Handbook of Regional and Urban Economics - VOL.3 PDF
Handbook of Regional and Urban Economics - VOL.3 PDF
The aim of the Handbooks in Economics series is to produce Handbooks for var-
ious branches of economics, each of which is a definitive source, reference, and
teaching supplement for use by professional researchers and advanced graduate
students. Each Handbook provides self-contained surveys of the current state of
a branch of economics in the form of chapters prepared by leading specialists on
various aspects of this branch of economics. These surveys summarize not only
received results but also newer developments, from recent journal articles and
discussion papers. Some original material is also included, but the main goal is
to provide comprehensive and accessible surveys. The Handbooks are intended
to provide not only useful reference volumes for professional collections but also
possible supplementary readings for advanced courses for graduate students in
economics.
CONTENTS OF THE HANDBOOK
Chapter 1
Advances in Regional Economics
P. NIJKAMP and E.S. MILLS
Chapter 2
The Location of Production Activities
M.J. BECKMANN and J.E THISSE
Chapter3
Residential Mobility and Household Location Modelling
W.A.V. CLARK and W.EJ. VAN LIEROP
Chapter4
Public Facility Location: A Multiregional and Multi-Authority Decision Context
B. JOHANSSON and G. LEONARDI
Chapter5
Spatial Equilibrium Analysis
T. TAKAYAMA and W.C. LABYS
Chapter 6
Regional Economic Dynamics
A.E. ANDERSSON and R.E. KUENNE
Chapter 7
Regional and Multiregional Economic Models: A Survey
P. NIJKAMP, P. RIETVELD and E SNICKARS
Chapter8
Regional, Interregional and Multiregional Input-Output Analysis
G.J.D. HEWINGS and R.C. JENSEN
viii Contents of the Handbook
Chapter 9
Spatial Interaction, Transportation, and Interregional Commodity Flow Models
D.F. BATTEN and D.E. BOYCE
Chapter 10
Regional Econometric and Dynamic Models
R.J. BENNETT and L. HORDIJK
Chapter 11
Qualitative Statistical Models for Regional Economic Analysis
N. WRIGLEY and F. BROUWER
Chapter 12
Multiple Objective Decision Analysis in Regional Economics
P. NIJKAMP and P. RIETVELD
Chapter 13
Regional Labor Market Analysis
A. ISSERMAN, C. TAYLOR, S. GERKING and U. SCHUBERT
Chapter 14
Regional Energy and Environmental Analysis
T.R. LAKSHMANAN and R. BOLTON
Chapter 15
Innovation and Changes in Regional Structure
E.J. MALECKI and P. VARAIYA
Chapter 16
Regional Policies in Developing Countries
H.W. RICHARDSON and P.M. TOWNROE
Contents of the Handbook ix
Chapter 17
Advances in Urban Economics
E.S. MILLS and P. NIJKAMP
Chapter18
The Theory of Urban Residential Location
M. STRASZHEIM
Chapter 19
Theories of Urban Business Locations
K. STAHL
Chapter 20
The Structure of Urban Equilibria: A Unified Treatment of the Muth-Mills Model
J.K. BRUECKNER
Chapter21
Computer Simulation Models of Urban Location
J.F. KAIN
Chapter22
Dynamic Urban Models
T. MIYAO
Chapter23
General Equilibrium Modeling of Systems of Cities
J.V. HENDERSON
Chapter 24
Economic Theory and Housing
R. ARNOTT
Chapter 25
The Demand and Supply of Housing Service: A Critical Survey of the Empirical
Literature
E.O. OLSEN
Chapter 26
Urban Transportation
M.E. BEESLEY and M.A. KEMP
x Contents of the Handbook
Chapter 27
Urban Public Facility Location
C. REVELLE
Chapter 28
Rural-Urban Migration in Developing Countries
D. MAZUMDAR
Chapter 29
Theoretical Analysis of Local Public Economics
D.E. WILDASIN
Chapter30
Financing Urban Public Services
R. PRUD'HOMME
Chapter31
Urban Amenities and Public Policy
T.J. BARTIK and V.K. SMITH
Chapter32
Urbanization in the Developing World: Patterns, Issues, and Policies
A.M. HAMER and J.F. LINN
Chapter33
City Size and Place as Policy Issues
G. TOLLEY and J. CRIHFIELD
Contents of the Handbook X1
VOLUME III
Chapter 34
Introduction: Applied Urban Economics
PAUL CHESHIRE and EDWIN S. MILLS
Chapter35
Trends in Sizes and Structures of Urban Areas
PAUL CHESHIRE
Chapter36
Urban Areas with Decentralized Employment: Theory and Empirical Work
MICHELLE J. WHITE
Chapter37
Quality of Life and Environmental Comparisons
JOSEPH GYOURKO, MATTHEW KAHN and JOSEPH TRACY
Chapter38
Agglomeration Economies and Urban Public Infrastructure
RANDALL W. EBERTS and DANIEL P. McMILLEN
Chapter39
Urban Labour Markets
GRAHAM R. CRAMPTON
Chapter40
Urban Housing Markets: Theory and Policy
CHRISTINE M.E. WHITEHEAD
Chapter41
Hedonic Analysis of Housing Markets
STEPHEN SHEPPARD
xii Contents of the Handbook
Chapter 42
The Land Market and Government Intervention
ALAN W. EVANS
Chapter43
Urbanization in Transforming Economies
CHARLES M. BECKER and ANDREW R. MORRISON
Chapter44
Economic Analysis of Housing Markets in Developing and Transition Economies
STEPHEN MALPEZZI
Chapter45
Poverty in Developing Countries
ERNESTO M. PERNIA and M. G. QUIBRIA
Chapter46
Urban Transportation
KENNETH A. SMALL and JOSE A. G6MEZ-IBANEZ
Chapter 47
Sorting and Voting: A Review of the Literature on Urban Public Finance
STEPHEN ROSS and JOHN YINGER
Chapter34
EDWIN S. MILLS
Northwestern University, Evanston
1323
1324 P. Cheshire and E. Mills
tion economies are the benefits of urban concentrations that accrue externally to
firms and possibly entire sectors. They result partly from profit-motivated clus-
tering of firms and partly from government capital facilities, typically referred
to as infrastructure-transportation and communication facilities, and some pub-
lic utilities. The hallmark of research during recent years has been the careful
measurement of returns to infrastructure capital.
Part II surveys literature on specific urban markets. Crampton (Chapter 39)
looks at studies of urban labour markets. This is a vast field where the boundaries
between urban labour market research, and labour market research in general,
are not always clearly drawn. The obvious spatial aspects of labour markets form
the first set of areas surveyed in this chapter, i.e., the interrelationship between
residential and workplace location, mobility and search. The second set of issues
surveyed relate to the various types of stratification of labour markets by skill,
gender and family status and how these factors influence labour market outcomes
and behaviour.
Whitehead (Chapter 40) reviews research on housing markets. This chapter
concentrates on the literature that has analysed the role of those specific attributes
of housing, including its locational specificity and durability, that render housing
markets both distinctive and urban. It focuses on housing market research in the
developed world. There has now been a substantial number of studies of hous-
ing demand of increasing sophistication. We cannot be sure, but it does seem
that quite secure knowledge has been gained in this area-in both how to do
such studies, and in terms of the magnitudes of the resulting estimates of key
relationships such as price and income elasticities of demand. The durability
of housing makes its supply distinctive with research distinguishing between
the determinants of new housing supply and investment in the existing stock.
After briefly surveying aggregate models of housing markets and the interaction
between housing and labour markets, the chapter concludes with a review of
housing policy research.
Hedonic analysis of housing markets is left to a separate chapter by Shep-
pard (Chapter 41). Although now associated with work on housing markets, this
technique had its roots in the Boston vegetable market in the 1920s. Sheppard
shows how the technique has developed into an invaluable and increasingly so-
phisticated analytical tool, not only for housing markets but also for markets in
urban land, the analysis of environmental goods and local public finance. Again,
this is an area in which research has developed and more recent research has
demonstrably learned from the collective efforts of earlier researchers. Certain
techniques and approaches have emerged as clearly to be favoured. Sheppard
Ch. 34: Introduction: Applied Urban Economics 1327
veloping countries, and serious poverty has almost disappeared in high income
countries of east and southeast Asia, such as South Korea, Taiwan, Singapore and
Malaysia.
Part TV contains papers on specific urban problems and sectors. There is
some arbitrariness in the assignment of papers between Parts II and IV, but all
subjects surveyed in Part IV have a strong government and government policy
component in most countries. Again, much of the content of the papers in Part
IV is relevant to developing countries, but the papers have a high income coun-
try orientation. Small and G6mez-Ilbfiez (Chapter 46) survey the sophisticated
literature on urban transportation. In no sector have US governments caused as
much resource misallocation as in urban transportation. The predominant wastes
result from overinvestment in fixed rail transit systems, underpricing of road use
and wasteful government ownership of bus systems. Yinger and Ross (Chapter
47) survey local public economics. They summarise what has been learned from
the voluminous literature on the Tiebout hypothesis and from the provincial ways
that local governments are financed in the US
What do we know now about urban economics as a result of applied research
undertaken during the last decade or two? The answer surely is 'a great deal',
although there is very obviously much more we would like to know. Even within
their own fields, our authors could only survey the body of research that had
been done. Evans, for example, reviews a great deal of work on the impact of
land use regulation on housing. There has been substantial progress here and
urban economists enjoy a significant degree of consensus on the economic impact
of this activity. Yet, Evans provides no account of the impact of regulation on
urban uses of land other than housing. These impacts are likely to be at least
as significant as they are on housing since we know that land cannot be cost-
lessly substituted out of either production or service activities. If we compare
communities in the US and UK that are as comparable as possible except for the
constraints their systems of land use regulation place on the supply of land, we
observe that the price of retail land is up to 100,000 times higher in the most
constrained community. Evans can provide no evidence on the economic impact
of such a level of constraint, however, because no research has been done.
In areas where applied research has been done, in contrast, there has been
much useful progress. On the basic spatial structure of urban areas, we have much
better measurement of the extent of dispersion or suburbanisation from urban
cores. We know that dispersion has been a worldwide phenomenon and, although
governments can influence it, their many attempts to stop it have so far done
much more harm than good. We have better estimates of the relative dispersion
of various employment sectors and the reasons for employment dispersion. We
Ch. 34: Introduction:Applied Urban Economics 1329
have important insights into the roles of subcentres in the dispersion process,
although there is still much to be learned about subcentres. We know that not
only radial highways but also radial fixed rail transit systems promote dispersion.
Indeed, American opponents of what they refer to pejoratively as 'sprawl' ignore
the fact that London was among the world's most dispersed metropolitan areas
long before auto commuting became important, and that an important reason for
this is its century-old fixed rail commuter system that extends far from the core.
Indeed, London if functionally defined, despite intensive regulatory efforts over
400 years, is more dispersed than Los Angeles, the favourite "whipping boy" of
US critics of low density development. The degree of relative-and in several
cases, absolute-recentralisation observed in many European cities during the
1980s did not reflect any new policies. The evidence suggests that it reflected
economic and social changes which generated new patterns of incentive.
Despite recent high quality research, there is as yet no consensus about mea-
sures of the quality of life in metropolitan areas. Indeed, the entire concept may
be a "will-of-the-wisp", and rankings of metropolitan areas may depend on de-
tails of people's utility functions. Although most explanations of the remarkable
stability of the size distribution of metropolitan areas in many countries have
been in terms of production cost, it is certainly consistent with the possibility that
preference functions for the disaggregated factors which contribute to peoples'
perceptions of the quality of life, at least in relation to the sizes of metropolitan
areas, differ among people and that preferences change only very slowly, at least
in a gross sense.
The precise nature of agglomeration economies is still something of a mys-
tery. Proximity of diverse activities and the densities permitted by the ability
to substitute capital for expensive land certainly economise on transportation,
including commuting, costs. Here, as with land use regulation, the research we do
not have could provide the key. For data reasons all quantitative studies have been
confined to production activities. Yet, some of the most pronounced agglomer-
ation economies (and 80% or more of economic activity in large cities) are-if
the actual location concentrations are a guide-in service sectors. Wholly new
"industrial districts" have emerged in the past 15 years in some cities-the media
district in London's Soho area is an example. But these new urban concentrations
have been in nonmanufacturing activities-the up to 80% of economic activity
that had to be excluded from the analysis of agglomeration economies.
Exactly how proximity economises on communication costs, especially in the
computer age, is still an open question, although it may be related to value added
at intermediate stages of production (with resulting "inventory" costs of trans-
port) and on the form of the communication. Bespoke information, particularly
1330 P Cheshireand E. Mills
and ghettos to the low-lying downwind, east of London. The reaction to these
developments was strict but ineffective regulation to try to impose containment.
The desire for containment was probably motivated as much by distributional
concerns-to protect the villa owners from the encroachment of the city-as
with anything else. In Paris, still threatened with siege, the walls expanded and
the rich retained their urban presence and developed a strong urban culture sup-
ported by urban amenities. The incentive for the influential to develop policies for
containment did not arise, and in the late twentieth century the rich of Paris are
still overwhelmingly concentrated in the urban core; and Paris, though still high
density, has grown by continuous expansion of its built-up area rather than by the
leap-frogging process that has ensured that functional London now covers most
of southeast England. The Netherlands came into existence through the public
provision of land via drainage. Local governments were obligated to provide
land for development and this is still reflected in the country's system of land
use regulation. Despite a density of population that is twice that of the UK, land
for housing remains, as a result, very much cheaper. US cities achieve their leap-
frogging and low densities by allowing new communities to incorporate (and
so retaining a near perfectly elastic supply of urban land) but then providing a
substantial incentive via the fiscal system to develop at the lowest density.
Relative to its status 20 years ago, urban research has perhaps made more
progress in understanding developing rather than high income countries. We now
have a much better understanding of the factors that generate rural-urban migra-
tion. Migration studies have shown that people move from rural to urban areas
in developing countries for much the same reasons that they do in high income
countries, i.e., to obtain better education and because more and better jobs can
be found in urban than in rural areas. Migration is on a much larger scale in
developing than in developed countries. We have also learned that migration and
urban growth have slowed as population growth has slowed and low productivity
agriculture has shrunk. The urban shares of total population in newly emerging
countries such as Brazil or Korea have recently approached those in high income
countries. Urban growth in such countries is hardly faster than in high income
countries.
A remarkable amount has been learned about urban housing in developing
countries. Urban housing prices and rents are typically much higher relative to
incomes in developing than in high income countries. The reasons are counter-
productive government housing policies. Some, such as controls on conversion
of land from rural to urban uses are shared with developed countries. Others,
such as a lack of legal infrastructure that permits binding contracts between
landlords and tenants, the refusal to permit private financial markets to develop
1332 P. Cheshire and E. Mills
enforceable modem mortgages, land use controls that prohibit the only housing
that low income residents can afford, and restrictions on development of pri-
vate housing development companies, are more specific to developing countries.
Governments, as they traditionally do, blame high rents and prices on greedy
landlords and speculators.
Urban poverty has declined even more than would be inferred from overall
measures of economic growth in many developing countries. Improved nutri-
tion and water quality and improved education and health care appear to be the
reasons. Poverty rates have stagnated or increased mainly in countries in which
violence and political breakdown have occurred--much of the former Soviet
Union, parts of tropical Africa and the southeast Asian peninsula. By far the
most important contribution to falling poverty rates is rapid economic growth.
The extent to which the poor share in the early stage of economic growth varies
among countries, but there is no record of economic growth causing a decrease in
the living standards of the poorest 10 or 20% of the population in any developing
country.
Urban transportation issues are also better understood than they were 10 or
20 years ago. On this subject, virtually all the progress since the early 1970s has
been applied, and there has been a great deal of it. Many careful studies have
estimated the extent and cost of road congestion. Congestion means the excess
of marginal over average social cost of additional road users and indicates a need
for additional investment and/or congestion pricing. All careful studies conclude
that additional fixed rail commuter systems are unjustified in the low densities
of US metropolitan areas, but there is disagreement as to what are optimum
investments: certainly much better traffic control systems and much better-
preferably privatised-bus systems. The extent to which new and expanded roads
are justified is controversial. Undoubtedly, the best strategy is careful benefit-
cost studies of each major proposal. Such studies are rare, in part because of the
ideological nature of the debate.
Poverty rates have also decreased in most developed countries during the last
decade or two-the US and UK being the most important exceptions. Much
more is now known about the socioeconomic characteristics of the poor in many
countries, certainly including the US. More is probably known in the US than in
any other country about the spatial and racial characteristics of the poor. Much
of the US research has focussed on the plight of the poor people, especially poor
minorities resident in central cities.
US studies have shown that the best nursery for future poverty is to have
poor and poorly educated parents, and especially to be raised in a single-parent
household. It is also now known that, other things being equal, the probability
Ch. 34: Introduction:Applied Urban Economics 1333
that children will be poor, be school dropouts, acquire criminal records, and have
teenage pregnancies is greater if they are raised in poor neighbourhoods. Aside
from improving educational programmes that poor children can access, not much
is yet known about government programmes that can alleviate poverty. There is
no academic consensus as to whether existing or modified government transfer
programmes increase or decrease poverty. Again, the ideological intensity of the
debate places objective analysis beyond mere mortals.
Yinger and Ross show that we now know almost all there is to know about
conceptual aspects of the Tiebout hypotheses. Applied work has been consider-
able; we know that typical suburban communities have Tiebout-like characteris-
tics, but we still do not know how Tiebout-like they are and what causes some
to be more Tiebout-like than others. An important outstanding issue is whether,
or not, suburban Tiebout-like characteristics depend on police-power land use
controls in relatively high income communities.
Housing is more than half of all urban real estate and endless research is
justified. More estimates of demand for aggregated housing, however, ought to
be assigned low priority in many countries. More important is what governments
should do, or stop doing, to improve the functioning of housing markets. Many
scholars worry about the availability of housing that relatively low income resi-
dents can afford. At upwards of 15%, housing constitutes the largest single item
of consumer expenditure in most countries. But while research into regulation of
utilities, energy and transport operation is voluminous, there have been far fewer
studies of how governments cause housing costs to be excessive: greenbelts,
growth controls, other land use controls, rent controls, controls on conversion
of land from rural to urban uses, excessive restrictions on private housing finance
organisations, failure to develop legal infrastructure for housing property rights
and contract enforcement, and so on.
In the US, securitisation has moved owner-occupied debt and rental debt and
equity financing from neighbourhood financial institutions to national and in-
ternational bond and stock markets. The benefits have been enormous. Benefits
presumably would be even greater in other countries, especially in relatively
small newly industrialising countries in east and southeast Asia. What is not
understood is why this innovation has been so slow to spread.
Every urban issue is better understood in high income rather than developing
countries. As noted, we can now be confident that most people, both the poor
and the nonpoor, migrate from rural to urban areas for a similar mix of reasons
in developing and high income countries. Most developing country governments
adopt many policies to slow rural-urban migration, especially of the poor. The
worst policies are those that try to increase labour productivity in agriculture. The
1334 P. Cheshire and E. Mills
result is sometimes to lower food prices, which is desirable, but often, because
they are accompanied by protection, such policies increase food prices. They
always reduce farm employment, as rapid agricultural productivity growth has
done in high income countries. High productivity agriculture is of course impor-
tant, but it is not an appropriate policy to slow rural-urban migration. The second
worst policies are those that promote rural industrialisation. Internal transporta-
tion is poor in most developing countries and most industries must be located in
large, preferably port cities to survive. A third set of misguided policies is to limit
the construction of the only housing that the urban poor can afford. Both govern-
ments and scholars should direct their attention to policies that can improve the
housing, health, education and productivity of the urban poor. Reduced controls
on private urban business development should be an important component of any
justifiable set of government actions.
As Pemia and Quibria show, we now know much more about poverty levels
and incidence in developing countries than we did a decade or two ago. Economic
growth is, of course, the best long run antipoverty policy. But the incidence
of poverty correlates only moderately with levels of development or income,
however measured. It should now be possible to undertake careful comparative
studies among developing countries to deepen insights as to the determinants of
poverty. The fact that high literacy rates and high levels of educational attain-
ment are important strategies for poverty reduction even in quite low income
countries is strongly suggested by historical data in Korea and Taiwan. Opening
up educational and employment opportunities to girls and women are impor-
tant in raising income levels, reducing birth rates, reducing infant mortality and
generally promoting family well-being.
Even among high income countries, poverty rates do not correlate strongly
with measures of real income levels or growth rates. Again, comparative studies
are needed. The US has been a puzzle for two decades. It has experienced almost
no decrease in poverty rates despite economic growth rates that are typical of
most other OECD countries, and very high income levels. Although the US prob-
ably spends as large a fraction of its GDP on transfers as most OECD countries,
the US safety net may leak more than those of some other countries. Although
studies by those who believe the US transfers promote dependence and poverty
have been justifiably criticised by economists, it seems premature to conclude
that transfer programmes do not have adverse incentive effects, either in the US
or other OECD countries. Presumably, careful cross-sectional studies could shed
light on the issue.
Could cross-national studies also shed light on the public economics of local
governments? US scholars undertake cross-sectional studies within the US, made
Ch. 34: Introduction:Applied UrbanEconomics 1335
PAUL CHESHIRE*
Contents
1. Introduction 1340
2. The distribution of city sizes 1342
2.1. Is there any underlying distribution of city sizes and, if there is, what form does it take? 1342
2.2. Explanations of the distribution 1348
2.3. Looser definitions of urban hierarchy 1351
3. Economic geographical approaches 1358
4. Intraurban patterns of development and change 1361
5. Some recent trends 1366
References 1369
* I am grateful to Charles Leven for reading an earlier draft of this chapter and for his helpful comments.
The author is naturally responsible for any remaining errors, whether of fact or omission.
Handbook of Regional and Urban Economics. Editedby P. Cheshire and E.S. Mills
( 1999 Elsevier Science B. V All rights reserved.
1339
1340 P Cheshire
Abstract
This chapter reviews the literature dealing with systems of cities and the patterns
of development within such systems. It starts with the longstanding question
of the distribution of city sizes, both in relation to how this distribution can be
described and, given the form that it takes, how that form can be explained. Such
explanations frequently invoke various sorts of agglomeration economies and so
some of the literature relating to these is included here. The chapter then surveys
the literature that examines patterns of development within urban systems, and
then work at a more disaggregated level on suburbanisation. The chapter con-
cludes with a summary of research into recent patterns of urbanisation, including
relative recentralisation.
1. Introduction
This chapter surveys a wide and disparate set of literature. One unifying theme
is that it is not concerned with the location or growth of individual cities but with
systems of cities; types of cities or settlements, such as large compared to small
cities, or metropolitan compared to nonmetropolitan areas; or core cities com-
pared to suburbs. It excludes such literature, therefore, as Krugman (1993), which
is concerned with modelling the factors that influence the growth of individual
cities. It also excludes studies of particular cities, their history or development.
Although it identifies the origins of some of the more important themes and
ideas, the main emphasis is on the relatively recent literature, since about 1975
or 1980. Given the very wide range of themes surveyed there is no pretence of
including all contributions; the aim has been to include the main ideas and the
main approaches and to evaluate some of the key contributions.
Section 2 starts with a review of the literature on the distribution of city sizes.
The first issue, addressed in Section 2.1, is whether, or not, the evidence shows
that there is such a distribution-at least in the sense of a universal distribution
conforming to some simple form. The Pareto distribution applied to the urban
Ch. 35: Trends in Sizes and Structures of Urban Areas 1341
hierarchy has a long history. The rank size rule--described by Tinbergen (1968)
as one of the most well defined of socioeconomic regularities, or more recently
by Krugman (1996) as one area where economists "have complex, messy models,
yet reality is startlingly neat and simple"-does not work perfectly but still haunts
the literature. To conclude that this is so simply because it works, would be too
simple, however. Section 2.1 reviews the literature that attempts to identify the
circumstances under which it could be said to work and tests the extent to which,
given the definition of "working", it does work. The evidence suggests the rank
size rule persists in working quite well.
This is followed by sections surveying the work that has tried to explain
why: (i) the size distribution of cities in an urban system conforms to a Pareto
distribution; (ii) there are cities in an urban system of differing sizes; or (iii)
the distribution of city sizes varies between countries. While much of the work
addressing the first of these questions relies on behavioural assumptions about
economic agents and uses what are recognisably economic tools, all of the litera-
ture addressing the second and third falls into this category. There is an important
body of literature attempting to explain why the size distribution of cities con-
forms to a Pareto distribution (or, even more specifically, obeys the rank size
rule) that relies on purely stochastic mechanisms or analogues with physical
phenomena. This body of work, together with other literature that attempts to
explain the conformity of urban systems to the Pareto distribution, is reviewed in
Section 2.2
Section 2.3 examines literature relating to the less strict definition of urban
hierarchy implied in questions (ii) and (iii) above. It includes some of the work
that has attempted to categorise and measure agglomeration economies since the
sources of these are widely perceived as explaining variations in city size. Further
material on this topic is also included in Section 4 where changing patterns of
agglomeration economies have been investigated in the context of the relative
growth and decline of different parts of the urban system.
There is also, however, a rather different approach to investigating urban sys-
tems and patterns of urban development. It is not directed towards investigating
the system of cities, in the sense of the distribution of city sizes, but, in the
context of an urban system, to explaining-or at least classifying-patterns of
urban development. This may be in terms of different relative rates of growth
of cities of different categories, or of core cities relative to their suburban rings.
This approach is the subject of Section 3 and, implicitly, is returned to in the final
section which surveys very recent trends in the patterns of urban development.
The chapter proceeds in Section 4 to analyse the factors underlying patterns
of differential spatial growth within city regions, and the growth of particular
1342 P. Cheshire
types of cities compared to others. Two main threads can be distinguished. There
is a line of work that has as its focus the changing distribution of people and jobs
within large metropolitan areas. The theoretical core for this is provided by the
model originating with Wingo (1961), Alonso (1964) and developed by Muth
(1969), Mills (1972) and Evans (1973). Explanations of intraurban employment
location are less complete, so the investigation of changes in the location of jobs
has tended to be more ad hoc. Broadly this body of work can be thought of as
the study of suburbanisation. One of its recurring questions has been whether,
in suburbanisation, the process is led by the decentralisation of jobs, or the de-
centralisation of people. The second thread of literature reviewed in Section 4
is that which studies changes in the distribution of people and jobs between
settlements of varying size, and between metropolitan areas on the one hand, and
nonmetropolitan-or rural ("ex-urban") areas-on the other. Theoretically, this
works tends to be more ad hoc than that on suburbanisation, but it does constitute,
nevertheless, a far larger volume of material.
The final section of the chapter is a brief review of the work that has doc-
umented and investigated recent changes in the distribution of population be-
tween large and small cities, between metropolitan and nonmetropolitan areas
and within the various components of city regions. "Recent" in this context is
defined as since 1980. The general conclusion is that despite the fact that the
experience of individual cities has become more varied, internationally (at least
within what might be called the mature economies) there is stronger evidence
of a predictable pattern of change, determined by common causal factors, than
might be expected given the diversity and variety of cities.
2.1. Is there any underlying distribution of city sizes and, if there is, whatform
does it take?
over 60 years ago, and both demonstrated that the city size distribution could be
represented as a Pareto distribution:
y = Ax-' (2.1)
or
what sensitive to sample size. The impact on the estimated value of a is not,
however, as consistent as it is with urban definition. This may reflect the rather
limited exploration of the issue, however, compared to some other studies such
as Malecki (1980). Guerin-Pace (1995), using an essentially identical form, for
example, but exploring varying size cutoffs from 2000 to 100,000 for the French
concept of agglomeration (the built-up area that, in France at least, is closely
correlated with a functional definition of "city"), finds that not only does the size
cutoff influence the estimated value of a, but also the conclusion as to whether
population was becoming more or less metropolitanised over time. Measured
population concentration (measured, that is, in terms of estimates of a) had in-
creased for each 50-year time period in France from 1831 to 1982 if all cities
over 2000 were analysed. If only cities greater than 50,000 (or 100,000) were
included in the sample, however, then at each of the four dates at 50-year intervals
population had apparently become less concentrated in the largest cities. 3
The more detailed studies seem, therefore, to show that estimates of a are
sensitive to the sample selection criteria. This implies that the Pareto distribution
is not precisely appropriate as a description of the city size distribution. The issue
of whether, or not, the Pareto distribution itself was an appropriate form was also
raised by Rosen and Resnick (1980) who explored adding quadratic and cubic
terms to the basic form. They found indications of both concavity and convexity
with respect to the pure Pareto distribution, with more than two-thirds of coun-
tries exhibiting an upward concavity. As Gutrin-Pace (1995) demonstrates, this
particular result is also sensitive to sample selection. Extending the sample to all
French cities over 2000, he finds no evidence of concavity for the appropriate
time period. Concavity was observed in the nineteenth century, but in the 20th
century it was only apparent if the sample was truncated to cities above 100,000.
