derecho de partidos. Madrid: Espasa-Calpe.
Revenga Sánchez, Miguel (1988). La formación del gobierno en la Constitución española de 1978.
Madrid: Cívitas.
Spain 593
Rodríguez Díaz, Ángel (1989). Transición política y consolidación constitucional de los partidos políticos.
Madrid: Centro de Estudios Constitucionales.
Ruiz-Jarabo Colomer. D. (1993). El juez nacional como juez comunitario. Madrid: Cívitas.
Sánchez Agesta, Luis (1980). El sistema político de la Constitución española de 1978. Madrid: Editora
Nacional.
Sánchez Morón, Manuel (1980). La participación del ciudadano en la Administración. Madrid:
Centro de Estudios Constitucionales.
Santaolalla López, Fernando (1992). ‘Partido político, grupo parlamentario y diputado’, in José
Juan González Encinar (ed.), Derecho de partidos. Madrid: Espasa-Calpe.
Share, Donald (1999). ‘From Policy-Seeking to Office-Seeking: The Metamorphosis of the
Spanish Socialist Workers Party’, in Wolfgang C. Müller and Kaare Strøm (eds.), Policy, Office,
or Votes? How Political Parties in Western Europe Make Hard Decisions. Cambridge: Cambridge
University Press.
Soler Sanchez, Margarita (2001). Campañas electorales y democracia en España. Castellón: UJI.
Soriano, José Eugenio (1991). Comunidades autónomas y Comunidad Europea. Madrid: Tecnos.
Terol Becerra, Manuel J. (1990). El Consejo General del Poder Judicial. Madrid: Centro de Estudios
Constitucionales.
Tomás Villarroya, Joaquín (1982). Breve historia del constitucionalismo español. 2nd edn. Madrid:
Centro de Estudios Constitucionales.
Torres del Moral, Antonio (1983). ‘La monarquía parlamentaria como forma política del Estado
español’, in Pablo Lucas Verdú (ed.), La Corona y la Monarquía Parlamentaria en la Constitución
de 1978. Madrid: Universidad Complutense.
—— (1992). Principios de Derecho Constitucional Español, Volumes I and II, 3rd edn. Madrid:
Servicio de Publicaciones de la Facultad de Derecho de la Universidad Complutense.
Vírgala Foruria, Eduardo (1988). La moción de censura en la Constitución de 1978. Madrid: Centro de
Estudios Constitucionales.
Zafra Valverde, José (1973). Régimen político de España. Pamplona: EUNSA.s
20
I N T RO D U C T I O N
In 1917 the Swedish King was forced to finally accept that the Riksdag, and not he, was
the ultimate authority upon which the cabinet’s political existence rested (Lewin 1992;
Stjernquist 1997; von Sydow 1997). While this event firmly established a parliamentary
democracy, it was not accompanied by a change of the written Constitution. Until
1975, Sweden was formally governed under the 1809 Constitution, according to which
the Parliament shared political power with the King and his cabinet. On 1 January 1975,
a new Constitution went into effect. In contrast to its predecessor, it stipulates an
almost ideal–typical single chain of delegation and accountability from the people to
civil servants. In this chapter, I present the institutional mechanisms that define this
chain and I show that it is challenged by new ex post institutional checks that have
grown stronger over time.
Let’s begin by examining the new Constitution and basic legislative procedures. One
impetus for reform was of course the large discrepancy between written statutes and
practice. Formally, the King still presided over cabinet sessions, appointed high civil
servants and could dissolve the Riksdag. As Head of State he was the supreme military
commander. All of these powers were already, before a series of reforms that took place
in the 1970s, primarily symbolic and obsolete. But many West European countries have
old-fashioned Constitutions so this alone was not the reason behind the reform. There
were also two other main reasons (Bergman 1995).
A second reason for reform was that the King still enjoyed a role, albeit very limited,
in cabinet formation, cabinet resignation, and parliamentary dissolution. This was a
source of embarrassment for the ruling Social Democratic Party. Its party programme
had since 1911 called for a Republic. While Social Democratic leaders knew that the
I would like to acknowledge the late Nils Stjernquist who was a constant source of knowledge and inspira-
tion on the chapter topic. I would also like to thank Hans Hegeland, Cynthia Kite, Ingvar Mattson, my fel-
low editors, and my colleagues at the research seminar at the Department of Political Science, Umeå
University for their helpful comments on various drafts of this chapter.
Sweden 595
idea of abolishing the monarchy was unpopular with most voters, the party’s official
support for a Republic provided the leadership with an incentive to favour a major con-
stitutional reform (Bergman 1995). A third reason behind the reforms was that the
opposition parties had their own motives for a large-scale reform. They wanted to abol-
ish the First (or Upper) Chamber and increase the proportionality of the electoral system.
In both cases, they argued, existing institutional arrangements favoured the Social
Democratic Party at their expense (von Sydow 1989).
With the 1970 elections, a partial constitutional reform enhanced the proportionality
of the electoral system. And on 1 January 1971, the upper (First) chamber of the
Riksdag was abolished (von Sydow 1989). The same constitutional reform also intro-
duced formal rules for a no-confidence vote against the cabinet. After that, the consti-
tutional debate focused on the monarch, resulting in a new Constitution that took away
most of his symbolic powers (Bergman 1999).1 In addition to performing the ceremonial
duties of a Head of State in international relations, the monarch now only opens new
sessions of the Parliament and chairs an advisory board on foreign affairs. The
Constitution places the Riksdag at the centre of representative democracy (Article 1: 4).
With these reforms, the legislative process also changed. Up to 1970, to legislate, a sim-
ple majority was needed in each of the two chambers. The cabinet (formally the King)
had to consent with these majorities for the law to pass. With regard to budget deci-
sions, taxes, and state expenditures, the Riksdag alone had decision-making
powers. If the two chambers disagreed, the total number of votes for and against in
each chamber were summed up and the combined majority won the vote. In addition,
under the 1809 Constitution, the cabinet had broad powers to legislate by decree. This
included laws that governed the public sector and such matters as citizenship. While
this legislative authority originally was quite broad, over time the cabinet delegated
more and more authority to the Parliament (Lagerroth 1955: 180–4; Andrén 1968:
270–2, 429–59; Kungl. Maj:ts proposition nr 90 år 1973: 201–26).
This reversed form of delegation, from the cabinet to the Parliament, ended with the
new Constitution. In 1975, all legislative and budgetary powers were taken over by the
Riksdag (Holmberg and Stjernquist 1980). Legislative initiatives from the cabinet or
from one or more MPs are referred to a Riksdag committee. This is mandatory. The com-
mittees can also initiate legislative proposals on their own. After committee deliberation,
the plenary chamber votes on committee proposals. The formal decision is taken by a
simple majority of all members present and voting. There is no quorum rule.
With the exception of European Union (EU) law, from January 1995 when Sweden
became a member, only the Riksdag can pass a law. Both the cabinet and state agencies
can issue binding statutes below the level of a law. However, on some specific matters,
such as Civil Law, Criminal Law, laws on Local Authorities, and Electoral Law, the
Riksdag cannot delegate to the cabinet the power to issue decrees. Over these matters
cabinet can only issue instructions on how the law should be implemented. In other
areas, the cabinet can issue decrees (regulations) within the framework of existing law.
State agencies can issue binding regulations only if such authority is explicitly delegated
1
Until 1980 the monarch had to be male. Since 1980 females of the royal family can also inherit the
throne.
596 Torbjörn Bergman
from the Riksdag and/or the cabinet (Holmberg and Stjernquist 1980: 244–5, 293–8,
2000: 123–44).
There is no constitutional court or federalism that limits the Riksdag’s power. There
is, however, an ex ante procedure that involves a Law Council (Lagrådet) consisting of
High Judges. The cabinet or a Riksdag committee can ask this Council to scrutinize a
legislative proposal. The Council can recommend that the Riksdag reject all or parts of
a government proposal because it conflicts with the Constitution, is inconsistent, or
because it conflicts with already existing law (Article 8: 18). The Riksdag is free to
decide whether or not to follow the recommendation. Because it is only advisory, the
scrutiny process is a relatively weak ex ante constraint. However, the mechanism can
become important ex post because courts and public authorities must not implement
laws or other regulations that conflict with the Constitution (Article 11: 14). The
prerogative to set aside unconstitutional statutes existed in practice until 1979 when the
Riksdag decided to incorporate it in the written Constitution (Holmberg and
Stjernquist 1980: 250). However, if the statute is contained in either a law—and there-
fore by definition issued by the Riksdag—or a regulation issued by the cabinet, it is only
if the conflict with the Constitution is ‘obvious’ that courts and other authorities can
refuse implementing the statute. It is hypothetically possible that a court could refuse
to implement a law on the grounds that the law council considers it unconstitutional,
even if the Riksdag had decided to adopt it anyway. In practice, however, courts and
public authorities have been very restrictive in invoking their right not to implement
laws and other regulations (Holmberg and Stjernquist 1980: 397–9, 2000: 144–7, 190–5).
A simple parliamentary majority can change any existing restrictions on the
Riksdag’s decision-making authority. All that is required is an intervening election
between decisions (Article 8: 15). From 1980, however, a minority of 10 per cent of the
MPs (currently thirty-five MPs) can ask the Riksdag to subject the constitutional
amendment to a decisive referendum. If a third of all MPs support such a motion, the
referendum must be held at the next general election. If a majority of those voting in
the general election also participate in the referendum and the anti-amendment side
wins, the proposed amendment is rejected. Otherwise, the incoming Riksdag is free to
adopt or reject the proposed amendment at will (Article 8: 15).
There are also other important constraints that are formally external to the
ideal–typical singular chain of parliamentary democracy. Two of the most important
constraints have to do with increasingly autonomous local and regional governments
and European integration. Neo-corporatist patterns of interest-intermediation and the
use of advisory referendums also impact on the chain. We will return to these con-
straints later on, but let us for now focus on the organizations that link one step to the
next in the democratic chain, that is, political parties.
P O L I T I C A L PA RT I E S
Political parties were first mentioned in the partial constitutional reform that went into
effect on 1 January 1971 (Holmberg and Stjernquist 1980: 143). The 1975 Constitution
Sweden 597
defines political parties as ‘any association or group of voters which appears in an elec-
tion under a particular designation’ (Article 3: 7).2
On the dominant Left–Right dimension, in the traditional Scandinavian five party
system (Berglund and Lindström 1978), the parties has been aligned as follows: the
Communist Party (from 1990 the Left Party), the Social Democratic Party, the Centre
Party (the former Agrarian Party) the Liberal Party, and the Conservative (or Moderate)
Party.3 In 1988, 1994, and 1998, the Green Party was elected to the Riksdag. In 1991,
when the Green Party lost representation, two other parties were elected. One was a
populist right-wing party, New Democracy, which lost all of its representation in the
following (1994) election. The Christian Democratic Party, on the other hand, managed
to stay represented in the Parliament. In 1998, with 11.8 per cent of the total vote, it
became the fourth largest party after the Social Democrats, the Conservatives, and the
Left Party. The Green, Centre, and Liberal Parties each received around 5 per cent of
the total vote (Allmänna Valen 1998).
The Social Democrats have controlled most cabinets, but not because of a lack of
party competition (Bergman 1995). Competition has in fact been fierce and the parties
have often split into two competing blocs, the Socialist (the Communist and Social
Democratic Parties) and the non-socialist blocs (the other three traditional parties). In
2003, the Green Party cooperates with the ‘Socialist’ bloc and the Christian Democratic
Party belongs to the ‘non-socialists’. A split among the voters has mirrored the party
blocs.
Traditionally, voters have had fairly stable party preferences and voted very much
according to class. However, in the early 1970s, voting patterns began to change. Fewer
voters have a stable preference for one party and many voters decide which party to
vote for late in the election campaign. At the same time there has been growing distrust
of political parties, political institutions, and politicians in general (Gilljam and
Holmberg 1995; Holmberg 1999). The share of the population with party membership
also used to be comparatively very high. In 1990, the political parties had a total of
1.4 million members. The Social Democratic Party alone had about one million mem-
bers out of a population of a little more that eight million. However, this was largely a
consequence of the fact that the Social Democratic Party allowed local unions to reg-
ister their members collectively as party members. Under pressure from the opposition
2
In Sweden, there are four documents with ‘constitutional’ status. In this study, the specific individual
Articles that are mentioned all come from the Instrument of Government, which is the act that defines the
main state institutions and the relationship between these institutions (The Constitution of Sweden 2000).
The other three fundamental laws are the Act of Succession, The Freedom of Press Act, and The
Fundamental Law of Freedom of Expression. For an amendment, parts of the Riksdag Act also require two
decisions and an intervening election. However, this Act can also be changed by one decision if more that
three-quarters of voting MPs vote in favour. For the change to be passed, however, the three-quarters major-
ity must also include more than half of all MPs, that is, currently more than 175 MPs must vote in favour.
3
Except for the relative positions of the parties of the ‘middle’, the Centre and the Liberal Parties, the
alignment has been very stable. However, the predominance of the Left–Right dimension has been weakened
by alternative conflict issues such as the nuclear power program in the 1970s and early 1980s, ecology more
generally in the 1980s, and conflict over EU-membership and an awakening moral–religious dimension in the
1990s (Oscarsson 1998).
598 Torbjörn Bergman
parties and union members, the Social Democratic Party abolished this practice in the
early 1990s. In 1999, the parties had slightly over four hundred thousand members, of
which about 40 per cent were Social Democrats. Most of the loss of members in the
1990s stems from the changed Social Democratic practice concerning union members,
but the numbers also reflect the trend that the political parties in the early and
mid-1990s lost members and that they have trouble attracting new ones (Bäck and
Möller 1997: 112; Gidlund and Möller 1999: 26). These problems do not necessarily
mean that the parties are moribund. They remain well organized and open to
membership influence, at least in cross-national comparison (Sannerstedt and Sjölin
1994; Widfeldt 1997). But the challenges to political parties are real enough and their
legitimacy has been questioned during the 1990s (Petersson et al. 2000).
In spite of their problems, political parties remain important for the democratic
chain. The parties nominate candidates for elections and make authoritative decisions in
the cabinet and Riksdag. They also select representatives to other bodies that participate
in the preparation and implementation of public policy. Cabinets have traditionally
relied heavily on government appointed commissions to prepare policy proposals
(Larsson 1994). In these commissions, representatives from all Riksdag parties—often
MPs—work with technical experts and sometimes with representatives of organized
interests to prepare the cabinet’s policy proposals. These proposals are published as
official reports to government. This policy-preparation process has been very important
and, even though its importance has declined in recent years, commissions remain influ-
ential. There is also a remiss system by which reports from these commissions are scruti-
nized by state agencies, other levels of government, and organized interests, before the
proposal returns to the cabinet and Riksdag (Heclo and Madsen 1987: 13; Ruin 1996).
Party representatives also serve on the boards of agencies, where they oversee public pol-
icy implementation. While board members are formally appointed by the cabinet, it is
usually the political parties, and often their parliamentary groups, who in practice
decide who will be appointed (Birgersson and Westerståhl 1992; Isberg 1999). The polit-
ical parties are thus involved at every step of the delegation chain. Their importance is
most obvious when they nominate candidates for political office, when they run elec-
toral campaigns, and in the relationship between the Riksdag and the cabinet. The role
and influence of political parties is less obvious and less important further along the
chain, but the Swedish parties are involved in all its steps. Well organized and cohesive,
they take part in the commissions that prepare cabinet proposals, in decision-making,
and in policy implementation. With this in mind, we now turn to the constitutionally
mandated chain of parliamentary democracy.
D E L E GAT I O N F RO M VOT E R S TO M P S
The system by which Swedish voters select their representatives has been proportional
since 1909. The First (or Upper) Chamber was indirectly elected. Election results to
regional assemblies determined each party’s share of the First Chamber seats. Since the
1970 elections to the first unicameral Riksdag, the multimember constituencies
elect between two and thirty-four representatives. A national pool of second-tier
Sweden 599
(or supplementary) seats is distributed among the parties to increase the overall
proportionality between votes and seats. The national pool is 39 of 349 seats
(Article 3: 6). The system excludes from the Riksdag parties that fail to get at least 4 per
cent of the vote nationally or 12 per cent in one constituency. With 4 per cent nationally,
a party shares in the distribution of the national pool of supplementary seats. A party
elected only because it is particularly strong (12 per cent or more of the vote) in one or
a few multimember constituencies is not included in the distribution of the national
pool of seats. The restrictive threshold for a constituency seat has meant that no party
has been elected on the basis of this rule alone.
Although voting is not mandatory, turnout reached over 90 per cent in the four
elections between 1973 and 1982. In recent elections turnout has tended to decline,
falling dramatically to 81.4 per cent in 1998. This is the lowest for any election since
1958, when turnout was 77.4 per cent (Allmänna Valen 1998).
With the 1998 elections, a ‘positive’ preference vote was introduced. The new election
law is designed to let voters pick a candidate from a party. If the individual preference
vote for a particular candidate is 8 per cent or more of the party’s total number of
constituency votes, then that candidate moves to the top of list (which otherwise
remains unchanged). If more than one candidate gets more than 8 per cent, the one
with the most preference votes moves to the top of the list and fills the party’s first seat
in that constituency. In the 1998 elections, about 30 per cent of voters used the prefer-
ence vote, but most of these went to candidates already at the top of the party list. The
direct effect of the preference vote was that 12 (of 349) MPs were elected in place of
candidates ranked higher on their parties’ lists (SOU 1999: 136).
While the new electoral law allows voters to have a say in which candidates are
elected from the party list, it also gave the parties full control over the list itself.
The new electoral system does not allow voters to add or cross-out candidates, which
the previous system did. Consequently, the new electoral law has also eliminated the
possibility for factions or groups within a party to present their own list, unless it is
approved by the party organization. In earlier elections, competing lists within the
same party could determine MP selection. Thus, while giving voters a greater role in
choosing candidates from parties’ lists, the new rules also give party organizations
more control over ballot access in the first place.
Once elected, the MPs mandate is individual. A party cannot remove an MP from his
seat. This means that he or she can leave the party, join another party, or work alone.
An MP cannot, however, simultaneously hold a Riksdag seat and serve in the cabinet
(or as Speaker). If an MP serves in the cabinet (or as Speaker), an alternate member is
temporarily called to fill the MPs seat. Aside from this, an MP can only be removed
from his position if the election is annulled or if a court of law finds him or her
unsuited to serve (Article 4: 7). This would happen if the MP has been convicted of a
serious crime. MPs have a limited form of immunity from prosecution, which covers
acts they commit when they vote and speak in the Parliament. This immunity can be
lifted by a majority of five-sixths of the MPs present and voting (Article 4: 8). In all
other circumstances, Swedish MPs are not protected from prosecution in regular courts
(Holmberg and Stjernquist 1980).
600 Torbjörn Bergman
an MP’s performance, they can demote or remove this person from the list in the next
election. Still, candidates who are highly ranked in one election tend to remain high in
subsequent ones.
In spite of constituency supremacy over nominations, once elected MPs see them-
selves more as party than as constituency representatives. Their main concern lies with
national level policy-making, and constituency interests are seen as complementary to
their main role as party representatives (Esaiasson and Holmberg 1996; Bäck and
Möller 1997: 119–20). In addition, the parliamentary party groups are both cohesive
and influential in shaping the policies that the MPs promote (Hagevi 1999; Isberg 1999).
Party competition over executive office provides strong incentives for such behaviour.
This leads us to the next step in the delegation chain.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
To varying degrees, depending mainly on whether or not the cabinet controls a Riksdag
majority, the opposition parties are able to influence policy decisions via parliamentary
committees. Among MPs, the three most prestigious committees are the Finance
Committee, the Foreign Affairs Committee, and the Committee on Constitutional
Affairs (Hagevi 1998: 69–82). Since most cabinets have not controlled a parliamentary
majority, the committees have often been arenas for building legislative majority coali-
tions. All standing committees have full time administrative staffs and are authorized to
take legislative initiatives of their own. They are specialized in areas that roughly
correspond to the jurisdictions of government ministries, and committee chairs are
distributed roughly proportionally among the parliamentary parties (Arter 1999).
Riksdag committees thus rank high in international comparisons of relative influence
(Strøm 1990; Mattson and Strøm 1995). In spite of the consensual style of policy-making
that these arrangements facilitate, government positions come with resources and
influence, and they are highly valued and sought after by the political parties.
Negative Parliamentarism
One of the most important institutional features (contract design) of Swedish
parliamentary democracy is its ‘negative’ parliamentarism (Bergman 1993). ‘Positive’
parliamentarism requires that an incoming cabinet demonstrate explicit parliamentary
support, typically by a simple majority in an investiture vote. In negative parliamen-
tarism, the coming to power of a new cabinet requires only tolerance by a parliamentary
majority, not active support. Typically, the Head of State appoints the new PM and no
investiture vote is held.
Within the category of negative parliamentarism, the Swedish formation rules are
unusual in two ways. First, since 1 January 1975 the Head of State (the monarch) is
excluded from the cabinet formation process. Second, Sweden has a ‘negative’ investiture
vote in the Riksdag. A candidate for PM is proposed by the Speaker of the Riksdag.
Before a new cabinet can assume power, a vote must be held, but the Speaker’s candidate
602 Torbjörn Bergman
for PM succeeds unless more than half of the members of the Riksdag (175 MPs) vote
against him or her. All candidates proposed by the Speaker since 1975 have been
approved on the first vote.
Compared with the formation process before 1975, the new voting rule has had two
important consequences. First, the parties are now forced to show openly whether or
not they tolerate a new cabinet before it forms. Abstention effectively means tacit
support for a new cabinet. Second, the party groups in the Riksdag have become more
involved in the formation process. Previously, the bargaining process was more easily
kept within a small circle of leaders. Since the MPs now have to attend a vote on the
new cabinet, party leaders consult more widely to ensure that their MPs are willing to
vote for or against a particular candidate. While such consultations are largely limited
to the parliamentary party groups, the trend has been towards broader consultations.
The Liberal Party has held explicit pre-electoral discussions of potential coalitions at
party conferences, and the Centre Party is known to have a time-consuming internal
decision-making process. Overall, however, cabinet formation and dissolution have been
decided largely by party leaders and not by party organizations, party conferences, or
membership votes (Bergman 1995).
No-Confidence Votes
As mentioned above, from 1917 to 1971, the principle that the cabinet must resign if the
parliamentary majority no longer tolerates it existed only by convention (Andrén 1968).
The partial constitutional reform that went into effect in 1971 introduced formal rules
for a no-confidence vote against the cabinet. According to the written rule, an absolute
parliamentary majority (50 per cent ⫹ 1 of all MPs) is needed to bring down the
cabinet (Article 12: 4). One practical effect of this voting rule is that it promotes
cabinet stability. Even minority cabinets with a relatively small support base in the
Parliament can be stable since abstentions are, in practice, counted as support for the
incumbent cabinet.4
A no-confidence vote can be directed at individual ministers, but if one is directed
against the PM the whole cabinet must resign if he loses (Article 6: 5–7). Attempts to
use this procedure against the PM occurred in 1980, 1996, 1998 and again in 2002, but
none was successful. In 1985 a vote was held on an individual minister, the Minister for
Foreign Affairs. This motion of no confidence also failed (Holmberg and Stjernquist
2000: 201). In fact, the opposition parties never expected to win. Instead, opposition
parties have used these motions to demonstrate their discontent with the cabinet or the
minister. A minister or a PM who believes that he might lose a vote of no confidence
is likely to resign before the vote is actually held. This happened in June 1988, when the
Minister of Justice resigned to avoid censure (Holmberg and Stjernquist 2000: 201).
4
Here cabinet stability should be understood only in a formal sense, that is, the voting rule makes it
difficult to unseat a cabinet already in power. This does not have to mean that a particular cabinet is free from
internal conflict or that it has a stable parliamentary base in the Riksdag.
Sweden 603
Confidence Votes
There are no formal procedures by which a cabinet can test or demonstrate its ability
to govern by winning a crucial vote in the Parliament. If a cabinet wants to show that
it can govern, it can attach such an importance to any policy vote in the Riksdag. In this
case, presumably the cabinet will choose a policy vote that is important enough to
motivate the resignation of the cabinet. If the cabinet loses the vote, it has two options,
neither of which is constitutionally mandatory but yet considered politically necessary.
It can call an extra election or it can resign, in which case the Speaker will start a new
government formation process.
The only explicit confidence vote held in the postwar era occurred in 1990. The
ruling Social Democratic cabinet linked its survival to an economic austerity package
designed to deal with inflation and a growing budget deficit. It lost the vote and
formally resigned. In this case, however, there was no alternative cabinet ready to be
formed, and after a brief cabinet crisis the same cabinet continued in power (Bergman
1995).5
Parliamentary Dissolution
Other mechanisms that govern the relation between an incumbent cabinet and the
Parliament are parliamentary dissolution rights. The most important cornerstone of
parliamentary democracy is the Parliament’s power to remove the cabinet from power.
Another way of terminating a conflict between the cabinet and the parliamentary
majority is an early parliamentary dissolution. Under the Swedish Constitution of 1809,
dissolving Parliament was one of the King’s prerogatives. Once the parliamentary
principle was established, it was accepted that the cabinet as a whole controls this
power. However, as late as the 1960s some constitutional experts argued that the King
could refuse a cabinet’s request to dissolve the Riksdag. Others disputed this claim (see
Bergman 1995). The constitutional reforms of the 1970s removed all doubt about the
King’s role in parliamentary dissolution by completely abolishing it.
There are two ways in which the Riksdag can be dissolved. The first is if the Speaker
proposes four successive prime ministerial candidates and each is voted down by more
than half of the MPs. In this case, new Riksdag elections must be held within three
months (Article 6: 3). Second, the cabinet has discretionary power to dissolve the
Riksdag by a regular cabinet decision (Article 3: 4) though there are some restrictions
on this power. The cabinet cannot dissolve Parliament during the first three months
after an election. Nor can it dissolve the Riksdag after it has resigned and is serving in
a caretaker capacity (SOU 1967:26: 184–8, 1972:15: 126–7, 1987:6: 127–33; Holmberg
and Stjernquis 1980).
In Sweden, even if an extra election is held the ordinary election must be held at the
constitutionally mandated date. This effectively decreases the incentive to hold extra
5
Because of the restrictive counting rule used for this volume, in comparative terms this formal resignation
and regaining of power does not count as a cabinet formation.
604 Torbjörn Bergman
elections. In the postwar period, only one extra election has been held. This was in 1958
when the parties were deeply divided over the creation of mandatory public pension
funds (see Molin 1965). In 1973, 1978, and 1981 extra elections were discussed as pos-
sible solutions to cabinet crises, but in each case this was ruled out since regular elections
would have to be held quite soon thereafter (Särlvik 1983: 130–2; Hadenius, Molin, and
Wieslander 1993: 263–81).
Table 20.1 summarizes the behavioural record until 1 January 2000 with respect to
cabinet investiture, no-confidence and confidence votes, and parliamentary dissolution.
Until 1975, no Riksdag vote was held when a new cabinet formed. From 1975, a cabi-
net investiture vote occurs only after the former PM has resigned—or has been voted
out of office—and the Speaker has suggested a new PM. Until 2000, inauguration votes
had been held only nine times. In three of these cases, the cabinet received fewer than
half of the votes cast, but in no case did the Riksdag reject a candidate presented by the
Speaker. As noted above, the no-confidence votes directed at PMs have all been unsuc-
cessful. The empty column under confidence votes is because Sweden does not have a
constitutionally mandated process for such votes. The informal practice concerning
such confidence votes was discussed above. Finally, the thesis (above) that the rules gov-
erning parliamentary dissolution make an early election unlikely is supported by the
fact that in the entire postwar period there has only been one early election (1958). That
decision was de facto made by the cabinet.
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv) Confidence Did this Who
formation votes (cv) cabinet end dissolved
No. of Voting results of No. of Cabinet Cabinet under specific with an early Parliament
unsuccessful successful iv ncv removed resigned institutional election? in this
investiture (1: pro gov., 2: against gov., by ncv to mechanism case?b
votes 3: abstentions, and 4: other) pre-empt n.a.
ncv
1 2 3 4
No. Cabinet Date of Investiture votes (iv) No-confidence votes (ncv) Confidence Did this Who
formation votes (cv) cabinet end dissolved
No. of Voting results of No. of Cabinet Cabinet under specific with an early Parliament
unsuccessful successful iv ncv removed resigned institutional election? in this
investiture (1: pro gov., 2: against gov., by ncv to mechanism case?b
votes 3: abstentions, and 4: other) pre-empt n.a.
ncv
1 2 3 4
While ministers are not formally required to do so, they answer almost all interpella-
tions and questions. If an interpellation/question is delayed or not answered, the
minister must inform the Riksdag as to the reasons why. An MP has no formal right to
get an answer from a particular minister. Even if the question is directed at a specific
minister, it can be answered by one of his cabinet colleagues. In the 1990s attempts have
been made to make the debates livelier. Since 1991 there are weekly televised sessions
in which questions are answered directly by a subset of ministers and since 1996 written
questions ( but not interpellations) only receive a written answer (Riksdagens årsbok
1990/91: 20–3; KU 1995/96: 18).
The Riksdag can also scrutinize cabinet reports. The practice of sending cabinet
reports to Parliament has increased over time. The Parliament now receives reports on
the various commissions that prepare government policy, the state-owned enterprises,
and on EU affairs. In addition, the Parliament has its own Parliamentary Audit Office
(Riksdagens Revisorer). While much smaller than the government’s audit agency, it
nonetheless provides the Riksdag with an autonomous source of information. The
Riksdag also appoints Ombudsmen who are entitled to scrutinize all state and local
government agencies, issue recommendations, and even to prosecute civil servants that
do not follow laws and regulations. The number of actual prosecutions has decreased
over time, but the number of cases examined by the Ombudsmen has increased.
Ombudsmen often issue critical reports. The reports are often influential, but because
their recommendations are not necessarily followed, complaints have been heard about
the relative ‘toothlessness’ of the Ombudsmen as an oversight mechanism (Petersson
and Söderlind 1993: 113, 115–8; Sterzel 1998: 226–9; Holmberg and Stjernquist 2000).
The Riksdag Committee on Constitutional Affairs produces annual reports on how
cabinet ministers conduct their official duties. The Committee is allowed access to cab-
inet minutes and documents. The Committee can also question individual ministers on
any matter that has to do with his/her portfolio. The plenary session of the Riksdag
receives the Committee report on the cabinet minister’s conduct and votes on whether
or not to support any critique that the Committee may direct against the cabinet or
minister. There is a certain amount of party posturing in these reports, but they also
serve an important function, both as a preventive measure and as a way to scrutinize
ministerial behaviour. The Committee has influenced how the cabinet fulfils its consti-
tutional duties, not least on more technical matters of cabinet procedure (Holmberg
and Stjernquist 1980: 401–4, 2000: 203–4; Sterzel 1998: 213–7).
Its legal authority to decide whether or not to prosecute individual ministers for
unlawful acts committed in office reinforces the authority of the Committee on
Constitutional Affairs. Such acts are narrowly defined as wrongful actions that
a minister could do only because he was a minister. Prosecution has never been an issue
in the postwar period (Holmberg and Stjernquist 2000: 204–5). Nonetheless, the
possibility of prosecution gives the Committee extraordinary measures and status as
the most important body for scrutinizing the conduct of cabinet and ministers.
In the step connecting MPs to the cabinet, Swedish institutional arrangements
(contract design) are such that they make it relatively easy to form a cabinet (i.e. the
absence of a positive vote requirement) and difficult to unseat a cabinet already in
608 Torbjörn Bergman
power (e.g. the mandated period between regular elections and the absolute
majority requirement for a vote of no confidence). Cabinet positions come with
administrative resources and agenda-setting powers and the political actions of
MPs are very much focused on influencing cabinet proposals (ex ante). Ex post instru-
ments, such as parliamentary questions, are also used, partly to demonstrate to the
home constituency that its concerns are taken care of, but play a minor role in the
Riksdag.
D E L E GAT I O N W I T H I N T H E C A B I N E T
Until 1975, the King formally approved all cabinet appointments that in reality were
made by the PM. Since 1975, constitutionally, the PM is free to appoint and dismiss
other ministers on his own and at will, although he can of course be constrained by
intraparty concerns or by coalition partners. In contrast, collegiality and consensus is
the practice (and the norm) in policy decisions within the cabinet.
Including the PM, the cabinet has had between sixteen and twenty-two ministers. As
of 2000 there were twenty cabinet ministers (Hadenius, Molin, and Wieslander 1993:
372–80; Holmberg and Stjernquist 2000: 109). It is the cabinet as a whole that rules the
country (Article 1: 6) and handles foreign affairs (Article 10: 1). Much of its influence
stems from its role as the body in which leaders of the governing party/parties prepare
most of the Riksdag’s decisions. The cabinet also has the final say over many top-level
state appointments and is in charge of the state bureaucracy. For issues that fall outside
the jurisdiction of the courts, it is also the highest body in the state-administrative
appeals process.
The Constitution does not provide much guidance as to cabinet organization. There
are a few broad rules. The cabinet collectively decides on the jurisdiction of each
ministry by decree (Article 7: 1). A quorum of five or more ministers is necessary to
hold a formal meeting of the cabinet (Article 7: 4). Finally, with the exception of some
issues concerning national defence, all cabinet level decisions must be taken at formal
cabinet meetings (Article 7: 3).
The Government Office (the Chancery) consists of the cabinet (or PM) office, the
individual ministries (currently ten ministries), and the Government Administration
Office. Excluding the Ministry for Foreign Affairs, there were about 1,800 employees in
the early 1990s. More than half were civil servants, legal experts, and similar employees.
About one-tenth were part of the commission system, while roughly one-quarter were
assistants and clerical staff. About a hundred were political appointees who resign with
cabinet. By 31 December 1999 the total number of employees was 2,281 while the
number of political appointees was seventy-five (Petersson 1994: 89; Holmberg and
Stjernquist 2000: 116–7).
Individual cabinet members (ministers) are normally heads of their departments.
The exception is that some ministers are responsible for preparing certain policy areas
(e.g. elementary and secondary schools within the Ministry of Education), but do not
head a ministry of their own. These are ministers ‘without portfolio’ (in 2000 there were
Sweden 609
nine such ministers). While such ministers used to be mainly responsible for the
formal–legal aspects of policy drafting, today these ministers have much the same tasks
as those who head a ministry.
Cabinet level decisions are made only at formal cabinet meetings. Unless individual
ministers explicitly reserve themselves against a particular decision, all ministers are
formally responsible for all cabinet decisions.6 Nonetheless, constrained only by law
and general cabinet procedure, the heads of individual ministries are in charge of their
ministries. In this capacity, individual ministers prepare many decisions that are never
discussed, but simply accepted by their fellow ministers at the formal cabinet meetings
(Andrén 1968: 299–311; SOU 1972:15: 150; Kungl. Maj:ts proposition nr år 1973: 179–86;
Holmberg and Stjernquist 1980; T. Larsson 1986: 181–6).
The decision rule within the cabinet is best described as consensus defined by the
PM. Under the 1809 Constitution, the King formally made all cabinet decisions himself,
which is one explanation for the traditional lack of a formal cabinet decision rule.
Establishing a decision rule was discussed during the constitutional reforms of the
1970s, but in the end it was left to individual cabinets to decide such matters (Holmberg
and Stjernquist 2000: 120–1; for a case study see S. E. Larsson 1986: 185–208). With
regard to cabinet decision-making, the PM obviously has considerable agenda power.
He has his own staff to initiate and coordinate policy-making.
Compared to the early postwar period, the PM’s resources to monitor other ministers
have increased quite dramatically. When PM Tage Erlander first arrived in his new
office in 1946, his entire staff consisted of a secretary and one assistant. In the mid-
1950s he began to have an advisory support staff of his own. However, it was not until
1964 that he began to build a formal office that included a junior minister (or state sec-
retary, statssekreterare). This followed a 1963 incident in which the PM only learned
about the capture of a Soviet spy after some of his ministers had already been
informed. By 1976, with the coming to power of the first postwar non-socialist coali-
tion, the PM’s office became an arena for coalition coordination and problem solving
(T. Larsson 1986: 181–95). During the 1991–4 coalition, in addition to civil servants and
the PM’s personal staff, the PM’s office contained a separate coordinating unit for each of
the four coalition partners and played a central role as a conflict-resolution mechanism
(Bergman 2000).