Another strand of work, associated primarily with Alperovich (1984, 1988,
1989) but also with Kamecke (1990), is the investigation of whether, or not, urban
systems really conform to the rank size rule. As Alperovich points out, this is not
only a question of whether the distribution of city sizes conforms to the Pareto
distribution, nor just, if it does conform, whether, or not, the value of a = 1-
the criterion adopted by Rosen and Resnick (1980). If the rank size rule is to
apply precisely, it also implies that the constant, A, is equal to the size of the
largest city. In successive articles, Alperovich derives a series of tests for rank
size rule conformity and applies them to data for administratively defined cities
over 100,000 in 15 countries (1984); to similar data for 17 countries (1988); and
3 This suggests that the results reported in Eaton and Eckstein (1997) may misrepresent the stability of the
French urban system over time and the extent to which it conforms to the rank size rule, since they confine
their analysis to only the 39 agglomerations that had populations exceeding 50,000 in 191 1.
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1347
to Metropolitan Statistical Areas (MSA) data for US cities at two dates (1989).
This last paper also explores the issue of the sensitivity of rank size rule tests to
the sample size. Using the tests suggested in Alperovich (1984), the strict form of
the rank size rule advocated is rejected. It is also found that the results are highly
sensitive to the sample size. Using the tests developed by Alperovich (1988),
framed in terms of relative rather than absolute city sizes, he finds there is less
sensitivity to the exclusion of smaller cities and that the rank size rule cannot
be rejected when applied to the data for US MSAs in either 1970 or 1980. In
absolute terms, however, omitting observations systematically causes estimates
of a to increase from 0.95 to 1.6.
An alternative procedure for testing whether the rank size rule is an appro-
priate statistical description of the city size distribution-or, more precisely, the
size distribution of 318 US MSAs in 1980-is proposed by Hsing (1990). He
suggests that instead of fitting the simple log linear Pareto distribution, the more
general functional form suggested by Box and Cox (1964) should be applied.
This reduces to a log linear form when the estimated transformation parameters
approach zero. At least for the data to which the Box-Cox function is applied
by Hsing, the transformation parameters prove to be significantly different from
zero and the implied "Pareto coefficient", a, is estimated to be 0.86 compared
to 0.97 for the simple log linear form. Thus, the evidence examined in this way
suggests both that the rank size rule does not (precisely) hold and that the simple
Pareto distribution itself is not entirely appropriate.
Cameron (1990), although writing at the same time as Hsing (1990), develops
both these ideas. Her main thrust, however, is that while refinement of the estima-
tion of the exact value of the a parameter of a Pareto distribution of city sizes is
a worthwhile endeavour, "not much structural modelling has been undertaken".
Wanting more than curve fitting, she argues that while a robust description is
desirable, it is mainly desirable so that we know more precisely what it is that
theory and empirical investigation should be seeking to explain.
Cameron (1990) raises a somewhat technical objection to the way of investi-
gating the reasons for variation in the form of the distribution of city sizes across
countries used by Rosen and Resnick (1980), and subsequently by Alperovich
(1993b). They made the estimated a values dependent on a number of indepen-
dent variables-such as income level, industrial value added as a proportion of
GDP, railway mileage and overall population density. Cameron's objection is that
this is statistically invalid. Since a values are points without variance, the "true"
values cannot be observed. Cameron's substantive contribution is, however, the
recognition that if just estimates of a are the variable to be explained, it is im-
possible to include characteristics of individual cities as part of the explanation
1348 P. Cheshire
Both Tinbergen (1968) and Krugman (1996), having noted the regularity of the
distribution of city sizes, grapple with the issue but fail to come up with a wholly
satisfactory explanation. Tinbergen (1968) develops a model derived from central
place theory but including manufacturing industry as well as services. Krugman
(1996) falls back on random processes, combined with the interaction of a non-
homogeneous physical landscape (that determines the size of hinterlands on the
basis of transport cost minimising routes) and economies of scale.
Alperovich (1993b) argues that there are two near-universal conclusions that
can be reached on the basis of the numerous studies of city size distributions. De-
spite the qualifications discussed above, it does seem reasonable to agree with the
first of these-that the Pareto distribution is a reasonably accurate representation
of the city size distribution for many countries--certainly for the upper tail of
the distribution. The evidence as to whether, or not, the exponent approximates
1 is not quite so clear-cut. Alperovich's view is that a second near-universal con-
clusion is that it varies considerably both between countries and through time.
Krugman (1996), however, concludes from the evidence that estimates of the
exponent tend to be quite close to 1 and-appealing to Rosen and Resnick's
(1980) results-get closer to 1 the more carefully metropolitan areas are de-
fined.4 Certainly, without having internationally comparable studies that have
used functionally defined metropolitan areas, it is hard to conclude positively that
4 Apart from Rosen and Resnick's (1980) results there are those of Osada (1997). In addition, it may be
worth noting that values of the Pareto exponent estimated for functionally defined core cities of the 10 county
European Community of 1981-treating all the core cities as a single urban system-were 1.04 for 1951 and
0.96 for 1981 (Cheshire and Gorla, 1987).
Ch. 35: Trends in Sizes and Structures of Urban Areas 1349
the exponent is often very far from 1. This implies a more substantive objection
to studies that have attempted to explain international variation in estimates of the
value of the exponent, based on administratively defined cities, than the technical
objection raised by Cameron (1990).
Explanations of the form of the city size distribution fall into two main types:
those that are based on economic relationships or models-urban systems
theories-and those that are essentially stochastic. The former, in turn, can be di-
vided between variants of central place theory and models that rest on economies
of scale, agglomeration and differential transport costs. Development of central
place theory as a formal explanation of the distribution of city sizes-associated
originally with Christaller (1933), and developed in a somewhat more formal
way, more in tune with economic modes of thought 5 by Lsch (1940)-seems to
have lapsed following the work of Beckmann (1958) and Beckmann and McPher-
son (1970). Certainly, central place theory implies the existence of cities of vary-
ing sizes but it is unclear what mechanism would drive a central place system to
generate a size distribution of the form observed (see, for example, Leven, 1968).
The same point can be made for the alternative approaches developed inde-
pendently by Evans (1972) and the more complete and better known work of
Henderson (1974, 1977, 1986, 1987, 1988). Evans's model has two sectors-
each of which is heterogeneous-manufacturing and business services with the
output of business services, entirely sold as an input to manufacturing. There
are costs of commuting so rents and labour costs vary with city size. Transport
costs of the output of manufacturing are zero, while those of business services are
assumed to be infinite. This is probably the original model in this vein to assume
two sectors-one of traded and one of nontraded, goods. The assumption of a
sector that is traded without transport costs is convenient in the study of urban
systems, although inappropriate in modelling the location of particular cities.
In Evans's (1972) model, business services are assumed to be produced sub-
ject to economies of scale but, because of the assumption with respect to trans-
port costs, they are produced in all manufacturing locations (i.e., cities). Man-
ufacturing firms are assumed to vary in their combination of inputs, and so the
profit maximising location for any firm will vary with respect to city size. Low
wage, labour intensive firms making limited use of business services-such as
textiles-will locate in smaller cities. At the other extreme, firms that use large
5 The only difference of substance, as opposed to treatment, that the present author has ever been able to
find between the models of Christaller (1933) and L6sch (1940) is that Christaller assumes that the spatially
distributed units of demand, the farmsteads, are continuously distributed through space, whereas, Lbsch as-
sumes that they are discretely distributed. This results in Christaller's market areas having an infinite set of
possible orientations, whereas, those of L6sch have only a finite set.
1350 P. Cheshire
quantities of high wage labour and business services and less space relative to
other inputs-such as headquarters functions, banking and financial complexes-
will locate in the largest cities. Cities are then viewed as coalitions of firms with
firms locating-either by a deliberate or a natural selection process-in the coali-
tion of firms (or "city") which minimises their costs. From this can be derived a
hierarchy of city sizes. Evans (1972) then explores the welfare implications of
the model and the impact of falling transport costs for commuting.
Henderson (1987) provides a review of models of urban systems in the con-
text of a general equilibrium model of an economy and synthesises much of
his own work to that date. His own model shares some features with that of
Evans (1972) but does not differentiate between "manufacturing" and "business
services", and allows residents to relocate between cities as they tradeoff the
diseconomies of city size against the higher wages generated in larger cities as
a result of economies of scale. Equilibrium is the outcome of real welfare being
equalised across all cities as a result of this process. If cities are too large for
optimal welfare then this creates the potential of development profit and an in-
centive to develop new cities. Cities of different sizes result from the assumption
that external economies are industry specific so that firms in a given industry
tend to cluster in cities of particular sizes as they, like residents, tradeoff the
diseconomies of larger cities against (in the case of firms) the lower operating
costs of clustering to obtain external economies from complementary activities.
"The size distribution of cities is not an accident of nature but is directly linked
to the regional composition of output and production conditions" (Henderson,
1988).
This model generates empirical questions: for example, whether, or not, par-
ticular industries tend to cluster in cities of a particular size. These are exten-
sively investigated-primarily in the context of the US and Brazil-in Henderson
(1988) and some supporting evidence is found. Nevertheless, the model does not
predict that the city size distribution will follow a Pareto distribution, far less that
the exponent will be close to 1. Nor does it plausibly explain the persistence of
a stable size distribution of cities over extended periods of time: in the US since
the nineteenth century for certain; in Japan for most of the twentieth century
(Eaton and Eckstein, 1997; Osada, 1997); and-with somewhat less stability-
in France (Guerin-Pace, 1995; Eaton and Eckstein, 1997) since the early years of
the nineteenth century, while in all three countries industrial mix and technology
changed radically.
Stochastic models could be said to take the opposite view to Henderson. The
size distribution of cities is an accident of nature. They come closer, however,
to producing a distribution of city sizes that conforms to a Pareto distribution,
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1351
although they run into trouble generating an exponent close to 1 and, less clearly,
in generating a stable Pareto distribution in which the individual components
change rank significantly as cities are observed to do. Various writers have pro-
posed a variant of the law of proportionate growth whereby the elements of a
system grow at a constant rate relative to the system as a whole (see, for example,
Nordbeck, 1971). To allow for changes in city rank, a stochastic component can
be incorporated in the growth rate of any given city. Vining (1977) showed that
a log linear distribution could result from a balancing of positive autocorrelation
of urban growth rates over time with a negative correlation between growth rates
and size. Simon (1955) suggested a model whereby urban growth occurred as a
series of discrete increments or discontinuous steps. Each increment might with a
fixed probability constitute a new city, or attach itself to an existing city with the
probability of it becoming a part of an existing city proportionate to the size of the
city in question. As Krugman (1996) points out, however, even this model does
not satisfactorily predict the observed outcome. Since the probability that a new
discrete increment to urban growth will form a new city must be close to zero,
and the exponent is close to 1, this appears to imply that the urban population
is infinity. This only does not actually happen because of an integer constraint.
There are no half cities twice as large as the largest, nor quarter cities four times
as large as the largest.
However, there would appear to be more fundamental objections to stochastic
models as explanations of the city size distribution. They are not only nihilistic,
in the sense that they entirely abstract from economic or social processes; but
they also amount to no more than saying, in the end, that the city size distribution
is, as it is, because it is. It would appear to be more realistic and no less helpful to
"explain" the distribution of city sizes in terms of people's preferences for living
in different sized cities. It would not be very hard to specify a distribution of
such preferences that explained the size distribution of cities; although since the
existence of such preferences would be as hard to validate as it would be easy to
specify, such an explanation would have a similar degree of elegant circularity.
The extent to which the literature briefly reviewed above has had as its explicit
focus explanations not just of why cities are of different sizes but why the distri-
bution of city sizes conforms to a particular form, varies from author to author.
Simon (1955), Beckmann (1958), Tinbergen (1968), Beckmann and McPherson
(1970) or Krugman (1996) were explicitly interested in finding explanations for
1352 P Cheshire
the "rank size rule". Henderson (1988) or Evans (1972) were interested in rather
less precise definitions of the "urban hierarchy".
Other authors, however, have been interested in what might be considered a
much simpler question: why are cities not all the same size? This question, in
turn, has been linked to investigations of the existence and forn of agglomer-
ation economies, subdivided between economies of localisation (internal to the
industry but external to the firm) and economies of urbanisation (external to both
the firm and the industry).6 A useful survey of the literature on agglomeration
economies and a comparison of empirical results is provided by Selting et al.
(1994). Many studies embodying measures of agglomeration economies, how-
ever, focus only tangentially on the issue of the distribution of city sizes. Some,
for example, such as Nakamura (1985), Henderson (1986) or Sveikauskas et al.
(1988), are really more relevant to industrial economics. They investigate why
productivity in given industries varies between cities of different sizes and try
to distinguish between the respective roles of urbanisation economies (resulting
from city size alone), and localisation economies (resulting from the size of a
given industry in a particular city).
These and other studies are faced with a significant problem of endogenous
variables. Finding suitable instruments is not straightforward. As Henderson
(1986) puts it, in the context of the model that he is estimating ".. . unfortunately
our situation is not a traditional one ... own industry employment, local wage
rates, taxes, and city population are jointly determined". He devises instruments,
but, given the spread of industries being examined, they are not industry spe-
cific and the 2SLS results are downplayed. In the OLS results on which the
conclusions are mainly based, localisation economies are significant across a
range of industries with urbanisation economies significant only in printing and
publishing in Brazil, and in nonmetallic minerals in the US. Nakamura (1985),
using a somewhat more elaborate set of instruments and with measures of capital
(unavailable to Henderson), finds evidence for both localisation economies-
particularly in heavy industries-and urbanisation economies in light industry.
Sveikauskas et al. (1988) focus exclusively on one industry-food processing.
Although restricting the scope of their study, they argue the gains more than
offset for this restriction. It allows the use of both a more appropriate form for
the production function-the translog form employed by Chan and Mountain
6 It has been argued that there are not just production economies of agglomeration but also agglomeration
economies in consumption and in the labour market. Alperovich (1993a) shows that not only is unemployment
in Israel negatively related to city size but so, too, is the incidence of prolonged spells of unemployment.
Equally, of course, it has been argued that there are diseconomies associated with increasing city size (see
Richardson, 1973).
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1353
(1983)-to distinguish the effects of technical progress from those of economies
of scale and the construction of instruments specifically appropriate to a par-
ticular industry (in this case the availability of materials inputs), as well as a
wider range of data. When applying the model estimated by Henderson (1986)
to their data, they largely replicate his results. Employing the translog produc-
tion function, and their more sophisticated instruments and data (including direct
measures of capital and output prices), however, produces sharply different re-
sults. Economies of localisation in food processing disappear but there is strong
evidence of significant urbanisation economies in the industry. Although highly
particular to their context, these results are valuable for the methodological care
and the attention to economic logic that they embody.
Others have employed a production function approach to measure productiv-
ity changes over time between size groups of urban areas, and between urban and
nonurban areas, to understand better the redistribution of manufacturing between
the largest, medium sized and nonurban areas that occurred particularly during
the 1970s. The work of Carlino (1985) and Moomaw (1985) falls into this cate-
gory. Fogerty and Garofalo (1988) straddle this literature and the investigation
of agglomeration economies in their own right. They find some of the most
persuasive evidence of any study of the importance not only of localisation and
urbanisation economies but also of the specific significance of density of employ-
ment in contributing independently to such agglomeration economies. As they
point out "since the two approaches [localisation and urbanisation economies] are
not mutually exclusive, our treatment of agglomeration economies incorporates
both". This possibility has been widely ignored by other writers.
Nearly all the studies attempting to estimate agglomeration economies have
used a production function approach. Some have been strictly cross-sectional on
the argument that it is only on that basis that agglomeration economies can be
reliably distinguished from technical progress. Others have been based on time
series data-an advantage of which may be that it is more reasonable to assume
that industrial composition is constant through time than invariant through space.
With very few exceptions-Nakamura (1985) who dealt with Japanese cities
is one such7 -the studies have been of US MSAs, although Henderson (1986,
1988) compared these with Brazilian cities. In almost all cases, evidence of
agglomeration economies has been found, although the estimates of the extent
of such economies and the comparative importance of economies of localisation
compared to those of urbanisation, has varied considerably. The more careful
7 BegoviC (1992), who analysed the cities of the former Yugoslavia, could be construed as another, but he
did not use a production function approach.
1354 P.Cheshire
studies seem to point towards the existence of both localisation and urbanisation
economies. The weight of evidence these studies provide is certainly consistent
with an equilibrium for urban systems in which there is significant variation in
the size of individual cities.
Apart from the normal problems associated with the estimation of produc-
tion functions, two particular problems dog efforts to estimate agglomeration
economies. The first is the simultaneity problem that makes it very difficult to
disentangle the separate contribution of city and industry size from the over-
all variation in productivity associated with increasing city size. The second
is that all econometric studies of agglomeration economies uniquely consider
manufacturing industry. Given the data problems involved in producing credible
estimates of production function parameters at the city level, this is understand-
able. Because of both the importance of the service sector in cities and its rapid
growth since the 1960s, however, the focus on manufacturing means that perhaps
three-quarters of the economic activity of urban economies in mature countries is
excluded from the analysis. In addition, more qualitative studies of certain service
activities, such as those of Dunning (1969) or Goddard (1973) of the City of
London or Saxenian (1993) of Silicon Valley, suggest agglomeration economies
are particularly important in some service sectors.
It is difficult to draw a precise line between work that should be classified
as attempting to explain city sizes and work that touches on the issue but is
primarily concerned with other issues. Nevertheless, if a line must be drawn, then
the studies briefly reviewed above probably fall into the latter category. So too,
must a number of other studies that have explored other aspects of agglomeration
economies. BegoviC (1992), for example, claims that the positive relationship
between city size and the diversity of its industrial structure reflects the existence
of economies of localisation which vary between, but not within, industries. The
increasing diversification of urban economies as city size increases is an idea
that goes a long way back in the literature. Clark (1945) appears to have been
the first to have taken a systematic interest. Although BegoviC (1992) does not
develop an explicit model, he does provide empirical evidence that not only did
economic diversification increase with city size in the former Yugoslavia but as
city size increased, the share of manufacturing in total employment significantly
decreased, while that of most services, especially business and technical services
and wholesaling, significantly increased. If business and technical services are in-
puts into manufacturing this suggests an analogy with the work of Sveikauskas et
al. (1988), who argued that in distinguishing between economies of localisation
and urbanisation in food processing it was necessary to offset for the availability
of raw materials. That is true of all sectors, so in principle in all attempts to dis-
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1355
tinguish between localisation and urbanisation economies one should standardise
for the availability of business and technical services. The availability of these, it
would appear, is itself a function of city size.
Apart from this empirically oriented work directed to sources of difference in
city sizes, there is also a thread of theoretical work investigating the reasons why
cities differ in size. Fujita (1989), Abdel-Rahman (1990), Abdel-Rahman and
Fujita (1990, 1993), Alperovitch (1995) and Fujita and Mori (1997) are some of
the more recent contributors. The earlier work of Abdel-Rahman and of Fujita are
variations on previous themes: the varying role of urbanisation economies and
economies of localisation in determining city sizes. Abdel-Rahman and Fujita
(1990), for example, use a monopolistic competition approach, including exter-
nal economies, and work out a partial equilibrium set of results. Abdel-Rahman
(1990), again within a partial equilibrium framework, considers a model with two
industries; one with economies of urbanisation and one with economies of local-
isation, generating two city types-one with urbanisation economies dominant
and one with localisation economies dominant. The underlying assumptions are a
L6schian plane with unfilled space, spaceless production in city Central Business
Districts (CBDs), but land consumed for residential purposes with consequent
commuting costs. These costs play a significant role in determining city size
and interact with economies of scale or scope. In equilibrium, the two industry
("diversified") city will only be larger than the single industry ("specialised") city
if at least one of the two industries exhibits decreasing returns.
Abdel-Rahman and Fujita (1993) develop a rather similar model within a
general equilibrium framework. Equilibrium conditions are derived and their
features are explored. There are two industries each with an element of fixed
labour costs (labour being the only factor), but that element is allowed to vary
if both industries are located within a single city because of potential economies
of joint production resulting from shared overhead costs, such as infrastructure
or a common input. Three possible city systems emerge: (1) only specialised
cities with only one good produced per city; (2) only diversified cities producing
both goods; or (3) both diversified and specialised cities (although the solution
is such that in this last case one of the two goods will always be produced in a
diversified city). The framework is one of competition, with developers maximis-
ing surplus but free entry ensuring that surplus is driven to zero. Total national
population is given and the familiar equilibrium conditions emerge with the value
of labour productivity and utility equal across households and cities. The model
is elegantly worked but the results are mainly unsurprising. Equilibrium city size
increases with the extent of "labour overhead" costs in the industries; the total
number of specialised cities increases as expenditure shares on the good pro-
1356 P. Cheshire
duced increases, total population increases and fixed costs of production ("labour
overhead") decreases. Equilibrium wages rise with city size as commuting costs
increase. In the outcome with both diversified and specialised cities, diversified
cities will always be larger than specialised cities, because if the fixed costs
associated with joint production were less than the fixed costs of either industry
individually, all production would occur in a diversified city.
The most novel contribution is the way in which the impact of variations in
commuting costs is explored. As commuting costs increase, the range of parame-
ter values within which diversified/specialised cities exist systematically varies.
As commuting costs decrease, diversified cities are more likely to emerge as the
cost of city size falls. It is shown that if commuting costs are set to zero-the
cities become spaceless-the results parallel those of the industrial organisation
literature and the conditions under which there will be combinations of single-
and multiproduct firms.
Fujita and Mori (1997) analyse a rather different problem: how new cities
come to be born in an expanding urban system and the form the resulting dis-
tribution of cities will take. The forces driving their model are growth and ag-
glomeration economies interacting with transport costs. As the system expands,
it becomes profitable at some point, for new cities to form so as to take advantage
of the increasing extent of uncontested market demand at the outer limits of the
current range of the initial city. Since the model is very much in the spirit of
central place analysis, it is perhaps not surprising that it can be shown that such
an expanding system would generate an urban system consistent with classical
central place theory.
The point of entry into the issue of city size adopted by Alperovitch (1995)
seems very different. It is to explore the relation between intraurban income
inequality and city size. This is done within the framework of a far less fully
developed general equilibrium model than that used by Abdel-Rahman and Fujita
(1993). The model builds on the results of Upton (1981) and Henderson (1988).
After reviewing the literature that empirically finds income inequality increasing
with city size in the US, but not varying significantly with city size in Canada,
Alperovich develops a model in which either outcome is possible but depends
on the consumption patterns of workers with varying endowments of human
capital. The model involves two types of industry, one producing traded goods
subject to industry but not firm level economies of scale, and the other produc-
ing nontraded goods such as housing and transport services (although it could
also, presumably, include urban amenities and cultural services). Competition
ensures that the price of traded goods is the same everywhere but, following
Henderson (1988), wages and the price of nontraded goods rise with city size.
Ch. 35: Trends in Sizes and Structures of Urban Areas 1357
Workers are then assumed to be divided between groups within which human
capital and consumption patterns are constant, but between which they vary.
Even though the proportionate representation of these groups may be constant
across cities, since the price of nontraded goods is higher the larger the city, if
consumption patterns vary systematically between groups, the equality of income
distribution-measured by the Gini coefficient-will also vary between cities. If
households with high incomes have relative preferences for nontraded goods (the
price of which rises with city size), but lower income households have relative
preferences for traded goods, inequality will rise with city size, and vice versa.
Thus, in a sense preferences enter into explanations of varying city sizes.
A very different approach to explaining variations in the size distribution of
urban systems is adopted by Ades and Glaeser (1995). This harks back to one
of the questions explored by Rosen and Resnick (1980) but avoids the technical
objection of Cameron (1990), by simply adopting the log of the population in
the main city in each country as their dependent variable. Although they do not
explore the relationship, the log of the population in the biggest city relative to
national urban population must be closely correlated with the estimate of a-the
Pareto exponent. It is also a direct measure of "metropolitanisation" or "prima-
cy". In an eclectic approach they combine insights from modern trade theory with
quantitative political science, commonsense and historical case studies. They
include in their cross-sectional analysis of 85 countries data for 1970-1985. 8
Their independent variables include a set similar to those used by Rosen and
Resnick (1980), such as the proportion of the country's labour force employed
in sectors other than agriculture and GDP per capita. They also include the log
of population in urban areas other than the main city, a dummy for whether, or
not, the main city is also the capital and the land area of the country. Drawing
on the argument advanced by Krugman and Livas (1992), they also include the
ratio of trade to GDP. Krugman and Livas argued that since international firms
sold on a more-or-less equal basis to both the hinterland and to the main city,
costs of imported goods would vary relatively less between the main city and
the hinterland than those of domestically produced goods. The latter, because
of economies of scale, agglomeration economies and transport costs, would be
sold at an advantage in the largest cities. Therefore, countries that engaged in
more trade protection would tend to have more dominant main cities relative
to their urban systems than countries with freer trade. Ades and Glaeser (1995)
then develop a series of arguments to the effect that the main city will tend to be
8 More detailed models are estimated for 70 and 50 countries for which a greater range of data are available.
These largely confirm the findings for the larger sample.
1358 P Cheshire
more dominant the more political instability there is in a country and the more
authoritarian is its regime. Factors cited include the comparative safety of life in
the main city compared to the outlying hinterland, the importance of proximity
to influence peddling (that increases with more dictatorial regimes), and the in-
centives for dictatorial regimes to favour their main cities at the expense of their
outlying hinterlands in both their tax treatment and supply of publicly provided
goods (circuses).
Strong empirical evidence is found to support their hypotheses. Main cities
tend to be 41% larger if they are also capitals; the size of the main city increases
with the total land area of a country; an increase in trade relative to GDP equal to
one standard deviation is associated with a 13% increase in the size of the main
city; countries with dictatorial regimes have main cities 45% larger than those
with nondictatorial regimes; and one coup per year is on average associated with
an increase in main city size of 2.4%. The more detailed analysis of the subsam-
ple of countries for which there was a wider range of data suggests that both the
level of tariff protection and trade as a proportion of GDP separately influence the
size of the main city in the expected way. They also support the hypothesis that
the development of the transport system is an influence on the degree of primacy
of a country's urban system. The size of the main city falls in countries as their
transport system becomes more highly developed. The implication appears to be
that lower transport costs are associated with a more dispersed urban population.
Because of problems of endogeneity and causality the authors re-estimate
some of their main models using instruments and testing the timing of events.
This last test suggests that if a country becomes a dictatorship then its main city
grows relative to the rest of its urban system. These relationships are then illus-
trated with case studies of archetypal main cities from different eras: classical
Rome, London in 1670, Edo in 1700, Buenos Aires in 1900 and Mexico City at
present.
The eclectic approach of Ades and Glaeser (1995) combines elements of eco-
nomic analysis, political theory and economic geography (old and "new"). There
is, however, also a rather different approach to investigating urban systems and
patterns of development. This has been adopted by some working within a broadly
defined economic framework, but is more associated with urban geographers.
It is not directed towards investigating the system of cities, in the sense of the
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1359
distribution of city sizes, but to explaining-or at least classifying-the patterns
of urban development.
As originally developed by Hall and Hay (1980), Klaassen et al. (1981), van
den Berg et al. (1982) or van den Berg (1987), it was conceived of as a theory
of the "stages of urban development". City regions were defined as functional
urban regions (FURs) using criteria developed from those used by Berry (1973)
to define his daily urban systems for the US. A set of core cities were defined
in terms of employment concentrations, including all contiguous local areas with
densities of jobs exceeding a given threshold; then a hinterland for each urban
core was defined on patterns of commuting. FURs were intended to coincide
with a city's field of economic influence to provide economically self-contained
city regions.
Hall and Hay (1980) argued that a given FUR passed through successive
stages of development. These stages were defined in terms of the relative rates of
population growth of the core city, its hinterland and the FUR as a whole. The
initial stages of urban development were seen as those associated with popula-
tion centralisation. During the first such stage the core city grew relative to the
hinterland, but both lost population in overall terms because of migration from
the FUR as a whole to the main metropolitan centre. Continued development
was seen as being characterised by centralisation within the FUR, first with the
core city gaining population in absolute terms, while both the hinterland and
the FUR overall continued to decline, then with the FUR and subsequently the
hinterland turning to positive growth but with the rate of growth in the core city
always exceeding that in the hinterland. These stages of centralisation were then
seen as giving way first to relative decentralisation, as the growth rate of the
hinterland exceeded that of the core, then to absolute decentralisation, and finally
to decentralisation and loss with overall decline in both core and hinterland, but
that decline being more rapid in the core.
This sequence was tested against data collected for all the FURs of Western
Europe for three periods, 1950-1960, 1960-1970 and 1970-1975. The analysis
showed that there was a tendency for FURs to exhibit such patterns of develop-
ment. Van den Berg et al. (1982) developed a very similar set of ideas, although
their "stages" were ordered differently, starting with the core city growing and
this growth more than compensating for population loss from the hinterland so
that the FUR overall showed population gain. They also argued that the pattern
was not so much one of stages that came to a halt with urban decline and decen-
tralisation, but one of cycles. Decline and decentralisation was then followed by
urban recentralisation and a new cycle begun. This idea was developed in more
detail by van den Berg (1987).
1360 P. Cheshire
Cheshire and Hay (1989) argued that the stages of urban development did
not constitute a theory, but rather a classificatory device. This revealed a set of
facts about the pattern of urban development that theory needed to explain. They
also updated the western European FUR analysis to 1980 and showed that the
patterns observed by Hall and Hay (1980) and van den Berg et al. (1982) had
continued. The urban system had continued to move towards decentralisation,
with decentralisation being observed in the largest city regions first and tending
to "move" from the cities of northern Europe to those of the south. They also
argued that these patterns did not reflect deterministic patterns inevitable in the
development of urban systems; but rather the particular balance of a set of so-
cioeconomic forces prevailing in western Europe during the time period studied.
These forces included the ongoing capitalisation of low productivity European
agriculture leading to a rural-urban flow of migrants, the size of which was
proportionate to the share of the agricultural labour force in the wider regional
economy. In the context of growing industrial employment, which halted around
1960 in northern Europe but continued through until the late 1970s in the south,
this produced urban growth and centralisation. A counteracting urban-rural flow
was associated with the forces driving the decentralisation of those sectors-
chiefly manufacturing-that handled goods in bulk; there was a further flow of
residential decentralisation, driven mainly by rising incomes and employment
decentralisation. As economies matured, which in a European context was equiv-
alent to moving northwards, these decentralising forces interacted with potential
flows of recentralisation driven by the relative growth of service activities. In
service sectors, relative price changes did not inevitably determine decentralised
locations. In addition, economic and demographic changes associated with ris-
ing female participation rates and falling average household sizes had played a
role. These changes meant that the advantages of decentralised locations relative
to centralised ones in northern European cities were, by about 1980, changing
as access to the labour market became relatively more significant and space
consumption relatively less significant in determining household locations.
Because the overall balance of these forces had altered by the mid-1980s, it
was argued that future patterns of European urban development would not nec-
essarily reflect the steady movement towards decentralisation that they had from
1950 to 1980. More varied patterns should be expected reflecting the particular
characteristics of individual cities. Cheshire (1995) then updated the analysis for
the FURs of western Europe to 1990. This showed that the regular trend towards
decentralisation had, indeed, changed during the 1980s, with much more varia-
tion and an increase in the number of city regions in northern Europe exhibit-
Ch. 35: Trends in Sizes and Structures of Urban Areas 1361
ing recentralisation. Where recentralisation had occurred it appeared possible to
relate this to the characteristics of the particular cities.
An alternative approach to understanding patterns of urban development was
developed by Andersson (1985) and Suarez-Villa (1988). Andersson (1985) ar-
gued that one should view urban development as an evolutionary process. He
identified four major stages in the modern era, each connected with a logistical
revolution that reduced key costs affecting transport, the costs of transactions, the
costs of coordinating production activities and, most recently, the costs of infor-
mation. These, he argued, were associated with a set of discontinuous extensions
of urban systems from the local, to the national, international and finally to the
emergence of a global urban hierarchy. Suarez-Villa (1988), in a wide ranging
but ultimately not wholly focussed review, attempted to link both the distribution
of city sizes (again measured by the Parato exponent, a) and their individual pat-
terns of centralisation and decentralisation to an evolutionary view of economic
development. These two papers, however, do not seem to have stimulated further
development of this approach.
so the investigation of changes in the location of jobs at this spatial level had a less
well articulated theoretical framework. Broadly, these two related bodies of work
can be thought of as the study of suburbanisation. One of its recurring questions
has been whether, in suburbanisation, the process is led by the decentralisation
of jobs, or the decentralisation of people (Steinnes, 1977, 1982).
The second main area of work is the study of changes in the distribution of
people and jobs between settlements of varying size, and between metropolitan
areas on the one hand, and nonmetropolitan areas-or rural ("ex-urban") areas-
on the other. Theoretically, this works tends to be more ad hoc than that on
suburbanisation but constitutes, nevertheless, a far larger volume of material.
It is the work on suburbanisation, however, that is reviewed first in this section.
As indicated above, a central question has been whether "people follow jobs" or
"jobs follow people". Perhaps the most complete investigation of this is provided
by Thurston and Yezer (1994). Building on the initial work of Bradford and
Kelejan (1973), Steinnes (1977, 1982), Cooke (1978, 1983) and Mills (1992),
they estimate a model using annual data in which the value of the population
density gradient at time t is determined by the annual percentage change in total
population, income and transport investment, the percentage of white population
in the central city in time t, by the value of the population and employment den-
sity gradients in the previous period and by a time trend. City specific effects are
eliminated by including a dummy for each city. l° Compared to previous studies,
the main missing independent variable is a measure of crime in the central city
(suburbs). This had been found to be significant in Bradford and Kelejan (1973),
Grubb (1982) and by Palumbo et al. (1990).
The strength of the Thurston and Yezer (1994) study, however, is that they
used as their dependent variable what is probably the theoretically most appropri-
ate measure-the value of the population density gradient; they employed annual
data, and had a much more detailed and exhaustive set of employment sectors.
They also had a carefully evaluated measure of transport investment. Not being
a direct measure of the theoretically appropriate variable, travel costs, however,
this still had problems. Although it had the expected sign (unlike in a number
of earlier studies), it was not significantly associated with the rate of population
suburbanisation. The advantage of using annual data is that it allows much more
precise measurement of lags and potentially gives greater insight into causality,
where the response in one sector to changes in another is rapid. The advantages
of greater and exhaustive sectoral disaggregation of employment are consider-
able. Location theory predicts that intraurban locational requirements will vary
10 Rather curiously Thurston and Yezer (1994) dismiss these as "of no particular importance".
Ch. 35: Trends in Sizes and Structures of Urban Areas 1363
between industries, as will the locational response to changes in transport tech-
nology/costs (see, for example Anas and Moses, 1978). Including an exhaustive
(or nearly-agriculture and extractive industry employment was excluded) set of
sectors greatly reduces the problem of omitted variables bias.
The findings of Thurston and Yezer (1994) are broadly supportive of theory
and much more successful in identifying the relationship between population
and employment suburbanisation than was the case with previous studies. They
found that in US cities rising incomes were significantly associated with popula-
tion suburbanisation, but that population growth and its ethnic composition was
not significantly related to changes in the population density gradient. They also
found evidence of significant relationships between employment and residential
segregation. There was only evidence supportive of the "jobs follow people"
view, however, with respect to employment in retail and services. Evidence for
the interrelationship between decentralisation of employment in construction and
transport, communications and public utilities, supported the interpretation that
investment in suburban infrastructure facilitated subsequent population subur-
banisation. In manufacturing, the results supported the earlier and apparently
anomalous results of Steinnes (1977). Suburbanisation of manufacturing em-
ployment was significantly related to slower population suburbanisation. This
is consistent with Steinnes's original interpretation: that manufacturing industry
generates negative externalities from which people tend to remove themselves
as their incomes rise, since an alternative way of representing the result is that
increased concentration of manufacturing in the central city increases the rate of
population suburbanisation.
The second body of work referred to above probably originates with Beale
(1975), although it was given an international dimension and much wider cur-
rency with the work of Berry (1976), Vining and Strauss (1977), Leven (1978),
Vining and Kontuly (1978) and Vining and Yang (1979). The literature consists of
a large number of empirically oriented studies that try to establish the facts; and a
far smaller body of more theoretically oriented work attempting to explain these
facts. In brief, the "facts" are that, starting sometime in the 1960s, a process began
of population dispersal from the large metropolitan areas. This was matched by
a reversal of previous patterns of depopulation in rural areas. In the 1970s, for
example, there was in Britain an almost perfect positive relationship between
measures of how rural a region was and its rate of population growth (Champion,
1989). This pattern went far beyond suburbanisation or decentralisation and was
associated with a strong growth of smaller cities and "ex-urban" communities.
Although originally noted in the US, it appeared also in other mature industrial
countries at about the same time (Leven, 1978; Vining and Kontuly, 1978).
1364 P Cheshire
It was realised quite early that there should be some connection between the
shift in population and employment towards smaller cities and nonmetropolitan
areas and changes in the spatial structure of costs (Anas and Moses, 1978) and/or
productivity (Vining and Kontuly, 1977). Efforts to estimate relative changes in
productivity were problematic, however, partly for lack of time series data on
capital stock by area type. This hindered the development of credible estimates
of total factor productivity growth. Such data were constructed for US cities and
states by Fogarty and Garofalo (1982) and Garofalo and Malhotra (1985), and
were used in studies of productivity growth in large MSAs between 1957 and
1977 in Fogarty and Garofalo (1988) and for 1959 to 1982 in Beeson (1990).
These studies all related only to manufacturing.
As discussed above, Fogarty and Garofalo (1988) found strong evidence sup-
porting the existence of both economies of scale internal to the industry (locali-
sation economies), and of economies of scale related to city size (economies of
urbanisation). In addition, they found evidence of specific economies of employ-
ment density as measured by the manufacturing employment density gradient and
the density of manufacturing employment in the central city. Together with this
evidence of various agglomeration economies, they also established a significant
negative relationship between city age and its manufacturing productivity as well
as significant regional variations. They did this on the basis of estimating a Vari-
able Elasticity of Substitution production function for manufacturing for a pooled
time series dataset across the 13 MSAs, for which they had both capital stock data
and manufacturing employment density gradients (from Mills, 1972). This fitted
the data well, with all coefficients having the expected signs. Since the functional
form estimated for both city size and employment density was quadratic, it was
possible to estimate values of the employment density gradient (-0.5) and city
size (population 2.9 million) which were associated with maximum total factor
productivity in manufacturing.
These strong results on agglomeration economies compared to previous stud-
ies (such as, for example, Moomaw, 1985) did not provide an explanation of
either decentralisation or the shift of manufacturing employment and popula-
tion to smaller cities. Indeed, they suggested these were seemingly "irrational"
and associated with a slowdown of productivity growth. As they acknowledge,
however, maintenance of productivity advantages in large cities could have been
offset by growing cost disadvantages in private inputs or public goods, or in
both. Their analysis related essentially to physical relationships between inputs
and outputs. Indeed, since manufacturing is a very heterogeneous sector, there
could have been composition effects within it with those subsectors enjoying
the strongest productivity advantages in large cities remaining there as other
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1365
investment to the areas outside the large MSAs partly caused their increased pro-
ductivity advantages because of the composition effect in manufacturing activity
that it induced. In that context, it may be worth noting that the timing of the
greatest gains in relative productivity in the large MSAs coincided with their
greatest period of relative loss of factor inputs.
Because of the limitations on data availability, the work done on recent trends in
cities is, not surprisingly, more descriptive than analytical. Demographic data-
especially for the small areas necessary to investigate changes at the urban level-
are the most widely available, up-to-date and straightforward to interpret. Since
recent trends necessarily rely on up-to-date information their study tends to rely
on demographic data. Except where noted, the work discussed in this section uses
straightforward tabulations, or variants of the devices employed by researchers
such as Vining and Strauss (1977) or Hall and Hay (1980). Nevertheless, analysis
of trends in sizes and the distribution of the urban populations to 1990 or 1991 are
available for western Europe (Cheshire, 1995), Australia (Maher, 1993), Canada
and the US (Bourne, 1995); and to 1994 for the US (Long and Nucci, 1995).
Although, as Bournme (1995) suggests, it represents a diverse landscape of
change, some patterns do appear. In mature economies, suburbanisation was
almost universal from the 1950s, but in the US and in some of the countries
of northern Europe, such as the UK, not only did decentralisation accelerate in
the 1970s but it took a new form. Tales of the revival of rural areas (starting
with Beale, 1975) may have turned out to be oversimplifications or related to an
ephemeral phenomenon. Much of the growth in nonmetropolitan areas in the US
appeared, on close examination (Vining and Strauss, 1977), to have been in rural
counties contiguous t,) metropolitan areas and, therefore, more of an extension
of suburbanisation to c:x-urban fringe counties. It was not so much a question
of a rural revival as ofl "the cities moving to the countryside" (Leven, 1978).
Nevertheless, even though it may not have lasted, there were remote rural areas
that experienced growth during the early 1970s for the first time in the twenti-
eth century. The growth of population in US nonmetropolitan areas combined
increased from 2.5% in the 1960s to 14.4% in the 1970s while, between the same
two decades, growth of population for the US as a whole declined from 13.4
to 11.4%) (Bourne, 1995). Within metropolitan areas, there was a substantial
dispersion of people to less densely populated counties (Long and Nucci, 1995)
and the central cities compared to the suburbs did markedly worse, not only in
Ch. 35: Trends in Sizes and Structures of Urban Areas 1367
terms of population growth but also on economic indicators such as employment,
migration and relative property price changes (Linneman and Summers, 1993;
Jensen and Leven, 1997). There was a sense in which, not only in the US, the
1970s represented a "clean break with the past".
Breaks, however, do not necessarily continue. The first commentators to note
some urban population revival were Cheshire (1987) in Europe, and Long and
DeAre (1988) and Frey (1993) in the US. In some of the large FURs of north-
ern Europe, such as London (Champion and Congdon, 1987) and Copenhagen
(Matthiessen, 1983), the maximum rate of outflow of migrants from the core
cities was observed in 1971 or 1972. Although loss continued, the rate of loss
from migration fell continuously until about 1980, and positive migration gain
was observed for a few years in the early 1980s. Dangschat (1993) reports that the
population of inner West Berlin, which had declined continuously from 1950, be-
gan to increase in 1985 and continued to do so until and including 1991. Annual
data on changes in the concentration of population for three categories of area
in the US-all counties, Bureau of Economic Analysis (BEA) Economic Areas,
and BEA Metropolitan Economic Areas (broadly identifiable with FURs)-show
similar patterns (Long and Nucci, 1995). From the first year for which the nec-
essary data became available, 1970, concentration was falling in all three types
of area. The county based index was the earliest to show a change of trend. It
began to move towards concentration in the second half of the 1970s. Positive
concentration began to occur between 1982 and 1983. This was quickly followed
by positive concentration appearing in the indices for the other two types of areas.
Concentration in all three, however, peaked in the mid- to late 1980s. By 1990 to
1991, all three were showing that population deconcentration was present again.
Somewhat more analytical approaches-but including data only to 1990 or
1991-have been adopted by Cheshire (1995) and Jensen and Leven (1997).
Cheshire showed that the steady movement towards decentralisation in the largest
FURs of western Europe (those with populations of more than one-third of a
million) that had been observed gathering pace in each decade from 1950 to 1980,
partially reversed in the 1980s. The percentage of core cities gaining population
increased between 1975-1981 and 1981-1991 from 22 to 47% in northern Eu-
rope and from 40 to 48% in France and northern Italy. Decentralisation continued
to spread to the cities of the south of Europe.
The change in trends in European cities was more in terms of relative changes
than of absolute ones. Looking at patterns of "turnaround"-that is, the differ-
ence in rates of population change between core cities and hinterlands in the
1980s compared to the 1970s-there were many cases where a strong turnaround
was observed but the FUR continued to lose population overall. Glasgow was
1368 P. Cheshire
such a city. In the 1980s, however, the rate of loss was much higher from the
hinterland or suburbs than from the city core. Nevertheless, those cities in north-
ern Europe with the strongest turnaround tended to share characteristics; they
were significantly more likely to be medium sized than the very largest cities
(defined as those FURs-with a core city exceeding 200,000). Cities with stronger
amenities, historic cores and strong service sectors were also more strongly rep-
resented. In the UK, for example, Oxford, Cambridge and Canterbury were the
three cities with the most significant turnaround. Overall, the message appeared
to be that urban recovery and recentralisation was partial and selective and likely
to be related to the growth of service employment, falling household sizes and
increased labour force participation rates.
Jensen and Leven (1997) show that on the basis of rather strong assumptions,
relative rates of change in the measure of quality of life (QOL), developed by
Blomquist et al. (1988), can be calculated from data available for central cities
and for the metropolitan area as a whole. They derived a simple model. This im-
plied that if house prices in the core are rising, or wages are falling relative to the
metropolitan area as a whole, then the QOL of the central city is relatively rising:
and when net migration into, or labour supply in, the central city is rising relative
to the metropolitan area as a whole, similarly the QOL of the central city is rising
relative to the metropolitan area. The advantage of this approach is that since it
rests on revealed preference, it can be applied in the absence of information on
amenities contributing directly to QOL. Not only are there strong assumptions,
however, but not all the appropriate data were available. Particularly, wages for
homogeneous labour were unavailable and they had to proxy this with family
incomes, offset with a crude measure of changes in human capital. They then
calculate the values for these changes in relative QOL for the 25 largest MSAs
combined and for each decade from 1950 to 1990.
The results are quite strong and consistent. All indicators except the wage
proxy point in the same direction and show statistically significant changes in the
relative position of central cities compared to the metropolitan areas. Between the
1970s and 1980s there was a sharp change of trend in central city house prices
compared to those in the metropolitan areas and rather less sharp changes in
trend in the other indicators. The Jensen and Leven (1997) analysis supports the
view that there was significant improvement in the central cities of the 25 largest
US MSAs during the 1980s compared to the 1970s. Their tests of significance are
invalid for individual MSAs, although the data suggests that there was substantial
variation.
The general conclusion of this section is that despite the fact that the expe-
rience of individual cities seems to have become more varied, internationally,
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1369
References
Abdel-Rahman, H.M. (1990), "Agglomeration economies, types and sizes of cities", Journal of Urban
Economics 27:23-45.
Abdel-Rahman, H.M. and M. Fujita (1990), "Product variety, Marshallian externalities and city size", Journal
of Regional Science 30:165-183.
Abdel-Rahman, H.M. and M. Fujita (1993), "Specialisation and diversification in a system of cities", Journal
of Urban Economics 33:189-222.
Ades, A.F. and E.L. Glaeser (1995), "Trade and circuses: explaining urban giants", The Quarterly Journal of
Economics CX:195-227.
Allen, G.R.P. (1954), "The 'courbe des populations': a further analysis", Bulletin of the Oxford University
Institute of Statistics 16:179-189.
Alonso, W. (1964), Location and Land Use: Toward a General Theory of Land Rent (Harvard University Press,
Cambridge).
Alperovich, G.A. (1985), "Urban spatial structure and income: new estimates", Journal of Urban Economics
18:278-290.
Alperovich, G.A. (1988), "A new testing procedure of the rank size distribution", Journal of Urban Economics
23:251-259.
Alperovich, G.A. (1989), "The distribution of city size: a sensitivity analysis", Journal of Urban Economics
25:93-102.
Alperovich, G.A. (1993a), "City size and the rate and duration of unemployment: evidence from Israeli data",
Journal of Urban Economics 34: 347-357.
Alperovich, G.A. (1993b), "An explanatory model of city-size distribution: evidence from cross-country data",
Urban Studies 30:1591-1601.
Alperovich, G.A. (1995), "The relationship between income inequality and city size: a general equilibrium
model of an open system of cities approach", Urban Studies 32:853-862.
Anas, A. and L. Moses (1978), "Transportation and land use in the mature metropolis", in: C.L. Leven, ed.,
The Mature Metropolis (D.C. Heath, Lexington).
Andersson, A.E. (1985), "The four logistical revolutions", Presidential address to the 32nd North American
Congress, Regional Science Association.
Auerbach, F. (1913), "Das Gesetz der Bevblkerungskoncentration", Petermanns Geographische Mitteilungen
59:74-76.
Beale, C.L. (1975), The Revival of Population Growth in Non-Metropolitan America (US Department of
Agriculture, Washington DC), ERS 605.
1370 P Cheshire
Beckmann, M.J. (1958), 'City hierarchies and distribution of city size", Economic Development and Cultural
Change VI:243-248.
Beckmann, M.J. and J.C. McPherson (1970), "City size distribution in a central place hierarchy: an alternative
approach", Journal of Regional Science 10:25-33.
Beeson, P.E. (1990), "Sources of the decline of manufacturing in large metropolitan areas", Journal of Urban
Economics 28:71-86.
Begovi, B. (1992), "Industrial diversification and city size: the case of Yugoslavia", Urban Studies 29:77-88.
Berry, B.J.L. (1961), "City size distributions and economic development", Economic Development and
Cultural Change 9:573-587.
Berry, B.J.L. (1973), Growth Centres in the American Urban System (Ballinger, Cambridge, MA).
Berry, B.J.L. (1976), Urbanisation and Counterurbanisation (Sage, Beverly Hills).
Berry., B.J.L. and W. Garrison (1958), "Altemative explanations of urban rank-size relationships", Annals of
American Geographers 48:83-91.
Blomquist, G.C., M.C. Berger and J.P. Hoehn (1988), "New estimates of the quality of life in urban areas",
American Economic Review 78:89-107.
Bourne, L.S. (1995), Urban Growth and Population Redistribution in North America: A Diverse and Unequal
Landscape, Major Report 32 (Centre for Urban and Community Studies, University of Toronto).
Box, G.E.P. and D.R. Cox (1964), "An analysis of transformations", Journal of the Royal Statistical Society,
Series B 26:211-243.
Bradford, D. and Kelejian, H. (1973), "An econometric model of flight to the suburbs", Journal of Political
Economy 81:566-589.
Cameron, T.A. (1990), "One stage structural models to explain city size", Journal of Urban Economics 27:294-
307.
Carlino, G.A. (1985), "Declining city productivity and the growth of rural regions: a test of alternative
explanations", Journal of Urban Economics 18:11-27.
Champion, A.G. (1989), Counterurbanisation: The Changing Pace and Nature of Population Deconcentration
(Edward Arnold, London).
Champion, A.G. and P. Congdon (1987), "An analysis of the recovery of London's population change rate",
Built Environment 13:193-211.
Chan, M.L.W. and D.C. Mountain (1983), "Economies of scale and the tomquist discrete measure of
productivity growth", Review of Economics and Statistics 65:663-667.
Cheshire, P.C. (1987), "Economic factors in urban change: European prospects", Discussion Papers in Urban
and Regional Economics, Series C, No. 30 (University of Reading).
Cheshire, P.C. (1995), "A new phase of urban development in Western Europe? The evidence for the 1980s",
Urban Studies 32:1045-1063.
Cheshire, P.C. and G. Gorla (1987), "The effects of Western European integration on Europe's urban system",
paper given to the XXVII European Congress of Regional Science, Athens.
Cheshire, P.C. and D.G. Hay (1989), Urban Problems in Western Europe: an Economic Analysis (Unwin
Hyman, London).
Christaller, W. (1933), Die zentralen orte in Sfiddeutschland, Jena: Verlag Gustav Fischer: published in English
(1966) as The Central Places of Southern Germany (Prentice-Hall, Englewood Cliffs, NJ).
Clark, C. (1945), "The economic functions of a city in relation to its size", Econometrica 13:97-113.
Clark, C. (1951), "Urban population densities", Journal of the Royal Statistical Society, Series A 114:490-496.
Cooke, T. (1978), "Causality reconsidered: a note", Journal of Urban Economics 5:538-542.
Cooke, T.W. (1983), "Testing a model of intraurban firm relocation", Journal of Urban Economics 13:257-282.
Dangschat, J.S. (1993), "Berlin and the German system of cities", Urban Studies 30:1025-1051.
Denison, E.E (1979), Accounting for Slower Economic Growth: The United States in the 1970s (The
Brookings Institution, Washington DC).
Dunning, J.H. (1969) "The city of London: a case study in urban economics", Town Planning Review 40:207-
232.
Eaton, J. and Z. Eckstein (1997), "Cities and growth: theory and evidence from France and Japan", Regional
Science and Urban Economics 27:443-474.
Evans, A.W. (1972), "The pure theory of city size in an industrial economy", Urban Studies 9:49-77.
Ch. 35: Trends in Sizes and Structures of Urban Areas 1371
Lbsch, A. (1940), Die raumliche ordnung der Wirtschaft, Jena: Gustav Fischer, Published in English as L6sch,
A. (1954), The Economics of Location (Yale University Press, New Haven).
Madden, C.J. (1956), "Some indicators of stability in the growth of cities in the United States", Economic
Development and Cultural Change 4:236-452.
Maher, C. (1993), "Recent trends in Australian urban development: locational change and the policy quandary",
Urban Studies 30:797-825.
Malecki, E.J. (1980), "Growth and change in the analysis of rank-size distributions: empirical findings",
Environment and Planning B 12:41-52.
Margo, R.A. (1992), "Explaining the postwar suburbanisation of population in the United States: the role of
income", Journal of Urban Economics 31:301-310.
Matthiessen, C.W. (1983) 'Settlement change in Denmark", in: C.W. Matthiessen, ed., Urban Policy and Urban
Development in the 80s; Danish Experience in a European Context (University of Copenhagen).
Mills, E.S. (1972), Studies in the Structure of the Urban Economy (John Hopkins University Press, Baltimore).
Mills, E.S. (1992), "The measurement and determinants of suburbanisation", Journal of Urban Economics
32:377-387.
Moomaw, R.L. (1985), "Firm Location and City Size: Reduced Productivity Advantages as a Factor in the
Decline of Manufacturing in Urban Areas", Journal of Urban Economics 17:73-89.
Muth, R.F. (1969), Cities and Housing (Chicago University Press, Chicago).
Nakamura, R. (1985), "Agglomeration economies in urban manufacturing industries: a case of Japanese cities",
Journal of Urban Economics 17:108-124.
Nordbeck, S. (1971), "Urban Allometric Growth", Geografiska Annaler 53B:54-67.
Osada, S. (1997), "The development of the Japanese urban system, 1970 to 1990", paper presented to Economic
Geography Graduate Workshop (London School of Economics and Political Science).
Palumbo, G., S. Sacks and M. Wasylenko (1990), "Population Decentralisation within metropolitan areas:
1970-1980", Journal of Urban Economics 27:151-167.
Parr, J.B. (1985), "A note on the size distribution of cities over time", Journal of Urban Economics 18:199-212.
Richardson, H.W. (1973) The Economics of Urban Size (Saxon House, Faroborough).
Rosen, K.T. and M. Resnick (1980), "The size distribution of cities: an examination of the Pareto law and
primacy", Journal of Urban Economics 8:165-186.
Saxenian, A.L. (1993), Regional Networks: Industrial Adaptation in Silicon Valley and Route 128 (Harvard
University Press, Cambridge, MA).
Selting, A., C. Allanach and S. Loveridge (1994), "The role of agglomeration economies in firm location: a
review of the literature", Staff papers (Department of Agriculture and Applied Economics, University of
Minnosota).
Simon, H. (1955), "On a class of skew distribution functions", Biometrika 42:425-440.
Singer, H.W. (1936), "The 'courbe des populations': a parallel to Pareto's law", Economic Journal 46:254-263.
Steinnes, D. (1977), "Causality and intraurban location", Journal of Urban Economics 4:69-79.
Steinnes, D. (1982), "Do 'people follow jobs' or 'jobs follow people'?: a causality issue in urban economics",
Urban Studies 19:187-192.
Suarez-Villa, L. (1988), "Metropolitan evolution, sectoral economic change, and the city size distribution",
Urban Studies 25:1-20.
Sveikauskas, L., J. Gowdy and M. Funk (1988), "Urban productivity: city size or industry size", Journal of
Regional Science 28:185-202.
Thurston, L. and A.M.J. Yezer (1994), "Causality in the suburbanisation of population and employment",
Journal of Urban Economics 35:105-118.
Tinbergen, J. (1968), "The hierarchy model of the size distribution of centres", Papers and Proceedings of the
Regional Science Association 20:65-68.
Upton, C. (1981), "An equilibrium model of city sizes", Journal of Urban Economics 15:15-36.
Van den Berg, L. (1987), Urban Systems in a Dynamic Society (Gower, Aldershot, UK).
Van den Berg, L., R. Drewett, L.H. Klaassen, A. Rossi and C.H.T. Vijverberg (1982), Urban Europe: A Study
of Growth and Decline (Pergamon Press, Oxford).
Vapnarsky, C.A. (1969), "On rank size distributions of cities: an ecological approach", Economic Development
and Cultural Change 17:584-595.
Ch. 35: Trends in Sizes and Structures of UrbanAreas 1373
Vining, R. (1955), "A description of certain spatial aspects of an economic system", Economic Development
and Cultural Change 3:147-195.
Vining, D.R. (1977), "The rank size rule in the absence of growth", Journal of Urban Economics 4:15-29.
Vining, D.R. and T. Kontuly (1977), "Increasing returns to city size in the face of an impending decline in the
size of large cities: which is the bogus fact?", Environment and Planning A 9:59-62.
Vining, D.R. and T. Kontuly (1978), "Population dispersal from major metropolitan regions: an international
comparison", International Regional Science Review 3:49-73.
Vining, D.R. and A. Strauss (1977), "A demonstration that the current deconcentration of population in the
United States is a clean break with the past", Environment and Planning A 9:751-758.
Vining, D.R. and C.H. Yang (1979), "Population dispersal from core regions: a description and tentative
explanation of the patterns of 17 countries", in: T. Kawashima, ed., Urbanisation Processes: Experiences
of Eastern and Western Countries, vol. 2 (Pergamon Press, Oxford).
Wheaton, W.C. and H. Shishado (1981), "Urban concentration, agglomeration economies and the level of
economic development", Economic Development and Cultural Change 30:17-30.
Wingo, L. (1961), Transportation and Urban Land (Resources for the Future, Washington DC).