In a formal sense, all ministers, including the PM, have the same weight in cabinet
decision-making. The 1975 Constitution allows the PM to appoint a Deputy (Article 7: 8).
This possibility is used only occasionally and the position is primarily honorary. Instead,
after the PM, it is the Ministry of Finance that has had a dominant role relative to
other departments (Larsson 1994; Mattson 1998). In fact, the Finance Minister, who is
6
The importance of collective decision-making in the cabinet can go beyond the role of the PM and
individual ministers. A tradition of collective cabinet responsibility may force coalition partners and individual
ministers to defend cabinet policies with which they disagree. This can have practical importance and, for
example, create a disincentive to join a coalition unless there is an assurance that its policies will be acceptable
(e.g. see Laver 1992: 47–8, on the case of Ireland).
610 Torbjörn Bergman
responsible for preparing the state budget, generally has had a better capacity to monitor
other ministries than the PM himself (Larsson 1990: 154). It is only in the 1990s that there
has been a major shift in resources and monitoring capacity in favour of the PM’s Office.
In sum, as the first among equals in the King’s cabinet, before 1975 the PM had to act
through the King to appoint his ministers. Today the screening and selection
procedures are directly in the hands of the PM. In addition, the PM has many informal
ways and nowadays a large staff to help steer the cabinet. Constitutional design also
limits the autonomy of individual ministers who must pass all major decisions by their
colleagues in the cabinet. This constitutional requirement is likely to favour policy-
cohesive cabinets.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
In contrast to most other countries, individual ministers are in Sweden not in charge of
the state agencies that implement public policy (Holmberg and Stjernquist 1980, 2000).
State agencies fall under the authority of the cabinet, and not under an individual
ministry responsible for a particular policy area (Article 7: 3). In addition, neither the
cabinet nor the minister can stipulate how an agency or a civil servant should in
particular cases implement law or exercise powers of the state over individuals. This is
a matter reserved for the civil servants that implement law and exercise state authority
(Article 11: 7). Matters of misuse and abuse by agencies or civil servants are to be
handled by the courts or by the cabinet in an appeals process. However, the boundary
between the cabinet and the ministries on the one hand and the ‘autonomous’ agencies
on the other is not clear and precise, but an area of constitutional debate (Holmberg
and Stjernquist 1980, 2000; Petersson and Söderlind 1993: 73–9).
Ministers have been known to complain about their inability to direct agencies and
about occasional resistance among the ministerial civil servants and state agencies to
cuts in funding or major policy shifts. For example, the 1976 shift from a Social
Democratic cabinet to a three party (non-socialist) coalition cabinet faced resistance
from some civil servants (Pierre 1995). In general, however, individual ministers ‘talk
to’, ‘consult with’, and ‘exchange information’ with the agencies that implement policy
in their area. In practice they also control the preparation of policy decision and the
agency budgets within the framework decided by the Riksdag. Ministers also control
nominations to the highest positions in the agencies, all of which gives them ample
opportunity to influence civil servants (Petersson and Söderlind 1993).
Until 1965 it was not possible to fire high civil servants. Today this constitutional
protection remains for judges in a limited form only. It is possible for the agency super-
vising the court system to fire judges with the support of a court of law (Article 11: 5).
Under the old (1809) Constitution, some high civil servants, such as heads of central
agencies and county governors, were appointed in the name of the Head of State,
while other appointments were made by the state agency itself (Andrén 1968: 385–93).
The hierarchical distinction between different categories of civil servants in terms of
appointment still exists, but now the dividing line is between those appointed by the
cabinet and those who are not (Article 11: 9). For the latter there is a merit system
and regular career positions for which one applies. For the former, the cabinet has
Sweden 611
the discretion to appoint a qualified candidate of its choice. There is a tendency for these
positions, for example as county governors, to be a reward for long and faithful service,
although not necessarily to a person representing the party in power. Governments fill
these positions with candidates from each political camp, but are also sometimes cri-
tiqued by the opposition parties (be they Socialist or non-socialist) for over-representing
those with partisan connections to the government (e.g. see, KU 1995/96:30: 52–6).
Further evidence that the dividing line between ministries that prepare and agencies
that implement is not completely strict is that the cabinet can decide that minor policy
decisions or allocation of funds will be handled by a ministry instead of an agency. One
such example is aviation regulation, which grew over time from a desk in a ministry to
its own agency (Andrén 1968: 339). This practice continues and it is not unusual that
the cabinet delegates to one of its ministries the task of deciding on the distribution of
funds according to principles laid down by the cabinet (KU 1991/92: 30: 78–80). Taken
together, individual ministers have numerous channels through which to enforce their
will. Nonetheless, the utility of a particular ministerial portfolio might be more
restricted in Sweden than in a system in which each minister exercises a more
formal and direct control over implementing agencies.
In addition, civil servants exercise more rather than less influence over policy today
than they did early in the postwar period. In fact, when asked what group in society
over time had increased its power the most, MPs mentioned civil servants second only
to the mass media (Esaiasson and Holmberg 1996: 200–2). Civil servants now take more
policy initiatives and their influence has been increasing both at the proposal stage and
in implementation (Wallin et al. 1999).
In spite of the scandals and corruption cases that do occur in the public sector, in
comparative perspective the Swedish bureaucracy remains largely corruption-free
(Andersson 1999). Agencies and administrative agencies that exercise ‘police-patrol’
ex post monitoring facilitate this. The National Audit Office (Riksrevisionsverket) is
empowered to scrutinize both how state institutions manage their funds and how they
perform their duties. The cabinet also appoints an Ombudsman (Justitiekanslern) to
supervise the state administration. His office investigates complaints by the public and
he serves as a prosecutor in cases concerning civil servants accused of abusing their
public duties and in cases concerning free speech (SOU 1993: 37).
While agencies under both the cabinet and the Riksdag serve as ex post controls of
policy decisions and their implementation, the cabinet agencies are better funded and
staffed than the Riksdag’s Parliamentary Audit (Riksdagens revisorer) Office (Petersson
and Söderlind 1993). The main audit office, the National Audit Office, is both under the
control of the cabinet and dependent on a budget allocated via the cabinet. It has also
been shown to be more sensitive than the Riksdag’s Parliamentary Audit Office to min-
isterial interests. The Parliamentary Audit also has constitutional protection (Article 12: 7).
The problem with this Audit, however, is that the Riksdag does not have well-developed
routines for making use of its findings (Ahlbäck 1999).7
7
The Riksdag has decided that in 2003 there will be one major audit agency in Sweden. It will be placed
under the Riksdag and not under the cabinet.
612 Torbjörn Bergman
‘Fire-alarm’ type monitoring of civil servants is made easier by constitutionally
guaranteed access to public documents. There are of course exceptions; the publication
of some material can lead to prosecution. This includes material that threatens national
security, harassment of a population group, child pornography, and libel. Government
documents that refer to sensitive matters, for example in foreign policy, defence,
commercial relations, and personal integrity, are to be kept secret. Nonetheless, the
main principle is free public access to all documents in the public sector. In addition, the
Freedom of the Press Act, which has constitutional status, gives every citizen the right
to publish any written material. The same Act allows all citizens, including public
officials, freely to provide information to the mass media (Petersson 1994: 173–8).
Public access to government documents helps prevent misuse and abuse of government
offices and government power. With the help of such sources, media can draw attention
to acts of potential wrongdoing. One area in which this has occurred is with regard to
cases of immigration and asylum-seekers, where public attention has led authorities to
change earlier decisions and foster public sector reforms. Other such areas include the
export of weapons and the practice of state agencies conducting some of their
business through ‘private’ companies that they fully or partially control via stock hold-
ing (Petersson and Söderlind 1993: 266, 273–6).
In sum, civil servants have served their political masters rather well. There has been
an element of ex ante screening and selection of civil servants on the basis of their pref-
erences. Some of this has to do with the long tenure of power enjoyed by the Social
Democrats. But more important and influential is the strong tradition of relative
administrative autonomy and bureaucratic integrity. This relatively favourable evalua-
tion is helped by constitutional guarantees for freedom of information and access to
public sector documents for both the citizens and the press. There are also working ex
post supervising institutions such as the Ombudsmen. However, it is noticeable that
elected representatives (MPs), the cabinet, and individual ministers play a rather limited
role in ex post monitoring. As principals they have tended to rely on civil servants for
policy initiative and development. Monitoring of civil servants that implement policies
has been less of a concern.
E X T E R NA L C O N S T R A I N T S
Parliamentary democracy is embedded in a context in which organized interests
influence the democratic policy-making chain. In Sweden, as in other small West
European nation-states, a system of democratic corporatism has been very important
(e.g. see Katzenstein 1985). While the literature has often focused on the organizations
representing labour and capital, a broad variety of organized interests have been
influential through their close ties with the state administration. The particularly
strong ties between the Social Democrats and the blue-collar union federation, the LO
(Landsorganistationen), is an often noted feature of Swedish politics (Back and
Berglund 1978; Bäck and Möller 1997; Feldt 1991; Strøm and Bergman 1992; Petersson
1994; Teorell 1998).
Sweden 613
There are actually two articles in the Constitution that might be considered to
support a corporatist system. One (Article 7: 2) states that cabinet decisions must be
properly prepared and that this can involve the consultation of both individuals and
associations. The other (Article 11: 6) provides the possibility of permitting individuals,
companies, associations, and foundations to manage the implementation of state
policy. If allowed by law (i.e. by the Riksdag), this can include decisions concerning the
benefits and/or obligations of individual citizens. Yet, these provisions do not mean
that Sweden has a corporatist Constitution. On the contrary, as noted above, Sweden
has an almost ideal–typical parliamentary Constitution. What the two statutes actually
show is that the corporatist system was long seen as a natural and uncontroversial part
of Swedish politics.
Corporatism of course deviates from the ideal democratic chain of delegation and
accountability, and a problem has been the lack of public accountability. That is, rather
anonymous interest group representatives serve on government commissions and on
the boards of state agencies, yet these groups are not accountable to the electorate.8 In
the early 1990s, the corporatist system is reported to have diminished in importance
(Lewin 1994; Petersson 1994: 162–5), partly as a consequence of new relations between
the major organized interests in the labour market. These new relations are in turn,
very much a consequence of the fact that groups representing business have opted out
of the system by refusing to appoint representatives to the commissions that prepare
cabinet proposals and the boards that implement public policy (Rothstein and
Bergström 1999).
While the impact of neo-corporatism as a constraint in many respects has weakened,
the importance of other constraints has increased. The direct and single chain of
parliamentary democracy has been affected by increased local government spending
and by the new power relations between central and local governments that have
accompanied this expansion (Petersson 1994: 122–40; Häggroth and Peterson 1999;
Stjernquist 1999). The importance of local governments in welfare spending has roots
back before the Second World War. But since the early 1970s, it has been national pol-
icy to give local (and regional) governments increasingly greater responsibility and
greater autonomy for welfare provisions in the social sector (schools, health, etc.). In
terms of social expenditure and personnel, the Swedish welfare state is very much a
function performed by local governments. In fact, about 70 per cent of total govern-
ment consumption occurs at these levels. Simultaneously, the national level govern-
ment has moved from steering by detailed instructions to a system of general policy
targets. In the process, it has become increasingly difficult for the national government
to promise certain services or public sector reforms. Today it is very much up to the
local governments to actually decide and implement these promises.
Referendums also constrain the parliamentary chain of delegation. Advisory refer-
endums were held both in 1955 (left- or right-hand side traffic) and in 1958 (public and
mandatory pension funds). Prior to this, the referendum had only been used in 1922
8
It is, however, not entirely clear that a system more characterized by lobbying and pluralism guarantees
that elected officials are more accountable to the general public.
614 Torbjörn Bergman
(prohibition of alcohol). In 1980 the national nuclear power programme became the
next subject to be tried in a national referendum. In 1994 membership in the EU was
decided in a referendum. Formally, these referendums have all been advisory but each
time the political parties have promised to follow the wishes of the majority—at least
to the extent that it has been possible to identify such a majority. However, a problem
with advisory referendums is that it is up to the political parties to interpret and imple-
ment the advice they receive. This includes the issue of for how long a referendum
result is valid. To make the referendums mechanism decisive would not solve these
problems. Instead the use of both kinds of referendums tends to make the power rela-
tions that are implied by the ideal type of parliamentary democracy more diffuse.
Other important trends work in the same direction.
Indirectly, constitutional adherence to the ideal type of parliamentary democracy
assigns a particular role to the courts. While few would dispute the importance of
autonomous courts that are able to resist pressure from powerful politicians, an ideal
parliamentary democracy assumes that the courts have a more limited role than they
have in an ideal–typical separation-of-powers system. Traditionally, the distinction
between state administration and the court system has been weak in Sweden. This is at
least partly because of roots that go back to a system in which the Kings’ rule over the
country involved both having power over the state administration and being the admin-
istrative appeal authority (Brunsson, Sonneyby, and Wittenmark 1990). Both courts and
judges have been reluctant to play an active role relative to the parliamentary system of
government. As mentioned, the courts’ have constitutional authority to refuse to allow
implementation of laws that they find are unconstitutional (Article 11: 14). However,
until the 1980s, the option was hardly ever used as courts and judges have wanted not
to appear political in what was a system where democratic politics had priority over
legalism (Board 1991; Petersson et al. 1999).9
There is now more of a constitutional debate over the proper role of the courts. Much
of the impetus for an increase in judicial activism seems to stem from developments
outside Sweden, such as the role of courts in European integration. One important
European development is associated with the Convention on Human Rights and
the European Court of Human Rights in Strasbourg (Petersson 1994: 120–1). The
European Court of Human Rights has questioned the traditional Swedish system of
cabinet-appeal rather than court-appeal in administrative matters such as the appeals
process for professional licensing and procedural rules concerning the expropriation of
property. Following European Court decisions, appeals on such matters have been
transferred from the cabinet to administrative courts of justice (Petersson and Söderlind
1993: 262–4; Holmberg and Stjernquist 2000: 195–6).
An obviously significant external constraint is the EU. On 1 January 1995, Sweden
became one of fifteen EU member states. The strong and central position of the
9
This debate has been fuelled by domestic legal cases in the 1990s. Two examples are from 1996. In
one decision a court of first appeal declined to implement a 1946 law about the complete ban of political
uniforms. In the other, the court at the last (second) level of appeal (the Supreme Court) decided not to
recognize a punishable offence on the grounds that the law was not based on proper delegation from the
Riksdag (Holmberg and Stjernquist 2000: 190–5).
Sweden 615
European Court of Justice within European integration has helped foster a new and more
active role for national courts. EU membership also has other political and constitutional
consequences. Sweden stayed outside of the Economic and Monetary Union (EMU) in
1999 and prior to a referendum in September 2003 opinion polls show that a majority of
Swedes oppose membership. However, the Social Democratic cabinet and most opposi-
tion parties have been eager to meet the strict budget and finance criteria for EMU mem-
bership. They have also decided to change Central Bank rules to give it greater autonomy
from direct and short-term influence by elected representatives (Mattson 1998).
EU membership was accompanied by an institutional innovation in the Riksdag.10
The European Affairs Advisory Committee (EU-nämnden) was created for the purpose
of monitoring the cabinet when it represents Sweden in the EU Council of Ministers.
This committee allows the Riksdag to scrutinize ministers, but it does not have the full
formal status of a regular Riksdag committee. The Committee does not have the right
to make policy initiatives on its own or to refer issues to the full chamber, but is still
considered important. This is shown by the fact that opposition parties send high-
ranking MPs to meetings between cabinet ministers and the Committee. Together
with the important developments reported above, EU membership is reshaping the
Swedish chain of parliamentary democracy. In fact, today the de facto Constitution
exhibits a ‘dualism’ between the ideal–typical national constitution and the treaties of
the EU (Algotsson 2000).
C O N C LU S I O N
While far from perfect, for much of the postwar period the Swedish chain of demo-
cratic delegation and accountability has not been affected by serious agency problems.
Fierce electoral competition between two clearly defined blocs and two alternative
visions of society allowed voters to be reasonably sure that elections would impact on
the direction of national politics. At the same time, the minority status of most cabi-
nets allowed for moderation in policy decisions. Many interests have been included in
the process and well-organized and influential parties have linked together the various
steps of the constitutionally mandated parliamentary democracy. In addition, ex post
mechanisms have worked fairly well. The courts, Ombudsmen, audit offices, and
specific agencies have uncovered misuse of state authority and funds. Policy implementers
are far from flawless, but there has been comparatively little patronage or outright
corruption. Public access to state documents and media scrutiny has helped secure a
strong tradition of openness and transparency. This has meant that supervision
through active ‘police patrolling’ and ‘fire alarm’ reporting from voters and media have
been met by the formulation of new policy in the cabinet and in the Riksdag.
Since the late 1980s, however, Swedish politicians have increasingly been faced with
distrust, lower electoral turnout, and a loss of party members. The parties have tried
10
For more on the Swedish committee see Hegeland (1999), Hegeland and Mattsson (1997); for a comparison
with the Danish committee, see Hegeland and Matsson (1996); for a broad EU-comparison, see Bergman
(1997); for a study of the impact of the EU on national democracy in Sweden and the other Nordic countries,
see Bergman and Damgaard (2000).
616 Torbjörn Bergman
to respond by increasing their state funding and allowing voters more say in the choice
of individual MPs. It does not seem to have reversed the trend. This, in turn, might have
something to do with the increasingly complex and diffuse political system. It is possible
that the growing discrepancy between de facto power relations and the ideal–typical
Constitution contributes to a declining popular trust in politicians and political parties.
Early in the twentieth century, parliamentary democracy developed within a
Constitution based on separation of powers. By the mid-1970s the last remnants of this
Constitution disappeared. At about the same time, a new separation-of-powers system
began to develop. There is once again a considerable discrepancy between the written
rule and the ‘living constitution’ (constitutional practice). At the beginning of the
new century, national politics has lost a considerable amount of leverage over local
governments, which show growing signs of developing disparate solutions to local
problems and economic scarcity. In addition, European integration has furthered the
development of a pattern of governance more characterized by power sharing with the
courts. And important parts of national-level decision-making authority have been
transferred to the EU. As a result of these tendencies, the formerly (and still constitu-
tionally) sovereign Riksdag is no longer the supreme focal point in a unitary chain.
These developments do not render parliamentary democracy unimportant or uninter-
esting. On the contrary, the challenge is to develop a further understanding of how the
recent changes impact on the national chain of delegation and accountability, which
continues to be the fundamental basis of democratic legitimacy.
REFERENCES
Ahlbäck, Shirin (1999). Att kontrollera staten. Den statliga revisionens roll i den parlamentariska
demokratin. Ph.D. Dissertation, Department of Political Science, Uppsala University.
Algotsson, Karl-Göran (2000). Sveriges författning efter EU-anslutningen. Stockholm: SNS Förlag.
Allmänna Valen (1998). Del 1: Riksdagen (1999). Stockholm: Statistiska centralbyrån (SCB).
Andersson, Staffan (1999). Hederlighetens pris: en ESO-rapport om korruption. Stockholm:
Departementsserien 1999:62.
Andrén, Nils (1968). Svensk statskunskap. Stockholm: Liber.
Arter, David (1999). Scandinavian Politics Today. Manchester: Manchester University Press.
Bäck, Mats and Möller, Tommy (1997). Partier och organisationer, 4th edn. Stockholm: Publica.
Back, Pär-Erik and Berglund, Sten (1978). Det svenska partiväsendet. Stockholm: Almqvist och Wiksell.
Berglund, Sten and Ulf Lindström (1978). The Scandinavian Party System(s). Lund:
Studenlitteratur.
Bergman, Torbjörn (1993). ‘Formation Rules and Minority Governments’. European Journal of
Political Research, 23: 55–66.
—— (1995). Constitutional Rules and Party Goals in Coalition Formation: An Analysis of Winning
Minority Governments in Sweden. Ph.D. Dissertation, Department of Political Science, Umeå
University.
—— (1997). ‘National Parliaments and EU Affairs Committees: Notes on Empirical Variation and
Competing Explanations’. Journal of European Public Policy, 4: 373–87.
—— (1999). ‘Trade-Offs in Swedish Constitutional Design: The Monarchy Under Challenge’, in
Wolfgang C. Müller and Kaare Strøm (eds.), Policy, Office, or Votes? How Political Parties in
Western Europe Make Hard Decisions. Cambridge: Cambridge University Press.
Sweden 617
Bergman, Torbjörn (2000). ‘Sweden: When Minority Cabinets are the Rule and Majority
Coalitions the Exception’, in Wolfgang C. Müller and Kaare Strøm (eds.), Coalition Governments
in Western Europe. Oxford: Oxford University Press.
—— and Damgaard Erik (eds.) (2000). Delegation and Accountability in European Integration: The
Nordic Parliamentary Democracies and the European Union. London: Frank Cass. (Also published
as a special issue of the Journal of Legislative Studies, 6(1).)
Birgersson, Bengt Owe and Westerståhl, Jörgen (1992). Den svenska folkstyrelsen, 5th edn.
Stockholm: Publica.
Board, Joseph B. (1991). ‘Judicial Activism in Sweden’, in Kenneth M. Holland (ed.), Judicial
Activism in Comparative Perspective. London: Macmillan.
Brunsson, Karin, Sonneby, Claes, and Wittenmark, Lars (1990). Beslutsmaskinen—en bok om
regeringskansliet. Lund: Studentlitteratur.
The Constitution of Sweden (2000). The Fundamental Laws and the Riksdag Act. Stockholm:
Riksdagen.
Esaiasson, Peter and Holmberg, Sören (1996). Representation From Above: Members of Parliament
and Representative Democracy in Sweden. Aldershot: Dartmouth.
Feldt, Kjell-Olof (1991). Alla dessa dagar . . . i regeringen 1982–1990. Stockholm: Norstedts.
Gidlund, Gullan M. (1983). Partistöd. Lund: Liber CWK Gleerup.
—— and Möller, Tommy (1999). Demokratins trotjänare. Lokalt partiarbete förr och nu. Stockholm:
Fritzes offentliga publikationer (SOU 1999: 130).
Gilljam, Mikael and Holmberg, Sören (1995). Väljarnas val. Stockholm: Norstedts Juridik.
Hadenius, Stig, Molin, Björn, and Wieslander, Hans (1993). Sverige efter 1900: En modern politisk
historia, 13th edn. Stockholm: Bonnier Alba.
Hagevi, Magnus (1998). Bakom riksdagens fasad. Göteborg: Akademibolaget Corona.
—— (1999). ‘Parliamentary Party Groups in the Swedish Riksdag’, in Knut Heidar and Ruud
Koole (eds.), Parliamentary Party Groups in European Democracies. London: Routledge.
Häggroth, Sören and Peterson, Carl-Gunnar (1999). Kommunalkunskap. Så fungerar din Kommun.
Stockholm: Hjalmarsson & Högberg Bokförlag.
Heclo, Hugh and Madsen, Henrik (1987). Policy and Politics in Sweden: Principled Pragmatism.
Philadelphia: Temple University Press.
Hegeland, Hans (1999). Riksdagen, Europeiska unionen och demokratin. Licenciatavhandling,
Statsvetenskapliga Institutionen, Lund: Lunds Universitet.
—— and Mattson, Ingvar (1996). ‘To have a Voice in the Matter: A Comparative Study of the
Swedish and Danish European Committees’. The Journal of Legislative Studies, 2: 198–215.
—— —— (1997). ‘The Swedish Riksdag and the EU: Influence and Openness’, in Matti Wiberg
(ed.), Trying to Make Democracy Work: The Nordic Parliaments and the European Union.
Stockholm: The Bank of Sweden Tercentenary Foundation & Gidlunds Förlag.
Holmberg, Erik and Stjernquist, Nils (1980). Grundlagarna med tillhörande författningar.
Stockholm: PA Norstedt & Söners förlag.
—— —— (2000). Vår författning, 12th edn. Stockholm: Norstedts Juridik AB.
Holmberg, Sören (1999). Representativ demokrati. Stockholm: Statens offentliga utredningar
1999: 64.
Isberg, Magnus (1999). Riksdagsledamoten i sin partigrupp. 52 riksdagsveteraners erfarenheter av parti-
gruppernas arbetssätt och inflytande. Stockholm: Gidlunds Förlag.
Katzenstein, Peter J. (1985). Small States in World Markets: Industrial Policy in Europe. Ithaca:
Cornell University Press.
KU (1991/92:30). Konstitutionsutskottets betänkande 1991/92:KU30 (1992 Granskningsbetänkande).
Stockholm: Riksdagen.
618 Torbjörn Bergman
KU (1995/96:18). Konstitutionsutskottets betänkande 1995/96:KU18 (1995 Reform av spörsmålsinsti-
tuten). Stockholm: Riksdagen.
KU (1995/96:30). Konstitutionsutskottets betänkande 1995/96:KU30 (1996 Granskningsbetänkande).
Stockholm: Riksdagen.
Kungl. Maj:ts proposition nr 90 år (1973). Förslag till ny regeringsform och ny riksdagsordning, m.m.
Stockholm: Riksdagen.
Lagerroth, Fredrik (1955). Moderna författningar mot historisk bakgrund. Stockholm: P. A. Norstedt
& Söners Förlag.
Larsson, Sven-Erik (1986). Reagera i koalition. Den borgerliga trepartiregeringen 1976–1978 och
kärnkraften. Stockholm: Bonniers.
Larsson, Torbjörn (1986). Regeringen och dess kansli. Lund: Studentlitteratur.
—— (1990). ‘Regeringens och regeringskansliets organisationsstruktur, berednings-
och beslutsformer under 150 år’, in Att styra riket—regeringskansliet 1940–1990,
Departementshistoriekommittén. Stockholm: Allmänna Förlaget.
—— (1994). ‘Cabinet Ministers and Parliamentary Government in Sweden’, in Michael Laver and
Kenneth A. Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge:
Cambridge University Press.
Laver, Michael J. (1992). ‘Coalition and Party Policy in Ireland’, in Michael J. Laver and Ian Budge
(eds.), Party Policy and Government Coalitions. London: Macmillan.
Lewin, Leif (1992). Ideologi och strategi, 4th edn. Stockholm: Norstedts Juridik.
—— (1994). ‘The Rise and Decline of Corporatism: The Case of Sweden’. European Journal of
Political Research, 26: 59–79.
Mattson, Ingvar (1998). Den statliga budget processen—rationell resursfördelning eller meningslös
ritual? Stockholm: SNS Förlag.
——and Strøm, Kaare (1995). ‘Parliamentary Committee’, in Herbert Döring (ed.), Parliaments and
Majority Rule in Western Europe. Frankfurt am Main: Campus and New York: St. Martin’s Press.
Molin, Björn (1965). Tjänstepensionsfrågan. En studie i svensk partipolitik. Göteborg:
Akademiförlaget.
Myrdal, Gunnar (1982). Hur styrs landet? Stockholm: Rabén & Sjögren.
Oscarsson, Henrik (1998). Den svenska partirymden: Väljarnas uppfattningar av konfliktstrukturen i
partisystemet 1956–1996. Ph.D. Dissertation, Department of Political Science, Göteborg University.
Petersson, Olof (1994). Swedish Government and Politics. Stockholm: Publica.
——, Hernes, Gudmund, Holmberg, Sören, Togeby, Lise, and Wängnerud, Lena (2000).
Demokrati utan partier? Stockholm: SNS.
—— and Söderlind, Donald (1993). Förvaltningspolitik, 2nd edn. Stockholm: Publica.
——, von Beyme, Klaus, Karvonen, Lauri, Nedelmann, Birgitta, and Smith, Eivind (1999).
Democracy the Swedish Way. Stockholm: SNS.
Pierre, Jon (1995). ‘Governing the Welfare State: Public Administration, the State and Society in
Sweden’, in Jon Pierre (ed.), Bureaucracy in the Modern State. An Introduction to Comparative Public
Administration. Cheltenham: Edward Elgar.
—— and Widfeldt, Anders (1992). ‘Sweden’, in Richard S. Katz and Peter Mair (eds.), Party
Organizations. London: Sage.
Riksdagen i siffror 1996/1997 (1997). Stockholm: Riksdagen.
Riksdagens årsbok. Riksmötet 1990/91 (1991). Stockholm: Riksdagen.
Rothstein, Bo and Bergström, Jonas (1999). Korporatismens fall och den svenska modellens kris.
Stockholm: SNS.
Ruin, Olof (1996). ‘Sweden: From Stability to Instability’, in Jean Blondel and Maurizio Cotta
(eds.), Party and Government. London: Macmillan.
Sweden 619
Sannersted, Anders and Sjölin, Mats (1994). ‘Folkstyrets Problem’, in Anders Sannerstedt and
Magnus Jerneck (eds.), Den moderna demokratins problem. Lund: Studentlitteratur.
Särlvik, Bo (1983). ‘Coalition Politics and Policy Output in Scandinavia: Sweden, Denmark and
Norway’, in Vernon Bogdanor (ed.), Coalition Government in Western Europe. London:
Heinemann.
SOU (1967:26). Partiell författningsreform. Betänkande av grundlagberedningen.Stockholm:
Justitiedepartementet. Statens offentliga utredningar (Official report to the government).
SOU (1972:15). Ny regeringsform. Ny riksdagsordning. Betänkande avgivet av Grundlagberedningen.
Stockholm: Justitiedepartementet. Statens offentliga utredningar (Official report to the
government).
SOU (1972:17). Nomineringsförfarande vid riksdagsval. Undersökning utförd av Dan Brändström på
uppdrag av Grundlagberedningen. Stockholm: Justitiedepartementet. Statens offentliga utred-
ningar (Official report to the government).
SOU (1987:6). Folkstyrelsen villkor. Betänkande av folkstyrelsekommittén. Stockholm:
Justitiedepartementet. Statens offentliga utredningar (Official report to the government).
SOU (1993:37). Justitiekanslern—en översyn av JK:s arbetsuppgifter m.m. Stockholm:
Justitiedepartementet. Statens offentliga utredningar (Official report to the government).
SOU (1999:136). Personval 1998—en utvärdering av personvalsreformen. Stockholm:
Justitiedepartementet. Statens offentliga utredningar (Official report to the government).
Sterzel, Fredrik (1998). Författning i utveckling. Uppsala: Iustus förlag.
Stjernquist, Nils (1997). Tvåkammartiden. Sveriges riksdag 1867–1970. Stockholm: Riksdagen.
—— (1999). ‘Huruledes särskilda menigheter må för egna behov sig beskatta’, in Lena Marcusson
(ed.), Festkrift till Fredrik Sterzel. Stockholm: Iustus Förlag.
Strøm, Kaare (1990). Minority Government and Majority Rule. Cambridge: Cambridge University Press.
—— and Bergman, Torbjörn (1992). ‘Sweden: Social Democratic Dominance in One Dimension’,
in Michael Laver and Ian Budge (eds.), Party Policy and Government Coalitions. Basingstoke:
Macmillan.
Teorell, Jan (1998). Demokrati eller fåtalsvälde? Om beslutsfattande i partiorganisationer. Ph.D.
Dissertation, Department of Political Science, Uppsala University.
von Sydow, Björn (1989). Vägen till enkammarriksdagen. Demokratisk författningspolitik i Sverige
1944–1968. Stockholm: Tiden.
——(1997). Parlamentarismen i Sverige. Utveckling och utformning till 1945. Stockholm: Gidlunds förlag.
Wallin, Gunnar, Ehn, Peter, Isberg, Magnus, and Linde, Claes (1999). Makthavare i fokus.
Stockholm: SNS Förlag.
Wiberg, Matti (1994). Parliamentary Control in the Nordic Countries. Forms of Questioning and
Behavioral Trends. Helsinki: The Finnish Political Science Asociation.
Widfeldt, Anders (1997). Linking Parties with People? Party Membership in Sweden 1960–1994. Ph.D.
Dissertation, Department of Political Science, Göteborg University.
21
I N T RO D U C T I O N
The main features of Britain’s partly-written but uncodified, ‘ancient and ever-altering
constitution’ (Bagehot 1963: 59), sometimes summarized as the ‘Westminster model of
democracy’ (e.g. Lijphart 1999: 9), are highly correlated with Strøm’s (2000: 268)
ideal–typical definition of parliamentary democracy as a ‘single chain of delegation
with multiple links’ where ‘[i]n each link, a single principal delegates to a single or
multiple non-competing agents’. Voters delegate powers to Members of Parliament
(MPs) through elections, which—due to the plurality electoral system—are largely an
intense competition between two major parties for an overall majority in the House of
Commons and—consequently—control of the government. General elections are
practically contests between two rivalling teams supporting alternative candidates for
the post of Prime Minister (PM). The cabinet, dominated by the PM, is at the apex of
decision-making. A highly professionalized, non-partisan civil service with a low level
of autonomy is designed to avoid bureaucratic drift. As long as the cabinet enjoys
majority support in the House of Commons, the constitutional features of the
Westminster model (see below) ideally ensure that the government’s ability to carry
out its policies is maximized and not constrained by any significant domestic institu-
tional checks.
Although the evolving British practice has always differed from this ideal type, the
model does capture some essential elements of the Westminster system. The strictly
unitary chain of delegation and accountability, which is generally believed to be typical
of the Westminster model, is designed to minimize the efficiency and transparency
losses caused by multiple principals and agents observed in other Western democracies
(Moe 1984: 768–9). It avoids confusion about political responsibility, as power is not
shared among a large number of actors with veto powers (see also Thies 2000: 241–7).
I owe thanks to Bill Jenkins, Lord Norton of Louth, Colin Seymour-Ure, and the editors of this volume for
their helpful comments on an earlier draft. Any error or opacity remaining is the author’s responsibility alone.
The United Kingdom 621
At least since the 1970s a number of authors have argued, however, that the price of
these desirable properties is too high, that the system does not ensure accountability,
because it lacks checks and balances, does not sufficiently guarantee civil liberties, and
is, in effect, an ‘elective dictatorship’ (Hailsham 1976). I will argue in this chapter that
due to constitutional change since the 1970s the ideal type of the Westminster model
has become a less and less accurate description of British politics. The privatization and
marketization of the 1980s and 1990s, for example, has weakened the link of delegation
and accountability between ministers and executive agencies. Furthermore, the ambi-
tious constitutional reform programme initiated during Prime Minister Tony Blair’s
first term in office entails at least the potential for the creation of a number of veto
points and institutional checks that may dilute the single chain of democratic delega-
tion and accountability.
The main sources of Britain’s uncodified Constitution are statute law, common law,
constitutional conventions, and so-called ‘works of authority’, that is, constitutional
interpretations by leading scholars. The Constitution can therefore be amended by a
simple Act of Parliament1 or, simply, by incremental change in the interpretation and
application of constitutional conventions. The constraints on the government are
mainly informal: ‘Without a written Constitution’, as Michael Foley (1999: 50) points
out, ‘the British system has grown to be dependent upon a range of customs, protocols
and traditions that prevent the gaps and ambiguities of the constitution from being
exploited by governments. Even when government has the power to overcome the
system’s checks and balances, it normally remains obligated to the social restraints of
“club government” . . .’ The perception that the informal norms of ‘club government’
may no longer ensure accountability have partly been the impetus for the 1997–2001
reforms. Despite its partly-written nature, there is agreement on certain ‘pillars’ of the
Constitution. The first three pillars constitute and reinforce the single chain of delega-
tion and accountability outlined above, the fourth can be said to dilute it:
1. O. Hood Phillips (1978: 5) calls the sovereignty of Parliament ‘the one fundamental
law of the British Constitution’. According to Albert V. Dicey (1959: 39–40), parliament-
ary sovereignty means that Parliament defined as Crown-in-Parliament has ‘the right to
make or unmake any law whatever; and, further, that no person or body is recognized
by the law of England as having a right to override or set aside the legislation of
Parliament’. Given the dominance of the House of Commons in the trias ‘Crown-in-
Parliament’—the reduction of the monarch’s role to a ceremonial figure head and the
reduced role of the upper house as a revising chamber with limited delaying powers—
this means de facto sovereignty of the House of Commons and of the government its
majority supports.