Zipf, G.K. (1949), Human Behaviour and the Principle of Least Effort (Addison-Wesley, Reading, MA).
Chapter36
University of Michigan
Contents
1. Introduction 1376
2. Why do firms suburbanize? 1378
3. Theoretical models of decentralized cities 1384
3.1. Models with endogenously determined employment location 1385
3.2. Models with exogenously determined employment loations 1388
4. Rent and wage gradients in decentralized urban areas 1393
5. Empirical evidence concerning urban wage gradients 1397
6. Suburbanization: Do jobs follow workers or workers follow jobs? 1401
7. Are commuting patterns in decentralized cities "wasteful"? 1404
References 1410
* I am grateful for comments from Charles Becker, Paul Cheshire, Ken Small, John Yinger and Stephen
Ross. I am also grateful for research support from the Economic Policy Research Unit, Copenhagen Business
School, Copenhagen, Denmark, where I was a visitor while writing this chapter.
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P. Cheshire
() 1999 Elsevier Science B. V All rights reserved.
1375
1376 M.J. White
Abstract
1. Introduction
1 Mills also finds similar results for a smaller sample of US metropolitan areas over a longer time period.
1378 M.J. White
The original urban models assumed that all jobs were located at the CBD. This
assumption made for tractability-an important consideration. It was also ap-
pealing since, historically, CBDs tended to develop at a transportation node,
usually a port. Firms located at the CBD minimized the cost of goods trans-
portation since doing so was valuable because workers could walk but goods
could not. A CBD location also allowed firms access to power and utilities that
originally were only available near the CBD.
Now consider the basic factors that cause employment in cities to move out
of the CBD. Suppose a hypothetical firm is located at the CBD but is considering
moving to a more suburban location. Firms that consider moving out of the CBD
face tradeoffs since some costs rise while others fall. As long as the firm's move
causes its workers' commuting distances to fall, then workers save on commuting
costs and the firm can capture some of this savings in the form of lower wages.
Second, the price of land declines at a decreasing rate with the distance from
the CBD. Therefore, firms that move out of the CBD benefit from lower land
costs which allow them to trade capital for land, i.e., they occupy low, horizontal
buildings instead of tall, vertical buildings. Third, goods transportation costs may
decline since the firm avoids the traffic congestion of the CBD. Fourth, loss of
agglomeration economies at the CBD may cause the finns' productivity to fall.
Finally, other costs faced by firms may also change when they move to the sub-
urbs. Costs related to information technology-which are changing rapidly-are
an example. Consider these factors individually.
First, firms have an incentive to suburbanize because they can pay lower
wages, which workers are willing to accept because they commute less. Since
wages are the largest single cost for many firms, this is likely to be an important
consideration. However, the extent to which suburbanizing allows firms to pay
lower wages depends on labor demand and supply. Suppose all firms are initially
located at the CBD and workers commute along straight lines connecting their
homes and their workplaces.2 An arbitrary firm X moves from the CBD to a
new location five miles south of the CBD, shown as point A in Fig. 1. All other
firms still remain at the CBD. At its new location, suppose firm X hires only
workers who live further out than the firm in the same direction away from the
CBD, i.e., along the line segment Aa. Therefore, all the firm's workers save 10
miles of commuting per day by shifting from CBD jobs to jobs at firm X. If
2 A more realistic model would take the specifics of the transportation network into account, so that workers
would commute from their homes to their workplaces along existing road or rail networks. This modification
would not change the general results discussed here.
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1379
a
Fig. 1. Commuting regions when a firm moves from the CBD to point A.
the daily wage at the CBD is w* and the cost of commuting per mile is t, then
workers will be willing to work at A for a daily wage of w* - 10t and the firm
can save 10t per worker per day by moving. However, only workers who live
along line segment Aa will be willing to work at A for this wage; all others
prefer to continue working at the CBD. Now suppose firm X wishes to hire more
workers than are willing to work for it at the wage w* - 10t. If it raises its wage
above w* - 10t, then its commuting region will expand from the line segment
Aa in Fig. 1 to a larger region such as that enclosed by the line bbb. At the
higher wage, some workers commute to firm X from homes that are not on the
same ray from the CBD as firm X (i.e., they commute circumferentially) and a
few workers outcommute. 3 As firm X continues to raise its wage, its commuting
region continues to get larger. Now suppose the suburban firm pays the same
wage w* as CBD firms. Then its commuting region will be the area below the
3 Commuting is still along straight lines connecting workers' residences and their workplaces, but the
commuting routes are no longer radial. The outer boundary of the urban area also bulges outward in the area
closest to the suburban firm, but this effect is not shown in Fig. 1.
1380 M.J White
horizontal line ccc, which bisects the line segment connecting the CBD and point
A. If firm X pays a higher wage than w*, then its commuting region will be the
area below an upward curving line such as ddd. However, even if firm X pays
higher than the CBD wage w*, its commuting region will be smaller than the
CBD firms and restricted to workers who live in the southern region of the urban
area. Thus, firms that move out of the CBD can pay lower wages only if their
commuting regions shrink from covering the entire urban area to covering just
a region around their suburban sites. If suburban firms are relatively large and
have relatively high demand for labor, then they may have to pay wages as high
or higher than CBD firms. 4
Now suppose an additional firm Y moves out of the CBD. As long as there
are no agglomeration economies outside of the CBD, firm Y has an incentive
to locate north rather than south of the CBD. This is because if firm Y locates
north of the CBD, then it will gain from the same wage reduction (discussed
above) in connection with firm X. But if firm Y locates south of the CBD, then it
must compete with firm X for labor and both firms will have to pay higher wages.
Thus, as firms suburbanize, they have an incentive to locate in different directions
around the CBD and, in particular, to avoid suburban regions that already have
high concentrations of firms.
Now consider the firms' gain from suburbanizing due to the lower cost of
suburban land. The extent to which the firms' land costs fall in the suburbs also
depends on the workers' commuting patterns. Suppose we change the previous
model by assuming that all workers commute via a fixed rail network, that con-
sists of radial lines leading out from the CBD in different directions. Firm X
again plans to move out of the CBD. Only sites located near public transit stations
would be plausible suburban locations for firm X, since workers must be able to
walk to work from the station. But this means that the supply of suburban sites
suitable for use by firm X is limited to sites close to transit stations and the price
of these sites is high because their accessibility makes them valuable for high
density residential use. These factors reduce firm X's gain from suburbanizing.
In addition, the gain to firm X from reduced wages at suburban locations is also
small in this case, because only workers who live along the same radial transit
line as firm X's suburban location have shorter commuting journeys when they
commute to firm X rather than to the CBD. (If workers must travel to the CBD
along one transit route and then outcommute to firm X along another transit
route, then they will be unwilling to work at firm X if it pays less than the wage
5 See Capozza (1976), O'Hara (1977), and Ogawa and Fujita (1980).
1382 M.J. White
Suppose we compare two urban areas having a different number of firms. Then
the smaller urban area will have an advantage over the larger because total trans-
actions costs among a smaller number of firms are lower. Thus, larger urban areas
are predicted to have lower rather than higher agglomeration economies. But this
goes against the notion that higher agglomeration economies are responsible for
the existence of larger cities. An alternative approach, used by Henderson (1977)
and Straszheim (1984) in the urban context and also commonly used in other
fields, assumes that production is characterized by external increasing returns as
the number of firms or the number of jobs in the city rises.6 This gives larger
cities an advantage over smaller ones which offsets their disadvantage of higher
aggregate commuting costs. However, this approach has the drawback that the
level of agglomeration economies is the same all over the urban area, regardless
of where firms locate. It might be useful to combine these two approaches, since
the latter represents the gain from more firms being present in an urban area,
while the former represents the cost of capturing these gains through interactions
among firms.
What about the issue of how agglomeration economies vary within an urban
area? A variation of the external increasing returns approach, used by Wieand
(1987), makes agglomeration economies depend on the number of firms located
at particular employment sites. Thus, firms located at the CBD benefit from a
high level of agglomeration economies, but finns located at a suburban em-
ployment subcenter that is smaller than the CBD benefit from a lower level of
agglomeration economies. A more general version of this approach is used by
Fujita and Ogawa (1982), who allow the level of agglomeration economies to
vary continuously over space, depending on the density level of firms at each
location. These approaches are useful in exploring what type of firm location
pattern is efficient in a decentralized urban area.
It should also be noted that agglomeration economies may differ for different
types of firms. For example, computer firms benefit from locating in the "Sili-
con Valley" area of San Francisco/San Jose because these firms can hire skilled
computer engineers without having to bear the costs of their training. But the
computer firms themselves are stretched along at least a 20-mile region, which
suggests that they do not need to locate close together to benefit from agglomer-
ation economies. In other industries, firms may benefit from being close together
because individual firms can closely observe and react to the behavior of competi-
tors. Agglomeration economies also occur across types of firms, for example, job
6 Firms' production function is multiplied by a shift variable N'. where N equals the number of firms or
jobs in an urban area and a > 1.
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1383
Table 1
Profit variations for two firms locating at two alternative sites
Firm 2
A B
Firm 1 A 5, 2 1,2
B 1,1 3,3
sites are more attractive to workers when there are shops and restaurants nearby.
These agglomeration economies seem to require proximate location. In general,
there has been little research on how agglomeration economies operate within
urban areas and how they affect firm location patterns.
An important implication of agglomeration economies, when applied to issues
of location within an urban area, is that they may cause development to occur at
inefficient locations. Consider a simple model in which there are two alternative
sites for a subcenter in a particular urban area or portion of an urban area. The
two sites, denoted A and B, are both adjacent to freeway intersections. There are
two firms, denoted 1 and 2. While either or both firms may locate at either site,
agglomeration economies make both firms better off if they locate at the same
site. Table 1 shows both firms' profits from locating at each site. Firm 1 makes a
profit of 5 at site A and a profit of 3 at site B if both firms locate at the same site,
but firm I makes a profit of only 1 if the two firms locate at different sites. Firm
2 makes a profit of 3 at site B and a profit of 2 at site A if both firms locate at the
same site, but it makes a profit of only 1 if the two firms locate at different sites. If
firm 1 moves first and chooses site A, then firm 2 will also choose site A and the
outcome will be economically efficient since the sum of both firms' profits (7) is
maximized. However, if firm 2 moves first and chooses site B, then firm 1 will
also locate at site B. In this case, the outcome will be economically inefficient
since the sum of both firms' profits (6) is lower than if they both located at site
A. The game has multiple equilibria, of which only one is economically efficient.
If the game were played many times in different regions by different firms, then
we would expect subcenters to develop at a mixture of efficient sites like A and
inefficient sites like B. Because the model has multiple equilibria, it is difficult
to predict in advance where suburban subcenters will develop. 7
7 Obviously firm 1 can bribe firm 2 to choose site A even if firm 2 moves first. But firm 1 may not be
present when firm 2 makes its move and, once firm 2 has chosen site B, the costs of moving may exceed the
gains from both firms being located at site A rather than site B.
1384 M.J. White
Suppose suburban locations are more profitable for at least some urban firms
than CBD locations. In this case, what overall spatial pattern of employment
will develop?; is it economically efficient?; and what are its effects on other
aspects of resource allocation in urban areas? There are actually two versions
of these questions. In one, the main issue is how agglomeration economies and
commuting costs affect the optimal and actual spatial patterns of employment
and residences in urban areas. The city is assumed to be built from scratch, so
there is no presumption that a CBD will exist. In the second, some firms are
assumed to move to exogenously determined suburban locations, but the change
is incremental and the historic CBD remains. The main focus is on examining
the effects of firm suburbanization on residential location and other aspects of
resource allocation in cities. I refer to these two literatures as models of endoge-
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1385
nously versus exogenously determined employment location. They are discussed
separately below.
Consider the optimal spatial location pattern for firms in an urban model with no
history. The earliest approach to this problem was by Mills (1972: chap. 5). Mills
analyzed a model of an urban area in which identical firms produce a good using a
fixed amount of land, and housing is also produced using a fixed amount of land.
Identical workers commute to the firms at a constant cost per unit of distance trav-
eled. Output produced by firms is transported to the CBD, where it is exported,
and the cost of goods transport is also constant per unit of distance. The optimal
allocation of land to production and housing is the allocation which minimizes
the sum of goods transport costs plus workers' commuting costs. Mills (1972)
shows that there are two efficient solutions: the "segregated" solution in which
land around the CBD is devoted exclusively to production while land surround-
ing the production area is devoted exclusively to housing, and the "integrated"
solution in which production and housing are mixed at all urban locations. The
segregated solution holds when the cost of goods transport is high relative to the
cost of commuting, since locating production in the CBD minimizes the cost of
transporting goods. The integrated solution holds when the cost of commuting
is high relative to the cost of goods transport, since commuting is eliminated
when all workers work at home. Mills also shows that in this model, the market
equilibrium solution is economically efficient. 8
I documented above the fact that employment has tended to suburbanize more
rapidly than population in US cities over the past several decades. This suggests
that urban areas in reality have moved from approximating the "segregated"
solution to approximating the "integrated" solution in Mills' model. This sug-
gests that the cost of goods transportation must have fallen relative to the cost of
commuting-a testable hypothesis.
The paper by Fujita and Ogawa (1982) uses assumptions similar to those of
Mills (1972), but adds agglomeration economies to the model. Fujita and Ogawa
analyze a straight-line city. Identical firms are again assumed to produce goods
using fixed amounts of land and labor and to transport the goods to the CBD for
export. The level of agglomeration economies depends on the density of firms
at particular locations and may be constant all over the urban area or may differ
at different locations. Workers each occupy a constant amount of land and the
costs of commuting and transporting goods to the export node are both con-
stant per unit of distance. Fujita and Ogawa (1982) solve numerically for the
equilibrium outcome, assuming that firms enter the city until profits fall to zero.
Because the costs of commuting and of goods transport trade off against variable
agglomeration economies, a number of different land use patterns may occur. If
agglomeration economies are high at the CBD and decline with distance from
the CBD, then firms concentrate at a single CBD surrounded on both sides by
housing. If agglomeration economies are constant at all locations and the cost of
commuting is high relative to the cost of goods transportation, then a dispersed
land use pattern occurs in which all workers work at home. Another possible
outcome is an "incompletely mixed urban configuration", in which the center
of the urban area is occupied by mixed firms and housing, surrounded on both
sides by regions occupied exclusively by firms, while the outer regions of the
urban area are occupied exclusively by residences. Workers occupying the cen-
tral region work at home, while workers occupying the exclusively residential
regions commute to firms located in the exclusively business regions. Finally,
other possible outcomes include two employment subcenters without a CBD and
a CBD plus two subcenters. In all of the solutions, the left and right sides of the
urban area are symmetric. 9
In a model with agglomeration economies, equilibrium outcomes are likely
to be inefficient since individual firms ignore the effects of their behavior on the
overall level of agglomeration economies and therefore on other firms' costs.
Henderson and Slade (1993) extend Fujita and Ogawa's model by making it
into a game between two developers. This introduces another set of reasons
why the equilibrium outcome may differ from the optimal outcome.l 0 In their
model, one of the developers develops the lefthand side of the city and the other
develops the righthand side. Each builds a development that contains a residen-
tial neighborhood and a business district (land uses are not allowed to mix).
When the city is small, it is efficient for both developers to locate their busi-
ness districts at the inner edge of their respective territories, so that they merge
and the combined city has a CBD. As the city increases in size, the costs of
goods transportation and commuting rise faster than agglomeration economies,
so that eventually it becomes efficient for the business district to split in two.
9 Fujita and Ogawa (1982) do not investigate whether, or not, the equilibrium land use outcome differs
from the optimal outcome. But in a later paper, Ogawa and Fujita (1989) discuss the relationship between
equilibrium versus optimum land allocations in a similar model.
10 See also Tauchen and Witte (1984), who analyze equilibrium versus optimum land use allocations in
a model of a CBD with agglomeration economies. They show that an inefficient number of firms enters the
urban area in the equilibrium outcome. Helsley and Sullivan (1991) investigate the possibility that different
employment subcenters could have different production technologies or could have extemal effects.
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1387
At that point, each developer locates its business district at approximately the
center of its territory, with residential neighborhoods on both sides, so that there
are two equal sized business districts-one on each side of the city. However,
Henderson and Slade (1993) show that when there are two developers, they have
an incentive to split the CBD prematurely, i.e., the split occurs at a lower than
optimum population level. The reason is that each developer takes account of the
agglomeration economies realized by firms in its half of the city, but ignores the
agglomeration economies realized by firms in the other half. Thus Henderson and
Slade's model provides an example of how strategic considerations, combined
with agglomeration economies, may cause the spatial layout of the urban area
to be inefficient. In a sequential version of their model, they provide another
example of how strategic considerations may distort the spatial layout of the city.
In that model, the first developer to enter makes its development inefficiently
large in order to capture first mover advantages; while the second developer then
makes its development inefficiently small. The result is that the two sides of the
city are asymmetric, which is inefficient.
A recent model by Anas and Kim (1994) also examines the equilibrium spatial
location pattern in a discrete version of a straight-line city. Anas and Kim do
not assume that there are external agglomeration economies, but they make an
assumption that urban firms sell their output directly to the households who live
in the urban area. These shopping interactions between firms and households are
similar to the transactions between pairs of firms that formed the basis of early
models of urban agglomeration economies. Workers (or their families) thus take
shopping trips to buy from firms, as well as making commuting trips. Because
goods produced at different locations are assumed to be spatially differentiated,
workers demand them all, although they have the highest demand for goods
produced by nearby firms. The model also incorporates traffic congestion and
endogenous congestion tolls. The results is a dispersed land use pattern: both
jobs and housing are present in all regions of the urban area, although the density
level of both is highest at the center because of greater accessibility. The dis-
persed location pattern occurs for a combination of two reasons: first, congestion
and congestion tolls make it worthwhile to reduce travel costs by mixing firms
and households and, second, when firms transact with households, they have an
additional incentive to mix with households so as to reduce the length of shopping
trips.
These models have given us a much improved understanding of how ag-
glomeration economies, congestion, the costs of commuting and of goods trans-
portation, and strategic considerations interact to determine the spatial layout of
employment in urban areas. The basic factors that cause firms to benefit from
1388 M.J. White
In these models, the set of possible employment location within the urban area is
exogenously determined. Wages at suburban employment locations may be either
exogenously or endogenously determined. The models focus on how workers
decide where to live and work and the resulting spatial patterns of land rents, pop-
ulation densities and commuting regions. In one sense, models with exogenously
determined employment locations are special cases of the previous set of mod-
els in which the employment location pattern is endogenously determined and
optimal versus actual employment patterns can be compared. However, the dif-
ficulty of solving models with endogenous employment locations usually means
that these models focus on the tradeoff between agglomeration economies and
transportation costs and they assume away most other issues. Models with ex-
ogenous employment patterns often focus instead on modeling other issues in the
context of a decentralized urban area, such as the spatial location pattern when
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1389
there are multiple household types, the effect of gasoline taxes on employment
suburbanization or the effect of zoning regulations that limit development.
The earliest model of a decentralized urban area was that of White (1976). In
her model, firms may locate either at the CBD or a constant number of miles away
from the CBD in any direction. All firms are initially located at the CBD, but
some move to the suburbs because they export their output from the urban area
and the cost of doing so is lower in the suburbs. As discussed above, firms that
suburbanize have an incentive to spread themselves out in all directions around
the CBD, so that the suburban employment locations become a ring subcenter.
Workers are assumed to commute along straight lines between their homes and
their workplaces. If the wage at the ring equals the CBD wage minus twice the
workers' cost of commuting between the ring and the CBD, then all commuting
in the model is radial incommuting. All workers located between the CBD and
the ring commute to the CBD, while workers located further out than the ring are
indifferent between working at the CBD or the ring. But now suppose the wage
at the ring rises above this level. Then there will be three commuting regions: a
circular region around the CBD composed of workers who commute to the CBD,
a doughnut-shaped region around it but inside the ring subcenter composed of
suburban workers who outcommute to jobs at the ring, and an outer doughnut-
shaped region beyond the ring composed of workers who incommute to jobs
at the ring. As long as firms are spread evenly around the ring subcenter, no
circumferential commuting occurs. In this situation, the land rent gradient and
population density both fall with distance from the CBD, then rise to a local
maximum at the subcenter and then fall again beyond the subcenter. Because
White's (1976) model did not incorporate agglomeration economies, the optimal
location for the ring subcenter was entirely outside the CBD's commuting region.
Sullivan (1986) also considered an urban area with a CBD and a ring subcen-
ter. But rather than making the wage pattern exogenous, he assumed that firms
at both locations have downward sloping labor demand functions. This allows
wages to be determined endogenously by the condition that labor supply must
equal labor demand at each employment location. In Sullivan's model, firms
located at the CBD are assumed to be subject to agglomeration economies, while
firms at the subcenter are not. His model, which is solved numerically, shows
how wages at the two employment locations are linked via the land market. For
example, suppose demand for labor at the CBD shifts outward. Then the wage at
the CBD rises and some workers who live between the CBD and the subcenter
shift from working at the subcenter to working at the CBD. This causes the
CBD's commuting region to become larger and the subcenter's commuting re-
gion to become smaller. The backward shift in labor supply to the subcenter
1390 M.J. White
causes the subcenter's wage to rise. Thus, wages at the two employment locations
tend to move together."
Ross and Yinger (1995) consider an urban model in which there are firms
both at the CBD and at either a point or a ring subcenter. Like Sullivan (1986),
they assume that firms have downward sloping demands for labor, so that wages
at each employment location are determined endogenously. Ross and Yinger
solve for a closed form solution to their model and they focus on examining
comparative statics results and whether, or not, these results are the same in a
decentralized urban model as in a model with only CBD employment. Although
their model has no agglomeration economies, their comparative statics results are
similar to those found by Sullivan (1986), i.e., exogenous shocks cause wages at
both employment locations to move in the same direction and a change in wages
at one employment location causes wages to change in the same direction at
the other. An increase in the exogenous utility level of urban residents is shown
unambiguously to raise wages at both employment locations. But because the
direct effect of the increase in utility and the indirect effect of the increase in
utility via wages have opposing effects on land rents, the direction of the effect
on land rents cannot be signed. 12
Wieand (1987) analyzes a similar model with a CBD and a point subcenter,
but he assumes that both employment locations have agglomeration economies
and he allows the location of the subcenter to vary. He assumes that the level of
agglomeration economies at each subcenter depends on the number of jobs at
that subcenter, so that firms at the CBD are more productive as long as the CBD
contains more firms. Given that establishing a subcenter is worthwhile, Wieand
(1987) explores the question of the optimal subcenter location. He finds that if
the subcenter will contain only a small number of jobs, then it is optimal for it
to locate near the outer boundary of the urban area; while if the subcenter will
be large, then it is preferable for it to locate near the CBD. This is because the
further the subcenter is located from the CBD, the fewer the number of workers
willing to commute to it for any given wage. If the subcenter contains few firms,
then the small commuting region is not a drawback and the best location for it is
near the urban periphery where land costs are low. But if the subcenter contains
many firms, then it is better for it to remain near the CBD where it can attract
more workers for any given wage. Wieand (1987) also points out that if the total
population of the urban area is fixed, the establishment of a suburban subcenter
I Sullivan (1986) also considers the effects of land use controls in the CBD or the residential areas.
12 Yinger (1992) explores a similar model in which workers commute along a network of radial and circular
roads, rather than along straight lines between their homes and workplaces.
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1391
causes problems for the CBD since its loss of jobs causes loss of agglomeration
economies and the nearby population density also falls. 13
In White (1990), a simulation model is used to explore public policy concerns
about long commuting trips by analyzing the extent to which policy measures
that encourage more firms to suburbanize would reduce commuting. An urban
model that originally has two suburban point subcenters is assumed to add two
additional point subcenters located further out and, for comparison, an urban area
that originally has a ring subcenter is assumed to add an additional one located
further out. The model has no agglomeration economies, but congestion raises
commuting costs near the CBD and the subcenters. The model is designed to
allow any number and any spatial configuration of subcenters to be simulated.
The main result of the simulation is that adding two additional suburban point
subcenters or an additional suburban ring subcenter reduces workers' average
commuting journey length by about 15-50%-more for the ring subcenters than
the point subcenters. 14
In Hotchkiss and White (1993), a simulation model is used to explore an
urban area where there are multiple household types-two-worker households,
"traditional" households with one male worker, and female-headed households.
Wages at each of the employment locations are exogenously determined (male
workers are assumed to earn more than female workers). The purpose of the
model is to explore the spatial implications of such social trends as the high
divorce rate, the increasing rate of labor force participation by married women,
and the increasing dispersion of income within urban areas that these two trends
cause. In general, two-worker households outbid other households for sites that
are most accessible to their job locations, traditional households occupy the most
suburban sites and female-headed households occupy the intermediate distances.
Because high income households occupy central rather than suburban locations,
the spatial allocation more closely resembles a European rather than an American
city.15 An increase in the cost of commuting causes a reduction in the number
of female-headed households in the urban area, and a decrease in wages for
female workers causes traditional households to replace two-worker households
in the urban area. Thus, the model suggests that seemingly unrelated policy
13 The Wieand (1987) model implies a testable hypothesis that firm or subcenter size should vary inversely
with distance from the CBD. To the author's knowledge, this hypothesis has not been tested empirically.
14 Even four subcenters is probably too few to realistically represent a large urban area. See Giuliano and
Small (1991) for a discussion of how to identify employment subcenters. They find 32 subcenters in the Los
Angeles metropolitan area.
15 These results suggests that to get the typical spatial location pattern of a large US city, explicit
disamenities of living near the center of the city would need to be introduced into the model.
1392 ,I.J.White
changes can have important spatial implications. Hotchkiss and White (1993)
also introduce a random component to wages, that allows the model to rep-
resent worker/household heterogeneity. The effect of introducing even a small
random component is that the commuting region boundaries become very fuzzy.
All household types occupy sites all over the urban area (instead of occupying
sites only in their particular commuting regions) and the number of different
household types represented in the urban area increases. Randomness in income
greatly dampens the responsiveness of the model to exogenous shocks.
Finally, the paper by Sivitanidou and Wheaton (1992) also explores a model
in which there are two employment locations, but firms compete with households
for land. One subcenter is allowed to have a cost advantage in production over
the other and relative wages at the two subcenters are determined endogenously.
The goal of the model is to determine to what extent differences between the two
subcenters are capitalized as differences in commercial land rents versus differ-
ences in wages. The authors first show that if the two subcenters have the same
costs, then the urban area is symmetric and both subcenters have identical size,
wages and commercial land rents. Also the residential land rent patterns around
them are identical. Now suppose subcenter A has a cost advantage over subcenter
B. Then, firms in subcenter A expand by offering higher wages. This leads to
an expansion in subcenter A's commuting area, which bids up residential land
rents around subcenter A. Because firms compete with households for land, the
expansion of subcenter A also raises commercial land rents. Thus, the cost dif-
ferentials between subcenters lead to both higher wages and higher commercial
and residential land rents at and around the subcenter that has the cost advantage.
Now suppose again that the two subcenters have equal cost, but zoning imposes a
binding restriction on the land area of subcenter A. Then commercial land rents at
subcenter A rise above those at subcenter B, but subcenter A's wages fall below
those of subcenter B. Because of the zoning restriction, commercial land rent at
subcenter A is higher than (rather than equal to) land rent for residential land just
adjacent to subcenter A. Residential land rent around subcenter A is also lower
than that around subcenter B, which means that workers are willing to work
for lower wages at subcenter A because their land costs are lower. Thus, zon-
ing restrictions on one subcenter lead to that subcenter having higher rents, but
lower wages. The authors conclude that differences in commercial rents across
subcenters within an urban area may not accurately measure differences in the
subcenters' productivity.
Models in which the spatial pattern of employment is exogenously determined
obviously cannot be used to analyze what spatial location pattern is efficient, but
they can be used to analyze other implications of decentralized employment in a
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1393
more realistic setting. Future research, perhaps using numerical techniques, may
be able to combine the strengths of both sets of models.
When jobs decentralize, the structure of an urban area becomes much more com-
plicated. To start with, all urban areas-monocentric and decentralized-have
rent gradients that relate the price of land to distance from the CBD. Rent offer
curves describe how much each household is willing to pay for land at each
location, and the market rent gradient is the upper envelope of all households'
rent offer curves. In the monocentric city, rent offer curves and the market rent
gradient always decline at a decreasing rate with the distance from the CBD.1 6
But in the decentralized city, households' rent offer curves may be affected by
workers' job locations and this may affect the shape of the market rent gradient.
Decentralized cities also have a wage gradient, that relates wages to distance from
the CBD for identical jobs. Workers living at particular residential locations have
a wage offer curve that indicates the minimum amount they must be paid to be
willing to work at any job location. The lower envelope of workers' wage offer
curves is the market wage gradient. The characteristics of individual workers'
wage offer curves and the market wage gradient, both of which may be affected
by where workers live, need to be established. In this section I explore how rent
and wage gradients in decentralized cities are determined and how they relate to
each other.l7
Suppose an urban area has an employment pattern consisting of a CBD and
suburban firms that are dispersed at isolated locations in all directions around the
CBD. Residential distance from the CBD is denoted u and workplace distance
from the CBD is denoted v. All workers are assumed to incommute, so that
their commuting distance is u - v. Wages per day are w(v), the out-of-pocket
cost of commuting is m per mile, and the speed of commuting is 1/s. Since
there is no congestion, the cost of commuting a mile in each direction is always
2(sw(v) + m). Each household has one worker. Suppose the rent on land per
unit is denoted r(u, v), land consumption per household is l(u, v) and hours of
leisure consumption are h(u, v). Households' rent offer curves must satisfy the
property:
-2(sw(v) + m)
ru (U,v) = , (4.1)
l(u, v)
16 See Mills and Hamilton (1989).
17 The main references for this section are Straszheim (1984) and White (1988a).
1394 M.J White
where the subscripts denote partial derivatives. Equation (4.1) looks familiar
from the analysis of the monocentric city: it says that land rents fall with res-
idential distance at a rate equal to the cost of commuting a mile further from
the CBD divided by the households' demand for housing. However, the presence
of terms depending on v in Eq. (4.1) suggests that the rent offer curve may be
affected by workers' job locations. Equation (4.2) is the wage offer curve. It says
that the rate of change in the wage at which workers are willing to work for firms
located further from the CBD equals the cost of commuting a mile further per
day divided by the number of hours of work, where the latter equals 24 minus
time spent commuting (2s(u - v)) minus time spent in leisure. The presence of
terms in u in the equation suggests that the wage offer curve may be affected by
workers' residential locations.
Since the rent offer curve has the same form as in the monocentric city case,
individual households' rent offer curves in decentralized cities must always have
a negative slope and must always decline at a decreasing rate with distance. Now
consider whether and how individual households' rent offer curves vary with job
location by differentiating Eq. (4.1) with respect to v. The result is:
0 V' U' UV
Fig. 2. Rent offer curves of households whose workers work at the CBD and at v', and the market rent gradient.
Figure 2 shows the rent offer curve rl of households whose workers work at the
CBD, and the rent offer curve r2 of households whose workers work at v'. The
market rent gradient, shown as a dashed line, is the upper envelope of the two rent
offer curves. The boundary between the two rent offer curves occurs at u', where
u' > v'. Households whose workers hold CBD jobs live in the inner residential
area between u = 0 and u', while households whose workers work at v' live in
the outer residential area further out than u'. 18
If we combine the rent and wage offer curves, Eqs. (4.1) and (4.2), we can
determine the relative rate at which the wage offer curve falls with distance from
the CBD compared to the rent offer curve, or,
wv/w rl
=-,r (4.4)
ru/r wn
rents decline. This figure is probably higher in Europe. Nonetheless, because the
fraction of earnings spent on land rent is low, the wage offer curve is predicted
to decline quite slowly with distance from the CBD. Therefore, measuring it
empirically turns out to be difficult.
Finally, consider whether and how individual workers' wage offer curves vary
with residential location by differentiating Eq. (4.2) with respect to u. The result
is
awv/au nu (4.5)
--. (4.5)
W, n
The percentage change in the slope of the wage offer curve when workers move
their residential locations a mile further out, but keep their job locations fixed,
equals minus the percentage change in hours of work when workers move their
residential locations a mile further out. Since w, is negative, the sign of aw/,lau
is the same as the sign of n. The sign of nu is ambiguous, but the most likely case
is that it is negative, since when workers spend more time commuting (residential
location shifts outward while job location remains fixed), they are likely to com-
pensate by reducing both the number of hours of leisure and the hours of work.
In this case, awv/au must be negative and workers' wage offer curves become
steeper as their residential locations become more suburbanized. Suppose there
are only two residential locations, consisting of rings located at distances u* and
u** > u*, while jobs occur at all locations. Figure 3 shows the wage offer curve
wl of the group of workers living at u* and the wage offer curve w2 of the group
of workers living at u**. The market wage gradient, shown as a dashed line, is the
lower envelope of the two curves. Workers living at u* take jobs located between
v = 0 and v", while workers living at u** take jobs located further out than
v = v". The market wage gradient has a negative slope and becomes steeper
with distance from the CBD.19
We have shown that the decentralized urban area has a market wage gra-
dient that relates wages to workplace location, in addition to having a market
rent gradient that relates the price of land to residential location. The prediction
that urban wages vary with distance from the CBD is one of the major testable
hypotheses of the urban model. In the model just discussed, only incommuting
was assumed to occur and wage gradients were therefore downward sloping.2 0
However, as discussed in Section 1, when suburban firms have high labor de-
mand, they must raise wages in order to attract enough workers. This causes
19 See White (1988a) for further discussion.
20 Note from Fig. 3 that the theoretical urban model has no prediction for the sign of the second derivative
of the market wage gradient, although the wage offer curves have declining slopes.
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1397
Fig. 3. Wage offer curves of workers who live at u* and u**, and the market wage gradient.
wage gradients to become less negative, or even to turn positive, and it also causes
outcommuting to occur. Flat or positive wage gradients are most likely to occur
in a particular subregion of an urban area, since they are caused by firms at a
large subcenter offering high wages in order to expand their commuting regions.
One of the major empirical hypotheses generated by the urban model with decen-
tralized employment is that wages for otherwise identical jobs vary with distance
from the CBD. In general, the urban model predicts that wages decline with
distance from the CBD, but wages could be constant, or even rise, with distance
if there is concentrated suburban employment. Rising wage gradients are likely
to occur only in a particular direction from the CBD.
The theory predicts that researchers attempting to measure urban wage gradi-
ents will confront several problems. First, as discussed above, finding evidence of
wage gradients is likely to be difficult since wages decline at a much slower rate
than rent as distance from the CBD increases. Second, finding a sample of jobs
and workers that are identical except for distance from the CBD is difficult. But if
observations of nonidentical jobs or workers are used to estimate wage gradients,
then the results may reflect location-specific job differences or location-specific
differences among workers rather than a true wage gradient. For example, police
jobs are likely to be less dangerous in the suburbs than the central city because
suburban crime rates are lower. This factor would tend to cause police wages
in the central city to rise relative to police wages in the suburbs. An estimated
wage gradient for police wages that did not control for crime rates would there-
1398 .J. hite
fore decline too steeply. As another example, suburban teachers tend to be more
experienced than central city teachers and, as a result, their wages are higher. An
estimated wage gradient for teachers that did not control for experience might
therefore be flat or even positively sloped, when the true wage gradient would
have a negative slope. Third, suppose a metropolitan area has a negative wage
gradient in one direction away from the CBD, but a positive wage gradient in
another direction-perhaps reflecting the presence of a large subcenter in that di-
rection. Then, if a single wage gradient were estimated for the entire metropolitan
area, it would probably result in no gradient being found.
An additional problem with the empirical literature on urban wage gradients
is that it combines tests of two distinct hypotheses under the same name. One
literature tests for a relationship between wages and distance from the CBD.
The other literature tests for a relationship between commuting journey length
and wages. The latter relationship results from the hypothesis that workers are
willing to commute further in return for higher wages, just as they are willing
to commute further in return for lower housing prices. Evidence supporting the
commuting journey/wage relationship has sometimes been interpreted as provid-
ing support for the hypothesis that wages decline with distance from the CBD
(see, for example, Madden, 1985), but this conclusion is not always correct. For
example, suppose wages rise rather than fall with distance from the CBD in a
particular city. Then suppose a set of workers who outcommute to their jobs
is used to test the commuting journey/wage relationship. These workers would
have the predicted positive relationship between commuting journey length and
wages, even though wages rise rather than fall with distance from the CBD. An
alternate possibility is that wages fall and then rise with distance from the CBD
in a particular city. In this case, an empirical test of the commuting journey/wage
relationship using randomly selected workers might fail to find a relationship at
all. In what follows, I survey only papers that directly test the wage/distance from
the CBD relationship. 21
Because of lack of availability of data on workplace location, many of the
papers that test for the existence of wage gradients have used data that the authors
collected specifically for the purpose. An early effort was that of Rees and Shultz
(1970), who collected data on wages, job characteristics, and job location as part
of a study of the Chicago labor market. They compared wages in the CBD (the
Loop) to those in the region south of the CBD and in the regions north and west of
the CBD, for workers in both white and blue collar jobs. At the time of their study,
21 See Crampton (Chap. 39, this volume) for a discussion of papers on the commuting journey/wage
relationship.
Ch. 36: UrbanAreas with Decentralized Employment: Theory and Empirical Work 1399
the region south of the Loop contained most of the Chicago metropolitan area's
heavy industry, while the regions north and west were mainly residential. Rees
and Shultz (1970) found that wages in the north/west region were significantly
lower than at the CBD for both blue and white collar occupations. But wages in
the south region were higher than at the CBD for blue collar occupations and not
significantly different from the CBD for white collar occupations. Their evidence
is consistent with the theoretical model discussed above in that wages rise with
distance from the CBD in the region of Chicago where jobs are plentiful but
workers living nearby are scarce, so that suburban employers need to raise wages
in order to induce enough workers to commute to suburban jobs. In contrast,
wages fall with distance from the CBD in the regions of the metropolitan area
where workers are plentiful and jobs are scarce.
A more recent study by Eberts (1981) also used data from Chicago. Eberts
obtained wage data for employees of 100 municipalities in all regions of the
Chicago metropolitan area. He estimated regressions explaining the log of wages
as a function of the log of the municipality's distance from the CBD for five cat-
egories of public sector workers: police, fire, administration, clerical and public
works. These data have the drawback that they are for municipalities rather than
individual workers, but job characteristics and workers' average characteristics
may vary across municipalities. As discussed above, these variations may bias
the estimate wage gradients either upward or downward. In fact, Eberts (1981)
finds that that the wage/distance from the CBD relationship is negative for all
five categories of workers and statistically significant for all except fire. The
elasticity of wages with respect to distance was about -0.2 for public works
and police employees, and -0.3 for administration and clerical. The absolute
decrease in monthly wages per mile of distance was $24 for administration, $10
for clerical, $12 for police and $9 for clerical (in 1974 dollars). Despite these
problems, Eberts' results provide some support for the hypothesis that urban
wage gradients are negatively sloped.
More recent studies of urban wage gradients have taken advantage of the Pub-
lic Use Microsample (PUMS) of the 1980 US Census of Population and Housing.
This dataset includes all the information collected as part of the US Census
of Population for a large sample of individual households. For each worker, it
indicates whether the worker's job location is in the CBD, the rest of the central
city or in any of a set of suburban zones, where the number of suburban zones
varies between 1 and 28 for different metropolitan areas.
Ilhanfeldt (1992) uses PUMS data for Philadelphia, Detroit and Boston (the
metropolitan areas having the largest number of zones) to estimate urban wage
gradients separately for a variety of occupational groups and for white versus
1400 M.. White
At the beginning of this chapter, I documented the long-term trend toward sub-
urbanization of both population and employment. A controversy in urban eco-
nomics concerns the issue of causation: does the spatial pattern of population
in a metropolitan area depend on the spatial pattern of employment, or does the
employment pattern depend on the population pattern? As discussed above, the
urban models literature made widespread use of the assumption that the spatial
pattern of employment is exogenously determined-either because all jobs are
assumed to be located at the CBD, or because employment is decentralized but
its spatial pattern is exogenously specified. This implicitly assumes that the resi-
dential location pattern is determined by the employment location pattern, rather
than the reverse. The direction of causation issue is testable and a literature has
developed that attempts to test whether jobs follow workers or workers follow
jobs.
Before examining this literature, however, it seems worthwhile to note that
the assumption of the urban models literature that employment location is ex-
ogenously determined is mainly made for convenience. In the early models, un-
derstanding the economics of residential location within urban areas was difficult
enough by itself, and the problem was made much more tractable by assuming
away the need to explain employment location. Since then, economists have
worked on explaining the pattern of firm location within urban areas and have
made progress, but our understanding of firm location remains sketchy relative
to our understanding of residential location. Furthermore, theoretical models
that explain both simultaneously remain highly simplified and static rather than
dynamic, as the discussion in Section 3.1 suggests. Thus, while the empirical
question of whether jobs follow people or people follow jobs is interesting and
important in its own right, the empirical models do not provide a test of any hy-
pothesis of the theoretical model. The theoretical model itself has not developed
to the point where it provides clear predictions on this issue.
1402 M.J. White
Steinnes (1977) was the first to test the causation issue. His technique was
to estimate a two-equation simultaneous model using a pooled time-series cross-
section dataset. Suppose Pi,, denotes the proportion of metropolitan area popula-
tion located in the central city in city i in period t, and Ei,t denotes the proportion
of metropolitan area employment located in the central city in city i in period t.
In the first equation, the dependent variable is Pi,t and the independent variables
are the lagged values Pi,t_l and Ei,_l, plus other exogenous variables that affect
the population pattern. In the second equation, the dependent variable is Ei,t
and the independent variables are the lagged values Ei,t-l and Pi,l,plus other
exogenous variables that affect the employment pattern. The test of the "people
follow jobs" hypothesis is whether, or not, the coefficient of Ei,t_ - in the equation
explaining Pi., is statistically significant and positive, which would indicate that
greater suburbanization of employment causes greater suburbanization of popu-
lation in the following period. The test of the "jobs follow people" hypothesis
is whether, or not, the coefficient of Pi,t_l in the equation explaining Ei, is
statistically significant and positive.
Steinnes (1977) tested the model separately for manufacturing, service and re-
tail employment. The results for manufacturing and services supported the "jobs
follow people" hypothesis, while the results for retail employment supported the
"people follow jobs" hypothesis. Cooke (1978) redid Steinnes' estimation using
residential and employment density gradients (rather than proportion located in
the central city) as his measures of the extent of suburbanization. He also found
that employment follows population, although the results for manufacturing em-
ployment were more significant than those for retail and service employment.
Mills and Price (1984) also found that jobs follow people in a model explaining
population and employment density gradients and using data from the 1960 and
1970 Censuses of Population.2 3 However, Thurston and Yezer (1994), also ex-
plaining density gradients but using annual data from the Local Personal Income
Series of the US Department of Commerce, found evidence that jobs follow
people in the services and retail sectors, but did not find evidence supporting
causality in either direction for five other employment sectors
An interesting recent paper on this issue is that of Boarnet (1994). Rather than
using metropolitan areas as observations, Boarnet uses data from 365 munici-
palities in northern New Jersey and, rather than explaining levels of population
and employment, he explains the changes in municipal population and employ-
ment between the years 1980 and 1988. Thus, Boarnet focuses the analysis on
23 The main conclusion of the Mills and Price (1984) paper is that adding to the model a set of variables
measuring central city problems-crime, race and taxes-does not add any explanatory power to the basic
model explaining suburbanization of population and employment.
Ch. 36: UrbanAreas with Decentralized Employment: Theory and Empirical Work 1403
the submetropolitan area level. For each municipality i, Boarnet defines a labor
market consisting of all other municipalities in the sample. Each municipality
is weighted by ll/d, where dij is the distance between municipality i and mu-
nicipality j and B is a separately estimated parameter that indicates how labor
market relationships across municipalities decay with distance. Employment in
municipality i's labor market area is defined as the weighted sum of employment
in all the other municipalities in the sample, and population in municipality i's
labor market area is defined as the weighted sum of population in all the other
municipalities in the sample. Thus, an individualized labor market is defined for
each municipality.
In Boarnet's model, the change in population in municipality i between 1980
and 1988 depends on the change in employment over the same period in munic-
ipality i's labor market area, the lagged value of population in municipality i,
the lagged value of employment in municipality i's labor market area, and other
control variables. The change in employment in municipality i similarly depends
on the change in population in the labor market area surrounding municipality i,
the lagged value of employment in municipality i, the lagged value of population
in municipality i's labor market area, and other control variables. The two equa-
tions are estimated using simultaneous techniques. Boarnet's results indicate that
the change in the municipal employment level is significantly related to changes
in the population of the labor market area, but the change in the municipal pop-
ulation level is not significantly related to changes in employment in the labor
market area.
Thus, Boarnet's results, like earlier papers, provide support for the hypothesis
that jobs follow people, but not vice versa. Since Boarnet's test uses municipali-
ties within a particular metropolitan area rather than more aggregated metropoli-
tan area data, his results suggest that the hypothesis that jobs follow people holds
for explaining both overall spatial patterns across metropolitan areas, and more
detailed spatial patterns within metropolitan areas.
Boarnet's work is unlikely to represent the final word on this subject, but
nonetheless his results and those of previous authors seem sensible and intuitively
appealing. If we think about the problem from the viewpoint of the theoretical
considerations discussed above, it seems clear that the employers' location deci-
sions are tied to the spatial pattern of population. Employers benefit from moving
to the suburbs because they can pay lower wages, but only if the move results
in workers having shorter commutes. Therefore, the gain to an employer from
suburbanizing depends on the density of population near the intended suburban
location and on the employer's level of demand for labor. As shown by Wieand
(1987), a small firm can benefit from locating near the periphery of the urban
1404 M.J. White
area because it only needs to hire a small number of workers, but a large firm
has an incentive to remain close to the CBD, since otherwise the high cost of
attracting many workers to the suburban site may more than offset the large firms'
other gains from suburbanizing. Thus, the gains to firms from suburbanizing
are directly linked to the distribution of population: firms have an incentive to
follow-but not to lead-workers to the suburbs. In contrast, households gain
from moving to the suburbs regardless of the spatial distribution of employment.
Even if all firms were located at the CBD, households would still benefit from
moving to the suburbs because workers' longer commutes are compensated by
lower housing prices-the basic tradeoff of the urban economic model. Thus, it
should not be surprising that empirical studies have tended to find that jobs follow
people, while people do not follow jobs. The assumption in the theoretical urban
models literature that the spatial distribution of employment is exogenous and de-
termines the spatial distribution of population (i.e., people follow jobs) was, and
is, merely a convenient simplification that made difficult models easier to solve.
Future work in urban economics needs to devote more attention to modeling firm
location choice, so that the simplification will no longer be necessary.
houses to their jobs. But actually, workers must commute along the existing road
network. If roads are not straight, then the minimum amount of commuting in
the metropolitan area rises. The assumption that all commuting journeys occur
along straight lines also causes Hamilton's calculations of the proportion of urban
commuting that is "wasteful" to be biased upward. 2 4
White (1988b) proposed that both the problem of jobs and housing being
differently distributed around the CBD and that commuters must travel along the
existing road network, could be solved by using an assignment model to calculate
the minimum average commuting journey length. An assignment model requires
that a metropolitan area be divided into zones, and that we know (1) the commut-
ing journey length between each pair of zones, (2) the number of residences in
each zone, and (3) the number of jobs in each zone. The model then calculates an
assignment of each worker to a job and residence that minimizes the aggregate
amount of commuting for all workers in the metropolitan area. Because of the
zonal structure, the assignment model can take into account any spatial distrib-
utions of jobs and houses around the CBD. Also, because the characteristics of
the existing road network determine the commuting journey length between each
pair of zones, the assignment model takes account of the fact that commuting
journeys must be made along actual rather than straight-line roads.
Suppose we also know the number of workers who actually commute in each
direction between each pair of zones. Conceptually, we can think of the model
as starting from the existing allocation of workers to jobs and housing and then
making any trades of job or housing assignments that would allow aggregate
commuting to be reduced, keeping the number of jobs and houses in each zone
fixed. For example, suppose worker A lives at the CBD and works five miles
south, while worker B works at the CBD and lives 10 miles south. Then the
total commuting could be reduced by 10 miles if workers A and B traded either
jobs or houses. Efficient trades tend to reduce the amount of outcommuting and
circumferential commuting that occurs. Circumferential commuting cannot be
eliminated completely as long as the distributions of jobs and housing around the
CBD differ from each other. But outcommuting can be eliminated completely as
long as employment is less suburbanized than housing. Thus, efficient commut-
ing patterns tend to involve relatively more incommuting, less outcommuting and
less circumferential commuting than actual commuting patterns.
White (1988b) used data from the 1980 US Census of Population to calculate
the average minimum commuting journey length for 25 US metropolitan areas.
24 Hamilton's (1982) commuting calculations were in terms of distance, but later researchers used commut-
ing time. See Hamilton (1990) and Small and Song (1992) for discussion of the effects of using commuting
distance versus time.
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1407
For these areas, the total number of zones ranges from five to 32, where the CBD
is a separate zone for the purposes of workplace location, but not residential
location.2 5 Census data on how many workers actually travel from each zone
to every other zone are used to calculate the average actual commuting journey
length. Comparing these figures, White found that the average actual amount of
commuting was only about 10% greater than the average minimum amount of
commuting.
Table 2 gives the matrix of actual commuting and optimal commuting flows
for the Buffalo metropolitan area, which has four suburban zones plus the central
city (zone 1) and the CBD. The lefthand column in both panels gives the number
of housing units in each zone and the top row in both panels gives the number
of jobs in each zone. (All figures are in thousands.) The top panel gives the
optimal assignment of workers to jobs and housing units and the bottom panel
gives the actual assignment. In the optimal assignment, all jobs in the CBD are
occupied by residents of the central city, whereas, in reality 62% of CBD jobs
are held by suburban residents. Thus, the optimal assignment eliminates long
incommuting journeys. Except in zone 4, all jobs in the four suburban zones
are held by residents who live in the same zone, while the remaining suburban
residents mainly commute to jobs in the central city. 26 Thus, most outcommuting
and circumferential commuting journeys are also eliminated. Although Buffalo
has only a few zones, the results are typical of the pattern for cities having a
greater numbers of zones.
A problem with using the assignment model is that as long as commuting jour-
neys within zones are shorter than those between zones, within-zone commutes
are treated as efficient and the assignment model does not change them. However,
if individual zones are relatively large, then additional trades within zones would
probably reduce commuting further. But the Census provides commuting data
only for a relatively small number of zones within each metropolitan area and,
in addition, the central city is a single zone even when it constitutes a large
fraction of the metropolitan area. In later papers, Cropper and Gordon (1991)
and Small and Song (1992) applied the assignment model to much more detailed
transportation data for particular metropolitan areas. Cropper and Gordon (1991)
used data for Baltimore that divided the metropolitan area into 498 zones; while
Small and Song (1992) used data for Los Angeles-Long Beach that divided the
metropolitan area into 706 zones. Cropper and Gordon found that the actual aver-
age commuting journey length in Baltimore was two-and-a-half times as high as
25 The zones for these data are the same as the zones in the PUMS data.
26 Zone 4 includes Niagara Falls, a suburban subcenter.
1408 M.J. White
Table 2
Optimal versus actual commuting patterns in Buffalo (1980)
the minimum average commuting journey length, while Small and Song's results
for Los Angeles-Long Beach indicated a ratio of 3. Both of these results are be-
tween Hamilton's (1992) actual-to-minimum commuting journey length ratio of
10, and White's (1988b) actual-to-minimum commuting journey length ratio of
1.1. Thus, there is evidence that much more commuting occurs than the minimum
amount required for workers to commute between a metropolitan area's existing
houses and its existing jobs, but the best evidence suggests that the ratio of actual
to minimum commuting is 2.5:3 rather than 2.5:10.
But does any of this demonstrate that commuting beyond the minimum in
metropolitan areas is inefficient or "wasteful"? To address this question, consider
what the assignment model omits. One problem is that many households contain
two workers and they may choose seemingly inefficient commuting patterns in
order to live together. Suppose one spouse works at the CBD and the other in
the suburbs and they minimize their combined commuting by living between
their two jobs. Therefore, one spouse outcommutes. The assignment model is
likely to trade away the outcommuting journey, since the procedure does not take
account of the spouses' desire to live together. As another example, suppose an
African-American worker lives near the CBD but works in the suburbs. Again,
Ch. 36: UrbanAreas with Decentralized Employment: Theory and Empirical Work 1409
the assignment model will tend to eliminate the outcommuting journey, but in
actuality the worker's household may remain at its current residential location
because it faces race discrimination in suburban housing markets, or prefers to
live in a more familiar environment. Or suppose a worker who lives south of
the CBD finds a new job north of the CBD, but remains in the same residential
location. This might be because (1) the cost of moving is high, (2) the worker
does not expect to keep the new job for long, and/or (3) the worker's household
likes the schools or parks in its current neighborhood and makes many (nonwork)
trips to them. The assignment model ignores these factors and tends to trade away
long circumferential commuting journeys. Finally, all jobs and houses are not
identical. Some trades might reduce commuting but would allocate high income
workers to small apartments and/or low income workers to large houses; while
other trades would allocate high skill workers to low skill jobs and vice versa.
All these examples, with the exception of the one involving race discrimination
in housing markets, suggest that households choose some amount of commuting
beyond the minimum. Such choices are economically rational and efficient as
long as households receive other benefits in return for commuting more, and
that they voluntarily choose the combination of the longer commute and the
other benefits. Nothing in the "wasteful commuting" controversy proves that
commuting beyond the minimum is inefficient.
The first model that attempted to allow for any of these factors is Cropper and
Gordon (1991). They estimate a multinomial logit model explaining residential
location choice as a function of individual household and neighborhood char-
acteristics, and commuting journey length. Using this model, they calculate the
utility level of each household in their sample if the household were to locate
in each of a set of zones. Then, when the assignment model is used to calculate
minimum aggregate commuting, an extra constraint is imposed for each house-
hold that the household cannot be moved to a residential zone where it achieves
a lower utility level than at its current residential zone. Additional corrections
prevent African-American households from being reassigned to zones that are
less than 10% African-American, prevent two-worker households from being
split up, and allow for differences between the location choices of owner versus
renter households. Thus, the Cropper-Gordon model corrects for at least some
of the factors that in actuality lead households to choose longer commutes. How-
ever, the results suggest that these corrections make relatively little difference.
The ratio of actual-to-minimum commuting is 2.0 for owners and 2.4 for renters
when the relocation constraints are imposed, compared to 2.3 and 2.8, respec-
tively, when no constraints are imposed. One possible reason for the constraints
having little effect is that, in many cases, the same commute-reducing effect can
1410 M.J. White
References
Anas, A. and I. Kim (1994), "General equilibrium models of polycentric urban land use with endogenous
congestion and job agglomeration", Journal of Urban Economics 40:232-256
Boamet, M. (1994), "The monocentric model and employment location", Journal of Urban Economics 36:79-
97.
Braid, R.M. (1988), "Optimal spatial growth of employment and residences", Journal of Urban Economics
24:227-240.
Capozza, D. (1976), "Employment/population ratios in urban areas: a model of the urban land, labor and goods
markets", in: G.J. Papageorgiou, ed., Mathematical Land Use Theory (Lexington Books, Lexington, MA).
Cooke, T. (1978), "Causality reconsidered: a note", Journal of Urban Economics 5:538-542.
Crane, R. (1995), "The influence of uncertain job location on urban forn and the journey to work", Journal of
Urban Economics 39:342-356
27 Also see Crane (1995) for a discussion of how uncertainty concerning job locations would affect the ratio
of actual-to-minimum commuting.
Ch. 36: Urban Areas with Decentralized Employment: Theory and Empirical Work 1411
Cropper, M. and P. Gordon (1991), "Wasteful commuting: a re-examination", Journal of Urban Economics
29:2-13.
Eberts, R.W. (1981), "An empirical investigation of intraurban wage gradients", Journal of Urban Economics
10:50-60.
Fujita, M. and H. Ogawa (1982), "Multiple equilibria and structural transition of non-monocentric urban
configurations", Regional Science and Urban Economics 12:161-196.
Giuliano, G. and K.A. Small (1991), "Subcenters in the Los Angeles region", Regional Science and Urban
Economics 21:163-182.
Hamilton, B. (1982), "Wasteful commuting", Journal of Political Economy 90:1035-1053.
Hamilton, B. (1990), "Wasteful commuting again", Journal of Political Economy 97:1497-1504.
Helsley, R.W. and A.M. Sullivan (1991), "Urban subcenter formation", Regional Science and Urban Eco-
nomics 21:255-276.
Henderson, J.V. (1977), Economic Theory and the Cities (Academic Press, New York).
Henderson, J.V. and E. Slade (1993), "Development games in non-monotonic cities", Journal of Urban
Economics 34:207-229.
Hotchkiss, D. and M.J. White (1993), "A simulation model of a decentralized metropolitan area with two-
worker, 'traditional' and female-headed households", Journal of Urban Economics 34:159-185.
Ilhanfeldt, K.R. (1988), "Intra-metropolitan variation in earnings and labor market discrimination: an
econometric analysis of the Atlanta labor market", Southern Economic Journal 55:123-140.
Ilhanfeldt, K.R. (1992), "Intraurban wage gradients: evidence by race, gender, occupational class, and sector",
Journal of Urban Economics 32:70-91.
Kim, S. (1995), "Excess commuting for two-worker households in the Los Angeles metropolitan area", Journal
of Urban Economics 38:166-182.
Macauley, M. (1985), "Estimation and recent behavior of urban population and employment density gradients",
Journal of Urban Economics 18: 251-260.
Madden, J.E. (1985), "Urban wage gradients: empirical evidence", Journal of Urban Economics 18:291-301.
McMillen, D. and L.D. Singell (1992), "Work location, residence location and the intraurban wage gradient",
Journal of Urban Economics 32:195-213.
Mills, E.S. (1972), Studies in the Structure of the Urban Economy. (Johns Hopkins University Press,
Baltimore).