2. Another important pillar of Britain’s Constitution is parliamentary government with
a ‘close union, the nearly complete fusion of the executive and legislative powers’
(Bagehot 1963: 65). In Britain’s parliamentary system of government, the cabinet
1
This includes legislation about the length of the parliamentary term. For example, the maximum length
of time a Parliament can sit was reduced from seven to five years in the Parliament Act 1911.
622 Thomas Saalfeld
depends on the confidence of Parliament, although under normal circumstances the
cabinet clearly dominates Parliament (Lijphart 1984: 7).
3. Unitary government is a logical consequence of the sovereignty of Parliament,
which can devolve powers to subnational tiers of government and take away those
powers whenever its majority chooses to.
4. Norton (1994: 69) also considers the treaties constituting the European Union (EU)
to be part of the Constitution. Although British ministers (answerable to the British
Parliament) are crucially involved in the passage of EU legislation, the increasing use of
qualified majority voting and the incorporation of the European Convention of
Human Rights into British law can potentially undermine Parliament’s sovereignty.
The vast majority of bills (approximately 80 per cent) passed by the House of
Commons are government bills. Whoever initiates a bill—the government, back-
benchers, or a member of the upper house—the government is the key legislative
actor. Parliamentary deliberations on most bills, and on all finance bills, begin in the
House of Commons.2 A bill is typically subject to three readings. It is usually examined
by a parliamentary committee following the second reading, that is, after the stage at
which a vote is taken on the principle of the bill (see Norton 1993b: 75). After the third
reading, the bill goes to the House of Lords (if it originated in the Commons). Despite
some evidence of increased assertiveness of the House of Lords since the first stage of
its reform in 1999 (cf. Blackburn and Kennon 2003: 710), in practice it is still generally
accurate to say that ‘most amendments prove acceptable to the other chamber and, in
the event of a clash, the House of Lords usually defers—though not always immedi-
ately—to the elected chamber’ (Norton 1993b: 77). Once both Houses have approved a
bill, it is sent to the Queen who, by constitutional convention, gives her Royal Assent.
The policy areas subject to a significantly different legislative process are those that
fall under Royal Prerogative powers, the budget, money bills, and delegated legislation.
Royal Prerogative powers vested in the monarch include the right to declare war or peace,
the power to pardon convicted criminals, the right to dissolve Parliament, the power
to approve international treaties3, and the right to confer Honours upon citizens. In
practice, ministers acting as advisers to the Crown exercise these powers. So-called
money bills, as certified by the Speaker, become law one month after leaving the
Commons, whether approved by the Lords or not. This also applies to the budget,
whose passage largely follows the regular legislative procedures albeit with a number
of qualifications: First, all proposals for expenditure or taxation must be demanded or
recommended by the government (formally by the Crown) before they can be con-
sidered. Second, all charges must first be considered by the House of Commons but
must also be embodied in legislation for approval by both Houses of Parliament. Third,
2
Nevertheless, a number of bills each session are introduced in the House of Lords, mainly to reduce the
imbalance in the workload of the two Houses.
3
Treaties do not require approval by either House of Parliament, and in the great majority of cases they
are not debated in Parliament. However, under what is known as the ‘Ponsonby rule’, when a treaty requires
ratification the government does not usually proceed with ratification until more than twenty-one days have
passed since the date on which the text of the treaty was laid before Parliament.
The United Kingdom 623
supply must be voted annually (Blackburn and Kennon 2003: 730–1). Delegated legisla-
tion is another area in which the process of legislation differs from ordinary public bills.
Delegated legislation is law made by ministers exercising powers given to them by a
parent Act. A joint committee of both Houses of Parliament considers delegated
legislation laid before Parliament, with a view to determining whether the attention of
the House should be drawn to it. With very few exceptions, delegated legislation
cannot be amended by either House. The government may, however, be persuaded to
withdraw and reintroduce a defective measure. The Parliament Acts do not apply to
delegated legislation, so the Commons cannot override the Lords in this field (Griffith
and Ryle 1989: 244–6).
P O L I T I C A L PA RT I E S
Political parties are key actors in the British political system, which has classically been
characterized as ‘responsible party government’ (Pulzer 1987). Parties offer choices
over competing policies and leaders. They have a virtual monopoly in the recruitment
of government personnel and play an important role in the screening of parliamentary
candidates. The parliamentary majority party delegates powers to its leadership, which
forms the government, and ensures a degree of internal accountability. The opposition
parties have incentives to subject the government’s activities to continuous criticism
and thereby contribute to a degree of public accountability (for a general discussion see
Müller 2000). Since the 1960s, three important developments have been particularly
important as far as the role of political parties is concerned:
1. Since the early 1970s the British party system has changed significantly as a result
of changes in voting behaviour. This had serious implications for the electoral account-
ability of the Conservative governments of the 1980s. The decline of class voting,
greater electoral volatility, regional disparities in voting behaviour, and the crisis of the
Labour Party all led to a drastic decline in the combined vote share among the two
major parties. The main beneficiaries have been the Liberal Party (until 1981), the
SDP/Liberal Alliance (1981–7), and, since 1988, the product of the merger of these two
parties, the Liberal Democrats. In addition, the regionalist parties in Scotland and
Wales have gained as a result of the declining electoral fortunes of the two major par-
ties (Kavanagh 1994). Under a first-past-the-post electoral system these developments,
especially the split in the anti-Conservative vote from 1974 onwards, led to a
Conservative hegemony which lasted for almost two decades (cf. Kavanagh 1994:
601–2, 608–9). The Conservative Party was protected from the consequences of
adverse electoral swings, although its vote share never surpassed 43 per cent. Today the
situation is reversed. The Conservative defeats of 1997 and 2001 were of a magnitude
that it may take several Parliaments to draw the appropriate organisational lessons and
develop a credible electoral strategy, with similar negative implications for the (ex post)
electoral accountability of the Labour government.
These developments have had obvious, and serious implications for democratic
delegation and accountability. The threat of losing the majority—which had been high
624 Thomas Saalfeld
until the end of the 1970s due to an electoral system capable of transforming relatively
small electoral swings into considerable changes in seat allocation—ceased to be as
powerful a constraint on government parties. Frequent alternation of power—one of
the conditions that lend legitimacy to strong majority rule in the absence of extensive
checks and balances (Lijphart 1984: 22)—was no longer guaranteed. Especially the
eighteen years of uninterrupted Conservative government (1979–97) eventually provided
powerful arguments for the constitutional reforms of 1997–2001, particularly with
regard to limiting ‘governmental excess’. In the eyes of many critics the scale and radi-
calism of the Thatcher government had turned ‘the anxiety over “ungovernability” in the
1970s...to a fear of governmental over-capacity as Margaret Thatcher appeared repeat-
edly to prevail over the system’s checks and balances’ (Foley 1999: 50).
2. Since the 1960s, both major parties have undergone a number of radical organ-
izational changes affecting intraparty democracy. These changes did not result from
changes in the law, but from internal developments. Following the electoral defeat in
1979 the Parliamentary Labour Party (PLP) lost its monopoly of selecting the Labour
leader, initially to an electoral college including representatives of the PLP, trade unions
and other affiliated organizations, and the Constituency Labour Parties (1980–92), and
later to the Labour Party membership.4 In 1994, Blair was the first party leader to be
elected by a ballot of all individual party members, Labour MPs, and MEPs (cf. Seyd
1998). Also, sitting Labour Members of Parliament were made more accountable to the
members of their respective Constituency Labour Party (CLP). Mandatory reselection
of sitting Labour MPs between 1980 and 1990 meant that sitting Labour MPs were rou-
tinely obliged to go through the process of being reselected by their constituency party
before each general election. Although the compulsory reselection of sitting MPs was
abolished in 1990, the involvement of affiliated organizations (mainly trade unions) and
constituency activists in candidate selection and leadership election was significantly
enhanced through the 1980 changes to the Labour Party constitution. Labour leaders
Neil Kinnock (1983–92), John Smith (1992–94), and Tony Blair (1994–) pushed through
a number of further organizational reforms strengthening the direct link between lead-
ership and members at the expense of intermediary bodies and organizations such as
the National Executive Committee (NEC) and, more importantly, the trades unions
(see Kelly 1994; Fisher 1996; Seyd 1998).
In the Conservative Party, the main organizational innovation was the introduction
in 1964 of leadership elections by the parliamentary party and the establishment
of formal rules for the dismissal of the Conservative leader by the parliamentary party.
As a result, the Conservative Parliamentary Party became the ‘principal’ who elected
and dismissed Conservative leaders. After the 1997 electoral defeat, the new leader of
the Conservative Party, William Hague, initiated further reforms introducing the first
unified party organization which integrates all three levels of the party: parliamentary
party, Central Office, and extra-parliamentary party organization (see Webb 1994: 110).
In 1998 the Parliamentary Conservative Party decided that future party leaders
4
Between 1981 and 1993, the PLP had a weight of 40% the affiliated organizations of 30%, and the CLP
30%. Since 1993, each of the three electoral bodies has had one-third of the votes.
The United Kingdom 625
D E L E GAT I O N F RO M VOT E R S TO M P s
The 659 Members of the House of Commons (2001) are elected in single-member dis-
tricts using a simple plurality ‘first-past-the-post’ electoral system with powerful, well-
known implications for party competition and government accountability. It reduces
the number of parties with effective parliamentary representation to just two, one of
which tends to get a clear overall majority of seats. If the parties are almost equally
strong, this tends to enhance competition and electoral accountability. Voters have a
relatively direct influence on the composition of the next government. They do not have
to grant a great deal of discretion to their agents in Parliament. In multiparty systems in
which no single party enjoys an overall majority, MPs may have some discretion in
selecting a government (De Winter 1995: 117–19). In Britain, MPs may influence the
choice of parliamentary leaders, but the voters effectively retain the right to elect the
government. Yet, critics of the first-past-the post system point at least to three problems:
1. The system is widely seen to be unfair (even amongst those who defend it) and
distorts electoral competition in favour of the two major parties. The threshold
for entry to the House of Commons is very high. This ‘enables national
parties to ignore many political movements and the demands they voice’ (Budge
1996: 25).
2. If the vote against one of the major parties is split, this major party is likely to be
in a hegemonic position. This reduces the chance of government alternation with
problematic consequences for democratic accountability, as the long years of
Conservative dominance (1979–97) demonstrate.
626 Thomas Saalfeld
3. The aggregate outcome of elections may be so irrational that some authors have
described elections as ‘the other National Lottery’ (Weir and Beetham 1999: 45).
The usual examples are the 1951 and 1974 elections, where the party with the
highest aggregate number of votes lost the election.
One of the most important advantages of the first-past-the-post system is the potentially
close personal link between MPs and their constituents. This is particularly important
in the absence of administrative courts and other redress mechanisms for citizens. In
theory, the system would also allow voters to hold individual MPs accountable. Yet,
although Members’ constituency work has increased in the past decades and may
significantly effect their re-election chances, the effect is generally moderate (Norton
and Wood 1993: 141). Voters cannot effectively punish MPs except at the high cost of
deserting their preferred party.
Although divided and hesitant on changes to the electoral system the Blair govern-
ment set up an Independent Commission on the Voting System under Lord Jenkins to
make recommendations for electoral reform. This commission reported in 1998 and
recommended a form of proportional representation (alternative vote with top-up
seats), but the government did not decide to adopt the recommendations. For all
regional (Scotland, Wales, Northern Ireland, Mayor of London) and European elec-
tions, different forms of proportional representation are now used. This constitutes a
significant departure from the Westminster model at the regional and European levels.
One of the peculiarities of British bicameralism is the fact that the upper house has
no direct agency relationship with the electorate, although as a revising and delaying
chamber it can have considerable influence on legislation. It consists of hereditary
peers, persons who have been awarded a lifetime peerage by the monarch on the PM’s
recommendation (‘life peers’), and Bishops of the Anglican Church, as well as current
and former Law Lords, the most senior judges of the country. The Conservative Party
has traditionally held at least a plurality of the seats in the chamber. In January 1999 the
House of Lords Act, the first stage of the Lords reform proposed by the Blair govern-
ment, removed the automatic right of the hereditary peers to sit in the upper house.
More than 700 hereditary peers lost their seats overnight, although under the so-called
‘Weatherill amendment’ ninety-two hereditary peers retained their seats until the
second stage of reform which failed in 2003 when none of the six alternative proposals
for the second stage put to both Houses achieved a parliamentary majority in both the
Commons and the Lords. The status quo of a largely appointed chamber with residual
hereditary element prevailed therefore. The rejection of all proposals stipulating vari-
ous percentages of elected members in these votes led to an outcome close to the Blair
government’s ideal point of an entirely appointed Upper House. The House of Lords
is therefore likely to remain less accountable to the voters and hence of inferior demo-
cratic legitimacy.
The Constitution and electoral laws do not call for primaries. Nor do electoral laws
prescribe intra-party candidate preference voting. The maximum length of the term
of the House of Commons is five years. However, the PM can request the monarch to
dissolve the House of Commons at any time before the end of its term. There is no
The United Kingdom 627
term limit for MPs and no mandatory retirement age. Aliens and persons under 21
years of age are disqualified from being elected. Further disqualifications include
lunacy, bankruptcy, treason, criminal convictions with a sentence of more than 1 year,
conviction of corrupt practices at elections, holding certain offices (e.g., in the civil
service or the armed forces), and membership of the House of Lords. There is no
legally enforced recall mechanism, though the House of Commons will expel a
Member if one of the qualifications listed above applies (Blackburn and Kennon 2003:
67–8). Members lose their seats if they are convicted of criminal charges or treason, if
they are certified insane for more than six months, or if they go bankrupt. Members of
the House of Commons who, by inheritance or appointment, become Peers of
England, Scotland, Great Britain, or the United Kingdom must resign their seats (see
Griffith and Ryle 1989: 92).5
The rules of electoral competition are relatively simple. According to the
Representation of the People Act of 1918, section 21(3), twenty full days must pass
between the proclamation of the dissolution of the old Parliament and the first meeting
of the new. During this interval, a general election is held (Griffith and Ryle 1989: 180).
In 1983, this law was amended to reduce the minimum period to sixteen days. This
implies an even greater advantage for a PM who has the power to dissolve Parliament
at any time. Individual candidates’ campaign spending is severely limited. Until 1998
there was no legal cap on spending by the parties. The reforms of 1998 introduced lim-
its for the parties that will apply for the 365 days prior to a general election and the four
months prior to other elections (Blackburn and Kennon 2003: 69). During the cam-
paign, both parliamentary and non-parliamentary parties receive free broadcasting
time. The allocation is largely proportional. The main opposition party (Her Majesty’s
Official Opposition) receives the same amount of time as the party of government.
Non-parliamentary parties that contest a minimum of fifty constituencies are allocated
broadcasting time in proportion to the number of candidates fielded.6 There is no
unrestricted purchase of broadcasting time. The publication of opinion polls is not
restricted except for the publication of exit polls on polling day.
Public party financing is limited. Until 1975, there were no public subsidies to either
parliamentary or extra-parliamentary parties, except the salaries (since 1911) and
allowances (especially for office costs) paid to MPs, and salaries paid to the Leader of
the Opposition both in the Commons and the Lords, the Opposition Chief Whip,
(Commons and the Lords) and two Assistant Whips from the main Opposition party in
the Commons (Blackburn and Kennon 2003: 173). Since 1975, the opposition parlia-
mentary parties have received some public subsidies. The distribution of these subsidies
between the opposition parties is roughly proportional to their share of seats now cal-
culated on the basis of a formula consisting of the number of votes and seats (for data
see Blackburn and Kennon 2003: 171). The amounts were increased in the 1998 party
5
Irish Peers may sit in the House of Commons.
6
In the 1997 general election there were seven such parties: the Referendum Party, the Natural Law Party,
the Pro-Life Alliance, the Liberal Party, the British National Party, the Green Party, and the Socialist Labour
Party.
628 Thomas Saalfeld
finance reform, which also grants both government and opposition parties support for
the development of policies for inclusion in their election manifestos from an annual
fund of £ 2 million (Blackburn and Kennon 2003: 172; Hazell et al. 2000: 255).
Candidate selection is mainly governed by the parties’ constitutions. None of the
major parties constrains candidate selection through term limits or a mandatory retire-
ment age. Both parties rely partly on formal procedures and, most importantly in our
context, partly on informal screening devices before the actual selection process starts.
A prospective candidate’s activities in the party, previous political record outside
Parliament (e.g. in local government), and general political and personal reputation
help the parties assess his or her suitability and reduce the danger of adverse selection
(cf. e.g. Rose 1987: 75). In both parties, the formal selection process is carried out by the
local Constituency Parties, whereas the party leadership retains certain rights to screen
and/or veto candidates. The Conservative Party’s Central Office controls admission to
the pool of prospective candidates, but has no veto over candidates once they have been
adopted by the constituency party (see Fisher 1996: 45). In the Labour Party central
ex ante controls are weaker than in the Conservative Party. Yet, Labour’s National
Executive Committee has the power to veto candidates after they have been adopted
by the Constituency Labour Party, a right that they have used with greater frequency
since the mid-1980s (Kelly 1994: 34; Fisher 1996: 73–4). Thus, ex post central control
over candidate selection seems to be more far-reaching in the Labour Party than in the
Conservative Party. The choice of a closed-list system of proportional representation
for elections to the European Parliament in 1999 has enhanced the powers of the party
head-quarters in such elections.
D E L E GAT I O N F RO M PA R L I A M E N T TO C A B I N E T
In the delegation of powers from Parliament to cabinet, the former is in a weak
position. The information asymmetry between government ministers and backbench
MP is extreme. The government is aided by the civil service, which has a vastly supe-
rior capacity to generate and process information. Ordinary MPs, by contrast, have
comparatively little independent information. Their resources in terms of research and
secretarial assistance are modest by international standards, let alone in comparison to
the government. While MPs may be aided by consultants, interest groups, or press
coverage, their ability to generate information independently remains severely limited,
especially on the opposition benches (cf. Commission to Strengthen Parliament 2000:
34). Although a great deal of government information is disclosed voluntarily, access to
information remains restricted by far-reaching secrecy laws and norms and depends, to
some extent, on the government’s good will. Until 1989, section 2 of the Official Secrets
Act of 1911 protected all government information from unauthorized disclosure. Thus,
in principle all unauthorized disclosures of official information were criminal offences.
This section was repealed in 1989. Although the Official Secrets Act of 1989 decrimi-
nalizes most disclosures, ‘Government practice, the civil service code of discipline and
the operations of the civil law of confidence continue to promote a highly secretive
ethos’ (Oliver 1991: 173). The Freedom of Information Act passed in 2000 has been
The United Kingdom 629
7
As long as the Conservative leaders ‘emerged’, the monarch had also some discretion in situations in
which there were doubts about the majority party’s leadership. In 1923, after the death of Andrew Bonar Law,
the monarch effectively determined the leadership of the Conservative Party by summoning Stanley Baldwin
(and not Lord Curzon).
630 Thomas Saalfeld
and Ryle 1989: 42). Votes of no confidence are directed against the PM and the cabinet
collectively, not against individual ministers. A single MP can propose such a vote. By
convention (as with any substantive motion), at least one day’s notice is required for the
motion to appear on the Order Paper. Once put to the House, it is then debated and, if
necessary, settled by a division. A debate on any no-confidence motion will normally
take precedence over the normal business of the day. There are no formal constraints
on the number of no-confidence proposals that can be proposed per session. Quorum
requirements are the same as for any other motion (Limon, McKay et al. 1997).
The government itself can call for a vote of confidence. This has happened only
three times between 1945 and 1999, while there were twenty-seven votes of no confi-
dence called by the opposition (Butler and Butler 2000: 201). The motivation for a vote
of confidence has usually been to consolidate a government’s parliamentary support in
the face of considerable opposition within its own ranks. The rules guiding the use of
confidence motions are similar to the ones for no-confidence motions. They are largely
guided by convention. The object of the vote is the PM and his or her whole cabinet,
not an individual minister. The right to propose a confidence motion rests, in effect,
with the PM. By convention, at least one day’s notice is required for the motion to
appear on the Order Paper, although in exceptional cases it could be called without
notice. The motion is then debated and settled using a division, if necessary. There is
no specific quorum requirement.
Table 21.1 summarizes some behavioural data on the life of cabinets and
Parliaments. The dissolution of Parliament is partly governed by law, partly by cons-
titutional conventions and the Royal Prerogative. A Parliament ends by a Royal
Proclamation, which also orders the issue of writs for the election of a new Parliament.
In most cases, ‘the time chosen is one which the Prime Minister considers most likely
to result in a victory at the election. But the Prime Minister may advise dissolution
because the Government is defeated when seeking a vote of confidence. Alternatively
the Opposition may carry a motion of no-confidence in the Government or of censure.
Also the Government may say in advance that it regards a particular vote as one of con-
fidence’ (Griffith and Ryle 1989: 43). By convention the monarch dissolves Parliament
on proposal of the PM only, and not in his or her own right. Although only the PM can
request dissolution, he or she cannot dissolve Parliament unilaterally but needs the
assent of the monarch. While the monarch by convention accedes to this request, it
cannot be taken for granted. The monarch must decide for herself that no one else
could form a government without dissolution, particularly when Parliament was
elected very recently.8 Applying the cross-national counting rules employed in this
volume, a total of nine cabinets ended in early elections after the PM dissolved
Parliament. Note, however, that British standards are much more relaxed—cabinets
are not expected to sit until the very end of the parliamentary term. Hence, by con-
ventional British standards only the cabinets Attlee II, Wilson I, and Wilson III were
ended by early parliamentary dissolution.
8
For example, it has been suggested that in 1974 Wilson might have been denied a dissolution if he had
asked for it only a month or two after forming a government.
Table 21.1. Cabinet Formation, Cabinet Termination, and Early Parliamentary Dissolution: United Kingdom 1945–2000 a
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the cabinet Who
formation votes (iv) votes (cv) under end with an dissolved
n.a. No. of ncv Cabinet Cabinet specific early election? Parliament
removed resigned to constitutional in this
by ncv pre-empt ncv mechanism case? b
n.a.
1 Attlee I 45/07/26 1 No
2 Attlee II 50/03/01 0 Yesc PM
3 Churchill II 51/10/26 1 No
4 Eden I 55/04/06 0 Yes PM
5 Eden II 55/06/07 2 No
6 Macmillan I 57/01/10 0 Yes PM
7 Macmillan II 59/10/20 2 No
8 Douglas-Home 63/10/13 0
9 Wilson I 64/10/16 3 No Yes PM
10 Wilson II 66/04/18 3 No Yes PM
11 Heath 70/06/29 2 No Yes PM
12 Wilson III 74/03/04 0 Yes PM
13 Wilson IV 74/10/22 0
14 Callaghan 76/04/05 3 Yes
15 Thatcher I 79/05/04 4 No Yes PM
Table 21.1. (contd.)
No. Cabinet Date of Investiture No-confidence votes (ncv) Confidence Did the cabinet Who
formation votes (iv) votes (cv) under end with an dissolved
n.a. No. of ncv Cabinet Cabinet specific early election? Parliament
removed resigned to constitutional in this
by ncv pre-empt ncv mechanism case? b
n.a.
9
I am quoting passages from the 1994 version. The Blair government has amended some of the details of
this document.
634 Thomas Saalfeld
days’, debates on motions of censure, parliamentary questions, debates on private
Members’ Bills, adjournment debates, motions for recess, and departmental select
committees. The Queen’s Speech from the Throne sets out the government’s business for
the session. On the second day, the Leader of the Official Opposition determines the
issues for debate in his reply, which is usually a general attack on the government’s pol-
icy. More important in quantitative terms are the twenty Opposition Days per session.
On seventeen of them, the largest opposition party (Her Majesty’s Official Opposition)
chooses the topics for debate, on three days the choice of topics is made by the second
largest opposition party. Again, these debates allow the Opposition to break the
government’s usual agenda-setting monopoly and to force the government to account
for its policies on issues determined by the Opposition. By convention, if the Official
Opposition tables a motion of censure on the government, the government provides time
for it to be debated. Finally, there are emergency adjournment debates under Standing
Order No. 20. These allow backbenchers and the opposition leaders to bring urgent
issues of their choosing before the House. While the introduction of private Members’
bills is severely limited (there is a ballot each session and there are bills introduced under
the ‘ten minute rule’ and there is a possibility for ‘ordinary presentation’ of private
Members’ bills), their first (and, if granted, second) reading gives Members an oppor-
tunity to bring an issue onto the parliamentary agenda.10
While the instruments mentioned above help to limit the government’s monopoly
control over the parliamentary agenda, their information value is limited. Yet redress-
ing the information asymmetry between government and opposition is critical for
effective scrutiny and accountability. Although the number of oral questions to ministers
has increased dramatically during the postwar period, only a minority of MPs think
they are well-suited to get hard-to-obtain information (Franklin and Norton 1993: 107).
Ministers answer oral questions in a rota from Mondays to Thursdays.11 Until 1997,
PM’s Questions were scheduled from 15:15 to 15:30 every Tuesday and Thursday. In
1997, the new Labour government changed this to thirty minutes every Wednesday.
Although Prime Ministers and ministers often fear Question Time, its value as a source
of information is limited. In particular, ministers have ample opportunity to give evasive
answers. They can also refuse to answer on grounds of national security and often encour-
age government backbenchers to ask so-called ‘planted questions’ with no other purpose
10
Standing Order No. 14 allows twenty private members’ bills per parliamentary session to be drawn in
a ballot. The ballot establishes an order of priority enabling those successful in it to use the limited private
Members’ time on Fridays. Private Members’ bills can also be introduced by ‘ordinary presentation’ according
to Standing Order No. 57. Such bills are not likely to be high up the list on private Members’ bills on Fridays.
Finally there is a possibility to introduce private Members’ bills under Standing Order No. 23, otherwise
known as the Ten Minute Rule. The Ten Minute Rule allows a brief speech in favour of the Bill by the
Member introducing it as well as a speech by one Member opposing the motion. This has to be made in the
House after Question Time (at or shortly after 3.30 p.m.) on Tuesdays and Wednesdays.
11
A ‘rota’ simply means that not all ministers are questioned on the same day. Ministers from two
Departments answer questions concerning their departments on one day. On the following day, questions to
two other departments are being answered by ministers from these departments. And so on. So there is
a relatively established (if not inflexible) ‘sequence’ in which departments can expect to have to answer
questions, that is, each department has to answer questions approximately once every two weeks.
The United Kingdom 635
than to allow a government minister to put his or her own view on the record and—as
the time for questions is severely limited—‘squeeze out’ critical questions. Refusal or
inability to give an answer will, however, invariably be exploited by the Opposition and
may lead to considerable public embarrassment in front of television cameras. Answers
to oral questions demonstrate debating skills, but the scoring of partisan points domi-
nates over the extraction of information. Since the number of questions usually
exceeds the time available for oral answers, many oral questions have to be answered
in written form. In addition, MPs may ask questions for written answer, which are printed
in the Official Report (‘Hansard’) together with a department’s answer. Despite their
inflationary use and limited effectiveness as a device to obtain information, back-
benchers’ questions may serve a very important, unintended function: they often provide
ministers with low-cost information about policy implementation (Chester 1981: 189).
Thus, parliamentary questions fulfil a purpose similar to McCubbins and Schwarz’s
(1984) ‘fire-alarm controls’, a function that is often under-estimated in the literature.
The resources of departmental select committees have been improved over the years,
most notably under Robin Cook as Leader of the House and the cabinet minister in
charge of the modernisation of Parliament in 2002 (Blackburn and Kennon 2003: 770).
The House’s most important instruments for holding the government accountable are
select committees. There are different kinds of select committees (cf. Norton 1981: 127–37).
We are particularly interested in those specializing in ex post scrutiny of the policies and
administration of a particular department. In 1979, for the first time in the history of
the House of Commons, a near-comprehensive system of departmentally related select
committees was set up. These committees have the authority to call persons to appear
before the committee, to request papers and records, and to scrutinize government
policy. They conduct their inquiries largely in the style of McCubbins and Schwarz’s
(1984) ‘police-patrol controls’ and are (with an occasional exception) not involved in
legislation (see Drewry 1989).
Although the departmental select committees have had little impact on specific
government policy, they have carried out thousands of investigations into the govern-
ment’s and government agencies’ policy implementation and expenditure. These
reports provide the Members of the House of Commons and the interested public with
highly valuable, ‘low-cost’ information and have greatly improved government
accountability. Nevertheless, there are severe limits on their activities. The select com-
mittees can invite Members of either House of Parliament to appear before them, but
cannot compel them to do so. As a consequence of Parliamentary Privilege, the select
committees may order a private individual to answer its questions (and enforce its
order), yet it has no authority to order a Member to appear. This authority resides with
the full House of Commons. Of course, a minister’s refusal to cooperate may be
brought to the attention of the House and thus cause embarrassment. Although select
committees often send for specific civil servants, the minister usually decides which
civil servant will represent the department in a committee hearing. Also civil servants
are guided by the rather restrictive Memorandum of Guidance for Officials Appearing before
Select Committees (the so-called ‘Osmotherly Rules’). These guidelines limit, among other
things, the disclosure of advice given by civil servants to ministers, interdepartmental
636 Thomas Saalfeld
exchanges concerning policy, and alternative policy options (see Woodhouse 1994:
177–217). In sum, therefore, Woodhouse concludes:
The reform of the select committee system has increased the amount of government information
made publicly available; it has also provided a routine departmental accountability, and at times
allows a parliamentary input into policy-making. However, in situations where the government
considers it politically expedient to withhold information, even if this appears to obstruct a
committee’s inquiry, then the power of select committees to send for persons, papers, and
records is of little consequence. (Woodhouse 1994: 217)
D E L E GAT I O N W I T H I N T H E C A B I N E T
The mechanisms of delegation and accountability within the cabinet have long been
interpreted in different, sometimes contradictory ways, although there seems to be
agreement that the traditional interpretation of the Constitution as ‘cabinet government’
( Jennings 1959) is outdated:
1. One influential interpretation is that postwar British politics has witnessed a
transformation from cabinet government to prime ministerial or even ‘quasi-
presidential’ government (cf. Crossman 1963: 51). The PM is de facto seen as the
ministers’ principal with considerable autonomous ex ante and ex post control
mechanisms at his or her disposal.
2. The opposite conclusion is often reached by scholars who emphasize the
tendency towards fragmentation of decision-making and departmentalism. The
superior expertise in the departments and policy networks around them renders
any hierarchical principal–agent construction an illusion (cf. Rhodes 1997).
3. A number of authors (e.g. Norton 1993a) have tried to resolve the apparent
tension between the first two interpretations by identifying different prime
ministerial leadership styles, some more autocratic, others more reliant on cabi-
net as a ‘team’.
The United Kingdom 637
4. The research on the ‘core executive’ in Britain in the 1990s (e.g. Rhodes 1997;
Smith 1999) criticizes the traditional focus on the respective resources of PMs and
cabinet ministers in isolation from each other and further actors in the political
system. Smith (1999: 105) argues that ‘[t]he Prime Minister can neither govern...nor
actually make decisions without dependence on a whole range of other actors
and institutions . . . What the Prime Minister does is play a tactical game that
creates the necessary alliances for achieving goals’.
D E L E GAT I O N F RO M M I N I S T E R S TO C I V I L S E RVA N T S
In British constitutional theory, politicians decide on policy, whereas officials give
advice and implement policy (Griffith and Ryle 1989: 27–8; Smith 1999: 107). The
model of the British civil service is very much one of ‘neutral competence’ (Moe and
Caldwell 1994: 187), in which highly professional civil servants serve the party in power.
In such a system, based on merit and seniority, ministers’ ex ante powers in the form of
political appointments, screening, or contract design are generally low, despite a
stronger effort under Thatcher and Blair to politicize key civil service positions. Since
British ministers tend to be ‘generalists’ (Rose 1987: 81) and—due to relatively frequent
cabinet reshuffles (Hennessy 1989: 492)—hold a portfolio for only a few years, the infor-
mation asymmetry between ministers and civil servants, and the scope for hidden
The United Kingdom 639
E X T E R NA L C O N S T R A I N T S
It is the essence of the Westminster model that there are hardly any formal institutional
checks on the democratically elected government. The monarch, as Head of State, is no
longer a constraint. The doctrine of sovereignty of Parliament implies that the Courts
are not able to challenge Acts of Parliament. Nevertheless, the centrifugal pressures of
the New Public Management have increased the use of contracts and legal or quasi-
legal agreements and led to an increased role of the judiciary as an enforcement mech-
anism in the UK since the 1980s. In addition, judges—who had traditionally acted within
the limits set by parliamentary legislation and sought to ensure that ministers have not
acted beyond the powers Parliament has delegated to them—have become more active
since the 1980s and decided cases by reference to fundamental rights rather than
upholding the will of Parliament. The Human Rights Act 1998—incorporating the
The United Kingdom 641
European Convention of Human Rights 1951 into British law—has reinforced this
trend ‘with potentially fundamental consequences for traditional mechanisms of
accountability’ (Flinders 2001: 155) in the Westminster system replacing, as Flinders
(2001: 158) claims, parliamentary with judicial accountability in some areas. Although
the Act does not empower the courts to strike down legislation, they can issue ‘decla-
rations of incompatibility’. The exercise of such declarations is identical in process, if
not in effect, to a judicial review of primary legislation (Blackburn and Kennon 2003:
759). Thus, the developments of the 1980s and the Human Rights Act 1998 (which has
come into force in 2000) have strengthened judicial accountability significantly and
increased the demands for a separation of judiciary and parliament and even the
creation of a supreme court (cf. Blackburn and Kennon 2003: 759–60).
The three major devolution Acts of 1998 set the framework for new institutions in
Scotland, Wales, and Northern Ireland. In 1999 the citizens in Scotland and Wales
followed Northern Ireland in electing regional assemblies by proportional representa-
tion, which in turn (delayed and repeatedly interrupted in the case of Northern Ireland)
elected regional executives.12 The Scottish Parliament and the Northern Ireland
Assembly have law-making and limited tax-varying powers. The National Assembly of
Wales, by contrast, only enjoys powers of secondary legislation and is entirely depend-
ent on an annual block grant from London. Initially these assemblies and the regional
governments emanating from them are unlikely to be a serious institutional check on
the sovereignty of Parliament. The piecemeal nature of this reform is indicated by the
fact that to date (2003) the English regions do not have equivalent representative assem-
blies. To what extent devolution will generate a momentum of its own and the assem-
blies will incrementally develop into serious institutional checks remains to be seen.
Between 1948 and 1997, the Bank of England was not an independent actor as is the
American Federal Reserve or the German Bundesbank, constraining the government’s
economic policy. The Governor of the Bank of England was and still is, appointed by
the government. Monetary policy and interest rates were controlled by the Treasury,
which instructed the Bank of England. One of the first decisions of the incoming
Labour government in 1997, however, was to transfer the determination of interest
rates to a Monetary Policy Committee, which is chaired by the Governor of the Bank
of England and includes representatives of the Bank of England as well as academic
experts. Since May 1997, therefore, the Bank of England has had a more independent
role in the setting of interest rates, although the Chancellor of the Exchequer still sets
global targets for inflation. Critics of this transfer of authority have pointed
out that it delegates key economic decisions to an unelected and democratically
unaccountable body. Yet, as William Bernhard (1998: 314) emphasizes, an independent
central bank can be an important source of information on the direction and
12
In none of the three assemblies does a single party control an overall majority. The Northern Ireland
Assembly is characterized by a complicated power-sharing system between the main Unionist and Roman-
Catholic parties. At the time of the writing (2003), the Scottish and Welsh Executives are based on a formal
coalition between the Labour Party and the Liberal Democrats.