Mills, E.S. and B. Hamilton (1989), Urban Economics (Scott, Foresman, Glenview, IL).
Mills, E.S. and R. Price (1984), "Metropolitan suburbanization and central city problems", Journal of Urban
Economics 15:1-17.
Muth, R. (1969), Cities and Housing (University of Chicago Press, Chicago).
Ogawa, H. and M. Fujita (1980), "Equilibrium land use patterns in a nonmonocentric city", Journal of Regional
Science 20:455-475.
Ogawa, H. and M. Fujita (1989), "Nonmonocentric urban configurations in a two-dimensional space",
Environment and Planning A 21:363-374.
O'Hara, J.D. (1977), "Location of firms within a square central business district", Journal of Political Economy
85:1189-1207.
Rees, A. and G.P Schultz (1970), Workers and Wages in an Urban Labor Market (University of Chicago Press,
Chicago).
Ross, S. and J. Yinger (1995), "Comparative static analysis of open urban models with a full labor market and
suburban employment", Regional Science and Urban Economics 25:575-605.
Sivitanidou, R. and W.C. Wheaton (1992), "Wage and rent capitalization in the commercial real estate market",
Journal of Urban Economics 31:206-229.
Small, K. and S. Song (1992), "'Wasteful' commuting: a resolution", Journal of Political Economy 100:888-
898.
Steinnes, D. (1977), "Causality and intraurban location", Journal of Urban Economics 4:69 79.
Straszheim, M. (1984), "Urban agglomeration effects and employment and wage gradients", Journal of Urban
Economics 16:187-207.
Sullivan, A.M. (1986), "A general equilibrium model with agglomerative economics and decentralized
employment", Journal of Urban Economics 20:55-75.
1412 M.J. White
Tauchen, H. and A.D. Witte (1984). "Socially optimal and equilibrium distributions of office activity: models
with exogenous and endogenous contacts", Journal of Urban Economics 15:66-86.
Thurston, L. and A. Yezer (1994), "Causality in the suburbanization of population and employment", Journal
of Urban Economics 35:105-118.
Wieand, K. (1987), "An extension of the monocentric urban spatial equilibrium model to a multi-center setting:
the case of the two-center city", Journal of Urban Economics 21:259-271.
White, M.J. (1976), "Firm suburbanization and urban subcenters", Journal of Urban Economics 3:323-343.
White, M.J. (1988a), "Location choice and commuting behavior in cities with decentralized employment",
Journal of Urban Economics 24:129-152.
White, M.J. (1988b), "Urban commuting is not 'wasteful' ", Journal of Political Economy 96:1097-1110.
White, M.J. (1990), "Commuting and congestion: a simulation model of a decentralized metropolitan area",
Journal of the American Real Estate and Urban Economics Association 18:335-368.
Yinger, J. (1992), "Urban models with more than one employment center", Journal of Urban Economics
31:181-205.
Chapter37
JOSEPH GYOURKO*
MATTHEW KAHN
Columbia University
JOSEPH TRACY
Contents
1. Introduction 1414
2. Theoretical underpinnings 1416
3. Recent econometric developments in estimating the quality of life 1422
3.1. Specification bias problems and the interpretation of capitalization results 1422
3.2. Group effects in the wage and housing data 1424
4. New estimates of the quality of life: results, comparisons, and
analysis 1429
4.1. Blomquist et al. (1988), "New estimates of the quality of life in urban areas" 1429
4.2. Gyourko and Tracy (1991), "The structure of local public finance and the quality of life" 1430
4.3. Stover and Leven (1992), "Methodological issues in the determination of the quality of
life in urban areas" 1433
4.4. Conclusions 1435
5. The value of the environment and the urban quality of life-a detailed
analysis 1436
5.1. Environmental hedonic prices: cross-city and within-city data and results 1436
5.2. Limitations of capitalization studies of environmental quality 1440
5.3. Econometric issues 1442
5.4. Alternative valuation methods 1443
* We thank Glenn Blomquist, Charles Leven and Jerry Hausman for helpful comments on previous drafts.
Naturally, the usual caveat still applies.
1413
1414 J. Gyourko et al.
Abstract
Recent research into the urban quality of life (QOL) is reviewed and analyzed,
with a special emphasis on the estimation of implicit prices of environmental
attributes. New work has incorporated traditional concerns of urban theory into
QOL analyses, as well as increased our understanding of specification bias prob-
lems in hedonic estimations. However, empirical research into the QOL finds
itself at a crossroads, as the large city-specific error components in the underlying
wage and housing expenditure hedonic specifications result in imprecise mea-
surement of overall QOL values and rankings. Amassing higher quality databases
to deal with this problem should be high on the agenda of those interested in this
research program.
1. Introduction
Measuring the quality of life (QOL) across urban areas has been an important
research program in urban economics, since Rosen (1979) and Roback (1980,
1982) showed how to identify implicit market prices of local amenities that can
serve as the weights in the construction of QOL indexes. Bartik and Smith (1987)
provide an excellent review of the early theoretical and empirical work from this
research project in the Handbook of Regional and Urban Economics (see Chap.
31, this volume). Our review takes up the story where theirs left off, beginning
with work published in the mid-1980s.
Ch. 37: Quality of Life and Environmental Comparisons 1415
Significant conceptual advances in this literature have been made over the last
decade. They include the incorporation of traditional concerns of urban theory
such as compensation for distance from the central business district (Hoehn et al.,
1987) and agglomeration effects (Blomquist et al., 1988) into the Rosen/Roback
framework. There also have been contributions showing how local governments
can influence the QOL through their own taxation and service provision decisions
(Gyourko and Tracy, 1989b, 1991). While interesting in their own right, these
theoretical advances have also helped improve our understanding of specification
bias problems, that are particularly important when the researcher's interest is on
specific capitalization results, rather than the overall QOL. In addition, this new
work has expanded our knowledge of the extent of differential capitalization in
land versus labor markets.
We also focus on research into the estimation of implicit prices of environ-
mental attributes. This work is particularly relevant to QOL research for several
reasons. First, the environment is a significant component of the local trait set.
Second, both cross-city and within-city data have been used to generate implict
price estimates, and the differences in results are helpful in understanding the
strengths and weaknesses of the empirical work on the overall QOL that relies
exclusively on cross-city data. In addition, the greatest experimentation with non-
hedonic valuation methods has occurred in the environmental area. Assessing the
relative strengths and weaknesses of the varying valuation methodologies yields
interesting insights into the QOL literature that, heretofore, has relied on revealed
preference techniques.
Despite all these developments, we believe that empirical research into the
QOL presently finds itself at a crossroads. This is because recent work by Gy-
ourko and Tracy (1989a,b, 1991) reports the presence of large city-specific error
components in the underlying wage and housing expenditure hedonic specifica-
tions estimated to determine the local trait prices. After carefully controlling for
these error components, it turns out that overall QOL values and rankings are
imprecisely estimated. The level of imprecision is such that much better descrip-
tions of local amenity and fiscal conditions, plus superior controls for housing,
worker and job quality, are needed to minimize the impact of these city-specific
error components. Amassing higher quality databases in the face of budget cut-
backs at the key data collection agencies of many national governments will be
a difficult and time-consuming process, but it should be high on the agenda of
those interested in such research.
Nevertheless, we are far from pessimistic about the prospects for this research
program. Progress can be made in other ways and on other fronts. One that we
discuss at length in Section 6 involves relaxing the key equilibrium assumption
1416 J. Gyourko et al.
underlying virtually all existing work in the area. The assumption that the re-
searcher is viewing local land and labor markets in states of long-run equilibrium
in any cross-section of data is what permits the equating of estimated trait prices
with their true market values. The violence done to QOL estimates by this very
convenient assumption is unknown, but Cheshire and Hay (1989) and Greenwood
et al. (1991) have taken the first steps towards informing us.
The chapter is organized as follows. A brief exposition of the Rosen/Roback
model is provided in Section 2. Section 3 then discusses the implications of
recent econometric developments in estimating the QOL. This is followed by
detailed analyses of three major studies that estimated the QOL across a number
of metropolitan areas. Section 5 provides a separate review of the role of the
environment in the QOL, as well as a discussion of alternative valuation tech-
niques. Section 6 then focuses on the challenge of integrating mobility into QOL
analyses. A brief summary concludes the chapter.
2. Theoretical underpinnings1
Amenities enter the indirect utility function only through their impact on a worker-
resident's utility. Services, in contrast, enter the indirect utility function both
through their impact on a worker-resident's utility and through their associated
impact on the gross- and net-of-tax prices faced by the worker-resident.
Assuming costless mobility and full information about the amenity and fiscal
attributes of each city, long-run equilibrium requires that the marginal worker-
resident be indifferent to her city location, 4 with wages and land rentals adjusting
so that,
where Yi is the total revenues that are a function of the city's amenity and fiscal
attributes through their impact on the production (and distribution) function, Li
2 For simplicity, all workers are treated as equally productive. Empirical applications include variables that
attempt to control for productivity differences across workers.
3 For simplicity of exposition, this specification of the budget constraint assumes that the state taxes only
wages. No fundamental conclusions are altered if more complex descriptions are employed.
4 If workers are homogeneous in preferences, then the arbitrage condition implies equalizing differences
that will be exact for all worker/residents. If workers are heterogeneous in preferences, then the equalizing
differences will be exact only for the marginal worker. See Roback (1988).
1418 J. Gyourko et al.
(I+t)n
·
V
(I+t)n.
W
W
Fig. 1.
represents the firm's labor usage, and Mi is the firm's intermediate input us-
age. 5 Given a city location, the firm's maximization problem yields the following
indirect profit function,
Again, assuming costless mobility and full information, the long-run equilib-
rium requires that the marginal firm is indifferent to its city location. This requires
that wages and land rentals adjust to impose the following arbitrage condition:
The long-run equilibrium wage and land rentals are found by solving the two
arbitrage conditions as illustrated in Fig. 's familiar representation.
The reduced form wage equation is obtained by isolating the gross-of-tax land
rental in Eq. (2.4) and Eq. (2.7), equating the two expressions, and solving for
Wg so that,
The reduced form equation for the gross-of-tax land rental (1 + tj)nj, which we
denote as Rj, is obtained in a similar fashion so that,
5 Equal residential and commercial property tax rates are presumed in Eq. (2.2) and Eq. (2.5), but the model
is unaffected if different tax rates are allowed. We also abstract from restrictive zoning that might introduce
a wedge between residential and commercial rents. In this sense, the model probably is more applicable to
larger, more heterogeneous cities than to smaller, relatively homogeneous suburbs.
Ch. 37: Quality of Life and Environmental Comparisons 1419
(l+t)n V"
n1
I'-
W*' W*'
Fig. 2.
The comparative statics of the model yield several insights into wage and
land rental capitalization. The first implication of the model is full capitalization
of property tax differentials into land rentals. That is, increases in the property
tax, t, that are not offset by added services or amenities (or other tax reductions)
are fully capitalized into land prices:
z,s.I,A,G
(2.10)
at IA =0.
-
z,s,I,A,G
Pure amenities by definition have no explicit market price. They are implicitly
priced in the labor and land markets through capitalization into wages and rentals
as shown in Eq. (2.11),
1420 J. Gyourko et al.
(1+t)n
.1
W's.=W*' w
Fig. 3.
VA VW IA (1-z)
anl| VR VR nW >
t,s,z,,G
(2.11)
aw -) [
VWOI-Z) J >>,0,
ts,, and
-VG + VW G (1 -z)
an VR VR W >0
t,s,z,l,A
-ci D
(2.12)
i F--VG++VRrGl
t,s,z,l,A
Two more comparative static results relate to changes in the income and sales
tax rates. Higher income (z) and sales (s) tax rates, holding service levels con-
stant, lead to lower land rentals. Since the income tax rate does not affect the
firm's indirect profit function, a higher rate also leads to higher gross wages. In
contrast, if firms use intermediate goods, then higher sales tax rates can lead to
higher or lower gross wages. These results are summarized in Eq. (2.13).
Vw Wg
an = VR (1 +t)
az t,s,l,A,G D
Wg
aw a(- (1-z)
-Z) >0,
7Z t,s,l,A,G D
(2.13)
an I = vVR VR I
W (ll- <
aS t,z,l,A,G D
aw - z [vS
vW(z nR I >
aW t,zI,A,G D
Full prices for amenities and fiscal attributes of cities are constructed from
these capitalization results as follows. Define Pk as the full price for the kth city
attribute, Zk, with,
an aw
Pk -e (2.14)
3Zk aZk
1422 J. Gyourko et aL
where e denotes the "exposure" of the marginal household to the labor market. 6
The next step in constructing a QOL index value is to use the equilibrium as-
sumption to equate prices with values. In this case, the value of city j is the sum
of its attributes multiplied by their full prices.
Obviously, the link between the price observed for an attribute and the value
placed on the attribute is central to the QOL methodology (see Evans, 1990), and
is a topic we return to at the end of the chapter.
All empirical work is subject to specification bias, and recent research suggests
that there are special concerns in QOL analyses because of the difficulty of
properly controlling for all taxes and the levels of effective service provision
(Gyourko and Tracy (1989a,b; 1991)). For example, the reduced form compara-
tive statics for government services parallel those for amenities only if all relevant
taxes are included in the reduced form equations. In addition, public sector rent-
sharing can bias the capitalization results for amenities by creating explicit prices
for them if local unions appropriate some of the locational rents through the
collective bargaining process via agreements that permit overstaffing or pay wage
premia.
Consider how the comparative statics for G would change if all taxes are not
controlled for in the reduced form wage and land rental equations. To keep the
analysis simple, let the locality vary only the local property tax rate in order
to adjust its tax revenues. This gives rise to the following relationship between
government services and the property tax rate.
an an an I at
aG In t,A at ,AGaG
an n at
aG t,,A (1 - t aG
(3.2)
aw aw aw at
oG
aGI - oG
,A aI,A, at
atW
at, G ac
aw
aG t,l,A
Note that the land-rental capitalization effect now involves two terms. The first
reflects the pure capitalization of the government services holding their explicit
tax prices constant. The second captures the compound effect of a pure property
tax rate increase holding services constant multiplied by the increase in the prop-
erty tax rate due to the higher level of services. Given the increase in property
tax rates due to the higher service level, the sign of the second term is negative
for land rentals and zero for wage rates (because the property tax rate effect
is fully capitalized into land-rentals with no spillover to local wages-see Eq.
(2.10)). Thus, the second effect acts to offset the first for land rentals, lowering
the observed level of capitalization. However, there is no similar effect on wage
capitalization. A consequence of the failure to fully control for taxes, then, is to
bias the relative capitalization effects of locally produced services toward wages.
Bias can also result from failing to properly control for all taxes when public
sector unionization leads to the sharing of locational rents between union mem-
bers and other community residents. Assume that unions attempt to organize in
areas where the potential return is the highest, so that the level of public sector
unionization increases on average with the size of the locational rents. Further-
more, let amenities comprise a significant component of locational rents. Then, as
we increase the level of amenities by moving across cities, we also are increasing
the level of public sector unionization, a U/ a A > O,where U represents the local
1424 J. Gyourko et al.
The data used in QOL studies typically involve microlevel observations on many
workers and homes within a given labor or housing market area. Researchers
need to be cognizant of the possibility that the wage and house price hedonic
regression residuals may contain city-specific error components (i.e., group ef-
fects). Ordinary least squares (OLS) estimations, that do not control for these
group effects, will lead to reported standard errors of the trait prices that are
biased downward (see Moulton, 1986, 1987). Naturally, this is important for
testing specific hypotheses about capitalization, and we show below that it is
also relevant for calculating the standard errors of the QOL rankings.
Consider the following specification forwages for worker/resident i in city j,
where Hi is a vector of individual housing structural traits, and vij is the compos-
ite error term. The composite error term again is the combination of a
city-specific error component, j, and a house-specific error component, qri. The
city-specific error component is common to all houses in the city, and repre-
sents systematic uncontrolled differences in house quality across cities, system-
atic uncontrolled differences in amenity/fiscal characteristics across cities, and/or
common demand/supply shocks to the local housing market.
If c 2 > 0 and o 2 > 0, then the composite error terms across workers and
across houses within the same city are correlated, violating the OLS indepen-
dence assumption. The magnitude of the bias imparted to the standard errors of
the estimated trait prices when independence is violated depends on the "design
effects" present in the data (Kish, 1965). The design effect for the jth city is
defined as follows:
where mj is the number of workers (houses) in the sample from the jth city, and
p is the common correlation coefficient between the composite errors for workers
(houses) within the city. In our example, these correlations are given by the ratio
of the variance of the city-specific error components to the total variance of the
composite error terms.
The simplest case to analyze is where the size of each city is the same and the
righthand side variables vary only between groups (i.e., there are only Z variables
in the model). In this case, the ratio of the true variance of fi2 to the OLS variance
estimate equals the design effect. 8 The presence of city-specific error components
generates design effects greater than one. The extent of the downward bias in the
8 See Tuenkloek (1981) for the derivation of the correct variance/covariance matrix.
1426 J. Gyourko et al.
OLS-based standard errors depends on the size of the city groups in the data and
the size of p.9
A second implication of the presence of group effects is that the definition
of a city's QOL becomes ambiguous. Assume the city-specific error components
represent left-out amenity/fiscal attributes of a city. If data were available on
these attributes, then implicit prices could be estimated and the attributes would
be included in the calculation of the QOL. In the absence of data on the attributes,
their impact on a city's QOL could be determined as follows:
( m j j) 2 -
E _ u 2
j i j
LM = 1/2' (3.8)
o2 2 2- N)j
11 Very briefly, since V(p) is a positive semidefinite matrix, there exists a unique lower triangular matrix
C such that V(B) = CC'. Let pi denote the ith simulated coefficient vector and ~i a vector (with dimension
matching ) of standard normals. Then /B' can be simulated as, fi' = + C i, where B are the original
hedonic coefficient estimates. The number of simulations should be increased until the implied standard errors
of the QOL rankings do not change at an acceptable level of significance between separate simulation exercises.
1428 J. Gourko et al.
12 A plausible story is that workers of higher quality will demand higher quality housing. This would imply
a positive correlation between the city en-or components. Left-out amenity/fiscal variables that are valued by
workers but not firms would imply a negative correlation. However, if firms also value the amenity/fiscal
variables, then the sign of the correlation is indeterminate.
Ch. 37: Quality of Life and Environmental Comparisons 1429
contaminated with measurement error. Panel data, then, do not provide a useful
alternative to improved cross-sectional data.
Over the past decade there have been three major QOL estimation projects within
the Rosen/Roback tradition that use market prices of local traits as weights in in-
dex construction: Blomquist et al. (1988), Gyourko and Tracy (1991) and Stover
and Leven (1992). Other recent efforts to estimate compensating differentials
arising from city attributes include Hoehn et al. (1987), Roback (1988) and Voith
(1991). However, they are not discussed here because they did not compute a sin-
gle QOL index value for each locality. In addition, research on other approaches
(e.g., Cheshire and Hay (1989)) and on mobility and the QOL (e.g., Greenwood
et al. (1991)) is dealt with in Section 6. Various organizations also publish QOL
rankings that use survey data or some ad hoc weighting scheme. Boyer's (1983)
PlacesRated Almanac and the Mobil Oil Corporation's 1989 Mobil Travel Guide
are good examples of that genre. 13
4.1. Blomquist et al. (1988): "New estimates of the quality of life in urban
areas"
This study spurred renewed interest into research on the urban QOL. The empiri-
cal work is based on a large cross-section of 253 counties within 185 metropolitan
areas in the US taken from the 1980 Census of Population and Housing. The
authors' underlying theoretical model is one in which each urban area has two
counties, and they derive new and interesting comparative static results with re-
spect to the impact of agglomeration effects associated with productivity effects
of the size of the urban area. While Blomquist et al. (1998) report substantial
intrametropolitan area differences in the QOL for 10 of the 38 urban areas for
which they have observations on multiple counties, it is unclear how having
amenity variation both within and across urban areas impacts on either the in-
dividual trait valuations or the overall QOL rankings. It would be useful to know
if the price of (say) their superfund site trait would be substantially different if
13 It should be noted that the three studies to be reviewed below all estimate reduced form wage and
housing hedonic price equations. Consequently, they do not test the basic assumption of the Rosen/Roback
model that labor and land markets are interrelated. Haughwout (1993) estimates structural versions of these
equations using three-stage least squares and finds strong evidence supporting the Rosen/Roback assumption.
The elasticity (standard error) of the wage with respect to changes in housing costs is 0.59 (0.02), while the
elasticity of housing costs with respect to the wage is 0.73 (0.02).
1430 J. Gyourko et al.
they had only used metropolitan area-level data for the 185 urban areas in their
sample, and whether any such differences were large enough in aggregate to
materially impact the QOL values or rankings of these metropolitan areas.
Blomquist et al. (1998) amassed data on 13 local traits, that included 11 cli-
mate and environmental attributes that could be considered pure amenities, and
two others (a teacher-pupil ratio and violent crime rate to proxy for the quality
of educational services and public safety, respectively) that are better described
as locally produced characteristics. Their six environmental traits included a rich
set of controls for superfund sites, landfills, and waste treatment and discharge
sites that had not been used in previous QOL analyses.
The authors employ a Box-Cox search procedure in estimating reduced form
hedonic wage and housing expenditure equations within an OLS framework. The
results indicate that their 13 local traits are jointly significant, both statistically
and economically. The full range of QOL values is $5146, implying that in equi-
librium the marginal household residing in the lowest rated county (St. Louis,
MO) requires that amount in terms of higher wages and/or lower housing expen-
ditures per year to be indifferent to living in the highest rated county (Pueblo,
CO).
Their results for individual trait prices almost always have the correct sign in
the sense that locating in an area with more of some obvious "bad" is associated
with compensation in the form of higher wages and/or lower land prices. Their
pollution-related findings are discussed in some detail in the next section, and we
refer the readers interested in other specific trait prices in this (and the other two
articles) to the papers themselves.
Blomquist et al. (1988) also report substantial labor market capitalization of
the local amenities, suggesting that firm and worker competition for scarce sites
does impact on both local land and labor markets. They also present the results of
calculations of subindexes based on subsets of variables. Those findings strongly
suggest that a top- (bottom-)ranked county need not be highly (poorly) rated on
all amenity dimensions. For example, the rank correlation between the subindex
based on their seven climate controls and the overall index is 0.63. The analogous
correlation between their environmental subindex and the overall index is only
0.21.
4.2. Gyourko and Tracy (1991), "The structure of local publicfinance and the
quality of life"
This paper uses the same 1980 Census of Populationand Housing data and many
of the same climate, but not environmental, amenity variables in Blomquist et al.
Ch. 37: Quality of Life and Environmental Comparisons 1431
(1988). An important difference in the two studies is that the geographic unit of
observation in Gyourko and Tracy (1991) is the central city of the metropolitan
area. This was chosen so that a richer set of tax and service fiscal controls could
be matched to a specific political jurisdiction.14 Gyourko and Tracy (1991) also
differs from Blomquist et al. (1988) in the use of a random effects estimator to
account for potential group effects in the data.
Despite these differences in data, specification and econometric technique,
the signs on the individual trait prices, especially the pure amenity variables,
are almost always the same as in Blomquist et al. (1988). The amenities also are
jointly significant in statistical and economic terms in both papers. And, both sets
of results indicate that substantial compensation for amenity differentials occurs
via wages in the labor market. The full range of QOL values across the 130 cities
in Gyourko and Tracy's (1991) sample is $8227. This is larger than in Blomquist
et al. (1988), but outliers drive that difference. The interquartile range of QOL
values is only $1484.
An important contribution of this paper is its documentation that local fiscal
environments also are important determinants of the QOL, suggesting that cities
may have more control over their local QOL than was previously thought. Con-
sistent with the basic theoretical prediction of the expanded model, variations
in effective property tax rates appear to be fully capitalized into land prices.
Except for the impact of corporate taxes, their results provide empirical support
for the tax capitalization comparative statics in Eqs. (2.10) and (2.13). The full
prices estimated for police, fire and health services each have the theoretically
anticipated signs, with the police and health service proxies being statistically
significant. Of the service vector variables, only the price of the student-teacher
ratio has the wrong sign and its value is very low (-$27).1 5 The impact of rent-
sharing with successful public sector unions on the local QOL is found to be
small on average, but other work suggests substantial capitalization effects into
land prices for cities with very high unionization rates among their public sector
employees. 16
14 The material differences in fiscal conditions that existed between a central city and its typical suburb
in 1980 (not to mention the extensive heterogeneity across suburbs) would have led to significant errors-in-
variables problems had counties or metropolitan areas been the geographical unit of observation.
15 Unfortunately, the different data, geographical units of observation, and specifications make it infeasible
to do a simple comparison of the full prices of the two service controls common to Blomquist et al. (1988)
and Gyourko and Tracy (1991). Such a comparison would be helpful in order to gauge whether, or not, the
estimated full prices of produced services are smaller when taxes are not controlled for, as the model outlined
in Section 2 implies should be the case.
16 What little capitalization Gyourko and Tracy (1991) do find occurs in the land market, not the labor
market. Gyourko and Tracy (1989b, c) focus solely on the impact of local public sector unionization on land
1432 J. Gyourko et al.
The interquartile range for the aggregate contribution of the seven tax/service
fiscal variables in Gyourko and Tracy (1991b) is $1188, compared to $1372
for the 11 pure amenity variables. In addition, they find that fiscal differentials
account for at least one-fifth of the variation in quality-adjusted housing expendi-
tures that can be explained by all city-specific variables, and at least one-half of
the variation in quality-adjusted wages that can be explained by all city-specific
variables.
In an OLS regression estimation done for comparison purposes with the pre-
vious literature, Gyourko and Tracy find that including the fiscal and public union
controls is associated with a 16.2 mean absolute change in ranks compared to a
specification that only includes pure amenities. 17 The bias, then, on the overall
QOL ranking from omitting the fiscal variables is substantial.
Perhaps even more important are the findings of this paper suggesting how
fragile the QOL rankings estimates are. Moving from an OLS to a random effects
estimation is associated with a 10.2 mean absolute change in ranks. In QOL value
terms, the mean change is $391 and the standard deviation about that mean is
$320. Not only are the rankings different when group effects are controlled for,
they are much less precisely estimated.
The issue of how to treat group effects increases the uncertainty regarding
the reliability of the results. Gyourko and Tracy (1991) computed QOL rankings
that presumed the group effects were entirely composed of omitted city charac-
teristics. When compared to the random effects-based results that did not price
out the group effects, the mean absolute change in ranks associated with fully
pricing these group effects was 27, with a standard deviation of 20 about that
mean change. In general, the largest decreases (increases) in a city's rank were
due to observed wages being higher (lower) than predicted.
values, using different metrics for identifying cities that are likely to be subject to successful rent-seeking.
Gyourko and Tracy (1989c) look at median house prices for 36 cities using the 1976American Housing Survey
data. Rent-sharing cities are defined as those whose local public sector employees were paid wage premia in
excess of one standard deviation above the mean, with the wage premia being calculated using microdata on
public sector workers drawn from the May 1977 Current Population Survey. Median house values were found
to be 28% lower in the rent-sharing cities. Gyourko and Tracy (1989b) examined median house prices in a
larger sample of 90 cities using the 1983 County and City Data Book information. Rent-sharing cities in this
study were identified as those whose public sector union organization level was at least one standard deviation
above the average (i.e., above 67%). This study reports that house values were 12% lower on average in the
rent-sharing cities. We interpret the findings of these three papers as indicating that relatively little rent-sharing
with local public unions occurs on average (at least between 1976 and 1983), but that there are significant
effects on home prices in very highly unionized towns.
17 The standard deviation about that mean change was 15.2.
Ch. 37: Quality of Life and Environmental Comparisons 1433
4.3. Stover and Leven (1992), "Methodological issues in the determination of
the quality of life in urban areas"
Stover and Leven (1992) work within the Rosen/Roback framework to derive an
alternative functional form that is used to reestimate QOLs using the Blomquist et
al. (1988) data. A key motivation of their approach is concern about the reliability
of hedonic wage equation estimates due to the difficulty of controlling for job and
worker characteristics. Stover and Leven (1992) also believe that the dependent
variable in a housing hedonic is superior because it measures the service flow
from the durable good, while wages understate a worker's total compensation.
These authors assume that both land and labor markets are in equilibrium
(a standard Rosen/Roback assumption), and that each market fully reflects the
values of local amenities as well as conditions in the other market. This gives rise
to wage and land rental specifications that take the following form: n = n{Z, WI
and W = w{Z, n}. These specifications differ from the reduced form specifica-
tions described earlier in that they include the price from the other market.
Stover and Leven (1992) demonstrate that the full price for the kth city at-
tribute, Zk, based on their single equation housing expenditure hedonic, is given
by the following expression:
dn
Pk = -nwwn -nWWZk, (4.1)
dZk
where the first term, -nwwndn/dZk, represents the direct impact of Zk on land
rentals. The second term, -nw wz, is the indirect impact of Zk on land rentals and
reflects feedback from the labor market.18
In interpreting the direct impact of Zk on land rentals, the partial derivative
nw reflects a pure income effect on the demand side because only amenities are
compensating in their model. Thus, it should be greater than zero if housing is
a normal commodity. The partial derivative w, is assumed to be positive as it
reflects a compensating differential for the increased cost-of-living associated
with an increase in land rents. Naturally, they are interested in land rentals that
rise due to a better amenity package (dnldZk). In interpreting the indirect impact
on land rentals reflected in the second term on the righthand of Eq. (4.2), an
increase in some Zk affects wages as shown in the partial derivative wz. This
feeds back into the land market through the n, term discussed above.
18 Obviously, a single wage hedonic also could be specified, but that is not the preferred option given Stover
and Leven's worries about the quality of wage data and job/worker trait controls.
1434 J. Gyourko et al.
Lj = Wij- ,(4.3)
pj .E(Wi)3
Li : -1 . Xil + Zj a2+
j+ i]
Pi I L xil
!9 Stover and Leven (1992) also estimate QOL values using a single wage hedonic (with an added control
for rent premiums analogous to L). The rankings generated from that estimation are negatively correlated
(Spearman rank correlation of -0.57) with those from their housing expenditure hedonic. This is troubling
given that their model predicts identical rankings from the two specifications. If the problem is that hedonic
wage estimates are noisy as the authors suggest, that would suggest little or no correlation between the rankings
generated from the housing versus wage specifications. It does not account for the strong negative correlation
between the two.
Ch. 37: Quality of Life and Environmental Comparisons 1435
1 1 1 Ei
= 1 + (Zj2 + j)
Pj Xil ++ pj Eli
.XiPi
Pj l' Xi
where the last step uses the fact that E(e) = 0 and that is independent of X.
Equation (4.5) shows Stover and Leven's (1992) "wage premium" variable to
be a function of the wage equation group error term (aj) and the city-specific
variable effects on wages (Zj 2 ). Note that misspecification of the wage regres-
sion through inadequate controls for worker quality will lead to a misspecifi-
cation of Lj if workers sort by quality into different housing markets. 2 0 Fur-
thermore, Stover and Leven are implicitly assuming that the a's reflect left-out
amenity/fiscal characteristics. If the a's arise from common demand and/or sup-
ply shocks to the local labor market or if they reflect systematic mismeasurement
of worker/job quality, then Li will not measure the pure impact of amenities on
wages as assumed in their model. Finally, if the group error components from the
wage and land rental regressions are correlated, then Li is endogenous, leading
to further bias in the QOL estimates.
4.4. Conclusions
Recent empirical work shows that there is still no quick technical fix to the prob-
lems of large group effects in the data or hedonic wage and housing equations
estimated with imperfect worker, job, and housing quality controls. These really
are data problems and probably will have to be solved by amassing higher quality
data. Given the considerable effort that Blomquist et al. (1988) and Gyourko and
Tracy (1991) expended to construct their local databases, and the fact that better
worker, job, or housing quality controls are not likely to be produced in an era
of declining budgets at key data collection agencies, the likelihood of superior
data soon appearing to solve the problem that empirical literature finds itself in,
seems low.
Amid that pessimistic conclusion, we can suggest a diagnostic that should
help in distinguishing whether, or not, the group effects reflect, on balance, the
unobserved heterogeneity in worker and housing quality or unmeasured city
traits. We begin by using the estimated hedonic coefficients on worker and hous-
ing quality to construct an estimate of the average observed quality of workers
5. The value of the environment and the urban quality of life-a detailed
analysis
Modem societies' growing concerns about the environment makes research into
the value of environmental attributes worthy of interest in its own right. Most
important for this chapter are the significant implications and insights this bur-
geoning literature provides for research into the urban QOL. More varied data
have been used to estimate the prices of environmental attributes. Section 5.1
presents a comparison of results generated from cross-city data typical of that
used in QOL research with those based on within-city data. In addition, the
environmental literature includes price estimates for a wider variety of attributes
than have been controlled for in the QOL literature. The findings here suggest
that QOL researchers should expand their list of environmental attributes to be
priced. Finally, there has been greater experimentation with nonhedonic valua-
tion methods in the environmental area. This provides the opportunity to assess
the benefits and costs of valuation methodologies different from the revealed
preference techniques followed in the QOL literature.
5.1. Environmental hedonic prices: cross-city and within-city data and results
21 Environmental attributes comprise roughly half of Blomquist et al.'s (1988) location-specific public
goods. They include particulates, superfund sites, effluent discharges, landfill waste, and treatment storage
and disposal sites. Of this broader set of environmental controls, Gyourko and Tracy (1991) only include a
control for particulate matter. However, Gyourko and Tracy (1991) report that the remaining five Blomquist et
al.'s environmental variables are jointly insignificant in a random effects specification estimated on a subset of
90 cities.
22 An earlier reference point is Roback's (1982) cross-city study based on the 1973 Current Population
Survey. Particulate levels were significantly higher in 1973 than in 1980. If people have preferences that feature
diminishing marginal returns with respect to air quality, then we would expect that the hedonic price would
have been higher in a study based on the 1970s data than the 1980s data. Roback (1982) reports OLS estimates
in which particulates are significant in two out of her four specifications, but the t-statistics indicate that all
four estimates would likely have been insignificant if city-specific group effects had been estimated.
1438 J. Gyourko et al.
monitoring stations was 14.2. Thus, these hedonic particulates estimates are very
small compared to the climate estimates.
In addition to climate and air quality variables, environmental researchers
have used the cross-city approach to measure the capitalization of hazardous
waste sites. For example, Clark and Nieves (1994) employed the 1980 Census of
Populationand Housing, to find that proximity to a superfund site is an positive
amenity with a price of $58. Blomquist et al. (1988) report that it is a disamenity
with a price of -$107-using the same data.
In sum, these cross-city studies provide strong evidence that environmental
variables are capitalized, but also suggest the price estimates are highly sensitive
to specification-an issue emphasized in Graves et al. (1988).23
The most obvious alternative to cross-city analysis is to use data for a single
city. If available, the latter is attractive because the researcher typically has access
to more disaggregated geographic information. For example, census tract-level
data allows the researcher to merge on tract-level environmental exposure, reduc-
ing the measurement error problems affecting inference from cross-city studies.
Another advantage is that working within a single city controls for local business
cycle effects (Topel, 1986).
One obvious disadvantage of within-city data is that only the prices of at-
tributes that vary within the city can be identified. Another is that intracity hedo-
nic wage regressions cannot be estimated because detailed data on where people
work within the city generally is unavailable. A third potential problem with
intracity studies is spatial sorting on unobservables. The best quality homes may
be in the best city neighborhoods. The typical hedonic rental study based on
the 1980 Census of Population and Housing data has rather crude controls for
housing structure. The data do not indicate the actual square footage or condi-
tion of the unit. If suppliers build nicer units in terms of unobservables in the
nicer parts of the city, then the econometrician will overestimate the value of the
QOL. Moreover, low environmental quality in a neighborhood may proxy for
low quality of housing structure.
A large number of within-city hedonic house price regression studies have
focused on the price of particulates. Smith and Hwang (1995) use meta-analysis
techniques to summarize multiple studies of housing capitalization between 1967
and 1988, with each city-specific hedonic home regression study representing a
data point to explain variation in estimated prices across cities over time. Based
on their sample of over 80 hedonic studies, Smith and Hwang (1995) report
23 Rosen (1979) stressed that multicollinearity between local public goods decreases the likelihood of
separating out marginal effects of individual local public goods.
Ch. 37: Quality of Life and Environmental Comparisons 1439
a median reduction in home prices of $20, and a mean reduction of $100 per
microgram per cubic meter ($1980).
Hedonic home regressions also are a natural methodology for quantifying
how capitalization of an attribute such as proximity to hazardous sites changes
over time. This environmental attribute has been examined using both cross- and
within-city data. The cross-city approach does not control for distance from the
site. Instead, a count of total sites within a city's borders is constructed. Thus,
noxious sites are assumed to be a pure local public bad (e.g., Blomquist et al.,
1988 and Clark and Nieves, 1994). Kohlhase (1991), Kiel and McClain (1995)
and Michaels and Smith (1990) offer intracity studies to test this. Given the size
of MSAs, the impact of such noxious facilities seems more likely to be seen in
intracity studies.
Kohlhase (1991) is particularly instructive in this regard. She uses the Houston
housing data of 1976, 1980 and 1985 to explore how the coefficient on distance
from a hazardous waste site changes over time as new information about the
site is revealed. Using price data before and after the hazardous site has been
placed on the National Priority List (NPL) in 1985, she shows that distance
was not a valued amenity until after the site was placed on the NPL. After the
announcement, distance became an amenity. The peak marginal price of an extra
mile of distance was $2364, which declines to zero at a distance of 6.2 miles. Her
findings illustrate the importance of not simply using how many toxic sites are in
a county as a proxy for toxic waste exposure-as is typically done in cross-city
studies.2 4
The Kohlhase (1991), Kiel and McClain (1995) and Michaels and Smith
(1990) studies employing within-city data have also made progress in investigat-
ing how "new news" about site toxicity gets capitalized differentially into hous-
ing prices at different radius distances from the site. Repeated cross-sectional
regression yields insights into the speed with which the perception of environ-
mental hazards is capitalized into price. The siting of a hazardous site can change
a community's QOL, induce outmigration, and depress home prices.
Kiel and McClain (1995) studied how the siting of an incinerator in north
Andover (Massachusetts), affected the prices of 2600 single family homes. They
collect data on the prices of homes sold in the area between 1974 and 1992.
24 Kohlhase (1991) also reports the interesting finding that a site's rank on the NPL list is not capitalized.
That is, higher ranked sites do not command a deeper discount. Kohlhase interprets this as evidence that people
are unable to differentiate information. A different possible interpretation is that being ranked high on the NPL
list has two offsetting effects. A high rank should indicate that your home is near a more dangerous site, but
counterbalancing this is an increased likelihood that the site will be cleaned up before sites that are lower on
the list.
1440 J. Gyourko et al.
They then partition calendar years 1974-1992 into several subintervals. The first
interval represents the prerumor (no site) stage. In the next stage, news of the
proposed project leaks, so that with positive probability the incinerator will be
sited in the vicinity. Next the project is constructed but is not online. In this phase
the probability of start-up is one, but the exact timing is uncertain. Following this
stage, the plant comes online. The impact of each of these stages is explored
by estimating a hedonic specification for each. Distance from the incinerator is
statistically insignificant through the rumor stage, and is insignificant at the 5%
level in the construction phase. Only when the incinerator is online and ongoing
is it capitalized. In addition, Kiel and McClain (1995) find a larger peak impact
than Kohlhase (1991) at $8100 per mile. 25
Only a few hedonic studies exist of compensation for water pollution, un-
like air quality or the location of superfund sites, no national monitoring system
exists to measure water quality differentials across space. 2 6 Feenberg and Mills
(1980) use the Harrison and Rubinfeld (1978) within-city data on the Boston
area and augment their specification to include proxies for water pollution that
include the water's oil level and turbidity. They find that both have a statistically
significant impact on home prices. Surprisingly, Feenberg and Mills do not find
strong evidence that beach-front property values decrease the most when nearby
water becomes polluted. At the cross-city level, Blomquist et al. (1988) proxied
for water pollution using NPDES effluent discharges and found that it lowered
rentals. Both Gyourko and Tracy (1989b, 1991b) and Blomquist et al. (1988)
found a large coastal affect in their cross-city studies. 2 7
In estimating Eq. (3.4), researchers assume that people take the stock of envi-
ronment in a location as exogenously determined. For climate, this is certainly
a reasonable assumption. Yet environmental quality, such as clean streets and
low smog, may be produced through high expenditures on local services and
25 Use of within-city data also permits examination of an interesting incidence question-whether, or not,
proximity to a hazardous site has different effects depending on the quality of the house. In a study of the
Boston real estate market, Michaels and Smith (1990) examine whether, or not, the distance from a hazardous
site (interacted with whether, or not, it has been placed on the NPL list) is differentially capitalized into four
submarkets based on home quality.
26 Water regulators have focused on measuring whether individual polluters are in compliance with
emissions rather than with measuring spatial variation in water quality.
27 In addition to quantifying the value of reduced air, waste and water pollution, the environmental QOL
literature has also focused on the impact of low probability environmental outcomes. Interesting examples of
such small probability events are earthquakes or cancer caused by proximity to electric lines. See, for example,
Brookshire et al. (1985) for an analysis of earthquake risk premiums in San Francisco and Los Angeles.
Ch. 37: Quality of Life and Environmental Comparisons 1441
resulting higher local taxes. Examples include frequent garbage collection, large
expenditures for vehicle emissions testing, or transfers to high polluting junk-car
owners in return for scrapping the vehicle (see Kahn, 1995). Unfortunately, no
QOL or other hedonic papers have explicitly modeled that environmental quality
is a byproduct of economic activity, regulation and citizen actions. 2 8
The implications for the QOL is that the researcher may not estimate the full
value for environment if the empirical specification ignores taxes and limitations
on behavior. Some environmental attributes are produced through costly actions
by local populace, with citizens paying via taxes, lost time or lost utiles due to
limitations on one's activities. If such unfunded mandates are not demanded by
the local citizens, then in the extreme case, wages could be high and rents low in
clean areas to compensate for the pain of achieving the standard.
In a similar vein, citizens have gained more control over their environmental
consumption through the growth of explicit markets. The Clean Air Act Amend-
ments of 1990 created a pollution-permit market in sulfur dioxide. Before this
market existed, real estate prices around the Adirondacks in upstate New York
might be lower because of Ohio-produced acid rain. If citizens in this area could
buy pollution permits from Ohio power plants and retire them, this yearly expen-
diture would lead to improved acid rain levels in New York. However, rents may
not rise by much because people who move there would recognize that they must
pay a per-capita share for the permits. If permits expand for other environmental
factors such as vehicle emissions, implicit capitalization should fall and hedonic
techniques would only reveal a fraction of the full payment for environmental
goods.
This discussion also highlights the limitations of hedonic prices for issues
such as policy analysis. Simply put, the hedonic price indicates how much one
pays, conditional on a spatial distribution of environmental quality. It provides
no insight as to who paid explicitly or implicitly to achieve any given spatial
distribution of environmental quality. Only if the distribution of environmental
quality were exogenously determined would the hedonic price indicate total envi-
ronmental payment. When environmental quality is produced through regulation,
taxes and limitations on individual behavior, or increased international trade with
nations that specialize in the polluting sector, then it is quite possible that "hedo-
28 For example, changes in air quality may reflect transport regulation or the decline of manufacturing.
Henderson (1996) and Kahn (1997) document the relationship between air quality levels and manufacturing
activity. If industry declines over time in a major city, then air quality could increase sharply while house prices
in a vicinity around these plants might decline due to the decrease in employment opportunities.
1442 Z Gyourko et al.
owners. However, there certainly are environmental amenities that will be ex-
pected to change over time (e.g., noxious facility sitings and water or air quality).
Given data on rentals and home prices for the same geographical locations, future
work should separately estimate rental and home price regressions and explicitly
test the cross-equation restrictions that the implicit prices are equal.30 Datasets
such as the Toxic Release Inventory could be used to study how new information
about exposure to environmental hazards is capitalized in the rental and housing
markets.
33 An interesting extension of this merger between stated willingness to pay through contingent valuation
and hedonic estimates would be to survey how people actually vote on environmental initiatives (Deacon
and Shapiro, 1975; Kahn and Matsusaka, 1997). With microdata on individual voting patterns and a person
characteristics, a structural discrete choice model could be fitted to obtain estimates of the marginal willingness
to tradeoff income for publicly provided environmental goods.
34 Portney and Mullahy (1986) report estimates of ozone's t-statistic that vary between 1.97-3.2. While
apparently statistically significant, the authors are likely to have underestimated their regression's standard
errors because they have not controlled for the fact that observations from the same MSAs are likely to have
correlated error terms (Moulton, 1986).
Ch. 37: Quality of Life and Environmental Comparisons 1445
Krupnick et al. (1990) create a 1979 database consisting of 290 families who
have children in school. These families kept a diary for 182 days. The authors'
goal was to quantify daily ozone's impact on the probability of having a respi-
ratory symptom the next day. The authors aggregate 19 respiratory diseases into
a single dummy variable indicator and estimating a Markov transition matrix
of the probability of having a respiratory condition on any day as a function of
air quality, personal characteristics (such as smoking), and previous day's health
and time spent inside. The authors find a small, statistically significant effect
that a 1% increase in ozone raises the frequency of symptoms by 0.11%. In
1979, the average Los Angeles resident suffered 76 days of having some cold
symptom. Their 0.11 elasticity indicates that if ozone levels fell by 50% between
1980 and the present, then the average Los Angeles citizen has experienced a
4.2 day reduction in respiratory symptoms. If individuals are willing to pay an
average of $25 to avoid such an episode, then the per-capita yearly benefits of
the reduction ozone would be $100 each. The relationship between direct health
study estimates and hedonic valuation estimates would be an interesting research
path.
A second literature directly connecting health and pollution has studied pol-
lution's impact on seeking medical care (Gerking and Stanley, 1986; Dickie and
Gerking, 1991). These authors estimate discrete choice models of whether, or not,
an individual sought medical care in the last year. Controlling for demographics,
preexisting conditions and the price of health care, the authors document the
positive impact of ozone on seeking medical care. Gerking and Stanley (1986)
find that a bid of $24.5 engendered a 30% reduction in ozone, based on the St.
Louis sample. Dickie and Gerking (1991) use Los Angeles data in Glendora and
Burbank to estimate a similar medical care demand probit with the econometric
innovation of allowing for person-specific random effects. The sample consists
of full-time workers who are not current smokers and oversamples people with
respiratory problems. Dickie and Gerking (1991) find that reducing Glendora's
ozone pollution from 117 days to 0 pphm willing to pay $170-210 dollars a year.
While health-based methods have increased our understanding of pollution
valuation, they complement but should not replace the hedonic approach. Es-
timation of health production functions requires representative samples of the
population and information on population self-protection. For example, if the
econometrician does not observe that people stay inside on highly polluted days
then she may conclude that ozone has very low costs for society as measured by
sickness. The health production literature has not modeled individual locational
choice and population sorting. Such sorting is a function of the equilibrium hous-
ing price gradient and individual preferences. A further issue with health studies
1446 J. Gyourko et al.
is that to create a proxy for health capital (the dependent variable) is not easy.
Krupnick et al. (1990) aggregate over 15 different health conditions into a single
health proxy. If the person's specific losses from these various conditions differ,
then this increases the difficulty of assigning a dollar value to the health losses
from increased pollution exposure. Given estimates of pollution's health impact,
a researcher would still need to impute the value of lost time and borrow from
the value of life literature to arrive at the dollar cost of pollution.
A key feature of the QOL methodology is the assumption that measured implicit
prices reflect marginal valuations for marginal worker/households in the data.
This assumption hinges on the equilibrium assumption that these worker/house-
holds are indifferent to their choice of location at the current implicit prices.
Given the methodological importance of this assumption, it is vital to test its
validity, adjust price estimates for any possible departures from equilibrium, and
explore estimation strategies that do not rely on the equilibrium assumption.
Few papers have addressed this issue. Cheshire and Hay (1989) propose a
new methodology that does not rely on the equilibrium assumption. As in the
traditional hedonic approach, their first step is to identify a set of Z variables
that are believed to affect the QOL. Their next step involves a significant depar-
ture from the hedonic approach in that wages and house prices are not used to
calculate weights for the Z variables. Instead, a group of "experts" is selected
who are asked to identify cities having the "worst" and "best" urban environ-
ments.35 Discriminant analysis is used to determine weights for the Z variables
that best explains the experts' classifications. The coefficients from the discrim-
inant analysis were then used to create an index for all cities in the sample. This
approach remains agnostic as to how the experts form their opinion about the
cities they are asked to classify. An important issue for study is how sensitive
the weights are to the particular set of experts selected, as the Cheshire and Hay
(1989) methodology represents a significant departure from the original goal of
Rosen/Roback to rely on market data rather than expert opinion.
Researchers who wish to continue in the Rosen/Roback tradition will need to
develop tests of the equilibrium assumption, and possibly methods for relaxing
the assumption. The first step in that process involves a careful understanding
of the likely implications of the equilibrium assumption, which would seem to
35 European experts were asked to classify cities in their home country.
Ch. 37: Quality of Life and Environmental Comparisons 1447
suggest that systematic patterns of intercity migration should not appear in the
data. However, the Rosen/Roback framework does not assume that households
are homogeneous in their preferences and/or exposure to the labor market. Thus,
the presence of substantial intercity household migration does not necessarily
indicate that the key underlying equilibrium assumption is being violated.
In fact, the equilibrium price assumption is consistent with three life-cycle
motivations for migration (see, Linneman and Graves, 1983). First, household
preferences for amenities and services such as mild weather and education ser-
vices are likely to vary over the life-cycle, so that if attribute prices remain
relatively constant through time, certain households may relocate to a new city
in order to achieve a more preferred local trait bundle. Second, changing house-
hold exposure to the labor market over the life-cycle can also be associated with
equilibrium migration. 36 An incentive exists for households that are reducing
their exposure to the labor market to relocate in cities with a higher mix of those
attributes that are capitalized relatively more in the labor market. Third, capital
market imperfections make it difficult for the household to perfectly smooth its
consumption as its real income changes over the life-cycle. Consequently, as
real income rises, a household may choose to relocate in order to consume a
more desired amenity/fiscal bundle.3 7 An additional equilibrium explanation for
migration is spatial mismatch. As individuals enter the labor market and form
new households, they may find that their current location is not optimal. Given
the equilibrium set of prices for amenity/fiscal attributes, a move is required to
36 If all city attributes were proportionally priced in both the land and labor markets, then changing labor
market exposure would change the price level but not the relative price levels of city attribute bundles. House-
holds, therefore, would face no incentive to migrate. Such proportionality is unlikely to exist in practice. The
basic model suggests that nonproductive (to firms) attributes will be relatively more capitalized into wages.
37 Using the 1980 Census of Populationand Housing,Graves and Waldman (1991) and Gyourko and Tracy
(1991) report findings consistent with the life-cycle labor market exposure motivation for migration. Graves and
Waldman (1991) regress the net inmigration (and net inmigration rate) of the elderly on the wage capitalization
variable from the QOL calculation for each of the 253 counties in Blomquist et al. (1988), and report positive
and significant coefficients. When they repeat this exercise using the net inmigration of prime age households
they find either insignificant or negative coefficients. Gyourko and Tracy (1991) calculate the implied net
subsidy (positive or negative) to a retired household for each of 130 cities based on the assumption that retired
households value the amenity/fiscal bundle in the same manner as the marginal household. They then construct
the share of "retired" households for each of the cities in their sample. The simple unweighted correlation
between the retirement employment share and the net subsidy is 0.15 (with a probability value of 0.08 under
the null hypothesis that p = 0). Weighting cities by their total number of labor market participants increases
the correlation to 0.35 (with a probability value of 0.001). Linneman and Graves (1983) find empirical support
for migration induced by likely changes in preferences for amenity/fiscal attributes and changes to real income.
Their study examines the determinants of both job and geographic mobility using microdata drawn from
the Panel Study of Income Dynamics. For the 1971-1972 period, they find that 13% of households changed
residences while 9% changed jobs. Roughly one-third of job changes also involved a residence change. They
model the decision to change job and/or residence using a multinomial logit framework, and find significant
mobility effects induced by changes in family size and real income.
1448 J. Gyourko et al.
38 Linneman and Graves (1983) label this motive "residential search", and list it as a disequilibrium ex-
planation. We prefer to include this with the other equilibrium motives since it is consistent with equilibrium
pricing in the Rosen/Roback framework.
39 Sjaastad (1962) and Muth (1971) both argue that migration is primarily a response to economic incentives
in labor markets. Migration reflects the flow of human capital across geographic boundaries as it searches out
its highest valued use. More recently, Topel (1986) and Greenwood and Hunt (1989) arrive at the same basic
conclusion. Berger and Blomquist (1992) argue that QOL issues matter in the choice of a destination given
a move, but are secondary in the initial decision to move. Muesser and Graves (1995) argue that the relative
importance of economic incentives versus QOL are likely to vary by time period.
Ch. 37: Quality of Life and Environmental Comparisons 1449
Greenwood et al. (1991) start with the assumption that net migration to an
area is a function of the relative net present value (NPV) of labor market earnings
available in the area and the relative amenity bundle.
nm = h {NPV ( Y ) (A ) , (6.1)
where nmjt is the net migration to location j in year t, Yjt is potential labor
market earnings in location j for a standardized unit of labor in year t, Yt is the
average potential labor market earnings in the labor market in year t, Aj is the
amenity bundle in location j, and A is the average amenity bundle available in
the market. Amenity and economic motivations for migration suggest that each
first derivative of h is positive.
Several simplifying assumptions are made to allow estimation of this net
migration equation. First, growth rates of potential incomes across areas are
assumed to be equal so that all net present values can be replaced by current
relative incomes. Second, a Cobb-Douglas functional form is specified for h,
where the coefficient on relative amenities is assumed to equal one. Adding a
stochastic error term gives the following estimating equation for the log of net
migration:
Annual data for 50 states and Washington, DC between 1971-1988 were used
in this analysis. The net migration and relative income potential variables are
constructed by the authors. 4 0 The relative amenity set is left to be picked up
as a location's fixed-effect. The coefficient X is estimated using an instrumental
40 Net migration for year t is defined as the sum of the "natural labor force" (NLF) in year t - 1 plus
"economic migration" in year t divided by NFLt_ 1. Economic migration is defined to be the difference
between the actual civilian population under the age of 65, and the estimated civilian population under the age
of 65 in the absence of migration. This estimated population is calculated by using 1970 cohort counts by state
extrapolated for intervening years using cohort-specific birth and death rates (adjusted so that the increments
across cohorts match with the gross birth and death rates by state), apportioning international migrants by state,
and subtracting all military personnel and their dependents. The natural labor force is calculated by applying
cohort-specific labor force participation rates (adjusted to eliminate discouraged worker effects) to cohort-
specific population counts, and summing across cohorts. The state relative income variable is constructed by
deflating an estimate of the adjusted state nominal wage rate by a cost-of-living index. The state nominal
wage is constructed using the Bureau of Economic Analysis (BEA) data on total industry wages divided by
average annual employment. This nominal wage is then adjusted for any state and local income and sales
taxes. If income and sales taxes are capitalized in both land and labor markets, then fully adjusting wages
in the construction of the relative income measure involves an overadjustment. The cost-of-living index is
constructed by Greenwood et al. and include an adjustment for regional housing prices.
1450 J. Gyourko et al.
variables (IV) procedure to correct for any endogeneity problems. The set of in-
struments is taken from a larger model presented elsewhere, and is not discussed
in the paper. The IV coefficient estimate (standard error) for Xis 0.21 (0.01). This
is consistent with the extensive literature finding a positive connection between
migration and economic incentives in the labor market. The data also strongly
reject the restriction that all states share a common intercept.
To get at the issue of disequilibrium pricing, Greenwood et al. define the
"equilibrium relative income" (RY*) for a location to be the level of relative
income that would imply no net migration for the area. The difference between
RY's and the actual relative incomes is taken as a measure of the extent of
disequilibrium in the market. Using the state fixed-effect estimates, the mea-
sured relative income for the state and the estimate of A, an estimate (and 90%
confidence interval) for RY* is constructed for each location. For six states and
the District of Columbia, they find the actual relative income outside the 90%
confidence interval for the equilibrium relative income. 4 1 On face, little evidence
of disequilibrium pricing is suggested by the data.
It should be noted that a number of important assumptions are implicitly being
made in the above formulation. First, amenities are treated as time invariant.
While this is likely for true amenities such as climate, it is less likely for fis-
cal attributes. 42 The functional form assumptions also rule out the equilibrium
life-cycle explanations for migration. Amenities enter the migration decision
only as a single index, effectively ruling out equilibrium migration due to pref-
erence changes for specific amenities over the life-cycle. The Cobb-Douglas
assumption further rules out any interactions between income and the demand
for amenities-in particular, the potential that real income increases over the life-
cycle can induce migration as households purchase a better set of amenities. 4 3
The fixed effects estimation strategy is also quite restrictive. While it obviates
the need for data on specific amenities by location, it clouds the interpretation
of the findings because the fixed effects estimates will pick up, not only relative
differences in net migration due to differences in amenity/fiscal characteristics
across states, but also any other characteristics of states that generate systematic
differences in net migration that are unrelated to relative income differences. To
41 Greenwood et al. do not report an overall test for disequilibrium pricing. In addition, the confidence
intervals for the relative equilibrium income levels should be wider than reported by Greenwood et al. due to
the constructed nature of the relative income variable.
42 It is clear from the text that Greenwood et al. are thinking of "environmental" characteristics as the center
of their analysis.
43 The sample selection excludes individuals over the age of 65 years. This will mitigate some of the life-
cycle migration effects in the data used in the estimation.