642 Thomas Saalfeld
consequences of government economic policy. In their public justification of monetary
decisions, independent central banks frequently comment on government policy,
provide alternative economic analyses and forecasts, and force governments to be more
open about the inflationary effects of their policy choices.
Referendums were held by the Labour governments in the 1970s. Nonetheless, they
were exceptional instruments entirely in the hands of the government. The Blair
government has made use of the referendum (on the election of regional assemblies
for Scotland and Wales in 1997, and an elected Mayor for London as well as the peace
process in Northern Ireland in 1998) and has promised to do so in the future (Britain’s
entry into European Monetary Union). It is important to note, however, that a refer-
endum is held exclusively at the request of the government of the day. In addition,
a referendum in Britain is, strictly speaking, only advisory in character. Obviously,
however, a popular vote on a policy issue creates a certain degree of pressure on the
government.
A more serious domestic constraint on the government is the existence of ‘policy
communities’. Rhodes (1997: 3) argues that the Westminster model ‘no longer provides
either an accurate or comprehensive account of how Britain is governed’ and suggests
that ‘interdependence, a segmented executive, policy networks, governance and hol-
lowing out’ characterize the British state. This makes the British political system more
reminiscent of a ‘differentiated polity’ than the traditional Westminster model with its
‘strong cabinet government, parliamentary sovereignty, Her Majesty’s loyal opposition
and ministerial responsibility’ (Rhodes 1997: 7). From the 1970s onwards, empirical
research has generated considerable evidence (e.g. Richardson and Jordan 1979; Jordan
1981; Rhodes 1997) that this development, which accelerated during the 1980s and
1990s, has led to a situation in which more central control is exerted, ‘but over less’
(Rhodes 1997: 3), and where democratic accountability has become increasingly prob-
lematic. In particular, the Westminster model grossly overestimates government auton-
omy (Rhodes 1997: 3). Although Conservative governments between 1979 and 1997
attempted to eliminate all traces of neo-corporatism, interest groups and professional
bodies have remained important players in many policy areas simply because their
expertise and cooperation are needed. Policy is often made in sectoral policy networks
composed of central government departments, professions, voluntary organisations,
and other key interests. Privatization and the establishment of regulatory bodies super-
vising privatized industries, the establishment of Next Steps agencies, and a host of spe-
cial-purpose bodies with considerable resources and operational discretion have
replaced hierarchical coordination and have turned governing into a more complex bar-
gaining process. This poses problems for democratic delegation and accountability as
the bodies in charge of monitoring and regulating privatized services and semi-
autonomous agencies are highly dependent on information provided by the regulated
(Rhodes 1997).
A key international constraint on the British government is its membership in the
EU. Britain’s entry into the European Economic Community in 1973 has added a new
political and judicial dimension to its Constitution. The 1972 European Communities
The United Kingdom 643
Act provided that, in the event of a conflict between the provisions of EC law and
domestic UK law, the former was to prevail. This obviously sits ill with the doctrine of
parliamentary sovereignty and explains why many British politicians and constitutional
lawyers find it difficult to accept any strengthening of supranational elements in the
EU’s decision-making process. In order to scrutinize EC/EU legislation, both British
Houses of Parliament have set up scrutiny committees. The House of Commons has a
European Scrutiny Committee, which identifies politically significant EU documents
and European Standing Committees which carry out detailed scrutiny, often with rele-
vant ministers in attendance (Blackburn and Kennon 2003: 352). The European
Scrutiny Committee consists of sixteen MPs. It has the power to investigate Pillar I, II,
and III matters and can refer a matter either to a European Standing Committee or
(more rarely) the House as a whole. Reports from the European Standing Committees
are rarely debated. Legally, the reports do not bind the government. Nevertheless, the
government has to make sure that its positions on EU policies are, at a minimum,
acceptable to its own backbenchers. The Select Committee of the House of Lords, with
its six specialised sub-committees, has better resources with regard to members’ expert-
ise, staff, advisers, and time. Therefore, it has generally been considered the more effec-
tive scrutiny committee. Nevertheless, documents often do not reach the committees in
both Houses in time for them to be considered and reported on before agreement is
reached in the Council of Ministers. As a result, the theoretical control these commit-
tees have over ministerial actions often cannot be exercised in practice. The govern-
ment therefore agreed in 1980 that a minister would have to justify the adoption of a
document in the Council of Ministers without prior debate in the House of Commons
(cf. Griffith and Ryle 1989: 438).
C O N C LU S I O N
The ‘core’ of the Westminster chain of democratic delegation and accountability has
remained tremendously stable in a country whose economic and political institutions
have undergone considerable change since 1979. As a result of the electoral system, the
voters’ aggregate choice, mediated by disciplined political parties, continues to have a
very direct impact on the selection of the government of the day. Normatively, this may
justify the fact that government, as the agent of Parliament and the electorate, faces few
domestic political constraints. Accountability relies heavily on informal screening. MPs,
the voters’ agents, and government ministers, Parliament’s agents, undergo lengthy
screening outside and inside Parliament before they are elected or selected for public
office. Parties are the most important ‘screening agents’ in this context.
However, outside the core of the chain, British government has witnessed major
changes. Voting behaviour has changed since the 1960s. The Labour Party, as the
Conservative Party’s main opponent, slipped into a deep crisis during the 1970s and
1980s. These changes contributed to a long spell of Conservative hegemony between
1979 and 1992 with no alternation in government and negative implications for ex post
644 Thomas Saalfeld
electoral accountability.13 If it is true that ‘the basic check on any British Government
is the general election’ (Crick 1970: 38–9), then this development was highly problem-
atic given that the British political system lacks many of the usual formal constitutional
checks and balances found in more decentralised systems. In addition, the nature of
government has changed during the Thatcher and Major years (1979–97). The state
withdrew from many of its previous duties via privatization, market testing and
‘Compulsory Competitive Tendering’. The regulatory bodies that were set up to mon-
itor the performance of market arrangements and ‘framework documents’ defining the
duties of Next Steps agencies make fascinating objects of study for those interested in
delegation and accountability. These changes have led to stronger political control over
smaller government. Governing has become considerably more complex as a result of
the creation of highly inter-dependent networks of ministerial and semi-autonomous
actors. This effect was compounded by Britain’s entry into the EU and the latter’s own
institutional development towards more supranational decision-making.
With the election of the Blair government in May 1997, Britain seemed to be at the
beginning of a new phase of dramatic constitutional change: The newly created assem-
blies for Scotland, Wales, and Northern Ireland were elected by proportional representa-
tion, as were the British Members of the European Parliament (1999). Following the
referendum in May 1998 and elections in June 1998, Northern Ireland had an elected
assembly and an executive based on a broad coalition of all major protestant and Roman
Catholic parties with considerable powers until it was suspended in 2002. The first stage
of the reform of the House of Lords has been implemented, although the government’s
plan to transform it into an appointed chamber in the second stage failed in 2003 due to
lack of support in both Houses. Any ‘democratisation’ of the selection process to the
Upper House even if it is only the diminution of the hereditary principle will enhance its
democratic legitimacy and inevitably strengthen its role as a political actor. The Blair
government began to feel this with a large number of government defeats in the House
of Lords on important pieces of legislation. Attempts to modernize the House of
Commons have been more modest and largely in line with the traditional majoritarian
character of the Westminster Model (Blackburn and Kennon 2003: 741–78).
‘A political party when in opposition tends to become enthusiastic about constitu-
tional and parliamentary reform, the more so the longer they remain in opposition’
(Blackburn and Kennon 2003: 749). This may explain the initial reformist zeal of the
Labour government under Blair after 18 years of opposition. With the manifest weak-
ness of the Conservative opposition in the first two terms of the Blair government, the
latter’s short-term incentives to tie its own hands were quickly diminished, however.
Blackburn and Kennon (2003: 766) therefore spoke of the ‘halting nature’ of the reform
process. This observation primarily refers to parliamentary reform, but can be gener-
alised for the entire constitutional-reform process. The half-hearted Freedom of
Information Act, the piecemeal nature of devolution, the non-implementation of elec-
toral reform proposals at the national level, and the failed second stage of the reform
of the House of Lords are cases in point. Nevertheless, devolution may well generate
13
In this respect, the situation was similar to one that existed between 1931 and 1945.
The United Kingdom 645
a momentum of its own and lead to general decentralisation. The greater independ-
ence enjoyed by the Bank of England in setting interest rates and the Human Rights
Act have created serious new constraints on the British government. Combined with
the effects of the New Public Management and its repercussions for executive account-
ability, Europeanisation and globalisation, these constitutional reforms have moved the
United Kingdom decidedly away from the ideal type of the Westminster model.
REFERENCES
Bagehot, Walter (1963). The English Constitution. Introduction by R. H. S. Crossman. Glasgow:
Fontana (first published 1867).
Bernhard, William T. (1998). ‘A Political Explanation of Variations in Central Bank
Independence’. American Political Science Review, 92: 311–28.
Blackburn, Robert and Andrew Kennon with Michael Wheeler Booth (2003). Griffith and Ryle on
Parliament: Functions, Practice and Procedures. Second edition. London: Sweet & Maxwell.
Budge, Ian (1996). ‘Great Britain and Ireland: Variations on Dominant Party Government’, in
Josep M. Colomer (ed.), Political Institutions in Europe. London: Routledge.
Burch, Martin (1988). ‘The United Kingdom’, in Jean Blondel and Ferdinand Müller-Rommel
(eds.), Cabinets in Western Europe. Basingstoke: Macmillan.
—— and Ian Holliday (1996). The British Cabinet System. London: Harvester Wheatsheaf.
Butler, David E. and Butler, Gareth (2000). Twentieth-Century British Political Facts 1900–2000. Eighth
edition. Basingstoke: Macmillan.
Chester, Norman (1981). ‘Questions in the House’, in Stuart A. Walkland and Michael Ryle
(eds.), The Commons Today, rev. edn. London: Fontana.
Crick, Bernard (1970). ‘Parliament in the British Political System’, in Allan Kornberg and
Lloyd D. Musolf (eds.), Legislatures in Developmental Perspective. Durham: Duke University
Press.
Commission to Strengthen Parliament (2000). Strengthening Parliament. Report by Lord Norton.
London: Conservative Party.
Cowley, Philip (2001). Revolts and Rebellions. London: Politico’s.
–––– and Norton, Philip (1999). ‘Rebels and Rebellions: Conservative MPs in the 1992
Parliament’. British Journal of Politics and International Relations, 1: 84–105.
—— and Stuart, Mark (2000). ‘Can Sheep Bark? British Labour MPs and the Modification of
Government Policy’. Paper presented at the Fourth Workshop of Parliamentary Scholars and
Parliamentarians, Wroxton College, Oxfordshire, 5–6 August.
Crossman, Richard (1963). Introduction to Walter Bagehot: The English Constitution. London: Fontana.
Dahl, Robert A. (1967). ‘Patterns of Opposition’, in Robert A. Dahl (ed.), Political Oppositions in
Western Democracies. New Haven: Yale University Press.
De Winter, Lieven (1991). ‘Parliamentary and Party Pathways to the Cabinet’, in Jean Blondel
and Jean-Louis Thiébault (eds.), The Profession of Government Minister in Western Europe.
Basingstoke: Macmillan.
—— (1995). ‘The Role of Parliament in Government Formation and Resignation’, in Herbert
Döring (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt am Main: Campus and
New York: St. Martin’s Press.
Dicey, Albert V. (1959). An Introduction to the Study of the Law of the Constitution, 10th edn. London:
Macmillan (first published 1885).
Drewry, Gavin (1989). The New Select Committees, 2nd edn. Oxford: Clarendon Press.
646 Thomas Saalfeld
Drewry, Gavin (1994). ‘The Civil Service: From the 1940s to “Next Steps” and Beyond’.
Parliamentary Affairs, 47: 583–97.
Dunleavy, Patrick, Jones, G.W., Burnham, Jane, Elgie, Robert, and Fysh, Peter (1993). ‘Leaders,
Politics and Institutional Change: The Decline of Prime Ministerial Accountability to the
House of Commons, 1868–1990’. British Journal of Political Science, 23: 267–98.
—— —— , and O’Leary, Brendan (1990). ‘Prime Ministers and the Commons—Patterns of
Behaviour, 1868 to 1987’. Public Administration, 68: 123–40.
Fisher, Justin (1996). British Political Parties. London: Harvester Wheatsheaf.
Flinders, Matthew (2001). The Politics of Accountability in the Modern State. Aldershot: Ashgate.
Foley, Michael (1993). The Rise of the British Presidency. Manchester: Manchester University Press.
––––(1999). The Politics of the British Constitution. Manchester: Manchester University Press.
Forman, F. N. (2002). Constitutional Change in the United Kingdom. London: Routledge.
Franklin, Mark N. and Norton, Philip (1993). ‘Questions and Members’, in Mark N. Franklin and
Philip Norton (eds.), Parliamentary Questions. Oxford: Clarendon Press.
Gray, Andrew and Jenkins, Bill (1998). ‘Ministers, Departments and Civil Servants’, in Bill Jones
(ed.), Politics UK, 3rd edn. Manchester: Manchester University Press.
Griffith, J. A. G. and Ryle, Michael (1989). Parliament: Functions, Practice and Procedures. London:
Sweet & Maxwell.
Hailsham, Lord (1976). Elective Dictatorship. London: BBC.
Hazell, Robert et al. (2000). ‘The British Constitution in 1998–99: The Continuing Revolution’.
Parliamentary Affairs, 53: 242–61.
Hennessy, Peter (1995). The Hidden Wiring: Unearthing the British Constitution. London:
Indigo.
—— (1989). Whitehall. London: Fontana.
Hood Phillips, O. (1978). Constitutional and Administrative Law, 6th edn. London: Sweet & Maxwell.
Jennings, Ivor (1959). Cabinet Government. Cambridge: Cambridge University Press.
Jordan, A. Grant (1981). ‘Iron Triangles, Woolly Corporatism and Elastic Nets: Images of the
Policy Process’. Journal of Public Policy, 1: 95–123.
Kavanagh, Dennis (1994). ‘Changes in Electoral Behaviour and the Party System’. Parliamentary
Affairs, 47: 598–612.
Kelly, Richard N. (1994). ‘Power and Leadership in the Major Parties’, in Lynton Robins, Hilary
Blackmore, and Robert Pyper (eds.), Britain’s Changing Party System. London: Leicester
University Press.
King, Anthony (1994). ‘Ministerial Autonomy in Britain’, in Michael Laver and Kenneth A.
Shepsle (eds.), Cabinet Ministers and Parliamentary Government. Cambridge: Cambridge
University Press.
Krehbiel, Keith (1991). Information and Legislative Organization. Ann Arbor: University of
Michigan Press.
Lijphart, Arend (1984). Democracies: Patterns of Majoritarian and Consensus Government in Twenty-
One Countries. New Haven: Yale University Press.
—— (1999). Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries.
New Haven: Yale University Press.
Limon, Donald and McKay, W. R. et al. (eds.) (1997). Erskine May’s Treatise on the Law, Privileges,
Proceedings and Usage of Parliament. 22nd edn. London: Butterworths Law.
McCubbins, Mathew D. and Schwartz, Thomas (1984). ‘Congressional Oversight Overlooked:
Police Patrols versus Fire Alarms’. American Journal of Political Science, 28: 165–79.
Moe, Terry M. (1984). ‘The New Economics of Organization’. American Journal of Political
Science, 28: 739–77.
The United Kingdom 647
A NA LYS I S A N D C O N C LU S I O N
This page intentionally left blank
22
I N T RO D U C T I O N
Representative democracy means delegation, and delegation implies the risk of agency
problems. Parliamentary democracy is a particular delegation regime, a way to struc-
ture the democratic policy process, and an attempt to solve agency problems. We can
thus think of parliamentarism both as a set of constitutions that share a single, funda-
mental property (the cabinet’s responsibility to the parliamentary majority), and as a
broad ideal type. We have referred to the former conception as parliamentary govern-
ment and to the latter as parliamentary democracy. The preceding eighteen chapters
have demonstrated wide-ranging differentiation within the real world of parliamentary
government in delegation designs and accountability mechanisms. Moreover, they have
provided a stepping stone for further analysis of the relationship between democratic
institutions and political representation.
Chapter 4 accounts for the accountability mechanisms that permit direct or indirect
citizen control of their representatives. We refer to checks that are contained in this par-
liamentary chain of delegation, such as direct legislative oversight of the cabinet, as inter-
nal constraints on these various agents. The problem with parliamentary democracy, as
noted in Chapter 3, is that due to institutional weakness or the lack of incentives, such
internal constitutional devices are not always effective. Therefore, in this and the con-
cluding chapter, we focus on two additional mechanisms by which democratic societies
can contain agency problems: political parties and external constraints. Political parties and
external constraints can complement internal accountability mechanisms, as well as one
another. Yet, their respective functions and comparative advantages differ.
Political parties, as Chapter 3 has pointed out, are first and foremost a mechanism to
align preferences between voters and politicians, and between office-holders along the
chain of delegation. As pointed out in Chapters 3 and 4, modern parliamentary systems
rely on cohesive and centralized political parties to induce policy agreement along their
long and indirect chain of delegation. Parties earn their ability to induce cohesion
through their ability to capture and disperse parliamentary agenda powers and other
political rents (Diermeier and Feddersen 1998). This capacity is in turn influenced by
institutional and party system features (see Chapter 4). Cohesive parties serve especially
as an ex ante screening device, by which politicians with appropriate beliefs, values, and
652 Kaare Strøm et al.
skills can be selected for public service. Parties are therefore well suited to combating
the agency problems of adverse selection.
External constraints, on the other hand, variously check politicians once they have
been placed in office. In contrast to internal accountability mechanisms, external con-
straints do not derive their authority directly from the national representatives of the
people, the Parliament, and within the constitutional chain of delegation. We shall
explore their forms and functions below. The virtue of external constraints as a vehicle
of ex post accountability is precisely that preferences do not get aligned, so that the over-
seers retain some distance from the office-holders and some motivation to scrutinize
their behaviour. External constraints are therefore the weapon of choice against moral
hazard among elected politicians and civil servants. In reality, of course, it is more accur-
ate to think of the differentiation between internal and external constraints as a con-
tinuum, since some constraints are more directly related to the parliamentary chain of
delegation than others. Such finer distinctions will indeed be drawn below.
Our investigation of political parties and external constraints focuses on variation
across space and time. We will demonstrate substantial cross-national variation in the
range of accountability mechanisms employed in democratic representation. Whereas
some democracies primarily seek to contain agency losses between voters and politi-
cians through the alignment of preferences generated by political parties, others rely
much more on external checks.
Less obviously, perhaps, there is also clear-cut evidence of intertemporal variation.
The most important trend among West European democracies is a shift in the import-
ance of ex ante versus ex post controls. There has been a decline over time in the
importance of ex ante controls. At the same time, external constraints in general, and
ex post mechanisms in particular, have been enhanced. Yet, important cross-national dif-
ferences persist on these two dimensions as well as within the chain. Thus, contempo-
rary parliamentary democracies have almost as many differences as they have
commonalities.
Our exploration of these issues falls into three parts. All build on the knowledge that
we have accumulated in previous chapters, as well as on relevant and complementary
secondary sources. Thus, in the next section we examine the strength of political par-
ties and their value as vehicles of citizen control. In the following section we look at
external constraints on democratic agents. Finally, in the concluding section we provide
a joint and comparative assessment of the current state of political parties and external
constraints as instruments of citizen control.
E X A N T E C O N T RO L S A N D C O H E S I V E P O L I T I C A L PA RT I E S
Ex ante controls are typical of the classical model of parliamentary democracy, and a
feature closely bound up with strong political parties and cohesive party government
(see Chapter 3). For parties to help citizens control their agents, they must function as
vehicles of mutual gain between voters and politicians. On the one hand, parties must
lessen the agency problems that voters face. In other words, it must be easier to con-
tain agency loss within political parties than in the polity at large. That, in turn, requires
Dimensions of Citizen Control 653
that certain conditions be met. First, partisanship must be associated with systematic
and relatively transparent differences in the bundles of goods and policies that govern-
ments produce. Second, party labels and the policy differences that they represent must
bear some relationship to the preferences of the voters. Third, voters must have a way
of holding the representatives of any given party responsible for its performance in
office. We can think of these three conditions jointly as the voters’ incentive conditions,
the circumstances in which they are willing to rely on party labels in their voting deci-
sions. Only if all three conditions hold can the voters’ incentive condition be satisfied,
and only then does it make sense for them to attach themselves to partisan labels.
But parties must also serve the purposes of the politicians that would form them.
Thus, for politicians to submit to party discipline and invest in partisan brand names,
partisanship must offer them tangible and predictable benefits. Parties must help politi-
cians control the flow of governmental decisions and dispersals (Aldrich 1995), and they
must help politicians get re-elected (or, perhaps, elevated to higher office) (Schlesinger
1991; Cox and McCubbins 1994). These, in essence, are the politicians’ incentive condi-
tions. If all these conditions are met, then citizens can rely on parties to screen and
sanction candidates for political office and thus reduce their agency costs (and politi-
cians can count on citizens to support them). Indeed, these conditions will structure
our discussion of political parties below.
We should think of parties here in a broad sense that may include coalitions of for-
mally independent organizations. What matters is whether they provide reliable and
useful information to the voters and credible incentives to politicians. Thus, Christian
Democrats have arguably attained a higher degree of partyness in Belgium and
Germany, where they have been organizationally divided, than in First Republic Italy,
where they formed a single but uncohesive organization.
Parties structure the first step in the chain of delegation by offering programmes (if
only in the loosest sense) and candidates committed to these programmes. When par-
ties establish meaningful ‘brand names’ (reputations), they allow voters to economize
on the information they must acquire. To protect their reputations political parties have
a strong incentive to deliver once elected. Indeed, the claim that they can control the
‘downstream’ formulation and implementation of policy is part and parcel of the com-
mitment of parties to their voters. And cohesive political parties allow the voters to
observe the behaviour of their representatives once they are elected. Parties reward
politicians who toe the party line and discipline those who do not. Generally, without
parties, politicians would be less able to achieve their career and policy goals (Müller
2000a). If this were not true, we would see far more successful independent candidates.
Political parties thus play a critical role as ‘bonding arrangements’ (Cox and McCubbins
1994) in parliamentary systems.
Traditionally, parties effectively reduced agency costs by bringing together voters
and candidates with similar policy preferences. They reduced uncertainty about the
‘types’ of politicians on offer and thus helped citizens contain their adverse selection
problems. European mass parties had their origins in the distinctive subcultures that
had emerged in most industrializing nation states by the early twentieth century. Thus,
parties brought together candidates and voters from the industrial working class, the
654 Kaare Strøm et al.
Catholic subculture, particular language communities, or the agricultural sector, just to
name a few. The common identities and transparencies of those social segments helped
assure voters that they had representatives that shared their values. Thus, political par-
ties reduced the risks of adverse selection. As long as the issue dimensions of demo-
cratic politics largely reflected the same social cleavages, citizens could generally trust
their agents.
Yet, the ability of contemporary political parties to meet these conditions is a matter
of considerable scepticism (e.g. see Dalton and Wattenberg 2000). Sweeping and fun-
damental social change has eroded some of their foundations, such as the trust that vot-
ers traditionally placed in parties, elected representatives, and indirect delegation of
political power. As old communities have withered, societies have become more differ-
entiated, and new issue concerns have arisen; parties have trended toward ‘catch-all’
profiles (Kirchheimer 1966). While this has helped traditional parties stay electorally
competitive, it may have reduced the value of their ‘brand names’ to the voters.
Political parties have many facets that affect their effectiveness as instruments of
bonding and preference alignment. Following V. O. Key’s (1964) conception, we can dis-
tinguish between the party in the electorate, the party organization, and the party in
government. These are mutually interdependent aspects of modern political parties.
Yet, they are not equally important for the viability of parties as mechanisms of citizen
control. Instead, it is useful to think of their interdependence as hierarchical. Unless
parties impose some kind of cohesion on government policy-making, it would make
little sense for voters to identify with them, or for activists to offer them their services
or funds. The party in government is thus fundamental. Without governmental parties,
neither the politicians’ nor the voters’ incentive condition could be satisfied. Yet, the
electoral party is almost equally critical, at least from the politicians’ perspective. If vot-
ers paid no heed to party labels, politicians would find limited use for governmental
parties. The (extraparliamentary) party organization is less crucial to party govern-
ment, as it is simply one way to organize the exchange between politicians and voters.
While there is no doubt that extraparliamentary party organizations can play an impor-
tant role in screening political candidates, as well as in raising funds and providing cam-
paign labour, they are not essential to the core functions of political parties. Indeed,
cadre parties used to do without much in the form of extraparliamentary (voluntary)
organization, as do indeed many newer parties.
While Key’s three facets of political parties are thus at least partially interdependent,
and while each tells us something important about party strength, they are not equally
important for citizen control of their representatives. Hence, the party in government
will be our central concern, followed by the electoral party. In the next section, how-
ever, we begin by examining variation across time and space in the importance of con-
temporary parties, in their different facets, as agents of the citizens.
Electoral Parties
The principal normative role of political parties in democratic societies is to align prefer-
ences between citizens and their political agents. For this to happen, electoral parties must
Dimensions of Citizen Control 655
meet several conditions. Voters must attach themselves to them on the basis of common
policy preferences, and these citizen attachments must in turn drive voting behaviour.
Moreover, voters must be faced with a set of parties that is reasonably stable and transpar-
ent, and that permits the voters to hold the occupants of government office to account.
We turn first to the latter concern, the degree to which electoral parties facilitate sta-
ble and transparent citizen representation and accountability. For parties to be effective
instruments of citizen control, the party system needs some measure of stability over
time. Party systems in which individual parties are very short-lived, or volatile, do not
satisfy this condition. The greater the electoral volatility, of course, the lower the sta-
bility of the party system, and the less effectively voters can attach themselves to par-
ties with predictable policy consequences, or subsequently hold them to account.
Table 22.1 contains comparative data on our seventeen countries that enable us to
assess the evolution of electoral party strength. Thus, Table 22.1 reports net electoral
volatility for parliamentary seats between the first two postwar elections (or, in the
cases of newly established regimes, the first two elections to which this measure can be
applied, that is, the second and third democratic elections) and the two most recent
ones.1 In the majority of cases volatility has indeed increased over the second half of
the twentieth century, in some cases substantially (e.g. France and The Netherlands).
More specifically, in many countries volatility decreased in the early post-1945 period,
but has then trended clearly upwards since about the 1960s (see Dalton and Wattenberg
2000). All but one of the countries in which volatility has not increased are those with
relatively recent authoritarian pasts (Italy, Spain, Germany), in which the early turbul-
ence reflects party system consolidation. Because of this late democratic start, party
systems in these countries were not ‘frozen’ until the 1960s or later (Lipset and Rokkan
1967), and the onset of destabilization was delayed (to roughly the 1990s in Germany
and Italy, whereas in Spain it has not yet occurred). Finally, note that while in the early
period net volatility varied sharply across nations, there are now virtually no countries
in which volatility is very low.
A related concern is the extent to which the parties on offer are willing and reliable
vehicles of popular representation. Parties may fail in this respect either if they are pri-
marily protest movements with no serious intention of making policy, or if they have
not developed the organizational capacity to do so effectively and cohesively. If parties
that are uninterested in, or unsuitable for, executive office make up a large segment of
the party system, then the voters’ ability to hold governing parties to account is
severely curtailed. Interparty competition suffers, and electoral accountability is dimin-
ished. Of course, we cannot always identify which parties are suitable and reliable and
which ones are not. In the real world, experience is typically the best guide. Thus, the
larger the segment of the party system that has actually served in the executive branch,
1
We measure volatility in seats rather than votes because the former measure captures electoral system
effects and relates more directly to parliamentary bargaining. If we had comparable cross-national data on
gross (individual) volatility, it would certainly add to our understanding of party system stability. The evid-
ence we do have, however, is highly consistent with our results here in that it shows significant trends towards
instability (e.g. Dalton 2002).
Table 22.1. Party Strength in the Electorate
Net volatility (in seats) Share of parliamentary seats held by Party identification (or closest Deviation from
non-governing partiesb equivalent) uniform swing
First two Two most Change for cabinet
postwar recent 1950–4 or 1995–9 Change In first In recent Change partiesd
electionsa elections earliest period two survey (entire period)
(in 1990s) surveysc
Austria 6.97 8.20 1.41 10.91 (1953) 32.24 21.33 67 52 (2000) ⫺15 0.42
Belgium 7.93 9.39 1.46 3.30 (1950) 16.00 12.70 50* 55* (1992) 5 0.63
Denmark 14.56 11.73 ⫺2.83 12.58 (1950) 20.11 7.53 52 50 (1998) ⫺2 0.40
Finland 6.75 11.50 4.75 0.00 (1951) 5.00 5.00 57 51 (1991) ⫺6 0.50
France 22.37 43.16 20.79 14.31 (1959) 0.86 ⫺13.45 59* 52* (1992) ⫺7 0.23
Germany 14.19 8.67 ⫺5.52 5.95 (1953) 5.38 ⫺0.57 78 64 (1998) ⫺14 0.44
Greece 12.84 8.17 ⫺4.67 12.00 (1977) 6.67 ⫺5.33 — 63* (1992) n/d 0.22
Iceland 5.77 8.73 2.96 17.31 (1953) 41.27 23.96 80 73 (1995) ⫺7 0.52
Ireland 10.42 13.25 2.83 10.88 (1951) 8.43 ⫺2.45 61* 39* (1994) ⫺22 0.19
Italy 21.36 14.69 ⫺6.67 38.64 (1953) 47.14 8.50 78* 60* (1994) ⫺18 0.50
Luxembourg 7.49 7.61 0.12 7.69 (1951) 16.67 8.98 61* 48* (1992) ⫺13 0.57
The Netherlands 4.50 19.67 15.17 10.00 (1952) 16.00 6.00 38 28 (1998) ⫺10 0.70
Norway 9.00 18.18 9.18 2.00 (1953) 21.21 19.21 66 42 (1997) ⫺24 0.24
Portugal 6.74 15.03 8.29 19.20 (1980) 6.52 ⫺12.68 — 67* (1992) n/d 0.11
Spain 27.86 8.29 ⫺19.57 14.00 (1977) 15.14 1.14 — 41* (1992) n/d 0.0
Sweden 8.70 13.90 5.20 2.17 (1952) 28.94 26.77 64 42 (1998) ⫺22 0.16
UK 9.12 16.95 7.83 1.92 (1950) 11.53 9.61 93 91 (1997) ⫺2 0.0
*
Eurobarometer Data
a
For France we consider the Fifth Republic; for Greece, Portugal, and Spain the first two elections after the democratization of the 1970s.
b
Non-governing parties 1995–99 are those that neither participated in the most recent government (1 January 2000) nor participated in government in at least two legislative terms during the
entire postwar period. In the earlier period (starting 1950), the number refers to the share of seats belonging to parties that existed in Parliament in the year listed (1950 for Denmark, 1980 for
Portugal, etc.) and that have not been in government during at least two different parliamentary terms since then, nor were in government at the end of the period (1 January 2000).
c
These surveys begin between 1964 (UK) and 1983 (Iceland), but most are from the 1970s.
d
Calculated as the proportion of cabinets in which parties in government experience divergent electoral results in the subsequent election.
Sources: Party ID: Dalton 2000: 25 and personal communication with the author; Eurobarometer Data; Schmitt and Holmberg 1995; Mair et al. 1998: 200; Jenssen 1999; Holmberg 2000: 41.
Net volatility, share held by non-governing parties, uniform swing: Comparative Parliamentary Democracies Data Set ( January 2002).
Dimensions of Citizen Control 657
2
We have adopted this latter qualification to extend the benefit of the doubt to newer parties in states that
have either recently democratized or experienced substantial party system transformations (Italy).
3
Moreover, Jean-Marie Le Pen’s triumph in the first round of the 2002 French presidential election sug-
gests that non-established parties have prospered even here.
658 Kaare Strøm et al.
gain or lose votes relative to the previous general election), we score the election as 0
(no deviation from uniform swing). If, on the other hand, at least two governing par-
ties experience different fates (i.e. one loses and the other gains), then we score that
observation as 1 (non-uniform swing). For each country, we have then computed the
mean over all elections between 1945 and 2000. Because of the limited number of
observations in some countries, we have not attempted to identify a trend.
Quite obviously the United Kingdom and Spain, with their exclusive records of one-
party governments, exhibit no deviation from uniform swing and thus score 0. Other
countries that have very few coalition governments, such as Portugal and Greece, also
have very low scores. Conversely, however, it is not true that all coalitional systems exhibit
high scores. Our highest sample score, for The Netherlands, is 0.70. This means that in
more than two-thirds of all Dutch elections, the voters have rewarded the governing par-
ties differentially. Other countries that score relatively high are Belgium, Luxembourg,
Finland, Iceland, and Italy. In these countries, voters do not treat the executive parties as
a cohesive bloc and simply ‘throw the rascals out’ or keep them in, but discriminate
between them. In many cases, however, such discrimination on the part of voters coex-
ists with a low degree of responsiveness of cabinet formation to electoral verdicts (Strøm
1990). In other words, whether a particular party has just been rewarded or punished by
the voters has little effect on its prospects of gaining executive office.
On the other hand, Ireland, Norway, Sweden, and particularly France have signific-
ant records of coalition government, but much greater uniformity of swing than the
countries mentioned above. These differences basically reflect our distinction between
pivotal and alternational (bloc) party systems, as governing coalitions in the latter type
of system have greater electoral cohesion than those in the former. To the extent that
they do, they also facilitate electoral accountability.
In sum, the evidence in Table 22.1 suggests two things. First, there is obviously consid-
erable cross-national variation in the extent to which the electoral parties of Western
Europe facilitate accountability to the citizens in the ways of the Westminster model. Yet,
even though electoral parties in Spain, for example, come as close to the Westminster
ideals as they ever did, overall the established parties of Western Europe have lost a signif-
icant part of their grip on the voters. Parties have become less stable and reliable instru-
ments of popular governance, and the voters have become less loyal and committed. It is
difficult and perhaps futile to establish cause and effect, to determine whether voters have
become more fickle because the party system is less conducive to their influence, or vice
versa. It is perhaps more useful to say that to the extent that a Westminsterian equilibrium
once existed, it is now in most countries not much in evidence.
a
1980s for Greece, Portugal, and Spain.
b
2000 for Portugal and Spain.
n/a ⫽ Not applicable.
n/d ⫽ No data available.
Sources: Membership density: Scarrow 2000: 90; Mair and van Biezen 2001. State subsidies: Chapter 4, Table 4.6. Mean
subsidies 1980s (prices standardized at 1987 levels): Farrell and Webb 2000: 117.
660 Kaare Strøm et al.
political parties in a variety of countries have systematically over-reported their true mem-
bership. Sometimes public subsidy schemes or other public policies even give them incen-
tives to report inaccurately. Moreover, institutional reforms can render trends in reported
members misleading. In some countries, such as Sweden and Norway, the switch from
collective trade union affiliation to individual membership (in the social-democratic
parties) has thus exaggerated the magnitude of membership decline.
Such considerations notwithstanding, there can be no doubt that membership decline
is pervasive. Membership densities have declined almost everywhere. Moreover, this
decline has in many cases been rapid and seemingly accelerating. The drop-off in party
membership has been particularly severe in Italy, The Netherlands, and northern
Europe. Consider, for example, the dramatic and recent downturns reported in
Denmark and The Netherlands, where membership densities have shrunk to one-fifth or
less of early postwar levels. In Sweden it has been estimated that if membership num-
bers shrink at their current rate, there will be no party members left by 2013 (Petersson
et al. 2000). Only the more recent democracies of Greece and Spain, plus Germany,
report an overall increase of membership. In the latter case, the positive trend is caused
by the CDU’s organizational efforts in opposition in the 1970s. If we had considered only
the last two decades, we would have found a negative trend even for Germany.
Political parties rely on some mixture of different inputs such as voluntary (and often
free) labour on the one hand, and capital on the other. To some extent financial contri-
butions, which need not come from party members, can substitute for rank-and-file
labour. There can be no doubt that politics, and campaigning particularly, has become
more capital-intensive and less labour-intensive. Indeed, in Western Europe the state
has become the greatest and single most import-ant contributor to political parties.
Table 22.2 shows that by the end of the 1980s government subsidies to political parties
had been introduced in all parliamentary democracies of Western Europe (see also
Chapter 4). Yet, as the last column suggests, European democracies vary widely in the
volume of subsidies they provide.
While public funding may help political parties survive as campaign organizations, it
is less clear that they can sustain their role as vehicles of broad political socialization,
recruitment, and screening. The doubts that our data have nurtured concerning the
health of contemporary party systems therefore apply as much to extraparliamentary
as to electoral parties. Yet, in the former case it is less clear that these trends are inimi-
cal to citizen control of their agents.
Parties in Government
Yet, the most critical aspect of party strength lies in the governmental party. Many
Western European political parties began as parliamentary organizations (they were
‘internally created’, according to Duverger 1954), and this function is still critical to the
prevailing model of parliamentarism (Frognier 2000). In order for parties to deliver pol-
icy outputs or rents, they clearly have to control access to the highest political offices in
the land. And citizen control of public officials would be hollow indeed, if the parties
through which it operated excluded cabinet-level officials. To be sure, elite-level
Dimensions of Citizen Control 661
representative politics in Western Europe is party politics, and the closer we come to
the apex of national policy-making, the more clearly this is true.4 European parties are
strong in Parliament and in the cabinet, where they display a high degree of cohesion
and leave little room for non-partisanship. In this respect, there is not much cross-
national variation (see Strøm 2000).
Tables 22.3 and 22.4 present a variety of indicators of partisanship in the cabinet and
the rest of the executive branch. Two of these, reported in Table 22.3, refer to the com-
position of the cabinet itself. One is the overall incidence of single-party cabinets, meas-
ured in the proportion of time that such cabinets have held office. The second and
4
See Blondel and Cotta (1996, 2000) for attempts to disentangle party and government.
Table 22.4. Party Cohesion in Government
Mean coalition Mean coalition Cabinet termination Cabinet Cabinet termination Cabinet
discipline discipline over party conflict, termination over over party conflict, termination over
1950–74 1975–99 raw number party conflict, raw number party conflict,
(rank) (rank) 1950–74 % of all cabinets 1975–99 % of all
(total number of 1950–74 (total number of cabinets
cabinet terminations) (rank) cabinet terminations) 1975–99 (rank)
5
Note, however, that according to Manow (2002), interparty competitiveness has not depressed the level
of patronage politics in Western Europe.
Dimensions of Citizen Control 665
Summary
Overall, the available data suggests a general decline in the functions that political par-
ties play in facilitating citizen control of their agents. We can perhaps best capture the
extent of partisan decline by considering the overall changes that have taken place since
the early postwar period. For most of our indicators, we can easily identify how the
respective countries would have scored in the 1950s (or a similarly early date) relative
to the 1990s. Thus, with regard to party identification, seven countries—Austria,
Finland, France, Ireland, Luxembourg, Norway, and Sweden—that previously would
have ranked above the current cross-national mean (54.4) have now fallen below this
level. Only one country (Belgium) has moved in the opposite direction. Similarly, five
countries—Austria, Belgium, Iceland, Norway, and Sweden—have experienced a
strong increase in the electoral support drawn by non-established parties, whereas only
France and Portugal have seen a significant decline in the support for such parties.
Membership density has fallen in twelve of the fifteen countries for which we have
data, and in some of these the decline has been truly dramatic. Only two countries,
Greece and Spain, have experienced a moderate increase.
The evidence is thus consistent and fairly striking. Though on all three dimensions
examined here Western European party systems differ significantly from one another,
it is clear that parties are on the whole less well equipped to facilitate citizen control
than they were in the mid-20th century. This is most obvious with respect to electoral
and extraparliamentary parties. Political parties today have a weaker grip on the elec-
torate, and membership in party organizations is falling, in some cases drastically.
Apparently, citizens are less content to rely on party ‘brand names’ than they were in
the early decades after the Second World War. This trend is strongest in northern
Europe and much weaker or non-existent in those southern European countries that
have democratized in the latter half of twentieth century. Yet, it is less clear that the
party in government has lost in significance. Governmental parties maintain much of
their cohesion and thus their ability to bond and discipline politicians. Their influence
over legislators and cabinet ministers is still strong.
E X T E R NA L C O N S T R A I N T S
Principals and agents within the democratic chain of delegation, the representatives of
the people, are simultaneously bound by their popular mandates as well as by various
other obligations and restrictions, which we generically refer to as constraints. We refer
to those constraints that are formed by institutions that are themselves agents in the
parliamentary chain of delegation as internal. They have been detailed in Chapter 4. In
this section, we shall focus instead on constraints that are external to the parliamentary
chain of delegation. The country chapters (Chapters 5–21) demonstrate that such con-
straints on popular representatives are in fact quite important. In this section, we
review these constraints and summarize their importance, which, as we shall see, has
on the whole increased further.
The most important constraints that have been discussed by our contributors are
the following: international (here particularly the European Union (EU)), societal,
666 Kaare Strøm et al.
subnational, executive (i.e. those associated with the role of the Head of State), central
banks, judicial, and direct democratic. Below we proceed in the same order, from the
most general and indirect constraints to those that are more direct and specific. In each
case, we assess the relevance of these constraints cross-nationally and over time.
Differences between countries or differences over time result from two developments.
One is deliberate constitutional design: institutions that serve as external checks can
be created, abolished, or reformed. The second source of change is informal political
adaptation, which may result from political self-interest, from convenience, from band-
wagon effects, or from no obvious motivation at all. For example, packing institutions
intended to constrain parliamentarians with party appointees can have the effect of
aligning preferences and hence undermining the very rationale of these institutions.
Conversely, freeing institutions from such partisanship can make external checks effec-
tive without any change in formal institutions.
Societal Constraints
In this volume, our main focus has been on the constitutional structure of the national,
supranational, or subnational government. Of course, in a study of decision rules in
6
Certainly, the smaller member states have more representatives in the European Parliament and much
more voting power in the Council than they would have under proportional representation based on popu-
lation size. In that respect, Luxembourg, for example, is extremely advantaged compared to the largest mem-
ber state, Germany.
7
Clearly, policy-making in the EU is exceedingly complex, and as Garrett und Tsebelis (2001) point out,
this focus on the Council neglects the agenda-setting power of the Commission. Yet, we believe the Council’s
role to be critical and know of no simple and superior alternative measure of the voting power of the vari-
ous members. Hence, the index included in Table 22.5 is intended as a heuristic measure of ‘one of the most
crucial factors’ (Hosli 1996: 256) that explain EU decisions.
8
Felsenthal and Machover (1998) maintain that the Banzhaf index alone measures the a-priori influence
(I) of voters in a voting body (based on a voter’s ability to affect whether the bill will be passed or defeated).
In contrast, its main alternative, the Shapley–Shubik index, is said to measure a voter’s expected relative share
in some prize (P) a winning coalition can divide among itself. For a critique of this measure, see Garrett and
Tsebelis (2001: 99).
668 Kaare Strøm et al.
Table 22.5. International, Societal, and Subnational Constraints
European Union Index of Index of Subnational government
coporatism (0–3) federalism and consumption
Number of Voting power decentralization (1995–6) (0–100)b
votes in in the Council (1–5)
the under qualified
Council majority rulea
a
This measure is the standardized Banzhaf index based on the qualified majority rule (62 out of 87 votes).
b
This index is the proportion of local and regional government final consumption in relation to general government final con-
sumption.
n/a ⫽ non-applicable (not EU member states).
Sources: European Union: Nurmi and Meskanen 1999: 167; Index of federalism: Lijphart 1999: 189; Index of regional and local
consumption: Lane and Ersson 1999: 188. With the exception of Luxembourg (data from 1992) updated via personal commu-
nication with Ersson. Index of corporatism: Lane and Ersson 1999: 23.
national parliamentary systems, such an emphasis makes a great deal of sense. Most
formal constraints on the power of those authorized to govern through parliamentary
elections, or indirectly authorized through the former, can in fact be located in that
constitutional structure. Yet, we would be remiss in describing the institutions of post-
Second World War Western Europe without considering the constraining effects of
actors, networks, and agreements that extend beyond these formal institutions. Our
contributors have pointed to a range of social institutions, in many cases components
of civil society, that place important constraints on those empowered in the national
chain of delegation. These social institutions can roughly be divided into three cate-
gories: corporatist arrangements, consociational structures, and communities of faith.
Corporatism
During the 1930s, corporatism, or more specifically what Schmitter (1974) and others
call ‘state corporatism’, was seen as an alternative to parliamentary democracy, as a
form of political decision-making most fervently championed by those who least
favoured parliamentarism. After the Second World War, corporatism, now more likely
Dimensions of Citizen Control 669
9
Lijphart (1999) actually refers to his index as one of ‘interest group pluralism’ and clearly refers to cor-
poratism as the opposite pole. Our rankings thus represent an inversion of this scale.
670 Kaare Strøm et al.
Netherlands is placed thereafter, while Belgium, Germany, and Luxembourg are placed
with the fifth Nordic country (Iceland). Remaining countries are scored as having no sig-
nificant corporatist intermediation. While this classification is less discriminating than
some other rankings, it has the advantage of grouping countries in robust clusters.
While the contributors to this volume have not been centrally concerned with the
fate of corporatism per se, they do suggest that its trajectory is not uniform across
countries. In Austria and Sweden, corporatist practices seem to have fallen into long-
term decline. Privatization (in Austria) and European integration (in both countries)
have reduced the scope for government intervention in the policy areas most relevant
to producer groups. In Austria, the long-term decline in the legislative influence of
interest groups was accelerated in 2000 by the replacement in government of the pro-
corporatist Social Democrats by the anti-corporatist Freedom Party (Tálos and Kittel
2001). In Sweden, business associations have since the early 1980s opted out of the sys-
tem of interest representation through government commissions and executive boards,
and the whole remiss system has declined in importance (Rothstein and Bergström
1999; Chapter 20). Similar, though less dramatic, developments have taken place in
Denmark (Iversen 1999; Blom-Hansen 2001). In Finland, on the other hand, corporatist
practices evolved somewhat later and were particularly prevalent in the 1970s and
1980s, but have since fluctuated more. In Norway (ranked number one by Lijphart),
corporatist practices were reduced under the non-socialist governments of the early
1980s. In the 1990s, however, social-democratic governments once again invoked the
cooperation of unions and business in comprehensive wage bargaining and in the craft-
ing of tax and social security reform. This concertation was generally successful in the
early 1990s, but had lost most of its effectiveness by the end of the decade. At the same
time, the number of consultative committees and boards has continued to decline.
And, as in other Scandinavian countries, the effectiveness of corporatist bargaining has
been weakened by labour fragmentation, as white-collar organizations have grown at
the expense of the traditional blue-collar sector and gradually severed the unions’ orga-
nizational ties to social democracy. On balance, therefore, it seems fair to say that in
Norway, like the other leading countries in this respect, corporatism has since the 1970s
withered (though not completely died). Thus, cross-nationally, corporatism is an exter-
nal constraint whose importance has declined.
arrangements have not been numerically necessary. As Müller also stresses, however,
the constraints imposed by consociationalism (as well as by corporatism) are largely
voluntary ones adopted by parliamentary majorities, which parliamentary majorities
have also been able to discontinue. Following the surge in support for the populist-right
Freedom Party, Austrian politicians have indeed dismantled many of the edifices
of consociationalism as well as corporatism. The same decline is in evidence in other
formerly consociational countries (Luther and Deschouwer 1999). One reason may be
that traditional social cleavages have lost much of their meaning.
Other forms of societal constraint can in some societies be found in the armed forces
or in powerful communities of faith. In Portugal, the armed forces played a key role in
overseeing civilian institutions during the first few years after the 1974 Revolution. In
many Latin American countries, of course, the constraining role of the military has
been much more freely acknowledged, and many cases of military intervention have
indeed been defended on precisely such tutelary grounds (Loveman 1993).
In most of contemporary Europe, however, the military takes a back seat in import-
ance to civilian institutions such as the Church. In Greece, the Greek-Orthodox Church
ranks among the most important social institutions. The Church is an important
national symbol, which stands as a bulwark against Islam as well as against Greece’s tra-
ditional foe, Turkey. The Church is also important because of its financial power and
social work. Something quite similar could be said about the position of the Catholic
Church in Ireland. By most measures, Ireland has Europe’s most devout population, and,
given Ireland’s tragic and conflictual past, the Church is important politically as well as
spiritually. No major party in Ireland is openly anti-clerical, and only in recent years has
the Church lost some of its influence on moral and social issues (Coakley 1999).
Most societal constraints differ from other external checks in that their influence is
felt particularly at early stages of the policy process. They correspond to some form of
agenda control, rather than ex post veto power. Note also that corporatism and con-
sociationalism especially are associated with strong, rather than weak, political parties.
Hence, corporatism is probably best understood as a form of voluntary delegation by
political parties with close ties to particular economic producer groups. Finally, most of
these forms of societal constraint seem to have diminished in importance over the past
couple of decades, as societies have become more differentiated and social identities
more plastic. In sum, then, societal constraints have distinctively structured European
parliamentary politics. They have provided social foundations for political parties and
simplified citizen ex ante control by contributing to the alignment of preferences. Their
decline is therefore not evidence of a strengthening of party control of popular repre-
sentatives, but rather the contrary.
Subnational Constraints
The national chain of parliamentary delegation is impacted not only by supranational
constraints but also by subnational ones. The most common type of subnational con-
straint is federalism, in which there are at least two levels of government, each of which
has some constitutionally protected area of jurisdiction in which it is makes final
672 Kaare Strøm et al.
decisions (Riker 1964, 1975). States that do not meet these criteria are either unitary, in
which all final authority is vested in the central government, or confederal, if all ulti-
mate authority rests in the lower levels of government.10 Federalism has a number of
putative virtues, such as the ability to accommodate social diversity and particularly
geographically concentrated minorities (Lijphart 1977). It may also foster policy exper-
imentation and competition among its subunits and thus a choice of policy environ-
ments for its citizens. Such policy choice may yield benefits in economic efficiency
(Tiebout 1956). Weingast (1995) has very interestingly expanded on this point, arguing
that the economic benefits of federalism accrue specifically under ‘market-preserving
federalism,’ in which subnational governments have the primary regulatory responsi-
bility, the central government ensures a common market without internal discrimina-
tion, and subnational governments face a hard budget constraint (do not control the
money supply). On the other hand, federalism may exacerbate regional inequalities,
and some studies suggest that fiscal decentralization, at least in many less developed
societies, may be associated with corruption and poor service delivery rather than ben-
eficial competition (Treisman 2000).
According to the singularity principle (see Chapter 3), a parliamentary democracy
should effectively be a unitary state, with no constitutional constraints on the power of
the national government. Most Western European democracies are indeed unitary,
though we need to recognize exceptions as well as caveats. Table 22.5 presents
Lijphart’s (1999) index of federalization. Three states covered in this volume are cur-
rently federal: Germany, Austria, and Belgium. The former two have been so ever since
their current constitutions were adopted. Belgium, on the other hand, has followed a
gradual path towards devolution, with a fully federal constitution adopted in 1993. The
Belgian and German federalisms reserve significant exclusive or joint powers for sub-
national governments, whereas in the Austrian case federalism is more apparent than
real. Thus Lijphart scores Austria below the other two federal states on decentraliza-
tion. Although the formal Constitution reserves residual powers for the nine states
(Länder), actual constitutional practice has vested more and more power in the central
government (and, increasingly, EU institutions) (Chapter 5).
While not fully federal, the Italian and Spanish Constitutions reserve significant pow-
ers for subnational governments, whose structure is incongruent (i.e. constitutional
powers vary significantly across subnational units). Italy and Spain both have particular
regions that enjoy an enhanced, and constitutionally entrenched, autonomy. These
peculiarities of constitutional structure reflect deliberate compromises based on his-
torical diversity and fierce regional rivalries. In fact, demands for increased autonomy
or even secession were significant as late as the 1970s in Italy and the 1980s in Spain,
after constitutional autonomy had been secured. In both countries the regions were
able to make additional gains in their competences in the 1990s (Colomer 1998; Lane
and Ersson 1999: 175–8). And even in historically centralist France, the Socialist
governments of the 1980s introduced regionalization policies (Schmidt 1990).
10
Strictly speaking, since they lack internal sovereignty, confederal governments are not states. Since
they also are not represented in this volume, we shall in the following disregard them.
Dimensions of Citizen Control 673
Historically, the most purely unitary states, and hence the closest approximations to
our ideal type of parliamentary democracy, have been found in northern Europe: in
Britain and Scandinavia. While this picture remains largely accurate, even here the
supremacy of national political institutions has been challenged. The most obvious
case, of course, is the United Kingdom, where devolution of powers to directly elected
assemblies in Scotland and Wales was among the first notable achievements of the first
Blair government (1997–2001). The Nordic countries have not seen any similar devel-
opment. Yet, as Bergman (Chapter 20) points out, in their implementation of the wel-
fare state Swedish local governments have become increasingly autonomous from the
central government. And the central government’s share of tax revenue is no higher in
Sweden than in the United States (Lijphart 1999: 193).
In fact, the degree of fiscal decentralization in the Nordic countries is such that
Lijphart (1999) classifies all but Iceland as ‘unitary and decentralized’. Hence, Table 22.5
includes a measure of decentralization in public spending. The variation across
Western Europe in the degree to which the subnational rather than the national level
expends tax money is striking. More than half of total government consumption
occurs subnationally in all Nordic countries but Iceland, in The Netherlands, Spain, and
in two federal states: Austria and Germany. Ireland and Italy also spend about half of
the money at the regional and local levels of government. Belgium, the third federal
state, spent only about one-third of public moneys subnationally about the time that
the new federal Constitution went into effect (1995).
Presidential Powers
The parliamentary systems of Western Europe are roughly evenly divided between
monarchies and republican constitutions. In their formal constitutions, the Head of State
generally plays an important role. Recall that before parliamentary systems were democra-
cies, they gradually became parliamentary by wresting constitutional powers away from the
monarch and vesting them with the parliamentarians, however democratically the latter
may have been elected (Bendix 1978). Even in most contemporary European republics, the
Head of State is still a counterweight to the power of the parliamentary majority.
This constraint on popular representation is a feature of virtually all the countries
represented in this volume. Only in Sweden has the king been entirely emasculated, as
the 1975 Constitution does not even leave him a formal role in the policy process. In
most other monarchies (Belgium, Denmark, Luxembourg, The Netherlands, Norway,
Spain, and the United Kingdom), however, the role of the monarch is almost strictly
ceremonial. Despite the crucial role of the Spanish King in establishing and defending
the transition to democracy, none of these Heads of State is a serious policy-maker or
constraint on popular representation. Only in Belgium has the monarch actively
inveighed on national policy, and even there the 1950 referendum on the ‘king’s ques-
tion’ effectively relegated him to the sidelines of national politics (see Chapter 6).
Under republican constitutions, the constraining force of the Head of State is likely
to be more real (see Elgie 1999). Compared to other executives, directly elected
Presidents are at an advantage in popular legitimacy, as in this respect they can claim to
674 Kaare Strøm et al.
be on equal terms with Parliament. Given that Parliament normally does not speak
with one voice, a directly elected President may even see himself (or herself ) as supe-
rior, particularly if the parliamentary majority is small (or unstable) or the President
close to the median voter. We therefore expect popularly elected Presidents to be par-
ticularly significant constraints on parliamentary policy-making.
Presidential powers fall into a number of distinct areas. First, many Heads of State
are vested with more or less significant powers to appoint and dismiss the Prime
Minister and other cabinet members. Most Presidents have a variety of other appoint-
ment powers. Presidents may also be authorized to dissolve Parliament and call new
elections. Furthermore, the Head of State may be allowed to participate directly in
policy-making, through veto powers or the authority to issue decrees or other legisla-
tive instruments. Some Presidents can call a popular referendum, and some have the
power to initiate judicial review. Table 22.6 reports on the selection and powers of
Western Europe’s republican Heads of State.
In Table 22.7, we present a scoring scheme for presidential powers, building on
Shugart and Carey’s (1992: ch. 8) additive index, as further developed by Metcalf (2000).
Our scheme is based on a total of seven presidential powers relevant to parliamentary
systems. In each category, a constitution can score a maximum of four points if the
President can single-handedly employ the respective power (i.e. if the President is
essentially a dictator in this domain). The more the President is constrained, the fewer
points are given, and, of course, no points accrue if the President does not have the
respective authority at all. These scores are then added up across categories. As Shugart
and Carey (1992: 149) admit, such an additive index is far from perfect but, as they main-
tain, ‘it is preferable to a purely nonquantitative, impressionistic ranking or no assess-
ment of comparative presidential powers at all’.
As always, our concern is with enforceable powers, and we seek to discount those
that are not. In so doing, our point of departure is the various constitutions. Yet, iden-
tifying enforceable powers is a particular challenge with respect to parliamentary
Presidents. One reason is that the discrepancy between the letter of the constitution
and political feasibility can be astonishingly large. As Duverger (1980) and the country
chapter (Chapter 12) point out, it seems to be particularly huge in Iceland, where the
formal Constitution grants the President much broader powers than has been reflected
in political practice. France is at the opposite extreme, in that the President under uni-
fied government (in non-cohabitation situations) is stronger than the Constitution seems
to imply.
One cause of this discrepancy between the letter of the constitution and political
practice may be that presidential authority often includes dramatic and highly con-
sequential ‘emergency’ powers that are rarely exercised. Whether the President then
truly has the discretion to use them often cannot be determined through observation.
Yet, we should be careful not to jump to the conclusion that such powers cannot be
exercised. For example, many constitutions grant the President the right to refuse a
Prime Minister’s request for parliamentary dissolution. Yet, in practice we hardly
ever observe such a course of events. But as Strøm and Swindle (2002) argue, this non-
occurrence need not indicate that the President is powerless to refuse dissolution, but
Table 22.6. Presidential Selection and Political Powers (2000)
Election Right (with Right (with Right (with Other relevant powers
discretion) discretion) to discretion) to (in the President’s own
to dissolve appoint Prime dismiss right, i.e. without
parliament Minister government prior proposal)
Appointment
Even in parliamentary systems, the Head of State typically has a range of appointment
powers, the most important of which is the authority to appoint the Prime Minister
and other members of the cabinet. Presidential appointment powers in parliamentary
systems are by definition constrained by the fact that the President must appoint a cab-
inet that is at least tolerated by Parliament. Thus, the President cannot succeed if a
cohesive parliamentary majority firmly opposes his or her choice. Hence, few
European Heads of State have been able to pick and choose their Prime Ministers.
Yet, it has not been unusual for Presidents in some of our countries to exert signific-
ant influence on prime ministerial and cabinet appointments. This has most notably
been the case in the French Fifth Republic, when the President’s coalition has also con-
trolled the parliamentary majority (Thiébault 2000). But France is not unique. In Italy,
the appointments of the Amato (1992–3), Ciampi (1993–4), and Dini (1995–6) cabinets,
which were crucial in the transition from the ‘first’ to the ‘second’ republic, were
largely due to the President (Pasquino and Vassallo 1995; Verzichelli and Cotta 2000).
Italian Presidents had influenced prime ministerial selection even before that time,
Notes: The final column is an additive index of the other columns, which are scores of presidential powers (the higher the score, the more powerful the President). For
sources and coding explanation, see the chapter text.
678 Kaare Strøm et al.
exploiting bargaining failures and internal splits and rivalries in the now-defunct
Democrazia Cristiana (Hartmann and Kempf 1989: 140–2; Hine 1993: 99, 320 note 26).
Presidents representing smaller parties have often facilitated the cabinet participation of
their respective parties. Yet, the most notorious case of presidential intervention was
the premiership of Fernando Tambroni (1960), who was hand-picked by President
Giovanni Gronchi and survived precariously for four months with neo-Fascist support
until severe left-wing riots brought down his government (Mack Smith 1997: 439–40).
Finnish Presidents ‘have determined which parties are “eligible” for entrance into a cab-
inet, put pressure on parliamentary groups and hesitant ministerial candidates, and
expressed a clear opinion concerning the coalition base’ (Nousiainen 2000: 271). In
1987, President Koivisto decisively influenced the formation of the unusual coalition of
Conservatives and Social Democrats and the selection of Prime Minister Harri Holkeri.
Presidents have exerted particular influence over cabinet formation when the party
system has been fragmented or the dominant parties uncohesive. This has either led to
party-based cabinets, with the party composition strongly influenced by the President,
or to technical (non-partisan) cabinets. Portugal, for example, saw three presidential
cabinets in 1978–9 (Magone 2000). Yet, Presidents do not always succeed. Austrian
President Thomas Klestil tried to return a ‘grand coalition’ government in 1999–2000
but eventually had to swear in the cabinet he had tried to avoid. Italian President
Gronchi’s 1960 intervention, mentioned above, was a notorious embarrassment.
The list that we have compiled comprises most of the postwar exceptions to parlia-
mentary dominance over cabinet appointments. The Presidents of Ireland and Finland
(since March 2000) have at best formal duties in cabinet appointment. In the case of
Greece, Article 37 of the 1976 Constitution already tightly circumscribed presidential
appointment powers by providing more precise constitutional appointment instructions
than any other Western European constitution (see Chapter 11). The 1986 constitutional
amendment removed what limited discretion had remained in the hands of the
President (Article 37). Thus, the cross-national norm is that presidential appointment
powers are merely formal and only under special circumstances become more than that.
In Table 22.7, because of regime differences and some ambiguities in the original cod-
ing scheme, we score presidential appointment powers somewhat differently from
Shugart and Carey (1992: 152–3) and Metcalf (2000). One reason is that formal presi-
dential appointment powers interact with the rules of cabinet inauguration, specifically
the distinction between positive and negative parliamentarism. Presidents are most pow-
erful when they make cabinet appointments directly and when there is no required sub-
sequent investiture vote (negative parliamentarism). Thus, we score Presidents who can
at least formally select individual ministers four under negative and three under positive
parliamentarism, and Presidents who can appoint ministers only on the proposal of the
Prime Minister two under negative and one under positive parliamentarism. Figure 22.1
contains the appropriate placements of countries with scores between one and four.
Cabinet Dismissal
Cabinet dismissal is the flip side of cabinet appointment. Yet, while appointing cabinets
is a normal duty of Heads of State, terminating them at will is not. According to
Dimensions of Citizen Control 679
Parliamentary Dissolution
In all but one (Norway) of our seventeen countries, Parliament can be dissolved before
the end of its regular term, or, as it is sometimes called, the constitutional inter-
election period (CIEP). Though there are many forms of dissolution power (see
Chapter 4; Strøm and Swindle 2002), the Head of State often plays an important role,
680 Kaare Strøm et al.
at least formally. Yet, in most parliamentary democracies, the President’s dissolution
power is constrained. Moreover, in practice the President’s dissolution right is of
greater relevance if he or she is free to appoint and dismiss the government.
According to the formal constitutions, only the Icelandic, Greek (until 1986), and
Finnish (until 1991) Presidents enjoy unrestricted dismissal powers. The respective con-
stitutions simply assert the President’s right or are very unspecific about the conditions
that permit dissolution. Hence, the Greek 1975 Constitution allowed dissolution if the
President considered that Parliament ‘was not in harmony with popular feeling’
(Article 41). Between 1991 and 2000, the Finnish President was restricted by the require-
ment of a ‘reasoned initiative by the Prime Minister’ and consultations (Section 27).
Likewise, the Irish President can dissolve Parliament on the advice of the Prime
Minister (Article 13). The Austrian President faces only a very soft constraint in the
vague clause that Parliament can be dissolved ‘only once for the same reason’. The
President can do so only on the proposal of the cabinet (which, however, the President
appoints and can single-handedly dismiss). The Presidents of France, Italy, and Portugal
have to engage in non-binding consultations, but are more seriously constrained by var-
ious time restrictions (see Table 22.5). Finally, the only specific situation in which the
German President can dissolve the Bundestag is when it fails to elect a Chancellor by a
majority of all MPs (Article 63) or when the Chancellor fails to win an absolute major-
ity in a vote of confidence (as with Helmut Kohl in 1983). Since 1986 and March 2000,
respectively, the Greek and Finnish Presidents have no dissolution powers.
In accordance with Shugart and Carey (1992) and Metcalf (2000), in Table 22.7 we
score a President with unrestricted dissolution authority as four. If there are restrictions
with regard to frequency or timing within the President’s or Parliament’s term, the score
is three. A score of two means that parliamentary dissolution can occur only after a pres-
idential election. If dissolution is only allowed as a response to specific events (e.g. when
the cabinet loses the confidence of Parliament), we report the score as one. Finally, if the
President has no dissolution power, the score is zero. If a President can dissolve only
upon the request of the cabinet or Prime Minister, we reduce the score by 0.5 points.
could be several years. And, of course, there was the chance that a new parliamentary
majority would share the President’s preferences. Under the post-1987 rules, however, the
President can delay legislation for three months only. Yet, the Finnish President has not
let the veto fall into complete disuse. Anckar (1990: 36) reports sixty-four vetoes for the
1919–78 period; Nousiainen (1998: 354) seventy vetoes 1919–95. However, none of these
vetoes has caused great conflict. The most frequent presidential rationale has been formal
(relating to due legislative process or aspects of the bill’s drafting) or financial, and all but
six (four of which occurred in 1990) were sustained by Parliament.11
In Italy the President can veto (‘request a new deliberation’, Article 74), but a second
vote with exactly the same majority in both chambers is sufficient to override.
Although in practice Presidents make very limited use of their suspensive veto (specific-
ally, Sandro Pertini seven times in his presidential term 1978–85, his successor
Francesco Cossiga four times in the first two years of his term), their vetoes are gener-
ally sustained by Parliament and the bills amended accordingly. In most cases, Italian
presidents have intervened to block bills that lacked the necessary revenue provisions
(Hartmann and Kempf 1989: 149).
The Portuguese President has stronger veto powers. On certain issues specified
in the Constitution, his veto can be overridden only by a two-thirds parliamentary
majority.12 The 1982 constitutional reform actually enhanced presidential veto powers
by extending the number of policy areas in which an override requires a two-thirds
majority (Chapter 18). As noted in Chapter 18, the Portuguese President has made
relatively frequent use of the veto, particularly in situations of ‘cohabitation.’
Building on Metcalf (2000), in Table 22.7 we score a presidential veto that cannot be
overridden as four, whereas three means override by extraordinary majority, two over-
ride by an absolute majority of the whole parliamentary membership, one override by
a simple majority of quorum, and zero no veto powers. Iceland does not fit this
scheme. The President can veto legislation, but this does not prevent the legislation
from coming into force, and it triggers not another vote in Parliament, but a referendum.
Depending on the nature of the President’s conflict with Parliament, we therefore
score this power in the range between the one and three.
Decree Powers
A less dramatic, but often more significant, presidential prerogative is the power to
issue decrees or administrative orders. By presidential decree powers Shugart and
Carey (1992: 151) refer to the ability of the President ‘to make new laws or suspend old
ones without the power of decree first having been delegated through enabling legis-
lation’. Presidents in Western European parliamentary democracies generally do not
possess far-reaching decree powers. There are, however, two exceptions. Article 16 of
the French Constitution grants the President the right to ‘take the measures demanded’
when ‘the institutions of the Republic, the independence of the nation, the integrity of
11
Thanks to Matti Wiberg for clarifying this point.
12
A two-thirds majority is required for organic laws on foreign relations, issues of private or public own-
ership, and some election rules (Article 139).
682 Kaare Strøm et al.
its territory, or the fulfilment of its international commitments are under grave and
immediate threat and when the proper functioning of the Constitutional governmen-
tal authorities is interrupted’. And Article 28 of the Icelandic Constitution allows the
President to issue provisional laws ‘in case of urgency’ and when Parliament is not in
session. These must not contradict the Constitution and need to be submitted to
Parliament as soon as it convenes. Yet, both of these decree powers are comparatively
weak, as reflected in Table 22.7.
Referendum
In a later section, we shall examine direct democracy as an external constraint on the
parliamentary chain of delegation. Here, we note that under some constitutions
Presidents, as intimated above, are empowered to call such referendums. Thus, in
Greece until 1986, the President could single-handedly proclaim referendums ‘on crucial
national issues’ (Article 44). Since that time, however, the President has had only a for-
mal role in the referendum procedure. In France and Portugal, the President decides on
the holding of referendums but only upon a cabinet proposal. We have already consid-
ered the authority of the Icelandic Head of State: it is reactive only, conditional upon a
piece of legislation that does not find the President’s approval. Otherwise the Icelandic
Constitution features only automatic referendum provisions. Hence, in practice the
Icelandic President has monopoly agenda control over referendums. In Table 22.7 we
apply the codings established by Shugart and Carey (1992) as revised by Metcalf (2000).13
Judicial Review
Judicial constraints on parliamentary power will be discussed at length below.
Sometimes, a Head of State may be instrumental in imposing this constraint through
his power to submit legislation to review by a constitutional court prior to promulga-
tion. The Presidents of France, Ireland, and Portugal can thus initiate constitutional
review. Until 1974 the French President shared that right with the Prime Minister and
the Presidents of both chambers. Since then even a parliamentary minority enjoys this
power. In Ireland the President alone can appeal to the constitutional court, yet she is
restricted by the exclusion of certain matters from judicial appeal.14 The President of
Portugal has an exclusive but unrestricted right to pre-promulgation constitutional
appeal. In Table 22.7, we apply the same coding rules as Metcalf (2000).15
13
If the President acting alone can call a referendum (i.e. he is a dictator or at worst a decisive player), we
score it a four. The score is two if presidential powers are ‘restricted’. Metcalf (2000: 671) understands this as
a ‘shared’ power ‘where either the President or the assembly acting alone can propose referendums’, that is,
when each is a decisive player. The score is one if the President and the cabinet must agree on the use of the
referendum (both are veto players). Finally, zero means that the President has no authority to call a referendum.
14
Up to the end of 1998, the Irish President had appealed to the constitutional court on twelve
occasions (Gallagher 1999: 83).
15
If the President has exclusive power to refer legislation to a constitutional court prior to promulgation,
the score is four. If the right is shared and either the President, the cabinet, or a majority of the assembly (or
its presiding officer) can request judicial review, the score is two. If the right is shared so that the President,
the cabinet, or some parliamentary minority can invoke judicial review the score is one. Finally, if the
President has no such authority, the score is zero (Metcalf 2000: 672).
Dimensions of Citizen Control 683
Summary
Table 22.7 presents an overview of all these presidential powers, employing our revised
version of the indices of Shugart and Carey (1992) and Metcalf (2000). Note that few
Western European Heads of State have substantial legislative or judicial powers, but
that many are more significantly involved in the making and breaking of cabinets and the
calling of elections. According to formal powers, the presidential rank order sees the
Icelandic President on top, neatly reflecting the country’s geographical position
between parliamentary Europe and presidential America. The Portugese President is
next in formal powers and clearly outranks the Icelandic one in the actual exercise of
these powers. Even the Presidents of Finland and Greece had considerable powers in
earlier decades, but are now among those least favoured. In contrast, the French
President has held on to his powers and is well known to exercise them to the extent
that the political situation allows. The Presidents of Austria, Ireland, and Italy all hold
some important formal powers. In terms of their real impact on parliamentary politics,
however, the Italian President ranks first and the Irish last. Finally, the German Head of
state has been rendered particularly powerless because of the painful history of the
presidency during the Weimar republic. Overall, then, although the Constitutions of
newly democratic Greece and Portugal featured powerful Presidents, the longer-term
trend has been towards a weakening of the formal and real powers of Western
European Heads of State, including the Presidents.
Central Banks
The Head of State is not the only actor within the executive branch that may constrain
the parliamentary policy process. In a number of countries, various executive agencies
have been granted significant autonomy. Foremost among these, particularly since the
early 1990s, have been national banks. The Blair government’s move in 1997 to grant
effective autonomy to the Bank of England is perhaps most widely known, but it is not
the only example and certainly not the most far-reaching. The most consequential such
development, of course, is related to the growing scope of EU jurisdiction, and specif-
ically the introduction of the European Monetary Union (EMU) and the European
Central Bank, which is as independent as any national central bank (de Haan and
Eijffinger 2000). Yet, the evolution of more autonomous central banks is not restricted
to the ‘euro zone’ (the EMU states), as Norway and Sweden demonstrate (and we
might have been able to refer to more such cases, had our survey covered a larger num-
ber of non-EU countries).
Table 22.8 reports two existing measures of central bank independence, both
based on the Cukierman–Webb–Neyapti index for the 1960s and the 1980s. The index
numbers reflect very little change. Note, however, that more recent developments,
such as the reforms in Belgium and Spain in the early 1990s, and in the United Kingdom
in 1997, all of which enhanced the independence of central banks, have not been
captured in these data. Greater central bank independence is thus a significant, and
well recognized, institutional development that serves to constrain parliamentary
policy-making.
684 Kaare Strøm et al.
Table 22.8. Central Bank Independence
Judicial Constraints
Our ideal–typical conception of parliamentary democracy privileges the voice of the
people, in that we assume democratic representation to be rooted in popular sov-
ereignty, in the idea that government policy ultimately should reflect popular preferences.
Yet, there is a significant alternative or complementary conception of democracy as a
way of protecting and promoting a bundle of civic and political rights and liberties.
This vision privileges instead a constitutional order that protects and promotes the
choices and preferences of the constitutional framers. Among the most serious con-
straints on democracy as popular sovereignty are thus the institutions that serve to pro-
tect constitutional (or liberal) democracy. Foremost among them are constitutional and
other important judicial institutions.
During the second half of the twentieth century, Western states evolved toward a
rights-oriented vision of democracy, a trend that is very visible in our sample. This judi-
cialization of parliamentary democracy has taken a number of different but mutually
reinforcing forms. One is the increased access to constitutional courts, and the growing
number of constitutions that have established such courts. A second important trend has
been the increasing activism of new as well as established constitutional or supreme
Dimensions of Citizen Control 685
courts. A third has been the growing willingness of ordinary courts to challenge demo-
cratically elected authorities. Finally, but certainly not least, the EU member states, and
even the EEA countries, have been profoundly affected by the growing authority of the
European Court of Justice and other supranational judicial institutions, particularly those
entrusted with the enforcement of the European Convention of Human Rights (ECHR).
16
The respondents in this survey were asked the following question: ‘Do you think that your work
as a judge of the Constitutional Court consists primarily of norm enforcement or of the development of
the law?’
Dimensions of Citizen Control 687
sixty-day period. However, in 1996 it declared all recycled decree laws unconstitutional
(Volcansek 2000: ch. 2). And whereas as late as 1987 and 1991 the Court favoured the
government parties by blocking referendums that were likely to undermine party rule,
in 1993 it changed its ruling and allowed a referendum challenging the Senate’s elec-
toral system, eventually leading to the transition to the so-called second Italian
republic (de Franciscis and Zannini 1992: 77; Volcansek 2000: ch. 6).
In some cases, increased judicial activism reflects a greater willingness on the part of
judges to challenge politicians. In other cases, it may be due to an increasing incongru-
ence between the political orientation of the judges and the parliamentary majority. In
Austria, until the formation of the ÖVP–FPÖ coalition in 2000, all judges had been nom-
inated by one of the traditional government parties, the SPÖ and the ÖVP. Hence, their
growing willingness to challenge government policies and to give up judicial restraint
first and foremost indicates greater independence. In Italy, judges became more willing
to disregard the wishes of their nominating parties when the political system came
under siege and particularly after the traditional parties collapsed (Volcansek 2000). In
France, the various alternations in the partisanship of Presidents and parliamentary
majorities since 1981 have led to frequent political incongruence between the govern-
ment of the day and the Constitutional Council. Hence, a left-wing government may be
confronted with a right-wing Council, or vice versa (Wright 1999: 115). Finally, the
increased power of constitutional courts is reflected in the anticipation of court deci-
sions by Parliaments, and in the reduced scope for parliamentary decision-making under
the Constitution as interpreted by the courts. By these mechanisms, ‘legislators are
gradually placed under the tutelage of the constitutional court’ (Stone Sweet 2000: 194).
Ordinary Courts
Even ordinary courts may significantly constrain the parliamentary chain of delega-
tion, for example, by placing restrictions on individual politicians. Since the 1980s the
number of spectacular cases of legal action against incumbent politicians has dramatic-
ally increased. Without doubt Italy’s ‘clean hands’ investigation is the most prominent
case (Nelkin 1996a,b). Yet, there have been important cases even in France (Wright
1999: 104, 110–2), Germany, Spain, Belgium, and Austria.
Austria 3* 3 n/d
Belgium 3* 3 3.5
Denmark 2 1 n/d
Finland 1 1 n/d
France 3* 4 3.7
Germany 4* 4 3.46
Greece 2 2 n/d
Iceland 2 1 n/d
Ireland 2 2 n/d
Italy 2.8* 4 3.33
Luxembourg 1 1 n/d
The Netherlands 1 2 4.2
Norway 2 1 n/d
Portugal 2* 3 n/d
Spain 3* 3 2.0
Sweden 2 2 2.5
UK 1 2 2.1
Notes: Low scores indicate judicial restraint, high scores judicial activism or discretion; Lijphart scale: Belgium
since 1984, before that 1.5; Italy since 1956, before that 2.0. *Centralized judicial review by special constitu-
tional courts.
Sources: Lijphart 1999: 226; Alivizatos 1995: 575; Cooter and Ginsburg 1996: 300.
Table 22.9 presents three attempts to measure the relevance of judicial constraints
cross-nationally. Lijphart’s ‘strength of judicial review’ variable aims to capture both
judicial review and the degree of judicial activism (Lijphart 1999: 225–6).17 Alivizatos’
(1995: 567) ‘politicisation of courts’ measure (1975–94) is meant to capture ‘to what
extent they influence the decision-making process’ (with scores ranging from 1⫽judi-
cial self-restraint to 4⫽judicial activism).18 Finally, Cooter and Ginsburg’s (1996) ‘judi-
cial discretion’ scores are based on the judgements of ‘a small number of comparative
law scholars’ who have rated the judicial ‘adventurousness’ of various countries on a
five-point scale (with 5 being the highest and 1 the lowest score). There is a high degree
of consistency between the three measures, although Lijphart’s classification fits best
with the interpretations of our country experts.
17
Lijphart’s entries are scored as follows: 4⫽strong judicial review, 3⫽medium-strength judicial review,
2⫽weak judicial review, 1⫽no judicial review. An asterisk in the table indicates centralized judicial review by
special constitutional courts.
18
Alivizatos (1995: 567).
Dimensions of Citizen Control 689
Direct Democracy
The final constraint on parliamentary policy-making that our contributors have ident-
ified is direct democratic instruments. European states generally became parliamentary
long before they became fully fledged democracies: the cabinet’s accountability to the
parliamentary majority was established before popular sovereignty was truly
entrenched. A constitution based on parliamentary sovereignty is thus not necessarily
conducive to direct popular decision-making. The contrary might more commonly be
the case, as our two historical cases from Chapter 1 (Britain and Sweden) illustrate. In
the United Kingdom, the referendum has no well-established place in the constitutional
order, and it is only since Britain joined the EU that even consultative referendums have
been held. Although direct democracy has been somewhat more common in Sweden,
Swedish governments have on two different occasions, after 1955 over right-hand-side
driving and after 1980 over nuclear power, effectively ignored referendums that had
equivocal or inconvenient outcomes.
Conversely, Switzerland is clearly not parliamentary while at the same time the
undisputed European leader in its reliance on direct democracy. In fact, the tension
between parliamentary and referendum democracy has never been fully resolved. In
the 1920s, Hans Kelsen suggested direct democracy as a remedy against the democratic
deficiencies of parliamentarism (see Chapter 1), an idea that has more recently been
forcefully championed by Ian Budge (1996; see McLean 1989 for a more sceptical view).
f
In Iceland, any change of the status of the national Lutheran church must be approved in a referendum.
g
In Iceland, a referendum is held only in case of conflict over legislation between Parliament and the
President.
h
In Italy, an abrogative referendum is held on the demand of one-fifth of the members of either chamber
of Parliament or 500,000 voters or five regional Councils.
i
In Italy, 500,000 voters or five regional councils can demand a referendum on the total or partial repeal of
a law or of an act having the force of law.
j
In Italy, 50,000 citizens can initiate legislation.
k
In Portugal, a referendum is held only on the condition of approval by the Parliament.
l
In Spain, a partial revision of the Constitution is subject to a referendum when it is demanded by one-tenth
of the members of either chamber of Parliament.
m
In Spain, the Constitution permits 500,000 citizens to propose legislation. This initiative is not applicable
to organic laws, taxation, or international affairs, nor to the royal prerogative of pardon.
n
In Sweden, a referendum is held on the demand of one-third of all MPs.
Sources: Country chapters; constitutions; Gallagher and Uleri 1996.
Dimensions of Citizen Control 693
Austria 0 0 0 1 0 1 0 2
Belgium 0 1 0 0 0 0 0 1
Denmark 0 2 6 3 1 3 1 16
Finland 0 0 0 0 0 1 0 1
France 4 1 4 1 1 2 0 13
Germany 0 0 0 0 0 0 0 0
Greece 1 0 1 2 0 0 0 4
Iceland 2 0 0 0 0 0 0 2
Ireland 0 1 2 5 4 10 0 22
Italy 1 0 0 3 12 32 7 55
Luxembourg 0 0 0 0 0 0 0 0
The Netherlands 0 0 0 0 0 0 0 0
Norway 0 0 0 1 0 1 0 2
Portugal 0 0 0 0 0 2 0 2
Spain 1 0 1 2 1 0 0 5
Sweden 0 2 0 0 1 1 0 4
United Kingdom 0 0 0 1 0 0 0 1
Total 9 7 14 19 20 53 8 130
parliamentary
democracies
Switzerland 17 43 26 81 60 100 6 333
Total 26 50 40 100 80 153 14 463
No. of countries 6 6 6 10 7 10 3 14
using
referendums
Note: Only the last two Greek and the last three Spanish referendums were held during the most recent tran-
sition to democracy or under democratic Constitutions.
Source: Setälä 1999: 333 (Updated for 1999 and 2000 from various sources, including the EJPR Data Yearbook
and personal communication in the case of Ireland).
Source: Setälä 1999: 335 (Updated from various sources, see Table 22.11).
694 Kaare Strøm et al.
C O N C LU S I O N : C RO S S - NAT I O NA L PAT T E R N S
This chapter has further highlighted the substantial variation in accountability regimes
among real-world parliamentary systems. Indeed, there are three dimensions of citizen
control along which such regimes differ. One is the constitutionally mandated chain of
citizen control that was the focal point of Chapter 4 and the country chapters. Even
though all parliamentary democracies share the few essential features that were out-
lined in Chapter 3, by no means do they have identical working constitutions. In our
seventeen polities, there is everywhere an indirect and singular chain of national policy-
making, but these respective chains are made up of different electoral systems, differ-
ent types of parliamentarism (positive or negative), different rules for no confidence
votes (although such votes exist everywhere), and different distributions of dissolution
powers, to mention only some of the variations. The country-specific chains promote
different degrees of agenda control by the agents. Although traditionally, parliamentary
democracy has been characterized by strong ex ante partisan control of democratic
agents, and by comparatively fewer and weaker ex post constraints, there is significant
variation along these dimensions as well. Thus, a more complete picture of parliamen-
tary democracy in modern governance must, as our country chapters have done, also
take into account cross-national variation and intertemporal shifts in partisan influence
and external constraints.
Partisan cohesion and external constraints help us understand accountability under
contemporary parliamentary democracy. While, classically, the parliamentary chain of
policy-making was designed without much regard to political parties, in the present
world political parties play a crucial role everywhere. Indeed, the classic Westminster
model combines the ideal–typical parliamentary institutions with two strong and cohe-
sive parties competing to control them. With an indirect, single, and unidirectional chain
and two dominant and cohesive parties competing for power, voters can be reasonably
sure that they can ensure accountability by voting the current leaders out of power.
However, in polities in which the party systems are more fragmented and/or the parties
less cohesive, and where government portfolios are distributed through post-electoral
negotiations, citizen policy control through political parties may be harder to achieve.
The way in which political parties coexist with the parliamentary chain thus partly deter-
mines the way in which accountability works. Similarly, in a polity where there are few
subnational, societal, or international constraints on decision-making, the constitutional
chain will be more decisive for national policy-making than in a polity in which political
decision-making is heavily circumscribed by constraints outside of the parliamentary
chain. The partisan and external constraint dimensions are thus tremendously impor-
tant for how parliamentary government works in the real world.
While the partisan and external constraints dimensions are analytically distinct, there
is reason to believe that to some extent they are mutual substitutes (see Chapter 3).
Thus, under strong and cohesive party government, ex post constraints may be redun-
dant. Yet, as we have noted, political systems evolve only gradually and may not equi-
librate to a political optimum. Thus, for example, even when institutional checks are
deliberately designed to compensate for decaying partisan controls, it takes time both
to recognize this decay and to design and establish the appropriate institutional
Dimensions of Citizen Control 695
response. Conversely, when decision-making institutions are freed from partisan con-
trol, the occupants may need time to entrench and make use of their freedom. For
these reasons, we should not expect a neat trade-off between partisan control and
external constraints. We shall return to this point in Chapter 23.
Political Parties
It remains for us systematically to summarize the importance of the two additional
dimensions of citizen control. Consider first partisan control. Table 22.13 draws
together the cross-national variation presented earlier in the chapter, and particularly
the most recent observations. We have discussed partisanship under three rubrics: the
electoral party, the party organization, and the party in government. As argued above,
we consider the party in government the most critical facet of partisan control, fol-
lowed by the party in the electorate, while the extraparliamentary organization is the
most dispensable component. In Table 22.13, the seven indicators of the strength of
partisanship reflect this hierarchy: four pertain to the party in government, three to the
electoral party. Each indicator has been introduced previously in this chapter. Together
they provide a summary picture of the state of ‘partyness’ in our seventeen democra-
cies. We normalize each measure by taking the maximum observed value (showing the
greatest partisan influence) and setting it equal to one, whereas the minimum observed
value is set equal to zero. All observations in between are then scored proportionally.
We have generated a simple additive index of partisanship by summing the scores on
the seven indicators and dividing the sum by seven.
There is good reason in our data as well as in the literature to question the ability of
contemporary political parties, even in their erstwhile strongholds, to perform many of
their traditional functions as mechanisms of citizen control (Dalton and Wattenberg
2000). Earlier in the chapter we have shown that the party-in-the-electorate, in particu-
lar, has declined throughout Western Europe. Yet, the magnitude of this decline is by
no means the same everywhere. To put it bluntly, even if the parties are dying, they are
certainly not all dying equally fast! In fact, partisan rule remains strong in a number of
countries.
In our summary, one country stands out in terms of party strength: the United
Kingdom, which has the highest score on four of seven indicators of party strength and
cohesion. There are indications that the extraparliamentary organizations have eroded,
but otherwise the Westminster model remains strong. Only in one measure of partisan
influence is the United Kingdom weak, namely the extent to which partisanship per-
meates the public administration. This contrasts with the country that comes in a close
second, Greece, in which partisanship is also much more prominent in government
appointments. The other more recent democracies of southern Europe, Portugal,
and Spain, are next in line in partyness. All of these are clearly alternational party
systems.
The Netherlands and Italy, whose edifice of ‘partitocrazia’ virtually collapsed in the
early 1990s, receive the lowest scores. Interestingly, it is the smaller, multiparty
Table 22.13. Partisan Influence on Parliamentary Delegation: Summary
696
Country Party in the electorate Party in government Total
High party Low share of High degree High High Low High partisan Mean Ranking
identification parliamentary of uniform proportion coalition partisan influence of
seats for non- electoral of single- discipline conflict over of country
governing swing for all party (1975–99) (1975–99) government
parties cabinet parties majority country
cabinets appointments
Notes: For three variables (party ID; share for non-governing parties; and partisan influence over appointments), the entries are based on the most recent data for each
country. For the other four variables we use national means for all or parts of the postwar period (uniform swing; single majority cabinets; coalition discipline; and par-
tisan conflict). The entries are based on tables presented earlier in this chapter. Partisan influence over appointments is a composite of the three measurers that were pre-
sented in Table 22.3. For example, Belgium has a score of one (high influence) because partisan influence is coded as strong on all three measurers in Table 22.3.
Denmark, on the other hand, is scored at zero because there is no indication of partisan influence on appointments on any of the three measurers in Table 22.3. For all
seven indicators, numbers are normalized so that the country with the highest level of partisan influence in each category is scored 1, and the country with the lowest
is scored 0.
Dimensions of Citizen Control 697
democracies of the Belgium, Finland, and Luxembourg that come closest to Italy and
The Netherlands in the weakness of partisan control. This larger group of countries is
also furthest from the Westminster models ideal of a government controlled by either
of two alternating parties or blocs. Instead, these low-partisanship countries are clearly
pivotal.
In Greece and the other newer democracies of southern Europe, Portugal, and
Spain, political parties are thus comparatively strong. There may be many reasons that
party democracy thus seems to be more vital in southern Europe than in the north. In
states that have relatively recently experienced dictatorship, the collective memory of
less attractive alternatives to parliamentary democracy may still be fresh, and political
parties may benefit from the role that they played in opposition to dictatorship or in the
transition to democracy. And although some dominant parties, particular PASOK in
Greece and the PSOE in Spain (see Share 1999), have clearly established patronage net-
works, they may simply have had less time to sully their reputations. Moreover, party
attachments may be fed by societal conflicts that are still more present than the more
archaic cleavages of the longer lived parliamentary democracies. In any case, the real
impact of the strength or weakness of partisan influence can only be assessed in rela-
tion to the larger picture of citizen control.
External Constraints
The role of external constraints in promoting citizen control is more complex than that
of political parties. When political actors and institutions that are not directly or indi-
rectly the agents of the citizens impinge on the national policy process, they become
external constraints on parliamentary democracy. External constraints can serve any
number of purposes, some of them decidedly undemocratic. In this volume, however,
we have focused on how external constraints can help citizens control their agents.
Thus, the democratic purposes of external constraints include protecting rights and lib-
erties, facilitating the provision of public goods, checking the power of particular
agents, dispersing power in general, and enhancing political transparency. External con-
straints can take quite a number of different forms. Some, such as corporatism, are soci-
etal, whereas others, such as autonomous central banks, are technocratic, and yet others,
such as directly elected presidents and popular referendums, are democratic. Clearly,
only the last form represents control of elected politicians by ordinary citizens them-
selves. Yet even the other forms would be ineffective without citizen support.
We have throughout much of this book generally associated partisanship with
ex ante citizen control and external constraints with ex post mechanisms. Yet, external
constraints can in fact affect the national policy process both before and after the act of
delegation. For instance, when a court of law or the voters in a referendum strike down
a piece of legislation after it has been enacted and promulgated, they function as an
ex post institutional check (or veto player) on the parliamentary representatives. In other
cases, however, referendum voters and court opinions can have advisory purposes, or
they can narrow the range of options of the parliamentary majority (Stone Sweet
2000), or as powerful players influence the legislative process ex ante. Yet, most of the
698 Kaare Strøm et al.
external constraints that we have discussed are first and foremost
ex post institutional checks. Table 22.14, constructed in analogue fashion to Table 22.13,
contains seven indicators of external constraint.19
Note that the southern European countries are not particularly externally con-
strained. One might have expected their modern and largely republican constitutions
to feature many constraints (e.g. referendum devices) commonly associated with con-
temporary constitutional design. Instead it is the smaller countries of nothern and
central Europe that exhibit the most significant sets of external constraints on their
national policy processes. Austria, Finland, and Iceland are the most prominent cases.
Denmark, Ireland, and, as the first of the major countries, Germany, follow suit. The
United Kingdom is at the other end of the scale, quite apart from most other European
countries in its lack of external constraint, but followed most closely by Belgium.
As in the cases of partisanship, we see definitive temporal trends in the importance
of external constraints. Clearly, as shown by the tables presented earlier in the chapter,
the significance of judicial, direct democratic, subnational, and international con-
straints have increased. On the other hand, corporatism, consociationalism, and other
societal constraints have mostly been in decline, as have some measures of partisan
influence and organization. Interestingly, it is thus the constraints that are most inim-
ical to partisan control that have particularly grown in importance, whereas the societal
constraints, which have traditionally helped parties perform their screening functions,
have decayed with them.
To be sure, the two dimensions of citizen control (partisan preference alignment and
external constraints) are not mutually exclusive, since no modern democracy is
completely without either. And having more of one does not necessarily condemn you
to less of the other. For example, the United States increasingly seems to feature
strong partisanship as well as strong external constraint. And as noted above, partisan-
ship and societal constraints may in fact reinforce one another. Nevertheless, since
partisan ex ante mechanisms and external constraints are to some extent rival and sub-
stitutable mechanisms of citizen control, we expect an overall negative association
between the two.
Figure 22.2 plots the most recent observations of all of our seventeen countries on
both dimensions. Note that countries that score high on one dimension, such as Britain
and Greece on partisan influence, or Denmark on external constraints, tend to be low on
the other. Thus, the two countries that exhibit the highest degrees of partisan control are
19
On our first indicator, international constraints, the scoring of the non-EU member states, Iceland
and Norway, reflects our interest in the EU as a constraint on national governments. This constraint, we
argue, is least among states that do not belong to this organization. Therefore, we score Norway and
Iceland as minimally constrained on a par with the larger member states (France, Germany, Italy, and the
United Kingdom). While in the internal market they are significantly constrained by EU decision-making,
outside the internal market (and the EEA agreement) national policy-makers in these countries do not face
the same constraints as those in the EU member states. On the other hand, to a larger extent than for the
member states, the national government remains the highest level at which political accountability can be
exercised.
Table 22.14. External Constraints on Parliamentary Delegation: Summary
Country Restrictions Subnational Corporatism c Presidential Central bank Strength of Strength of Mean Ranking
on voting government powersd independencee judicial referendum of
power in EU consumptionb reviewf institutiong countries
Councila
Notes: The entries are based on tables presented earlier in this chapter. Numbers are normalized so that the country with the highest level of external constraint in each
category is scored 1, and the country with the lowest is scored 0.
a
Based on the Banzhaf index from Table 22.5. Norway and Iceland (not EU members) are coded as having minimal constraints in this category.
b
Based on the proportion of local and regional government final consumption in relation to general government final consumption, Table 22.5.
c
Based on the Lane and Ersson corporatism index, Table 22.5.
d
Based on the total presidential power score, Table 22.7.
e
Based on the central bank independence index, Table 22.8, 1980–9.
f
Based on Lijphart’s index of judicial review from Table 22.9.
g
Based on the number of types of referendum that are available in each country, Table 22.10. For example, Austria has all forms of referendums (coded as 1) while
Belgium (coded as 0) has no referendums at the national level.
700 Kaare Strøm et al.
UK
GRE
POR
SPA
Partisan influence
FRA GER
NOR AUS
SWE
ICE DEN
LUX IRE
FIN
BEL
ITA
NET
0
0 1
External constraints
Notes: Countries are listed using abbreviations that are identical to Lijphart’s (1999); The data is taken from
Tables 22.13 and 22.14; The measures are correlated at ⫺0.14 (Pearson’s r).
simultaneously at the low end of external constraints. Indeed, our two summary indices
correlate negatively, though at the relatively weak level of ⫺0.14 (Pearson’s correlation).
Yet, countries that are low on one dimension need not be high on the other. There are
countries, most notably Austria, that are well endowed with both dimensions of citizen
control. And, more ominously, there are those that are lacking in party cohesion as well
as external constraints. Italy, for example, scores comparatively low on both dimensions
of citizen control, and so does The Netherlands. While post-fascist Italy is one of the
clearest cases of the trend away from partisan control and towards external constraint,
it also appears that since the early 1990s the evolution of institutional checks has not
kept pace with the collapse of partisanship. On the other hand, the majoritarian Italian
electoral reforms of the 1990s represent a fascinating attempt to strengthen party
cohesion at the same time that institutional constraints are becoming increasingly
effective.
In addition to Italy, the Benelux countries are the most prominent cases of weakness
on both dimensions. While the level of partisan influence varies somewhat, external
constraints are in all these four countries rather weak, as they have largely ceremonial
Heads of State and mostly feeble judicial institutions. In fact, the weakness of both
dimensions of citizen control in these countries depresses the association between
these two dimensions in our entire sample. Outside of these four countries and Austria,
Dimensions of Citizen Control 701
REFERENCES
Aldrich, John H. (1995). Why Parties? The Origin and Transformation of Political Parties in America.
Chicago: The University of Chicago Press.
Alivizatos, Nicos (1995). ‘Judges as Veto Players’, in Herbert Döring (ed.), Parliaments and Majority
Rule in Western Europe. Frankfurt: Campus and New York: St. Martin’s Press.
20
Hadenius (2001) reaches similar conclusions on the cross-national variation but places less emphasis on
the trends that we identify here.
702 Kaare Strøm et al.
Anckar, Dag (1990). ‘Democracy in Finland: The Constitutional Framework’, in Jan Sundberg
and Sten Berglund (eds.), Finnish Democracy. Jyväskylä: Finnish Political Science Association.
Andeweg, Rudy B. (1997). ‘Institutional Reform in Dutch Politics’. Acta Politica, 32: 227–57.
Bekke, Hans A. G. M. and van der Meer, Frits M. (eds.) (2000). Civil Service Systems in Western
Europe. Cheltenham: Edward Elgar.
Bendix, Reinhard (1978). Kings or People: Power and the Mandate to Rule. Berkeley: University of
California Press.
Blom-Hansen, Jens (2001). ‘Organized Interests and the State: A Disintegrating Relationship?
Evidence from Denmark’. European Journal of Political Research, 39: 391–416.
Blondel, Jean and Cotta, Maurizio (eds.) (1996). Party and Government. Houndmills: Macmillan.
—— —— (eds.) (2000). The Nature of Party Government. Houndmills: Palgrave.
Bowler, Shaun and Donovan, Todd (2000). Demanding Choices. Opinion, Voting, and Direct
Democracy. Ann Arbor: University of Michigan Press.
—— —— and Tolbert, Caroline J. (eds.) (1998). Citizens as Legislators: Direct Democracy in the
United States. Columbus: Ohio State University Press.
Budge, Ian (1996). The New Challenge of Direct Democracy. Cambridge: Polity.
Coakley, John (1999). ‘Society and Political Culture’, in John Coakley and Michael Gallagher
(eds.), Politics in the Republic of Ireland. London: Routledge.
Colomer, Josep M. (1998). ‘The Spanish “State of Autonomies”: Non-Institutional Federalism’.
West European Politics, 21(4): 40–52.
Comparative Parliamentary Democracies Data Set ( January 2002). This data set combines the
institutional mechanisms detailed in Chapter 4 of this book with the behavioural data that is
published in Wolfgang C. Müller and Kaare Strøm (eds.) (2000). Coalition Governments in
Western Europe. Oxford: Oxford University Press.
Cooter, Robert D. and Ginsburg, Tom (1996). ‘Comparative Judicial Discretion: An Empirical
Test of Economic Models’. International Review of Law and Economics, 16: 295–313.
Cox, Gary W. and McCubbins, Mathew D. (1994). ‘Bonding, Structure, and the Stability of
Political Parties: Party Government in the House’. Legislative Studies Quarterly, 19: 215–31.
Cronin, Thomas E. (1989). Direct Democracy. The Politics of Initiative, Referendum, and Recall.
Cambridge, MA: Harvard University Press.
Cukierman, Alex, Webb, Steven B., and Neyapti, Bilin (1994). Measuring Central Bank Independence
and Its Effect on Policy Outcomes. San Francisco: ICS.
Dalton, Russell J. (2000). ‘The Decline of Party Identifications’, in Russel J. Dalton and Martin P.
Wattenberg (eds.), Parties without Partisans. Oxford: Oxford University Press.
—— (2002). Citizen Politics. 3rd edn. New York: Chatham House.
—— and Martin P. Wattenberg (eds.) (2000), Parties without Partisans. Oxford: Oxford University
Press.
de Franciscis, Maria Elisabetta and Zannini, Rosella (1992). ‘Judicial Policy-Making in Italy: The
Constitutional Court’. West European Politics, 15(3): 68–79.
de Haan, Jakob and Eijffinger, Sylvester C.W. (2000). ‘The Democratic Accountability of the
European Central Bank: A Comment on Two Fairy-tales’. Journal of Common Market Studies,
38: 393–407.
Diermeier, Daniel, and Feddersen, Timothy J. (1998). ‘Cohesion in Legislatures and the Vote of
Confidence Procedure’. American Political Science Review, 92: 611-21.
Drewry, Gavin (1992). ‘Judicial Politics in Britain: Patrolling the Boundaries’. West European
Politics, 15(3): 9–28.
Dubois, Philip L. and Feeney, Floyd (1998). Lawmaking by Initiative: Issues, Options and Comparisons.
New York: Agathon Press.
Duverger, Maurice (1954). Political Parties. London: Methuen.
Dimensions of Citizen Control 703
Page, Edward C. and Wright, Vincent (eds.) (1999). Bureaucratic Élites in Western European States.
Oxford: Oxford University Press.
Pappalardo, Adriano (1981), ‘The Conditions for Consociational Democracy: A Logical and
Empirical Critique’. European Journal of Political Research, 4: 365–90.
Pasquino, Gianfranco and Vassallo, Salvatore (1995). ‘The Government of Carlo Azeglio Ciampi’,
in Carol Mershon and Gianfranco Pasquino (eds.), Italian Politics. Ending the First Republic.
Boulder: Westview Press.
Petersson, Olof, Hernes, Gudmund, Holmberg, Sören, Togeby, Lise, and Wängnerud, Lena
(2000). Demokratirådets rapport 2000: Demokrati utan partier? Stockholm: SNS.
Powell, G. Bingham, Jr. (2000). Elections as Instruments of Democracy. Majoritarian and Proportional
Visions. New Haven: Yale University Press.
Qvortrup, Mads (2000a). ‘Checks and Balances in a Unicameral Parliament: The Case of the
Danish Minority Referendum’. Journal of Legislative Studies, 6(3): 15–28.
—— (2000b). ‘Are Referendums Controlled and Pro-hegemonic?’ Political Studies, 48: 821–6.
Riker, William H. (1964). Federalism: Origins, Operation, Significance. Boston: Little, Brown.
—— (1975). ‘Federalism’ , in Fred I. Greenstein and Nelson W. Polsby (eds.), Handbook of Political
Science. Volume 5. Governmental Institutions and Processes. Reading, MA: Addison-Wesley.
Rothstein, Bo and Bergström, Jonas (1999). Korporatismens fall och den svenska modellens kris.
Stockholm: SNS.
Scarrow, Susan (1996). Parties and Their Members. Oxford: Oxford University Press.
—— (1997). ‘Party Competition and Institutional Change: The Expansion of Direct Democracy
in Germany’. Party Politics, 3: 451–72.
—— (2000). ‘Parties without Members?’, in Russell J. Dalton and Martin P. Wattenberg (eds.),
Parties without Partisans. Oxford: Oxford University Press.
Schlesinger, Joseph. A (1991). Political Parties and the Winning of Office. Ann Arbor: University of
Michigan Press.
Schmidt, Vivien A. (1990). Democratizing France. Cambridge: Cambridge University Press.
Schmitt, Hermann and Holmberg, Sören (1995). ‘Political Parties in Decline?’, in Hans-Dieter
Klingemann and Dieter Fuchs (eds.), Citizens and the State. Oxford: Oxford University Press.
Schmitter, Philippe C. (1974). ‘Still the Century of Corporatism?’ Review of Politics, 36: 85–131.
Setälä, Maija (1999). ‘Referendums in Western Europe—A Wave of Direct Democracy’.
Scandinavian Political Studies, 22: 327–40.
Share, Donald (1999). ‘From Policy-Seeking to Office-Seeking: The Metamorphosis of the
Spanish Socialist Workers Party’, in Wolfgang C. Müller and Kaare Strøm (eds.), Policy, Office,
or Votes? Cambridge: Cambridge University Press.
Shugart, Matthew Soberg and Carey, John M. (1992). Presidents and Assemblies. Cambridge:
Cambridge University Press.
Smith, Gordon (1976). ‘The Functional Properties of the Referendum’. European Journal of
Political Research, 4: 1–23.
Sterett, Susan (1994). ‘Judicial Review in Britain’. Comparative Political Studies, 26: 421–42.
Stone, Alec (1992). The Birth of Judicial Politics in France. The Constitutional Council in Comparative
Perspective. Oxford: Oxford University Press.
Stone Sweet, Alec (2000). Governing with Judges. Constitutional Politics in Europe. Oxford: Oxford
University Press.
Storey, Hugo (1995). ‘Human Rights and the New Europe: Experience and Experiment’. Political
Studies (Special Issue), 43: 131–51.
Strøm, Kaare (1990). Minority Government and Majority Rule. Cambridge: Cambridge University
Press.
706 Kaare Strøm et al.
Strøm, Kaare (2000). ‘Parties at the Core of Government’, in Russell J. Dalton and Martin P.
Wattenberg (eds.), Parties without Partisans. Oxford: Oxford University Press.
—— and Müller, Wolfgang C. (1999). ‘The Keys to Togetherness: Coalition Agreements in
Parliamentary Democracies’. Journal of Legislative Studies, 5(3–4): 255–82.
—— and Swindle, Stephen M. (2002). ‘Strategic Parliamentary Dissolution’. American Political
Science Review, 96: 575–91.
Tálos, Emmerich and Kittel, Bernhard (2001). Gesetzgebung in Österreich. Vienna: WUV-Facultas.
Tiebout, Charles M. (1956). ‘A Pure Theory of Local Expenditures’. Journal of Political Economy,
64: 416–24.
Thiébault, Jean-Louis (2000). ‘France: Forming and Maintaining Government Coalitions in the
Fifth Republic’, in Wolfgang C. Müller and Kaare Strøm (eds.), Coalition Governments in Western
Europe. Oxford: Oxford University Press.
Treisman, Daniel (2000). ‘The Causes of Corruption: A Cross-national Study’. Journal of Public
Economics, 76: 399–457.
Tsebelis, George (2002). Veto Players: How Political Institutions Work. Princeton: Princeton
University Press.
Van Koppen, Peter J. (1992). ‘Judicial Policy-Making in the Netherlands: The Case-by-Case
Method’. West European Politics, 15(3): 80–92.
Verzichelli, Luca and Cotta, Maurizio (2000). ‘Italy: From “Constrained” Coalitions to
Alternating Governments’, in Wolfgang C. Müller and Kaare Strøm (eds.), Coalition
Governments in Western Europe. Oxford: Oxford University Press.
Volcansek, Mary L. (2000). Constitutional Politics in Italy. The Constitutional Court. Houndmills:
Macmillan.
Weingast, Barry R. (1995). ‘The Economic Role of Political Institutions: Market-Preserving
Federalism and Economic Development’. Journal of Law, Economics, and Organization, 11: 1–31.
Williamson, Peter J. (1989). Corporatism in Perspective. London: Sage.
Wright, Vincent (1999). ‘The Fifth Republic: From the Droit de l’État to the État de droit?’ West
European Politics, 22(4): 92–119.
Zimmerman, Joseph S. (1997). The Recall: Tribunal of the People. Westport: Praeger.
23
I N T RO D U C T I O N
In previous chapters we have examined, configuratively as well as cross-nationally, the
successive stages in the constitutional chain of delegation and accountability of Western
European parliamentary democracies. Throughout this book, we have deliberately
addressed accountability as a process of control more than accountability as an outcome
(agency loss). Yet, both aspects are obviously important. Hence, in the first section of
this chapter we take a closer look at democratic accountability outcomes. Subsequently,
we draw together the main lessons of this study. In the third and final section, we con-
clude by placing our study in the context of current large-scale societal challenges. As
we move from one section to another, we also move from specific to more general
problems. In all three sections, we aim to facilitate further scholarship on a number of
critical and still under-researched lines of inquiry.
U N D E R S TA N D I N G AG E N C Y LO S S
We begin with the perennial problem of agency loss. Agency loss is the damage suf-
fered by a principal because an agent lacks the skill or incentives to complete the tasks
delegated to him, or in other words the difference between the policy obtained through
delegation and the principal’s most desired (and feasible) outcome. Those who delegate
hope that their transfer of authority will be successful, but all forms of delegation have
their problems.
There are two basic ways in which agents can ‘go wrong’ and generate agency loss.
They can be unable to satisfy the principal and/or they can be unwilling to do so. In the
former case, the agent may simply be unfit, for example because the qualifications
required for getting the job are systematically different from the demands of the job
itself. Or perhaps the principal has inflated demands that no earthly agent could satisfy.
For example, mass surveys often show that the typical voter at the same time favours
larger government expenditures and lower taxes. There may indeed be a range of such
circumstances in which, for no lack of trying, the agent is incapable of meeting the
principal’s demands. But just as importantly, agents may fail to satisfy the principal
708 Kaare Strøm et al.
because they have their own goals and agendas. They may disagree with the principal
about policy objectives, or about the costs they are willing to incur to reach these goals,
or they may simply be more interested in their own private affairs.
In this section, we assess agency loss under parliamentary government in five prob-
lem areas: leisure shirking (though our discussion of this issue will be only prelim-
inary), policy divergence, policy coordination, rent extraction, and satisfaction with
democracy. Briefly put, leisure shirking means that the agent fails to achieve the best
possible outcome for the principal due to a lack of effort. Policy divergence means that
the agent pursues policies that differ from those that the principal most prefers. Policy
coordination problems arise when agents make mutually conflicting decisions, or when
the same agent makes inconsistent choices across policy areas or over time. Rent
extraction means that the agent exploits his political authority to obtain private bene-
fits for himself or others, rather than to further the principal’s policy agenda. Finally, a
low satisfaction with democracy simply means that, for whatever reason, the citizens
are dissatisfied with their agents. This is our broadest and more encompassing measure,
which serves to complement the previous and more specific ones. The level of popular
satisfaction presumably reflects a bundle of different assessments, as well as the implicit
citizen trade-offs between different values such as efficiency and accountability, and the
historically conditioned performance expectations of citizens in particular polities.
There are two basic yardsticks for assessing the extent of agency loss. One is to
compare the observed agency loss to some ideal analytical (or normative) standard, as
we have done in Chapters 2 and 3. The other approach is empirical and comparative.
In this chapter, we opt for the latter, comparative approach.
Our theoretical arguments in previous chapters generate two kinds of comparative
expectations. On the one hand, we expect the agency problems of parliamentary democ-
racies to differ systematically from such regime types as presidentialism. Thus we
expect, based on the argument in Chapter 3, that in parliamentary systems, the prob-
lems of policy divergence and rent extraction will be more salient than those of leisure
shirking and policy coordination. Its simplicity, and specifically the singularity of its
constitutional design and the agenda control it vests in the cabinet, render parliamen-
tary democracy susceptible to policy shirking and rent extraction. At the same time, the
same features make parliamentary democracy decisive, which should enhance policy
coordination, and incentive-efficient, which in turn should reduce the incidence of
leisure shirking.
On the other hand, within the class of parliamentary systems, we also expect sys-
tematic differences corresponding to the design of mechanisms of citizen control. Not
all parliamentary democracies have the same propensity for agency loss. Each country-
specific chain of parliamentary policy-making has specific features that affect the prin-
cipals’ ability to select, monitor, or sanction agents and thereby also the amount of
agency slippage. Since the combination of such factors is specific to each country,
the variation in agency performance may well be larger within the worlds of parlia-
mentarism and presidentialism, respectively, than between the two. We thus expect the
severity of different agency problems to differ according to the particular design of the
democratic chain of delegation and accountability, for example, the parliamentary
oversight capabilities.
Challenges to Parliamentary Democracy 709
more likely to be politically transparent than for example India, which should make it
systematically easier for principals to oversee their agents. Political philosophers going
back to Plato and Aristotle have praised the governance virtues of smallness.1 Although
contemporary scholarship on the systemic effects of size has produced inconsistent
results,2 we find it plausible for size to affect agency loss, not least though its effects on
informational transparency.
Our control measures in all the analyses below are therefore GDP per capita and
population size (logged). While numerous other controls could be imagined, the size
of our sample (seventeen countries at most) requires parsimony. Given this small sam-
ple size, and the fact that we essentially include the universe of stable European parlia-
mentary systems, we also relax our expectations concerning statistical significance. As
the degrees of freedom in our regression estimates can fall as low as ten
(in the estimate of errors), it becomes very difficult to reach conventional levels of stat-
istical confidence, and we shall comment on estimates that do not meet these standards
but still seem substantively meaningful. Of course, readers can consult our results and
judge for themselves how meaningful the various estimates are.
For each of our agency loss measures, we estimate six different models: one for each
of our four citizen control variables entered separately, one in which we include all four
of these plus our two control variables (population size and GDP per capita), and a final
model (model six) in which we delete all variables that in the full model (model five)
are not significant at the 0.1 level in a one-tailed test (or the 0.2 level in a two-tailed test).
This corresponds to a t-statistic of 1.37 in a model with ten degrees of freedom
(Gujarati 1995: 809).3
Leisure Shirking
Consider first the agency problem that is perhaps the most difficult to measure, and
which we have decided not to pursue in quantitative terms, namely leisure shirking.
Leisure shirking refers to situations in which the agent deliberately chooses not to exert
the effort necessary to succeed. This may, of course, be because he prefers a quiet life
of leisure to work. Putnam (1993: 5) describes the regional government of Puglia in
southern Italy as staffed by prototypical leisure-shirking officials. Reportedly, the
functionaries are ‘indolent’ and ‘likely to be present an hour or two each day and to be
unresponsive even then’ and, as a local mayor has put it, ‘they don’t answer the mail,
they don’t answer the telephone’.
1
In Federalist No. 10, however, Madison turned the argument on its head and argued that size, because it
is associated with a greater variety of parties and interests, helps protect against majority tyranny.
2
Dahl and Tufte (1973: 40) agree with Madison that increases in population size entail a greater
number of organizations and subunits and a more complex policy-making process, but warn that other
factors, such as socio-economic development, might interact. Political participation, or the citizens’ sense of
democratic efficacy, does not seem to vary cross-nationally (though possibly within countries) with size (Dahl
and Tufte 1973: 65; see also Anckar 2000).
3
In the estimates for two of our agency loss measures (Tables 23.2 and 23.4), model five has only one
significant independent variable. In these cases, we do not report model six simply because it would be
identical to one of the bivariate models.
712 Kaare Strøm et al.
Public officials in Southern Italy are not unique in this respect. The ‘prevalence of
absenteeism’ in the French Parliament, as noted, for instance, by Anthony King (1976),
is a closely related phenomenon. As Le Monde put it in the 1970s, some MPs arrive from
the provinces Wednesday morning and go home Thursday night. Not much seems to
have changed since these lines were written (Knapp and Wright 2001: 147). Absenteeism
has been a frequent phenomenon in Italy, too (Hine 1993: 170), and it seems to be even
more widespread in the European Parliament (Nugent 1999: 240).
Yet, low attendance rates need not be an indicator of leisure shirking in the narrow
sense. Many MPs serve large and distant constituencies. Very often French MPs have
been mayors who have devoted their time and energy to local rather than national
politics. Moreover, a study of the Austrian Parliament shows that those MPs who are
neither very active in Parliament nor in constituency work often have other demanding
obligations, for example, as interest group representatives (Müller et al. 2001). Thus, when
public officials shirk some political responsibilities, it may be because they are attend-
ing to others.
Nonetheless, these multiple demands on the time of MPs mean that they are not
fully living up to the obligations resulting from their role in the parliamentary chain of
delegation while typically still getting full financial compensation. It is also noteworthy
that the most notorious cases of political leisure shirking are found in constituencies
with low political transparency (as in the European Parliament), ineffective interparty
competition (as in most of the Italian south), or both.
Leisure shirking is in many cases an individual-level phenomenon without great
political ramifications. As with many forms of democratic agency loss, it may be par-
ticularly prominent in less transparent societies where citizens have little information
about their public officials. It is perhaps most politically troubling when coupled with a
clientelistic network of rent extraction, for example, when public officials shirk because
their appointments were never based on any consideration of their qualifications in the
first place. Yet, important as it may be, we have only scarce and anecdotal information
on leisure shirking by public officials and hence cannot attempt to assess its relative
incidence or importance in the countries under scrutiny in this volume. This again
underscores how empirically under-researched this phenomenon is. As Chapter 3 has
argued, due to the efficiency of individual incentive schemes, leisure shirking should be
less prevalent in parliamentary systems (particularly Westminster systems) than in
presidential ones. That is to say, because politicians in parliamentary systems are less
constrained, they should also be less tempted to slack off. Because of the asymmetries
of information, we expect that leisure shirking is also more characteristic of poor and
diverse societies than of relatively smaller states boasting high levels of public informa-
tion. While certainly a relevant concern even in Western Europe, we suspect that it is
not the region’s most critical agency problem.
Policy Divergence
Policy (or dissent) shirking means that the agent is implementing policies that are not in
line with the preferences of the principal not for lack of effort, but because the agent
Challenges to Parliamentary Democracy 713
pursues his own policy objectives that differ from those of the principal. This is
certainly a common complaint among citizens: that politicians, rather than being lazy or
incompetent, have their own policy agendas and ignore those of the voters. Yet, while
this has also been a concern among political scientists of many stripes, empirical docu-
mentation of such slack has been a stubborn problem. While analytically we can distin-
guish agency loss that occurs because of constraint from cases due to either laziness or
wilful shirking on the agent’s part, these phenomena are often indistinguishable to the
observer. Thus, while we may be able to identify policy divergence between principal
and agent, we cannot always tell whether it is due to shirking on the latter’s part.
One approach has been to identify specific commitments that particular govern-
ments have made to their voters and then examine, often in case studies, whether such
‘pledges’ are indeed kept (Rose 1984). This research points to considerable variation
between governments.4 A second approach has been to examine the policy process in
more aggregated and often comparative terms and investigate whether variation in
partisan control of the government (or parliament) can systematically be linked to
differences in policy output (e.g. see Castles 1982; Boix 1998; Garrett 1998).
As a rule in such studies, however, the parties’ differential policy preferences are
assumed rather than empirically established. Typically, authors assume that parties of the
left, for example, care more about fighting unemployment than about inflation, more
about providing services than about balancing budgets, more about social welfare than
national defence, more about redistribution than about tax relief. While all these assump-
tions may on the whole be correct, they remain rather stylized depictions of political par-
ties, and authors in this tradition rarely actually examine the kind of ‘contracts’ that
parties offer the voters. Klingemann, Hofferbert, and Budge (1993) seek to merge these
two traditions by first identifying government budget priorities from party manifestoes
and then investigating the impact of such programmatic concerns on budgetary allo-
cations. Although their research design has been controversial (Laver and Garry 2000),
they report significant cross-national differences.
A fourth approach, which has become increasingly common since the mid-1980s, is
to measure policy divergence between voters and governments based on expert or
mass survey placements on standardized policy scales. In one of the most notable
exemplars of this tradition, Powell (2000) measures the ideological distance between
government and citizens on a ten-point left–right scale (see Table 23.1). While each
4
Often researchers have looked to election platforms (‘manifestoes’ in British parlance) to identify such
policy commitments. The differences between governments in pledge fulfilment are enormous. There are
cases such as the Thatcher government that have a remarkable record of implementing fairly radical
programmes. Even with regard to these governments, however, careful observers can identify policy sectors
(such as the containment of social welfare expenditures) in which the government (for better or worse)
backed away from its initial policy commitments. On the other hand, some governments clearly and visibly
abandon their manifesto promises. The Spanish Socialists (PSOE) of the 1980s promised Keynesian economic
stimulation, significant income redistribution, vastly increased social expenditures, and withdrawal from
NATO. Once in power, the PSOE abandoned each of these commitments. Share (1999: 97) speaks of a
‘blatant abandonment of the 1982 electoral pledges’. Yet, both governments were returned to office several
times and Mrs Thatcher and Mr Gonzalez went on to become particularly long-lived Prime Ministers. Thus,
it is not obvious that religious commitment to campaign promises is critical to retrospective voter assessments.
714 Kaare Strøm et al.
Table 23.1. Policy Distance Between Government and Median Citizen in Western Europe
Note: Mean of citizen self-placement for Greece and Portugal. Powell (2000) compares voter self-
placements with expert placements of political parties. Since we lack expert placements for Greece
and Portugal for three of the four periods we use voter placements of political parties. The asterisk
(*) signifies that these are our calculations of the average difference, not Powell’s.
Source: Powell 2000: 180–1, 184–5; Gunther and Montero 2001: 104, 107.
placement on this scale may well mean different policies in different countries, it at least
allows comparison of citizen preferences with those of the parties competing in the
same country. The closer the citizens and government parties are on this scale, the
more successfully elections generate representational congruence (2000: 163). Powell
places parties on this scale with the help of expert surveys, and citizens by their self-
placement in mass surveys (see Table 23.1).
We have added Greece and Portugal to the data provided by Powell (2000), leaving
out only the two smallest states in our investigation, Iceland and Luxembourg. Since
these countries are not included in the expert surveys used by Powell (2000), namely
Castles and Mair (1984) and Huber and Inglehart (1995), our data for Greece and
Portugal are not identical to those for the other countries. Greece and Portugal are rep-
resented by the mean (rather than the median) citizen self-placement, and the positions
of the government parties are inferred from voter rather than expert placements.5
5
Voters and experts do not always place parties fully consistently. Thus, in the early 1990s the Portuguese
government party, the PSD, was rated 6.4 on the left–right scale by experts but 7.7 by voters (Huber and
Inglehart 1995; Gunther and Montero 2001). Yet, other party placements show much greater consistency
between experts and voters. Thus, the Spanish PSOE was placed at 3.6 by experts and 3.9 by voters in the
early 1980s, and at 4.0 by experts and at 4.3 by voters in the early 1990s. On the whole, the country-specific
data tend to be consistent over time.
Challenges to Parliamentary Democracy 715
Powell’s normative yardstick is the absolute policy distance between the government
and the median voter, who represents ‘the only policy that would be preferred to all
others by a majority of voters’ (2000: 163). In his calculation of government positions,
Powell weights parties by their numbers of seats and reports two figures for each coun-
try: one representing the early 1980s and the other the early 1990s. In our sample, the
countries included in Table 23.1 range from a smallest observed distance of 0.1 points
to a maximum of 2.5. Among the fourteen countries for which we have data for both
periods, Greece, Ireland, and Spain consistently exhibit low policy distance, whereas
France, Portugal, and the United Kingdom are at the opposite end of the scale. We find
no temporal trend in policy divergence. From the 1980s to the 1990s, policy divergence
increased in Denmark and France but decreased in The Netherlands and Norway. In
aggregate (giving equal weight to each country), there is almost no difference in repres-
entational congruence between the early 1980s and the early 1990s.
In Table 23.2, we present a standardized set of regressions results with this measure
of policy divergence (government–citizen policy distance) as the dependent variable.
compared with pivotal ones (such as Italy 1945–92 or Belgium). In alternational systems,
policy divergence at any one time may be substantial, but at least the direction of this
slippage varies over time. Thus, in the early 1980s left-wing British voters may have felt
as distant from their national government as conservative French voters felt from
theirs, but two decades later at least the tables in both places had been turned. In
pivotal systems, efficiency may be greater but bias more consistent. For close to a half-
century after Second War, Italian voters not enamoured with Christian Democracy had
little to be hopeful about. And although on economic issues, the Italian Christian
Democrats were probably a good deal closer to the median voter than either Thatcher
or the early Mitterrand, Italians who cared greatly about other policy dimensions, such
as family policy or clericalism, could understandably feel that they were forever saddled
with a national government that represented their interests very poorly.
Policy Coordination
Agents can fail their principals not only by pursuing policy positions at odds with them,
but also by failing to coordinate their various policy positions in a productive fashion.
There are, of course, many ways in which policy coordination can fail: mutually con-
tradictory decisions may be taken on related issues, government revenues and expend-
itures may be out of synchronization, decisions taken one day may counteract those
adopted the day before.
In this subsection we shall focus on one particularly salient and relatively well-
researched topic, namely budgetary policy coordination. One of the benchmarks of such
coordination is fiscal discipline, which we here take to mean the avoidance of any build-
up of government debt. While there may be circumstances in which it is in the short run
desirable, or at least not particularly troubling, to operate a government without a bal-
anced budget, there is ample reason to worry about the accumulation of budget deficits
and the long-term build-up of public debt. Specifically, government borrowing drives up
interest rates and hence ‘crowds out’ productive private sector investments, adversely
affecting economic growth.6 Moreover, public borrowing may have redistributive effects,
causing long-term income transfers from taxpayers to the holders of government bonds.
Finally, excessive debt burdens can impair the government’s capacity to make short-term
interventions in the economy even when otherwise this would be beneficial, for example
to smooth a sudden downturn in the business cycle (Franzese 2002: 51).
Thus, in most cases the accumulation of public debt is due to government weakness
and myopia and points to a lack of coordination. It therefore serves as our measure of
policy coordination. Mounting levels of public debt have been a serious concern in
many developing countries, as well as not a few prosperous European ones. Indeed,
concern with fiscal indiscipline was sufficient to cause the leaders of the European
Union to build fairly restrictive (albeit, as time has shown, not terribly enforceable)
requirements into the Maastricht Treaty that paved the way for the EMU.
6
See Franzese (2002: 50–1) for a theoretical discussion of the crowding-out thesis.
718 Kaare Strøm et al.
There is ample reason to believe that institutional design matters to fiscal perform-
ance. Thus, Persson and Tabellini (2000: 345) note that ‘political and institutional fac-
tors play an important role in shaping public debt policy’. The causes of high public
debt levels may include political instability, that is, frequent turnover in government,
which can have at least three detrimental effects. First, incumbents that feel threatened
are likely to resort to government borrowing to ‘create a constituency for re-election’.
Second, the costs of future spending cuts are not fully internalized by the political decision-
makers because they have good reason to assume that they will no longer be in office
when these costs come due. Third, incumbents who assume that they will not be
returned to office may attempt to pre-empt the spending decisions of incoming
governments by leaving them nothing but debt (Persson and Tabellini 2000: ch. 13; see
also Franzese 2002: ch. 3).
These institutional explanations relate partly to the agents’ inability and partly to
their lack of willingness to avoid accumulating government debt. Either way, we inter-
pret them as shortcomings in citizen control of their political agents. The ill effects of
instability can most effectively be counteracted if political agents can be induced to
adopt longer time horizons, or (second-best) if they can be constrained from acting on
their short-term interests. Stable, encompassing, and cohesive political parties are one
vehicle by which politicians may be induced to extend their time horizons. Institutional
features may similarly shift their focus to future concerns or at least prevent them from
acting solely on their contemporaneous ones. Thus, although the effects of political
instability afflict both parliamentary and presidential regimes, the fixed electoral terms
of the latter make them somewhat less vulnerable to short-term opportunism. Moreover,
Persson, Roland, and Tabellini (2000) argue that because they have more cohesive gov-
ernment structures that contain fewer checks against spending increases, parliamentary
democracies are more prone to overspending than are presidential ones.
Our main concern here, however, is with the variation that might exist among parlia-
mentary systems. Thus, we expect fiscal discipline to be positively correlated with both
dimensions of citizen control. Party influence should serve as at least a partial protec-
tion against the ills of government instability and opportunism, and external constraints
should strengthen government credibility. Of course, central bank independence is typ-
ically adopted expressly for the purpose of mitigating some of these parliamentary
credibility problems. Fiscal federalism may have similar effects, as it constrains the
spending power of central governments, and since sub-national governments are often
subject to more effective fiscal restrictions than national ones.
Our operational measure of budgetary coordination is the level of public debt in 1992,
the period just before the European Union’s transnational commitments to fiscal recti-
tude began to take effect. There was considerable variation in government indebtedness
among West European parliamentary systems. As Table 23.3 shows, Luxembourg had a
very small debt in relation to its GDP. For nine of the other countries, however, the debt
was greater than half of GDP. Among these, Belgium and Italy both had staggering
debts well above the value of their entire GDP (see Franzese 2002: 25).
Table 23.4 reports the relationships between our measures of citizen control and pub-
lic debt. Of the four independent variables, party conflict is by far the most influential
Challenges to Parliamentary Democracy 719
Austria 58
Belgium 129
Denmark 71
Finland 41
France 46
Germany 44
Greece 99
Iceland 47
Ireland 94
Italy 118
Luxembourg 5
The Netherlands 80
Norway 36
Portugal 60
Spain 54
Sweden 71
UK 47
Source: OECD Economic Outlook, June 1999. Luxembourg data from Eurostat:
http://europa.eu.int/comm/eurostat/Public/datashop/print-product/EN?
catalogue⫽Eu.
Note: *Central government gross financial liabilities as a percentage of GDP, 1992.
determinant. When entered separately, party conflict alone explains half of the cross-
national variation in public debt. The full model confirms the same pattern. Moreover,
adding other explanatory variables only weakens the adjusted coefficient of determi-
nation. Thus, cabinet cohesion powerfully constrains fiscal indiscipline and is far and
away our most powerful predictor. The effect of party competition (our first indicator
in the model) is in the same direction, but statistically insignificant.
Unexpectedly, external constraints such as Central bank independence and fiscal
federalism have much weaker and less consistent effects. When entered separately, both
have the expected negative associations with public debt, but neither effect is close to
significance. In the full model, however, the effect of central bank independence is
reversed, though still insignificant. Thus, when other effects are accounted for, central
bank autonomy is counter-intuitively associated with high (rather than low) levels of
public debt! We can only suggest two factors that may account for this surprising result
(but note again the lack of statistical significance). One is that in some countries
(the United Kingdom and Norway, for example) co-partisanship alone long sufficed to
discipline politicians. Thus, until the late 1990s strong fiscal discipline in these countries
coexisted with very weak central banks. A second possibility, which is not inconsistent
with the first, is endogeneity: perhaps central bank reform is disproportionately likely
720 Kaare Strøm et al.
Table 23.4. OLS Regressions on Public Debt (as a Percentage of GDP)
precisely in countries that have experienced some serious fiscal indiscipline. If this is
true, then we must be careful not to mistake cause for effect. Our data do not permit
us to examine these possibilities more closely, and we shall therefore close on our
main result in this section: we find significant evidence that policy coordination (here:
fiscal discipline) can be promoted by party cohesion, but much less reason to draw any
conclusion about the effects of external constraints.
most ardently seek power for the sake of self-enrichment. We refer to such action gener-
ally as rent extraction.7 Since rent extraction is generally considered suspect, if not illegal,
it is typically concealed. As Djankov et al. (2000: 17) note, ‘Measuring rents is inherently
extremely difficult, especially across countries.’ Yet, rent seeking is a massive problem
worldwide. Various studies have estimated that in some countries such distortions may
account for up to 50 per cent of GDP (for reviews see Tollison 1997; Mueller 2004).
Much rent extraction occurs through government regulatory activities. Even though
in recent years market imperatives have become much more important in the countries
included in this study, government intervention for the purpose of redistributing
wealth to political allies is still rampant in many societies. It has been argued, for
instance, that not only in Russia and the former Soviet bloc, but also in Western democ-
racies, industrial privatization has been carried out in ways that have systematically
benefited entrepreneurs aligned with the incumbent governments (Feigenbaum,
Henig, and Hamnett 1999).
One notorious avenue of rent extraction has gone through government control of
market entry. Government regulation of entry is, in principle, compatible with a num-
ber of theories about the role of government. It could function as a protection for con-
sumers against predatory competition (which Shleifer and Vishny 1998, refer to as the
‘helping hand’ theory), it could be a way for powerful interests to erect barriers against
competitors (the ‘grabbing hand’ theory) (Stigler 1971), or it could be a way for ambi-
tious politicians to milk affected interests for ‘contributions’ (the ‘tollbooth’ theory)
(Shleifer and Vishny 1998).
Empirically, Djankov et al. (2000) have marshalled evidence on market entry regula-
tion in seventy-five countries, seeking to ascertain how easy or difficult market entry
(i.e. starting a new firm) is and to establish the economic, social, and political causes of
the cross-national differences. Their indicators of regulation are the number of bureau-
cratic procedures that have to be completed in starting a new business (which in their
sample range from two to twenty overall), the number of workdays that have to be
spent (from 2 to 174), and the overall costs involved, measured relative to annual per
capita GDP (from 0.4 to 260 per cent).
Examining a wide range of policy output variables, the authors find that stricter
regulation of entry is not associated with higher-quality products or with better pollution
or health outcomes. Hence, the data do not support a ‘helping-hand’ view of entry regu-
lation. Nor do the authors find higher levels of regulation to be associated with greater
profits or with impediments to competition. Hence, the evidence does not support the
‘grabbing-hand’ view, either. Yet, Djankov and associates do find stricter regulation to be
positively associated with corruption as well as with the size of the unofficial economy.
These results support the tollbooth view of regulation, which, as the authors put it, means
that politicians and bureaucrats themselves are the beneficiaries of the market distortions
they introduce, because they receive campaign contributions, other ‘favours’ (for instance,
jobs put at their disposal to be allocated according to political criteria), or bribes.
7
Rent extraction is based on distortions of the market–government intervention that cannot be justified
as a remedy for market failure (e.g. see Mueller 1996: 11).
722 Kaare Strøm et al.
Among the countries included in Delegation and Accountability in Parliamentary
Democracies, differences in entry regulation exist and are in some cases considerable.
The number of bureaucratic procedures ranges from four to sixteen, the number of
work days required from 11 to 154, and the costs involved in starting a new business
from 0.1 to 0.48 per cent of per capita GDP. And for the fifteen countries on which we
have data, the bivariate (Pearson) correlations between the three measures are all
significant (at the 0.05 level) and vary between 0.64 and 0.75. Thus, these measures
consistently indicate significant and systematic variation in entry regulation even
within Western Europe.
One particularly serious form of rent extraction is corruption—the trading of public
benefits for private favours. Corruption has many causes (e.g. see Rose-Ackerman 1999;
Treisman 2000). It may be a product of adverse selection as well as moral hazard, as some
corrupt officer-holders may have intended all along to abuse their power for venal pur-
poses, whereas others probably at the outset had more admirable ambitions.8 The oppor-
tunity costs for public officials certainly seem to be a critical factor: where the alternatives
are plentiful and yield acceptable rewards, corruption is likely to be less prevalent. This
problem is particularly severe in societies in which the public sector is the only, or domi-
nant, avenue for social betterment. The great prevalence of corruption, particularly (but
certainly not exclusively) in poor societies, is a sad testament to the power of these forces
(see Ridley and Doig 1995; Della Porta and Mény 1997; Rose-Ackerman 1999).
Table 23.5 summarizes our data on entry regulations and corruption for the seven-
teen countries in our sample. The first two columns present each country’s score for
1998 and 2000 on the Corruption Perceptions Index developed by Transparency
International. The last three columns present three operationalizations of entry
regulations: the number of procedures required to start a new business, the time spent
in days, and the costs in percent of GDP per capita. For comparative purposes,
Table 23.5 also contains entries for five other Western democracies, three of which are
parliamentary (Australia, Canada, and New Zealand), one the archetypical presidential
system (United States), and the final one Switzerland with its unique directorial
government.
As suggested by Table 23.5, political regulation of economic entrepreneurship is
more extensive in Austria, France, Greece, Italy, Portugal, and Spain than in Denmark,
Finland, Sweden, and the United Kingdom, to list those countries that score most con-
sistently on the various indicators. In line with these results, Djankov et al. (2000) find
that countries with more open access to political power, greater constraints on the
executive, and greater political rights have fewer entry regulations. Yet, political rent
extraction is not the only possible explanation for the differences recorded. Differences
in legal traditions appear to have a strong impact, with English or Scandinavian legal
origin leading to fewer regulations than French or German origin.
8
As we have pointed out in previous chapters, the adverse selection problem is that powerful government
offices may attract particularly power-hungry aspirants. The moral hazard problem is that agents controlling
a large and powerful government may be tempted secretly to use these resources for purposes that the
principal never intended and would not have approved.
Challenges to Parliamentary Democracy 723
Austria 7.5 (11) 7.7 (9) 12 (12) 154 (15) 0.45 (14)
Belgium 5.4 (15) 6.1 (15) 8 (8) 42 (7) 0.10 (7)
Denmark 10.0 (1) 9.8 (2) 5 (4) 21 (3) 0.01 (3)
Finland 9.6 (2) 10.0 (1) 4 (1) 32 (6) 0.01 (2)
France 6.7 (12) 6.7 (13) 16 (15) 66 (9) 0.20 (11)
Germany 7.9 (10) 7.6 (10) 7 (6) 90 (12) 0.09 (6)
Greece 4.9 (16) 4.9 (16) 13 (14) 53 (8) 0.48 (15)
Iceland 9.3 (4) 9.1 (4) n.d. n.d. n.d.
Ireland 8.2 (9) 7.2 (11) 4 (1) 25 (5) 0.11 (8)
Italy 4.6 (17) 4.6 (17) 11 (10) 121 (14) 0.25 (12)
Luxembourg 8.7 (7) 8.6 (8) n.d. n.d. n.d.
The Netherlands 9.0 (5) 8.9 (6) 8 (8) 68 (10) 0.19 (10)
Norway 9.0 (5) 9.1 (4) 6 (5) 23 (4) 0.02 (4)
Portugal 6.5 (13) 6.4 (14) 12 (12) 99 (13) 0.31 (13)
Spain 6.1 (14) 7.0 (12) 11 (10) 83 (11) 0.13 (9)
Sweden 9.5 (3) 9.4 (3) 4 (1) 17 (2) 0.03 (5)
United Kingdom 8.7 (7) 8.7 (7) 7 (6) 11 (1) 0.01 (1)
Australia 8.7 8.3 3 3 0.02
Canada 9.2 9.2 2 2 0.01
New Zealand 9.4 9.4 3 17 0
Switzerland 8.9 8.6 12 88 0.13
United States 7.5 7.8 4 7 0.01
There are strong cross-national correlations between entry regulation and corrup-
tion. Countries deemed to be most corrupt are also much the same that have the most
restrictive entry regulation. For two entry measures, procedures and costs, there is a
high and statistically significant correlation (⫺0.69 and ⫺0.67, respectively) with the TI
corruption index. (The coefficients are negative because in the TI index high scores
mean low corruption.)9 Yet, there are certainly individual exceptions. On two of three
indicators, the United Kingdom is the least regulated of our countries, whereas it is not
rated as exceptionally ‘clean’. Austria and The Netherlands, on the other hand, appear
9
Political corruption (the 2000 index) co-varies least strongly with the measure for entry regulations in
days. The correlation is ⫺0.5 and barely significant at the 0.05 level (0.054).
724 Kaare Strøm et al.
more regulated than corrupt, whereas the Belgian case is the opposite. Switzerland is
similar to The Netherlands: it has about the average level of perceived corruption, but
demands much more of a start-up firm. The former ‘old dominions’ present a pattern
similar to Britain: comparatively clean (more so than Britain, actually) and modestly
regulated. By European standards, the United States is a middling country in perceived
corruption, but the regulatory burden on small businesses is light. Compare Austria
and the United States. They are almost identical with respect to perceived corruption,
yet the differences in market entry are enormous, with the regulatory process lasting
only 7 days in the United States, as against 154 days in Austria. If rent extraction is more
directly reflected in market regulations than in societal corruption, then such practices
appear to be much more prevalent in Austrian politics than they are in the United States.
In Table 23.6 we examine the predictors of rent extraction as measured by the
percent of per-capita GDP required to start a new business. Generally, external con-
straints have important and notable effects. Yet, our two measures of external con-
straints point in different directions. While central bank independence seems to
increase entry costs, fiscal federalism (sub-national government consumption) has the
opposite effect. The former result is unexpected, and we can only speculate as to why
central bank independence and rent extraction go together. Thus, note that countries
such as Austria and Greece, with comparatively independent central banks, lose a high
proportion of their GDP to regulation costs, while the three Scandinavian ones and the
United Kingdom have relatively little entry regulation and at the same time less inde-
pendent central banks (at least as of the 1980s). A time-series analysis might help sort
out such relationships that look anomalous in a purely cross-sectional regression, but
the question will have to await further study. In contrast, the second external con-
straints variable, subnational government consumption, has the expected negative
effect on political rent extraction, which is consistent across all the specifications in
Table 23.6. Thus, the evidence suggests that fiscal federalism helps contain rent extrac-
tion, a result that conforms well to our expectations.
Our results also suggest that strong partisanship tends to reduce rent seeking,
though these results are less striking. Inter-party competition significantly constrains
entry regulation even when we control for other predictors and thus does seem to help
726 Kaare Strøm et al.
make governments cleaner. Party cohesion in government, on the other hand, has only
a weak and insignificant constraining effect on political rent extraction.
Table 23.7 reports the determinants of perceived corruption, as measured by the
2000 Corruptions Perceptions Index. Note that on this index, a high score signifies a
comparatively ‘clean’ country. Table 23.7 reports significant effects for one indicator of
partisanship as well as for both measures of external constraint. Party cohesion has a
consistent and significant influence. The less cohesion in government, the more cor-
ruption. This is in line with our findings on entry regulation and matches our theoret-
ical expectations: governments harbouring severe internal conflict are less capable of
serving as citizen agents than governments that work together smoothly. When the
variables are entered separately, party cohesion also explains a greater portion of the
variance in corruption than any other predictor.
The estimates for central bank independence suggest a positive relationship with
agency loss which, again, is not in accordance with our expectations. The country-
specific factors discussed above may be relevant even here. Or one might suspect endogene-
ity problems. Given the substantial and increasing adverse effects of a loss of monetary
credibility, corruption-ridden countries may historically have taken the lead in trying to
inoculate themselves through policies of monetary autonomy.
On the other hand, as we expected, our second measure of external constraints, sub-
national government spending, is negatively associated with corruption. The relation-
ship exists both when the variable is entered by itself and in the full model. Thus, as in
several other domains, fiscal federalism helps contain democratic agency loss, here in
the form of outright (perceived) corruption.
Norway, trust in politicians has declined everywhere. In addition, Newton and Norris
(2000: 57) show that between the early 1980s and the early 1990s confidence in five
public institutions—police, legal system, armed forces, Parliament, civil service—
declined significantly in the United States, Japan, and all Western European parliamen-
tary systems except France (non-significant decline) and Iceland (statistically significant
increase in confidence). Newton and Norris (2000: 72) also conclude that ‘our research
provides substantial support for theories that focus on the performance of govern-
ments and political institutions to explain citizens’ declining confidence in them’. It is
possible that these trends in public confidence have since been affected by such cata-
clysmic events as those of 11 September 2001, but the well-documented longer-term
trajectory has for quite some time worried students of representative democracy.
Paradoxically perhaps, while Western publics have evidently become less trustful of
politicians and political institutions, they are at the same time more and more unani-
mous in their support for democracy. Klingemann (1999) argues that political support
for democracy, as a regime type, is analytically and empirically distinct from the distrust
that is attached to the performance evaluation of specific institutions. Analysing a
much larger set of countries than we consider here, he finds ‘no evidence of growing
dissatisfaction with democracy as a form of government’ (Klingemann 1999: 56).
Table 23.8 summarizes data from several surveys on public confidence in parliamen-
tary and judicial institutions in European parliamentary democracies. These data allow
us to compare these two types of institutions, as well as to track trends in their relative
support over time. If parties and the parliamentary channel of delegation have in fact
been weakened, we might expect this to be reflected in declining levels of public confi-
dence. In fact, although the overall trend is down, public confidence in parliamentary
institutions is in many places holding up quite well. When the latest available numbers
for the 1990s are compared to the first available numbers from the 1980s or early 1990s,
confidence in Parliament has actually gone up considerably in Denmark, The
Netherlands, and Portugal. It has elsewhere mostly stayed about the same, and only in
Germany (after re-unification) and Norway is the drop in double digits. The core
institution of parliamentary democracy, the national legislature itself (with the most
obvious and understandable exception of Italy in the early to mid-1990s), by and large
has held its position quite well (see Table 23.9). For sure, parliaments are in most
countries less trusted than the judiciary, but this discrepancy is much greater
in Northern Europe (e.g. Denmark, where it is about forty points) than in the South
(e.g. Spain, where it is virtually non-existent). Moreover, the differential in public
confidence between Parliament and the judiciary has declined just about everywhere.
Table 23.9 provides more summary information. Because we use slightly different
sources, the numbers reported in Table 23.9 are not identical to those in Table 23.8. In
most cases, the differences between the two sources are marginal, though in some,
such as Finland, the discrepancies are actually remarkably large.10 The advantage
of these additional data sources is that Table 23.9 also includes a particularly broad
10
If we compare the ‘confidence in Parliament’ column for 1996 in Table 23.8 with the equivalent column
for mid-1990s in Table 23.9, the difference between the two observations is 22 percentage points.
Table 23.8. Trust in Parliament and the Judiciary in West European Parliamentary Systems
Notes: Columns 1 and 2 are Eurobarometer data, from Klingemann, in Norris 1999: 50, 52; Columns 3–6 are World Values Surveys data, from Newton and Norris 2000: 56.
a
West Germany before 1990.
Sources: Norris 1999; Newton and Norris 2000.
n.d.⫽ No data available.
730 Kaare Strøm et al.
measure of citizen assessment, namely satisfaction with overall democratic
performance. Comparing these survey items across our various countries, the overall
picture is quite clear. Public distrust is directed more against parliamentary institutions,
and presumably even more against specific politicians and parties, than against democracy
as such. Under traditional forms of parliamentary democracy, where accountability has
largely been managed through ex ante mechanisms involving political parties, this is still
very problematic. Tables 23.8 and 23.9 also suggest that although confidence in specific
institutions (read: courts in particular) is down, we find no downward trend in overall
satisfaction with democracy (although the data in Table 23.9 cover only two time points
in the 1990s) or specifically with parliaments.
Table 23.10 presents regression models of our most recent data on satisfaction with
overall democratic performance (1995) in sixteen parliamentary democracies (exclud-
ing Iceland, for which data are missing). The models reveal a number of interesting
relationships. Of our four citizen control dimensions, fiscal federalism has the most
consistent and significant effect on popular satisfaction with democracy. The more
decentralized the polity, the higher the citizens’ satisfaction. Given that fiscal federalism
means that spending decisions are made closer to the citizen (e.g. in the city council
rather than the national parliament), but that policies may vary significantly from
one sub-national government to the next, this result suggests that citizens value local
control of the political decision-making process more than policy uniformity.
Central bank independence has the same positive sign but is not significant. Party
cohesion and competition both have the expected signs (that is to say, cohesive and
competitive parties are associated with high levels of satisfaction), but these results are
only significant in one model specification each. Note, however, the significant effects
of both control variables. Models five and six both reveal that citizens in small countries
are more satisfied than those in more populous ones, as indeed we expected. Moreover,
there is a strong tendency for people in richer countries to be more satisfied with
democracy than those who live in poorer ones. Altogether, Table 23.10 reports that
coupled with these control variables, our measures of citizen control explain a large
portion (up to 63 per cent) of the variance in democratic satisfaction.
Summary
We have seen that our measures of citizen control are in many cases predictably
associated with the various agency problems that we have examined. Even in our small
sample, the estimated effects of individual parameters are in many cases, but far from
always, statistically significant. For all performance (agency loss) criteria except one, the
explanatory power of our overall model is sizeable, as the adjusted coefficient of deter-
mination in at least one specification consistently exceeds 0.40. The exception is policy
divergence, where we find little systematic relationship between citizen control mech-
anisms and political outcomes. Interestingly, although there is some modest evidence
that polities with cohesive Westminster-like parties exacerbate problems of policy
distance between governors and the governed, no specification of our model is able to
explain any significant part of the variation in such policy divergence.
On the other hand, even though these effects are not always statistically significant
in our sample, we find quite systematic evidence that party cohesion and competition
contain other forms of agency loss. In fact, our estimate of the effect of executive
cohesion shows the expected sign (meaning that cohesion constrains agency loss) in all
nine of these specifications, whereas the direction of the estimates for inter-party com-
petition similarly conforms to expectations in nine out of ten cases.
In other words, the evidence strongly and broadly suggests that cohesive and
competitive political parties and governments help reduce the risks of democratic
delegation. Specifically, executive cohesion strongly and significantly reduces the risks
of corruption and fiscal indiscipline. Party competition, on the other hand, reduces rent
extraction and promotes general satisfaction with democracy. Such salutary effects,
then, are associated with the Westminster model of democracy, which according to
our results deserves much greater respect than it has received in much of the recent
literature. Yet, the positive effects of electoral competition and cohesion do not strictly
732 Kaare Strøm et al.
depend on a pure two-party format. Even alternational multiparty systems may under
certain circumstances fit the bill. Pivotal, polarized, and fragmented party systems,
however, do not.
Our estimates for external constraints do not both exhibit the same consistent and
plausible coefficients. Contrary to our expectations, central bank independence is in
many specifications associated with higher, rather than lower, reports of agency loss.
Above, we have speculated as to what these findings may reflect. Fiscal federalism, on
the other hand, constrains agency loss in every one of our thirteen specifications, even
though a number of these results are not statistically significant. On the whole, fiscal fed-
eralism seems most effective in containing rent extraction and corruption. Thus, external
constraints are probably most effective when they pit ‘ambition against ambition’, with-
out the constraining effects of party cohesion. This is generally more likely with respect
to rent extraction than, for example, policy divergence. Thus, Bagehot and Madison were
both right: party cohesion and external constraints each have their virtues.
Finally, our control variables demonstrate that country-specific and demographic
properties may have systematic effects on democratic delegation that should not be
ignored. Small and wealthy countries have populations that are more satisfied with
their democracies, and small countries also have fewer corruption problems. These are,
at least in part, rather straightforward effects of the greater political transparency that
such societies afford. These effects should not be mistaken for consequences of the polit-
ical institutions that these same societies may happen to have. Even under identical polit-
ical structures, we should expect Switzerland, Luxembourg, or Norway to solve their
democratic delegation problems more successfully than, for example, Russia or even
Spain. Such results do not suggest that political agents in the former sorts of countries are
necessarily more virtuous than those in the latter, nor do they in themselves prove the
superiority of their institutional arrangements. Constitutional designs that have proven
successful only in small and rich countries, such as social corporatism (Katzenstein 1985)
or perhaps consociational democracy (Lijphart 1977) may not pass muster under more
demanding circumstances.
W H AT H AV E W E L E A R N E D ?
Let us now look back at the broader lessons of this study. In the introductory chapters,
we identified the problems of democratic delegation in parliamentary democracies and
suggested that principal–agent models offer a coherent, rigorous, and parsimonious
way to understand a host of these challenges. A morass of competing definitions has
hitherto prevented students of constitutional design from fully understanding the
virtues and vices of parliamentary democracies. In contrast, by spelling out a few
analytical conceptions we have been able to demonstrate the main differences between
alternative constitutional designs. Working from the ideal types, we have developed
models that illustrate fundamental differences between parliamentarism and presiden-
tialism, as well as critical distinctions within the family of parliamentary systems.
Throughout the twenty-three chapters of this book, we have shown that this rational
choice-based approach can facilitate large-scale empirical and comparative research and
Challenges to Parliamentary Democracy 733
Parliamentary Democracy
As we have shown in the introductory chapter, students of democratic institutions have
long realized both intuitively and empirically that political delegation is problematic.
Yet, they have taken different positions on the severity of the problem and on whether
indeed it is possible to found popular government on delegation. John Stuart Mill (1984
[1861]) viewed political delegation through representation as preferable to other forms
of government, whereas Michels (1962 [1915]) thought delegation through party organ-
izations in the long run unsustainable, as political leaders will eventually cease to serve
those whom they ostensibly represent. Mosca (1947) saw parliamentary delegation as
feasible only if checked by outside actors and institutions of the kinds that we have
called external constraints (courts, audit offices, etc.). Kelsen (1926) likewise considered
electoral accountability by itself insufficient to solve the democratic moral hazard prob-
lem, but advocated instead the remedies of direct democracy and interelectoral partici-
pation. Ostrogorski (1907) and Schmitt (1969 [1923]) went so far as to warn against the
detrimental impact of political parties on the constitutional principles of democracy.
Our delegation approach captures many of these most central concerns of the clas-
sical writers on democratic constitutional design. Yet, showing that this particular con-
ceptual apparatus can be applied to parliamentary systems is not in itself the ultimate
scholarly achievement. There is in principle any number of ways to understand parlia-
mentary institutions. The challenge is not only to show that any particular framework can
faithfully and effectively represent the concerns of those who have thought long and hard
about the problems, but also to demonstrate the value it adds to the existing wisdom.
We find that value added especially in three properties of our approach: generality,
parsimony, and deductive rigour. In terms of generality, Chapter 1 shows that the prob-
lems associated with political representation or delegation at different stages of the
policy process have often been studied in mutual isolation and in different discourses.
We, instead, suggest that similar agency problems pertain to each of the links in the
chain of delegation that we have identified. Although specific problems of representa-
tion may be particularly acute at different stages of the policy process, they can all be
captured in a single, coherent framework.
Nor is this framework ominously complex and inscrutable, and this is where the
virtue of parsimony lies. The problems of delegation stem from two basic sources,
namely differences between principals and agents in preferences and/or information.
Similarly, we have argued that agency problems under incomplete information manifest
themselves in two critical forms: adverse selection and moral hazard. And our survey
of the institutional remedies have focused on ex ante verses ex post controls, and specif-
ically on the roles of political parties and external constraints in giving flesh and blood
(and enforcement power) to such control devices.
734 Kaare Strøm et al.
Finally, one of the chief virtues of the principal–agent approach is the fact that it
allows us to model political delegation relationship in precise and rigorous ways. Such
models in turn allow us to predict the consequences of specific institutional designs, or
to compare the performance of different institutions across a variety of assumptions
about the preferences and information of political principals and agents, as we have
begun to do in Chapter 3. While we hold this potential to be a profoundly important
virtue of principal–agent models, it is one that we have only begun to realize here. Yet,
we confidently expect such analysis to yield rich fruits in future research.
These are analytical results and expectations, and only future research can show
whether indeed these trade-offs exist between the virtues of parliamentary and pres-
idential constitutions. Yet, these trade-offs are not associated simply with the larger
choice between parliamentary and presidential constitutions, but pertain also to
institutional choices within the family of parliamentary systems. Our research shows
considerable institutional variation among parliamentary systems. The analytical
device of the classical Westminster model yields good insights into the parliamentary
regime type. But at the same time, it is only a rough approximation and one that most
parliamentary systems (and even the United Kingdom) resemble less and less. As the
summary table in Chapter 4 (Table 4.20) helps illustrate, parliamentary design entails a
number of variations and trade-offs. For each indicator, our countries manifest differ-
ent choices. While some choices promote expedient delegation, others provide
accountability to the principal and yet others support stability by promoting counter-
balancing powers in the hands of the Prime Minister.11
While we have not yet tested our expectations concerning the differences between
parliamentary and presidential systems, we have been able to examine the effects of
institutional variation within the parliamentary camp. Our results show that in most
contexts, institutional mechanisms of citizen control do have meaningful and often
predictable consequences. And some of the more precise patterns are noteworthy.
Parliamentary systems with cohesive, competitive parties and/or effective external
constraints generally perform better than those without these features. Inter-party com-
petition can apparently foster policy coordination, as evidenced by our data on public
debt. Together with external constraints on national governments, partisan cohesion
and competition can also contain rent extraction and enhance democratic satisfaction.
Our empirical results thus suggest that a government prone to infighting is more
likely to cause agency loss than one that is not. Parliamentary democracies that approx-
imate the Westminster or alternational multiparty model demonstrate that partisan
conflict in the executive, for example, is not an inherent problem in parliamentary
democracy, even though it may be a contingent feature of many parliamentary sys-
tems. Instead the agency loss experienced in a particular country is a result of its mix
of a particular constitutional chain, party system, and external constraints. Pivotal mul-
tiparty systems without effective external constraints would seem to be the least desir-
able form of parliamentary government.
11
To exemplify, a strong Prime Minister may promote efficient delegation, but accountability on a day-to-day
basis is perhaps better assured if he has fewer powers. Similarly, granting the PM liberal parliamentary
dissolution powers may ensure stability, but at the price of weakening his accountability to his immediate
democratic principal, the parliamentary majority.
736 Kaare Strøm et al.
The speed of change is not overwhelming, but voters in many countries have gained
more direct influence over candidate selection and parliamentary elections through
membership referendums within parties and preference voting in general elections.
Parliaments are moving in the direction of getting more power over less. There is in
many countries far more room for parliamentary questions and committee scrutiny
than there used to be prior to the 1970s or 1980s. But at the same time, many policy
issues have been shifted to the supranational arena, and thus national executives have
gained in relative power at the expense of their legislatures.
Some part of the power-shift in favour of the cabinet stems from the central role of
national executives in EU decision-making, but the trend was visible even before that EU
influence became salient. Above all it favours Prime Ministers, who throughout Western
Europe have acquired more staff and larger administrative resources than they had in the
early years after Second World War and who have become a greater focus of media and
voter attention even for purely domestic reasons. In this respect at least, it is possible to
talk about a ‘presidentialization’ of premierships all over Western Europe. The final step
in the chain of delegation is also changing as ministers keep their formal powers but in
many cases release their agents from the strict control and micromanagement implied
by more classical models of parliamentary democracy. We will return to this point later
when we discuss the contemporary challenges facing all representative democracies.
The second and third trends, weaker parties and stronger constraints, were discussed
in detail in Chapter 22. Suffice it here to say that the decline of political parties, espe-
cially in electoral and membership terms, entails a serious challenge to traditional par-
liamentary democracy and especially the Westminster model. So does the general
enhancement of external constraints. These developments fundamentally change the
constitutional principles of delegation and accountability. As Strøm explains in Chapter 3
and our empirical chapters demonstrate, cohesive and competitive political parties
can promote both efficiency and accountability. Weakened political parties first and
foremost mean that the link between citizens and their political agents becomes less
transparent. In some ways, weakened political parties may revive the spirit of many
constitutions and specifically the individual responsibility of political agents. Thus, this
development may enhance the independence of the civil service or the constraining
force of agents that are no longer ‘tied-in’ by partisan bonds. But at the same time the
atrophy of political parties can leave citizens with less information and fewer political
cues. Just consider the consequences of MPs regularly changing party affiliation
between elections, as has been pervasive in Italy since 1994.
consequences, as this will help us understand what institutions are and when they will be
effective. Institutions are consequential in large part because there are rewards and
punishments associated with the behaviour that they regulate. These rewards may be
impersonal and universal, as for example, when market institutions help buyers and sell-
ers engage in transactions for mutual gain. More generally, institutions may induce or sus-
tain particular beliefs and expectations that in turn sustain collective goods and/or have
distributional consequences. Or institutional consequences may manifest themselves in
much more personalized and particular forms, as when the occupants of public offices are
given discretionary power to reward, permit, block, or punish political behaviour.
Yet institutions not only have consequences, they also have purposes driven by the
preferences of their occupants. And unless one understands the purposes of political
institutions, one may well draw misguided inferences about their consequences. This is
less frequently recognized, even by institutionalists. Saying that institutions have pur-
poses is not tantamount to adopting a functionalist mode of analysis. We do not want
to argue that institutions invariably produce certain systemic goods, or that there is
some organic process by which institutions get ‘naturally’ selected to help polities
survive and prosper. Nor do we want to argue that institutions historically have been
selected in a fully instrumental fashion, or in the glorious light of a complete under-
standing of their consequences. Yet, institutions are chosen and sustained by political
actors for specific purposes, albeit often in haphazard and chaotic ways, and no doubt
often under expectations that later turn out to be wildly erroneous.
Whatever their genesis, one powerful purpose of modern democratic institutions is to
further citizen preferences and contain the behaviour of their political agents. Needless to
say, these purposes do not exhaust the consequences of democratic institutions. Nor are
these purposes necessarily stable over time. In some cases, such as the US Constitution,
the democratic purpose was present from the inception. In others, such as the British one,
it may have gradually imposed itself over the course of a long period of time.
While this commitment to institutional consequences and purposes may sound
exceedingly general and abstract, it does have quite specific implications for our under-
standing of parliamentary (and other) democracies. To help spell those out, let us briefly
consider two other deservedly influential and powerful accounts of contemporary demo-
cratic institutions, namely those of Lijphart (1977, 1984, and 1999) and Tsebelis (2002).
While we owe a great deal to each of these authors, there are also important ways in
which our understanding of democratic institutions differs from theirs, respectively.
Consider first Lijphart’s seminal work. While, like Lijphart, we find the Westminster
model a greatly helpful benchmark for institutional analysis, we miss in his analysis a
consideration of what we take to be one of the main purposes of such democratic
institutions—the containment of agency problems. In Lijphart’s analysis, democratic
institutions serve to give representation to popular preferences, and particularly it
seems, those associated with the interests of specific, often demographically defined,
social groups such as ethnic, linguistic, or religious minorities. But while we would
certainly not deny that political institutions of representation serve this purpose, it seems
to us imperative to consider the problems of agency loss that any such representation
may entail.
738 Kaare Strøm et al.
In Lijphart’s view, the virtues of democratic representation hinge decisively on the
access of agents from different social groups to the policy process. While this may be a
necessary condition for stable and successful democracy, it seems to us far from suffi-
cient. What is missing from this analysis is a consideration of the agency slack which
these agents may exploit. If adverse selection or moral hazard is a significant political
problem, then institutions of political representation need mechanisms by which the
democratic principals can control or even throw out the ‘rascals’—their political agents.
In his analyses, Lijphart fails to identify such mechanisms of within-group agency control,
or to differentiate between systems that have them and those that do not.
Consider the author’s conception of Westminster, or majoritarian, democracy. Lijphart
(1999: 31) considers its fundamental properties, majority rule and adversarial competi-
tion, principles of exclusion (see also Lewis 1965). We agree. But whereas Lijphart seems
to understand this principle as actual ex post exclusion of democratic principals, we see
it instead as ex ante excludability of agents. In other words, the key to Westminster
democracy is not that some voters must be kept out, but rather that some politicians can
be so treated. And this potential to exclude agents serves precisely to protect the ultimate
principals, the citizens. The competitive nature of Westminster democracy is its
strongest mechanism of political accountability. As this volume has amply demon-
strated, Westminster democracy certainly has its share of problems, but it is important
to recognize the democratic purpose of its party cohesion, competition, and hierarchy.
Agency problems are also missing from one of Lijphart’s (1977) contrasting models,
consociational democracy. In the plural societies in which consociationalism was orig-
inally observed, voters are divided into mutually distinct and often antagonistic social
groups or segments. The purported virtue of consociationalism is that within such a
system, group leaders are able to take more accommodating positions than those most
favoured by their respective supporters. But this is unlikely to be the case unless the
voters either do not know what their respective leaders are doing, or are unable to influ-
ence or displace them. As Tsebelis (1990, ch. 6) has shown in his game-theoretic recon-
struction, the benign effects of consociational democracy seem to hinge on high voter
information costs and/or a monopoly of political representation (i.e. no effective elec-
toral competition) within each social segment. In other words, consociationalism
works best if there are substantial barriers to entry for political entrepreneurs (new
parties) and/or regular citizens are systematically in the dark about elite decisions.
These are fragile and counter-intuitive conditions for a successful democratic order.
If political agents are at all self-interested, then a situation of poorly informed prin-
cipals with no effective choice over their agents is one that positively invites such agency
problems as rent extraction, leisure-shirking, and possibly corruption. It is therefore no
surprise that consociational or quasi-consociational countries such as Belgium, Austria,
and Italy score high on many of our indicators of agency loss, and especially those having
to do with rent extraction, corruption, and policy coordination. For much the same rea-
sons, we suspect, the same countries have also harboured some of the strongest populist
right-wing protest parties in Europe. While such parties no doubt feed on xenophobia
and alienation, their success is greatly enhanced when they can cast themselves in
the role of opponents of a detached and cozy cartel of career politicians (see Rose 2000).
Challenges to Parliamentary Democracy 739
The fact that out of these countries, Austria appears better governed overall, may
reflect its superior combination of effective external constraints and cohesive and
competitive political parties, compared with other consociational systems.
Let us now turn to Tsebelis’ (2002) powerful and already very influential veto player
analysis of political institutions. While we find the veto player approach a parsimonious
and highly effective way to analyse political institutions, we again believe that there is
more to political institutions than Tsebelis here allows, and that what is missing from
his analysis is not simply ‘embellishment and detail’. Tsebelis identifies two types of
veto players: institutional and partisan ones. He argues that if they are properly identi-
fied, partisan and institutional veto players are mutually substitutable. We believe not.
Chapter 3 in this volume critiques this assumption, which seems to neglect both the
differential consequences and purposes of parties and institutional checks.
On the consequence side, we need to recognize the existence of many kinds of priv-
ileged players, such as dictators, veto players, decisive players, and merely powerful players
(see Chapter 3). Strøm and Swindle (2002) have shown that at least for parliamentary
dissolution decisions, such finer distinctions do in fact systematically matter. Dictators,
for example, have stronger powers and effects than other veto players. More generally,
institutional effects depend on the actual rewards and punishments that different kinds
of veto players can impose, and here partisan and institutional veto players clearly dif-
fer (the latter also among themselves). Institutional veto players typically have consti-
tutionally guaranteed and discretionary powers. These have often been vested in them
precisely so that they can ignore the immediate desires of elected officials or the major-
ity of voters. They may even have been selected precisely on the basis of preferences
that deviate from those of the median voter. They may have a specific and well-defined
mandate, such as upholding the constitution or attaining a target rate of inflation. Party
leaders, on the other hand, have only such powers as their elected representatives are
willing to give them, which ultimately depend on the voters’ preferences and tolerance.
Thus, compared with partisan ones, institutional veto players typically have greater
authority as well as different purposes.
Tsebelis seeks to explain the consequences of different institutional designs for policy
stability or change. He argues that the greater the number of veto players, and the
greater the diversity of their preferences, the greater the constraints on policy change.
While we would add that the type and specific authority of the relevant players also matter,
we generally concur. But if one is interested not only in whether policy change is likely,
but also in whether in the process the policy makers would serve the interests of their
democratic ‘masters’, then it is not only helpful, but critical, to distinguish between dif-
ferent kinds of powerful (veto) players. Thus, if institutional and partisan veto players
were truly mutually substitutable, then a polity with a fragmented party system and
little external constraint could be equivalent to one with a small number of cohesive
parties and several strong external constraints. Both might have the same number of veto
players and the same dispersion of preferences. For democratic delegation, however,
there is every reason to believe that these two categories of political systems differ. In the
former case, we would expect to see very weak control of political agents and a high
risk of agency loss, in the latter very strong control and a much lower risk of slippage.
740 Kaare Strøm et al.
C O N T E M P O R A RY C H A L L E N G E S : M I N D T H E GA P
Parliamentary democracies are changing, in large part because their environment is
changing, too. We shall close this chapter (and book) by considering the major chal-
lenges to democratic delegation that lie ahead, as far as we can foresee the trends that
are under way. There are two basic variables that can drive a wedge between democratic
principals and agents, namely differences in preferences and information. These, then,
are the sources of the critical challenges to representative democracies. While preference
divergence between principals and agents has always been, and will no doubt continue
to be, a critical issue in all democratic polities, the most critical challenges ahead may
turn out to involve discrepancies in information. This is in large part due to the great
contemporary changes in the ways in which information is conveyed, accessed, and
evaluated.
As we mentioned in the introduction to this chapter, the classical design of parlia-
mentary democracy harbours a potential transparency problem that stems from the
very rationale for representative politics. One reason to delegate is that principals can-
not or will not pay the costs of acquiring all the information and skills necessary to per-
form specific tasks on their own. Less-informed principals leave matters in the hands of
better-informed agents. For delegation to be successful, principals want their agents to
be as well informed as possible so that they can make decisions that have the highest
possible likelihood of leading to beneficial results. Yet, things are not quite so simple.12
Because of the moral hazard problem, they also do not want the gap in information
(information asymmetry) between themselves and their agents to be insurmountable.13
Hence, adverse selection and moral hazard have been governance problems as long as
citizens have chosen representatives to make decisions in their place. They will remain
problems as long as representation continues to be a fact of life. But even though these
problems are indeed general and timeless, their particular forms, and their comparative
urgency, are not.
12
For example, principals may differ in their preferences over different informational desiderata. Some
may particularly value agents that are articulate, well educated, and knowledgeable (though social scientists
are probably prone to overestimate the importance of such concerns among the voters). Others want politi-
cians that are willing to go along and get along. Yet others put a premium on managerial efficiency.
13
Thus, several information-related conditions should hold. These conditions can occasionally be mutually
contradictory, or at least work at cross-purposes. If principals are too ignorant, or if their costs of acquiring
information are too high, then it is dangerous for them to delegate to agents that do not share their prefer-
ences. It may also be difficult for ignorant principals to figure out what the skill level of the agent actually is.
Challenges to Parliamentary Democracy 741
14
See Mair, Müller, and Plasser (2004) for a more comprehensive discussion of this phenomenon in
Western Europe.
Challenges to Parliamentary Democracy 743
effects of the former development were to trump those of the latter (as may well be the
case), democratic societies are not necessarily home free. Although reliable political
information is relatively inexpensive and widely available, it seems to be more unevenly
distributed among the citizens (Norris 2000).15 And in advanced democracies the
widening gap between the well informed and the poorly informed may become as
much of a challenge as any other inequality.
15
See, however, Lupia and McCubbins (1998) who argue that information shortcuts can compensate for
such lack of political knowledge.
744 Kaare Strøm et al.
process control but by improving its output efficiency (Majone 1998, 2001a, b; Scharpf
1999).16
As this volume demonstrates, the constitutions of the member states have to varying
degrees been adapted to the increasing importance of EU politics. Yet, even under the
best of circumstances reforms such as the introduction of parliamentary EU affairs
committees cannot fully compensate for the decline in parliamentary power. European
integration quite simply involves a decline in the capacity of national politicians to
affect policy outcomes that matter to their citizens. Even though the same trends, we
hope, increase the overall capacity of governments to deal with many serious problems
(environmental protection, international crime), growing agency problems could under
unfavourable circumstances consume the potential for gains in effectiveness.
All of this suggests that politicians at all stages of the chain of delegation suffer from
information deficiencies vis-à-vis the European Union. Of course, the information
problems of ordinary citizens are much greater, and politicians variously try to exploit
them. Hence, government members claim credit for popular EU policies, blame the
European Union for unpleasant developments or regulations, and leak fabricated infor-
mation about their heroic defence of national interests in Brussels. In contrast, opposi-
tion politicians take positions that are popular nationally but have no chance of finding
support in the European Union. Of course, this strategy is particularly attractive for
EU-sceptical parties.
Information problems can become even more acute when governments respond to
globalization by acting in international settings that are even less transparent than the
European Union. For example, decisions made in the World Trade Organization (WTO)
often have substantial domestic consequences but little transparency. Had not the WTO
become the bogeyman of some militant NGOs, even the best-informed voters might
never have learned about the relevance of decisions made there.
Finally, serious information problems may arise when decisions are made anony-
mously in world markets (Held et al. 1999: chs. 3–5). Such decisions, which are typically
exogenous to the national political decision-making systems and hence cannot be con-
trolled by politicians, may still be extremely consequential. Specifically, the liberaliza-
tion of financial markets since the 1980s has accelerated the speed by which such
market decisions impact on national polities, leaving politicians little scope and less
time for manoeuvre. Yet, it is the national political office holders that, if lucky, benefit
from market developments, or, if unlucky, pay the bill. In either case, however, demo-
cratic accountability is deprived of its rational basis. In a nutshell, more and often
particularly consequential decisions are taken outside the national chain of delegation.
At the same time citizens continue to hold politicians accountable for outcomes they
may not control, as indeed politicians claim credit for those outcomes that suit them.
16
Such proposals have included further strengthening the European Parliament, the direct election of the
Commission and Council Presidents, the creation of a second chamber of the European Parliament consist-
ing of delegations from national Parliaments, and the introduction of European referendums (e.g. see
Abromeit 1998; Grande 2000).
Challenges to Parliamentary Democracy 745
17
The trend towards greater reliance on parliamentary dissolution parallels what happened in Britain
during the evolution of parliamentary government (Cox 1987).
Challenges to Parliamentary Democracy 747
are becoming insulated from popular control. The challenges also include mastering the
institutional complexity and cognitive demands of Europeanization and Globalization.
At the same time, the decay of traditional information shortcuts means that
political information is cheapened in more ways than one, and that knowledge of the
past becomes systematically less helpful in controlling the representatives of the people.
While the decay of traditional information sources exacerbates the problem of adverse
selection, social disintegration may intensify moral hazard. And if traditional mech-
anisms of ex ante agent control are breaking down everywhere in the western world,
the resulting agency loss may be particularly severe under parliamentary systems that
have relied heavily on such devices. It is no great surprise that citizens are becoming
progressively more critical of those who hold power.
“These are daunting prospects, and the solutions are not obvious. Students of polit-
ical institutions and governance certainly cannot solve the perennial problems of polit-
ical delegation once and for all. It is unlikely that we even fully understand the
contemporary twists on these issues and their implications. Yet, we can remind policy
makers that the institutional menu is extensive, and that parliamentary democracies
have responded to serious challenges before.”
REFERENCES
Abromeit, Heidrun (1998). Democracy in Europe. New York: Berghahn Books.
Anckar, Carsten (2000). ‘Size and Party System Fragmentation’. Party Politics, 6: 305–28.
Bergman, Torbjörn and Damgaard, Erik (eds.) (2000). Delegation and Accountability in European
Integration: The Nordic Parliamentary Democracies and the European Union. London: Frank Cass.
––––, Müller, Wolfgang C., and Strøm, Kaare (eds.) (2000). ‘Special issue: Parliamentary democracy
and the chain of delegation’. European Journal of Political Research, 37: 3.
Behn, Robert D. (2001). Rethinking Democratic Accountability. Washington: Brookings Institution
Press.
Bendor, Jonathan, Glazer, Ami, and Hammond, Thomas H. (2001), ‘Theories of Delegation’,
Annual Review of Political Science, 4: 235–69.
Blondel, Jean and Cotta, Maurizio (eds.) (1996). Party and Government. Houndmills: Macmillan.
—— —— (eds.) (2000). The Nature of Party Government. Houndmills: Palgrave-Macmillan.
Boix, Carles (1998). Political Parties, Growth, and Equality. Cambridge: Cambridge University Press.
Castles, Francis G. (ed.) (1982). The Impact of Parties. London: Sage.
—— and Mair, Peter (1984). ‘Left-Right Political Scales: Some “Expert” Judgements’. European
Journal of Political Research, 12: 73–88.
Cukierman, Alex, Webb, Steven B., and Neyapti, Bilin (1994). Measuring Central Bank Independence
and Its Effects on Policy Outcomes. San Fransisco: ICS.
Cox, Gary W. (1987). The Efficient Secret. Cambridge: Cambridge University Press.
Dahl, Robert A. (1989). Democracy and Its Critics. New Haven: Yale University Press.
—— and Tufte, Edward (1973). Size and Democracy. Stanford: Stanford University Press.
Dalton, Russell J. (1999). ‘Political Support in Advanced Industrial Democracies’, in Pippa Norris
(ed.), Critical Citizens. Oxford: Oxford University Press.
—— (2002). Citizen Politics, 3rd edn. Chatham: Chatham House Publishers.
Della Porta, Donatella and Mény, Yves (eds.) (1997). Democracy and Corruption in Europe. London:
Pinter.
748 Kaare Strøm et al.
Djankov, Simeon, La Porta, Rafael, Lopez-de-Silanes, Florencio, and Shleifer, Andrei (2000).
The Regulation of Entry. National Bureau of Economic Research, Cambridge, MA, Working
Paper 7892.
Entman, Robert M. and Bennett, W. Lance (2001). ‘Communication in the Future of Democracy:
A Conclusion’, in W. Lance Bennett and Robert M. Entman (eds.), Mediated Politics.
Cambridge: Cambridge University Press.
Feigenbaum, Harvey, Henig, Jeffrey, and Hamnett, Chris (1999). Shrinking the State. Cambridge:
Cambridge University Press.
Franzese, Robert J. Jr. (2002). Macroeconomic Policies of Developed Democracies. Cambridge:
Cambridge University Press.
Garrett, Geoffrey (1998). Partisan Politics in the Global Economy. Cambridge: Cambridge University
Press.
Gilardi, Fabrizio and Braun, Dietmar (2002). ‘Delegation aus der Sicht der Prinzipal–Agent-
Theorie’. Politische Vierteljahresschrift, 43: 147–61.
Grande, Edgar (2000). ‘Post-National Democracy in Europe’, in Michael Th. Greven and Louis
W. Pauly (eds.), Democracy Beyond the State? Lanham: Rowman & Littlefield.
Gujarati, Damodar N. (1995). Basic Econometrics, 3rd edn. New York: McGraw-Hill.
Gunther, Richard and Montero, José R. (2001). ‘The Anchors of Partisanship’, in P. Nikiforos
Diamandouros and Richard Gunther (eds.), Parties, Politics, and Democracy in the New Southern
Europe. Baltimore: Johns Hopkins University Press.
Held, David, Anthony McGrew, Goldblatt, David, and Perraton, Jonathan (1999). Global
Transformations. Cambridge: Polity.
Hine, David (1993). Governing Italy. Oxford: Oxford University Press.
Hirschman, Albert O. (1970). Exit, Voice, or Loyalty. Cambridge, MA: Harvard University Press.
Huber, John and Inglehart, Ronald (1995). ‘Expert Interpretations of Party Space and Party
Locations in 42 Societies’. Party Politics, 1: 73–111.
Katzenstein, Peter J. (1985). Small States in World Markets. Ithaca, NY: Cornell University Press.
Kelsen, Hans (1926). Das Problem des Parlamentarismus. Vienna: Braumüller.
King, Anthony (1976). ‘Modes of Executive–Legislative Relations: Great Britain, France, and
West Germany’. Legislative Studies Quarterly, 1: 11–36.
Kirchheimer, Otto (1966). ‘The Transformation of the Western European Party Systems’, in
Joseph LaPalombara and Myron Weiner (eds.), Political Parties and Political Development.
Princeton: Princeton University Press.
Klingemann, Hans-Dieter, Hofferbert, Richard I., and Budge, Ian (1993). Parties, Policies, and
Democracy. Boulder: Westview Press.
––––. (1999) ‘Mapping Political Support in the 1990s: A Global Analysis’, in Pippa Norris (ed.),
Critical Citizens: Global Support for Democratic Governance. Oxford: Oxford University Press.
Knapp, Andrew and Wright, Vincent (2001). The Government and Politics of France, 4th edn.
London: Routledge.
Lane, Jan-Erik and Ersson, Svante (1999). Politics and Society in Western Europe, 4th edn. London:
Sage.
Laver, Michael and Garry, John (2000). ‘Estimating Policy Positions from Political Texts’.
American Journal of Political Science, 44: 619–34.
Levitt, Malcolm and Lord, Christopher (2000). The Political Economy of Monetary Union.
Houndmills: Macmillan.
Lewis, W. Arthur (1965). Politics in West Africa. London: George Allen and Unwin.
Lijphart, Arend (1977). Democracy in Plural Societies. New Haven: Yale University Press.
Lijphart (1984). Democracies. New Haven: Yale University Press.
Challenges to Parliamentary Democracy 749