Ch. 37: Quality of Life and Environmental Comparisons 1451
7. Conclusions
Recent research has clearly enriched our knowledge of the urban QOL. Concep-
tual innovations have incorporated traditional concerns of urban economics, such
as distance from the urban core and agglomeration effects into the Rosen/Roback
framework. The importance of the local fiscal conditions to the overall QOL and
to specific capitalization results has also been made clear. The burgeoning field of
environmental economics continues to contribute to our knowledge of the value
of key local traits, in addition to yielding insights about data quality and basic
strategy with respect to estimating the overall QOL.
Nevertheless, empirical research into the urban QOL stands at an important
crossroads. Controlling for the large location-specific group effects in the data
shows individual trait prices and overall QOL indexes to be much less precisely
estimated than suggested by the OLS-based results. Dealing effectively with the
challenge posed by this issue requires understanding that this primarily is a data
problem. In particular, there appears to be no econometric or methodological
solution on the horizon. Data on more urban areas will help, and richer databases
that more fully describe local amenity, environmental and fiscal conditions are
absolutely necessary.
More technical advances are needed with respect to relaxing the key underly-
ing equilibrium assumption in the Rosen/Roback framework. Greenwood et al.
(1991) have taken the first steps in this direction, and we believe that more fruitful
progress can be made while data quality are slowly improved.
References
Abelson, P.W. and A. Markandya (1985), "The interpretation of capitalized hedonic prices in a dynamic
environment", Journal of Environmental Economics and Management 12:195-206.
Bartik, T.J. and V.K. Smith (1987), "Urban amenities and public policy", in: E.S. Mills, ed., Handbook of
Regional and Urban Economics, Chap. 31 (Elsevier Science Amsterdam).
Beeson, P. and R. Eberts (1989), "Identifying productivity and amenity effects in interurban wage differentials",
The Review of Economics and Statistics 71:443-452.
1452 J. Gyourko et al.
Berger, M.C. and G.C. Blomquist (1992), "Mobility and destination in migration decisions: the roles of
earnings, quality of life, and housing prices", Journal of Housing Economics 2:37-59.
Blomquist, G.C., M.C. Berger and J.P. Hoehn (1988), "New estimates of quality of life in urban areas",
American Economic Review 78:89-107.
Boyer, R. (1983), Places Rated Almanac (Rand McNally, Chicago).
Brookshire, D.S., M.A. Thayer, W.D. Schultze and R.C. d'Arge (1982), "Valuing public goods: a comparison
of survey and hedonic approaches", American Economic Review 72:165-177.
Brookshire, D.S., M.A. Thayer, J. Tschirhard and W. Schultze (1985), "A test of the expected utility model:
evidence from earthquake risks", Journal of Political Economy 93:369-389.
Cheshire, P.C. and D.G. Hay (1989), Urban Problems in Western Europe: an Economic Analysis (Unwin
Hyman, London).
Clark, D. and L. Nieves (1994), "An interregional hedonic analysis of noxious facility impacts on local wages
and property values", Journal of Environmental Economics and Management 27:235-253.
Cummings, R.G., D.S. Brookshire and W.D. Schulze (1986), Valuing Environmental Goods: An Assessment
of the Contingent Valuation Method (Littlefield, Adams, Rowman and Allanheld, Totowa, NJ).
Deacon. R.T. and P. Shapiro (1975), "Private preference for collective goods revealed through voting on
referenda", American Economic Review 65:943- 955.
Dickie, M. and S. Gerking (1991), "Willingness to pay for ozone control:inferences from the demand for
medical care", Journal of Environmental Economics and Management 21:1-16.
Ehrenberg, R. and J. Schwarz (1986), "Public sector labor markets", in: O. Ashenfelter and R. Layard, ed., The
Handbook of Labor Economics (North-Holland, Amsterdam).
Epple, D. and A. Zelenitz (1981), "The implications of competition among jurisdictions: does Tiebout need
politics?", Journal of Political Economy 89:1197-1217.
Epple, D., A. Zelenitz and M. Visscher (1978), "A search for testable implications of the Tiebout hypothesis'",
Journal of Political Economy 86:405-425.
Evans, A.W. (1990), "The assumption of equilibrium in the analysis of migration and interregional differences:
a review of some recent research", Journal of Regional Science 30:515-531.
Feenberg, D. and E. Mills (1980), Measuring the Benefits of Water Pollution Abatement. Studies in Urban
Economics (Academic Press, New York).
Freeman, R. (1986), "Unionism comes to the public sector", Journal of Economic Literature 24:41-86.
Garen, J. (1988), "Compensating wage differentials and the endogeneity of job riskiness", Review of
Economics and Statistics 70:9-16.
Gerking, S. and L. Stanley (1986), "An economic analysis of air pollution and health: the case of St. Louis",
Review of Economics and Statistics 68:115-121.
Graves, PE. and D.M. Waldman (1991). "Multimarket amenity compensation and the behavior of the elderly",
American Economic Review 81:1374-1381.
Graves, P., J. Murdoch, M. Thayer and D. Waldman (1988), "The robustness of hedonic price estimation: urban
air quality", Land Economics 64:220-242.
Greenwood, M.J. and G.L. Hunt (1989), "Jobs versus amenities in the analysis of metropolitan migration".
Journal of Urban Economics 25:1-16.
Greenwood, M.J., G.L. Hunt, D.S. Rickman and G.I. Treyz (1991), "Migration, regional equilibrium, and the
estimation of compensating differentials", American Economic Review 81:1382-1390.
Gyourko, J. and J. Tracy (1989a), "The importance of local fiscal conditions in analyzing local labor markets",
Journal of Political Economy 97:1208-1231.
Gyourko, J. and J. Tracy (1989b), "Local public sector rent-seeking and its impact on local land values",
Regional Science and Urban Economics 19:493-516.
Gyourko, J. and J. Tracy (1989c), "On the political economy of land value capitalization and local public sector
rent seeking in a Tiebout model", Journal of Urban Economics 26:152-173.
Gyourko, J. and J. Tracy (1991a), "Public sector bargaining and the local budgetary process", Research in
Labor Economics 12:117-136.
Gyourko, J. and J. Tracy (1991b), 'Thestructure of local public finance and the quality of life", Journal of
Political Economy 99:774-806.
Ch. 37: Quality of Life and Environmental Comparisons 1453
Hamilton, J. (1993), "Politics and social costs. Estimating the impact of collective action on hazardous waste
facilities", Rand Journal of Economics 24:101-121.
Hamilton, J. (1995), '"Testing for environmental racism, prejudice, profits, political power?", Journal of Policy
Analysis and Management 14:107-132.
Harrison, D. and D. Rubinfeld (1978), "Household prices and demand for clean air", Journal of Environmental
Economics and Management 5:81-102.
Haughwout, A.F (1993), "Measuring the impact of fiscal policy:a theoretical and empirical examination",
working paper (Princeton University).
Hausman, J.A. and W.E. Taylor (1981), "Panel data and unobservable individual effects", Econometrica 49:
1377-1398.
Hausman, J. (1993), Contingent Valuation: A Critical Assessment (North Holland, Amsterdam).
Henderson, J.V. (1985), "The Tiebout model: bring back the entrepreneurs", Journal of Political Economy
93:248-264.
Henderson, J.V. (1996), "The effects of air quality regulation", American Economic Review.
Hoehn, J.P., M.C. Berger and G.C. Blomquist (1987), "A hedonic model of interregional wages, rents and
amenity values", Journal of Regional Science 27:605-620.
Inman, R. (1995), "Do you know how much money is in your public purse?" Federal Reserve Bank of
Philadelphia. Business Review, pp. 19-30.
Kahn, M. (1995), "A revealed preference approach to ranking city quality of life", Journal of Urban Economics.
38:221-235
Kahn, M. (1997), "Particulate pollution trends in the United States", Regional Science and Urban Economics.
February, Vol. 27.
Kahn, M. and J. Matsusaka (1997), "Who votes for the environment? Evidence from California voter
initiatives", Journal of Law &Economics.
Kiel, K. and K. McClain (1995), "House prices during siting decision stages: the case of an incinerator from
rumor through operation", Journal of Environmental Economics and Management 28:241-255.
King, M.L. and M.A. Evans (1986), "Testing for block effects in regression models based on survey data",
Journal of the American Statistical Association 81:677-679.
Kish, L. (1965), Survey Sampling (John Wiley, New York).
Kohlhase, J.E. (1991), "The impact of toxic waste sites on housing values", Journal of Urban Economics
30:1-26.
Kolstad, C. (editor) (1991), Measuring the Demand for Environmental Quality (North Holland Press,
Amsterdam).
Krupnick, A., W. Harrington and B. Ostro (1990), "Ambient ozone and acute health effects: evidence from
daily data", Journal of Environmental Economics and Management 18:1-18.
Linneman, P. and P.E. Graves (1983), "Migration and job change: a multinomial logit approach", Journal of
Urban Economics 14:263-279.
Michaels, R. and V. Kerry Smith (1990), "Market segmentation and valuing amenities with hedonic methods:
the case of hazardous waste sites", Journal of Urban Economics 18:223-242.
Mitchell, R.C. and R.T. Carson (1989), "Using surveys to value public goods: the contingent valuation method"
(Resources for the Future, Washington, DC).
Mobil Oil Corporation (1989), Mobil Travel Guide.
Moulton, B.R. (1986), "Random group effects and the precision of regression estimates", Journal of
Econometrics 32:385-397.
Moulton, B.R. (1987), "Diagnostics for group effects in regression analysis", Journal of Business and
Economic Statistics 5:275-282.
Muesser, P.R. and PE. Graves (1995), "Examining the role of economic opportunity and amenities in
explaining population redistribution", Journal of Urban Economics 37:176-200.
Muth, R.E (1971), "Migration: chicken or egg?", Southern Economic Journal 37:295-306.
Ostro, B.D. (1987), "Air pollution and morbidity revisited: a specification test", Journal of Environmental
Economics and Management 14:87-98.
Pfefferman, D. and T.M.F Smith (1985), "Regression models for grouped populations in cross-section
surveys", International Statistical Review 53:37-59.
1454 J. Gyourko et al.
Portney, P.R. (1994), The contingent valuation debate: why economists should care", Journal of Economic
Perspectives 8:3-17.
Portney, P.R. and J. Mullahy (1986), "Urban air quality and acute respiratory illness", Journal of Urban
Economics 20:21-38.
Ranson, M. and C. Pope (1995), "External health costs of a steel mill", Contemporary Economic Policy
April:86-101.
Roback, J. (1980), "The value of local urban amenities: theory and measurement", Ph.D. dissertation.
University of Rochester.
Roback, J. (1982), "Wages, rents, and the quality of life", Journal of Political Economy 90:1257-1278.
Roback, J. (1988), "Wages, rents, and amenities: differences among workers and regions", Economic Inquiry
26:23-41.
Rosen, S. (1974), "Hedonic prices and implicit markets: product differentiation in pure competition", Journal
of Political Economy 82:34-55.
Rosen, S. (1979), "Wage-based indexes of urban quality of life", in P. Mieszkowski and M. Straszheim, eds.,
Current Issues in Urban Economics (Johns Hopkins University Press, Baltimore).
Scott, A.J. and D. Holt (1982), "The effect of two-stage sampling on ordinary least squares methods", Journal
of the American Statistical Association 77:848-854.
Sjaastad, L.A. (1962), "The costs and returns of human migration", Journal of Political Economy 70:80-93.
Smith, V.K. and Ju Chin Huang (1995), "Can hedonic models value air quality? A meta-analysis", Journal of
Political Economy 103:209-227.
Stock, J. (1991), "Nonparametric policy analysis: an application to estimating hazardous waste cleanup
benefits", in: P. Barnett, ed., Nonparametric and Semiparametric Methods in Economics and Statistics.
Stover, M.E. and C.L Leven (1992), "Methodological issues in the determination of the quality of life in urban
areas", Urban Studies 29:737-754.
Topel, R.H. (1986), "Local labor markets", Journal of Political Economy 94:S111-S144.
Voith, R. (1991), "Capitalization of local and regional attributes into wages and rents: differences across
residential, commercial and mixed-use communities", Journal of Regional Science 31:127-145.
Chapter38
RANDALL W. EBERTS
DANIEL P. McMILLEN
Contents
1. Introduction 1457
2. Agglomeration economies and urban theory 1460
2.1. Descriptive analysis 1460
2.1.1. Internal scale economies 1461
2.1.2. Localization economies 1461
2.1.3. Urbanization economies 1463
2.1.4. Modeling implications 1464
2.2. Theoretical models of urban areas 1464
2.3 Microfoundations of agglomeration economies 1468
3. Estimates of agglomeration economies 1473
3.1. Functional form of the production function 1475
3.2. Level of aggregation 1478
3.3. Specification of agglomeration economies 1479
3.4. Alternative measures of urban size 1481
3.5. Factor prices and cost functions 1483
3.6. Omitted variables 1483
3.7. Wage and rent equations 1485
3.8. Spatial arrangement of cities 1486
4. Public infrastructure 1486
4.1. Infrastructure as a public input 1487
4.2. Estimates 1488
4.3. Studies including both agglomeration effects and public infrastructure 1490
5. Conclusions 1491
References 1492
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P Cheshire
© 1999 Elsevier Science B. V All rights reserved.
1455
1456 R. W. Eberts and D.P McMillen
Abstract
employment faster than smaller metropolitan areas (Moomaw, 1980; Vining and
Koutuly, 1977).
Observing that large US cities were stagnating in size, many researchers and
policy makers believed that some cities were too large, claiming that the nation's
largest cities were less livable and less economically viable and contributed more
than proportionately to the nation's social problems (Sundquist, 1970). They pro-
posed policies to give favorable tax breaks to businesses in smaller cities, and to
channel federal infrastructure dollars away from large cities to smaller ones.
Against this backdrop of population trends and policy proposals, a wave of
estimates of the net productivity advantages of large cities were conducted. The
first estimates showed a substantial productivity advantage of large cities over
smaller ones. Sveikauskas (1975) reported that a doubling of city size would
yield a 6% increase in manufacturing productivity. Estimates of Segal (1976)
and Fogarty and Garofalo (1978) were even larger-8 and 10%, respectively.
Such estimates suggest that the New York City MSA, with its 9.5 million people
in 1975, would be at least 50% more productive than an MSA of 50,000 people.
Some researchers used these estimates to infer an optimal size city based on
the economies and diseconomies associated with city size. Carlino (1982), for
example, estimated the optimal size city to be 3.6 million people during the
1957-1969 period and 3.4 million during the 1970s. Some researchers viewed
these estimates to be too large and inconsistent with the casual evidence of pop-
ulation and employment trends, and subsequent studies set out to resolve these
inconsistencies.
Interest in the empirical relationship between public infrastructure and eco-
nomic development in the US was first generated, not in the urban literature,
but in the national productivity literature. The 1970s and 1980s saw not only
slow growth or even declines in large urban areas, but also a sharp decline in
national productivity. Several articles and books were influential in associating
the national productivity decline to a decline in public infrastructure investment.
Choate and Walter's (1983) alarming commentary on US infrastructure, enti-
tled America in Ruins, fostered the public's perception of crumbling roads and
bridges. Aschauer's (1989) large estimates of the marginal product of public cap-
ital, ranging from 0.38 to 0.56, fueled the debate. His results implied that public
infrastructure was woefully neglected. The high estimated returns promised that
government capital would pay for itself in terms of higher output within a year.
To some, this finding appeared to be a costless panacea for future growth; to
others, these results engendered much scepticism and consequently generated a
series of papers seeking to understand what was behind these large estimates.
This "third deficit", as Munnell (1990) referred to it, was linked to the celebrated
Ch. 38: Agglomeration Economies and Urban PublicInfrastructure 1459
but perplexing productivity slowdown in the 1970s and to lagging real wages.
Congress established the National Council on Public Works Improvement (1988)
to conduct an in-depth assessment of the state of public infrastructure and come
up with policy recommendations. The Council recommended that the nation
should double its investment in public infrastructure. Other government agen-
cies, including the Federal Highway Administration and the Corps of Engineers,
sponsored research to explore the relationship between public infrastructure and
productivity, primarily at the state and national level.
During this period, few researchers explored the effects of public infrastruc-
ture at the metropolitan level. From a research and policy perspective, this ne-
glect makes little sense. First, public infrastructure is a key factor in facilitating
the benefits of agglomeration and in easing the congestion associated with large
cities. Second, public infrastructure is an important input into the production
process (and in the household utility function) independent of its associated
effects with agglomeration economies. Third, much infrastructure investment is
concentrated at the metropolitan area level, where the interaction between in-
frastructure and economic activity takes place. Infrastructure, like agglomeration
economies, is a spatial construct, and linking the actual public infrastructure fa-
cility (e.g., a highway segment and network) to its users is important. This spatial
correspondence is less distinct at the state and national level. Finally, state and
local governments are responsible for most of the expenditures on the nation's
infrastructure. Upwards of 85% of total US nonresidential public capital is put in
place by state and local governments, with local governments assuming a large
part of that responsibility. 2
This chapter reviews the theoretical and empirical literature on agglomeration
economies and urban public infrastructure. Theory links the two inextricably by
suggesting that agglomeration economies exist when firms in an urban area share
a public good as an input to production. Clearly, public infrastructure is one such
1 The exceptions were Dalenberg (1987), Eberts (1986) and Deno (1988). The primary reason for the
lack of research at the metropolitan level is the difficulty in constructing suitable measures of public capital
stock. Most research on the productivity effects of public capital stock uses the perpetual inventory technique
to estimate capital stock. This approach requires a long investment series, that is not readily available. The
only metropolitan and city public capital stock estimates using this approach for the US were constructed
by Eberts et al. (1986). Considerably more studies of infrastructure's effect on economic activity have been
conducted at state level. Most have used a dataset constructed by Alicia Munnell (1990) using Census of
Government data. However, for agglomeration studies, state-level estimates are considered too aggregated
given that agglomeration economies are urban specific effects (Calem and Carlino, 1991, among others also
make this argument).
2 Local governments have a similarly large role in many European countries. Seitz (1993) reports that
local governments are responsible for investing 70% of (the former West) Germany's share of total public
investment.
1460 R.W. Eberts and D.P McMillen
sharable input. Also, as argued above, public infrastructure facilitates the benefits
of agglomeration economies. For example, one suggested source of agglomera-
tion economies is the city's role as an "urban warehouse", which allows firms to
carry lower inventories. A warehouse hardly is feasible without public investment
in roads, bridges, and other local infrastructure.
The organization of the chapter is as follows. We begin in Section 2 with a
review of the theoretical literature. The review follows the literature's chronolog-
ical development from descriptive studies of agglomeration economies, to use of
agglomeration economies in explaining the existence and growth of urban area,
to studies of the microfoundations of agglomeration economies. In Section 3, we
review empirical studies of agglomeration economies. We begin with the most
common type of study, which uses production functions to provide quantitative
estimates of agglomeration economies. We then discuss econometric problems
with the traditional methodology and highlight alternative estimation strategies.
A review of the literature on public infrastructure-that is mostly empirical-is
presented in Section 4. Section 5 offers some conclusions and suggestions for
future research.
3 This categorization appears to have entered the urban/regional literature through Hoover (1937), who
uses the terms "localization" and "urbanization" and discusses their role in industrial concentration. Hoover
credits the idea to Ohlin (1935), who also defines three categories of economies that lead to concentration
of industry: "(1) economies of concentration of industry in general, (2) external economies of concentration
of a particular industry, and (3) internal large-scale economies of a producing unit" (Ohlin, 1935: p. 203).
Adam Smith (1776) and Alfred Marshall (1890) also have extensive discussions of the advantages of urban
areas for industrial location, although without a neat categorization and with an emphasis on internal scale and
localization economies.
Ch. 38: Agglomeration Economies and UrbanPublic Infrastructure 1461
2.1.1. Internal scale economies
Internal scale economies are the most familiar category to economists. They exist
when expanding production at some given site lowers a firm's unit costs. The
classic example is Smith's (1776) pin factory, in which a firm's growth enables it
to take advantage of division of labor. Workers' skills increase through repetition,
that lowers production costs. Internal scale economies also are attributed to the
existence of indivisible inputs, that occurs when an input has a minimum efficient
scale. Other examples include the ability to take advantage of bulk purchases at a
site and the more efficient use of specialized machinery. Internal scale economies
allow a large firm to underprice its smaller competitors, that may compensate
for the additional shipping costs that necessarily occur by concentrating produc-
tion in a single location. Large-scale manufacturing establishments such as steel
production and automobile assembly are obvious examples of industries with
significant internal scale economies. Such establishments lead to the formation
of a "company town": steel in Gary, rubber in Akron, glass in Toledo, chocolate
in Hershey or aircraft manufacture in Seattle.
4 The assumption that economies are external to the firm is important in making a competitive market
system compatible with aggregate increasing returns to scale. See Chipman (1970). Mergers can transform
external economies into economies that are internal to the new, larger firm.
1462 R.W Eberts and D.P McMillen
Sales of output and purchases of inputs fluctuate in many firms and indus-
tries for random, seasonal, cyclical, and secular reasons. To the extent that
fluctuations are imperfectly correlated among employers, an urban area with
many employers can provide more nearly full employment of its labor force
than can an urban area with few employers. Likewise, a firm with many
buyers whose demand fluctuations are uncorrelated will have proportion-
ately less variability in its sales than a firm with few buyers. It can hold,
therefore, smaller inventories and employ smoother production scheduling.
(Mills and Hamilton, 1994: p. 20)
1464 R. W1Eberts and D.P. McMillen
X = ALNfKY. (2.1)
More recent studies of individual urban areas have followed in Mills' tradition,
while being more precise about the source of the agglomeration economy. For
example, Sullivan (1983) develops a general equilibrium model of an urban econ-
omy, providing a detailed analysis of production in the export sector. He assumes
that the export sector's production is characterized by the following unit cost
function:
ployment is added to the central city. Later, employment growth is split between
the two locations. 7
Unlike many studies, Helsley and Sullivan's (1991) analysis is explicitly dy-
namic. The dynamics of agglomeration economies form a new growth area for
both theoretical and empirical research (e.g., Hanson, 1996; Henderson, 1988;
Mori, 1997; Palivos and Wang, 1996; Sasaki and Mun, 1996; Walz, 1996). Hen-
derson (1988) argues that urbanization economies are most important in the early
stages of an industry's development, causing the industry to locate in the largest
cities. As an industry grows, it may move to smaller cities, with localization
economies becoming the dominant force. An example is the Silicon Valley com-
puter industries, that began in the San Francisco Bay Area due to the urbanization
economies associated with a large city with electronic firms and major universi-
ties. As the industry grew, localization economies became significant. Now, the
industry is so large that ample levels of localization economies can be enjoyed in
other locations (e.g., Austin, Texas and Portland, Oregon).
Mori (1997) also analyzes the dynamics of urban development when firm
location is driven by agglomeration economies. In Mori's model, agglomeration
economies are generated by consumers' demand for variety and scale economies
in the production of manufacturing goods. Large cities offer more variety in
consumption, that attracts workers who increase the demand for manufactured
goods. The increased demand supports a greater number of specialized manufac-
tured goods in large urban areas. A large city develops as transport costs decline,
because manufacturing firms can support a larger market area while enjoying
the agglomeration economies of large cities. But large urban areas pay higher
shipping costs on average for agricultural goods. As the urban area grows, some
firms may find it worthwhile to locate in the agricultural hinterland, producing
another city. Other firms follow to take advantage of agglomerative forces in
the new, smaller city. The smaller city may continue to grow as the increased
agglomeration economies lead more firms to enter. Eventually, the cities may
merge, producing a "megapolis".
7 Anas and Kim (1996) take another approach to modeling subcenter formation. They assume that con-
sumers prefer to shop in areas with many stores. Subcenters develop when these "shopping externalities" are
strong relative to the cost of traffic congestion. Henderson and Mitra (1996) provide a fascinating analysis of
the development of "edge cities"--cities with large diversified employment centers in suburban areas.
Ch. 38: Agglomeration Economies and UrbanPublic Infrastructure 1469
8 "For any nontrivial partition of N over space, there are economies of agglomeration if and only if the cost
of producing the resources to be shared by others is strictly subadditive over space" (Goldstein and Gronberg,
1984: p. 101).
1470 R. W. Eberts a, d D.P McMillen
The Goldstein and Gronberg classification closely matches the previous de-
scriptive literature. Economies of agglomeration exist when an urban area pro-
vides an input that lowers costs for all firms. A key input is public infrastructure-
roads, airports, training facilities, universities, etc. If costs are lowered for only
one industry, we have localization economies. If costs are lowered for all firms,
we have urbanization economies. Note that the cost advantage may or may not
enter directly through productivity gains, as assumed in the preceding theoretical
models (other than Sullivan, 1983). For example, a sharable input that lowers
the cost of obtaining labor enters the cost and profit functions instead of the
production function. This line of reasoning suggests that empirical researchers
should consider using a cost function specification rather than the more tradi-
tional production function approach.
Other studies directly attempt to model the microfoundations of specific ag-
glomeration economies. An early example is Sullivan (1986), who models a city
with two employment sectors. Manufacturing enjoys a traditional localization
economy that depends on aggregate manufacturing output. Similarly, an office
sector is subject to a localization economy that depends on aggregate office-
sector output. However, another source of office-sector agglomeration economies
is incorporated into the model: workers in the sector must make costly trips to
a central market each day "to interact with the representatives of other firms;
these employees exchange information and market their products" (Sullivan,
1986: p. 60). The information is not available to a firm that locates outside of the
city, and is more costly to obtain the farther a firm is from the central business
district (CBD). Office workers have an incentive to locate near each other in
the CBD to reduce the cost of face-to-face contact; this proximity represents a
form of sharable input. Sullivan's (1986) model incorporates the "communica-
tion economies" that many authors use to justify the existence of agglomeration
economies. Sullivan does not attempt to explain why both manufacturing and of-
fices exist in the same city when separate cities would reduce overall commuting
costs. Perhaps manufacturing establishments are office customers, and nearby
locations reduce transaction costs.
Helsley (1990) models the role of cities in spreading information. In his sim-
ple linear city, each firm produces a single output q according to a constant-
returns firm-level production function with a Hicks-neutral shift factor that de-
pends on the level of knowledge:
Knowledge is a public input, but declines with distance according to the decay
function y (z, u), that represents the portion of x(u) that remains at z. This func-
tion reaches a peak at z = u (where y(u, u) = 1), and declines symmetrically
for z > u and z < u with Iz - ul. The level of knowledge at location u is
a(u) = fJo y(u, z)x(z) dz, where uo and ul are the endpoints of the linear CBD.
The key result is that "the level of knowledge, land rent, and output and factor
intensities achieve unique maxima at the center of the CBD in equilibrium and
decline as distance from the center of the CBD increases" (Helsley, 1990: p. 400).
Productivity is highest at the city center, and at every location is related positively
to the number of firms in the CBD.
Another often-cited source of agglomeration economies is improvements in
the labor market as city size increases: in large cities a firm is more likely to
find the skilled workers it requires, and workers are more likely to find suitable
employment. Helsley and Strange (1990) model the matching process between
firms and workers. 9 The productivity of workers, whose skills exactly match their
job is a, and a loss of /i per unit distance in the characteristic space, is caused
by having skills other than a. Let x be the address on the unit circle of firms' job
requirements, while y is the address of workers. Then the output of the match
(x, y) is a - B Ix - y . Firms do not observe the skills of workers before choosing
a city, and workers do not observe the job requirements of firms. Both groups
know the number of firms and workers in each city, and assume that skills and
job requirements are random draws from a uniform distribution on the unit circle.
The expected quality of the match between firms and workers increases with city
size, which confirms the intuition of previous descriptive studies.
This simple but clever formulation leads to important insights:
First, the agglomeration economy has the characteristics of a local public
good. A firm entering a city improves the expected quality of the match
between job requirements and skills for all workers, leading to a positive
9 Kim (1989, 1990, 1991) provides a similar analysis, while allowing firms to train workers who have
low skills. Abdel-Rahman and Wang (1995, 1997) develop a two-sector (high versus low skill) model with
a matching process in the skilled labor market. A hierarchical system is developed in which a single large
metropolis contains all skilled workers, while peripheral regions include unskilled workers. The advantage to
the high-skill firms of a single location is the increased efficacy of the labor market.
1472 R.WV Eberts and D.P McMillen
ex ante relationship between wages, productivity, and city size .... Sec-
ond, there are two externalities associated with firm location .... One is
the conventional productivity externality. If a firm enters a city, it improves
the productivity of all workers, but it considers only its own profits. This
externality causes a city to contain too few firms under free entry. The other
externality arises from spatial competition and the heterogeneity of workers
and firms. An entrant to a city reduces the labor market areas, and hence
profits, of incumbent firms, but considers only its own profits. This compe-
tition externality leads to too many firms under free entry. The competition
externality dominates in our model. Third, equilibrium city sizes are not
optimal ... [and] the zero profit number of firms is not efficient. (Helsley
and Strange, 1990: pp. 190-191)
In Goldstein and Gronberg's (1984) framework, the efficacy of the labor market
is a form of public input that is sharable by all firms in Helsley and Strange's
(1990) cities.
In a subsequent paper, Helsley and Strange (1991) provide a rigorous ex-
planation of the agglomeration economies that Mills and Hamilton (1994) call
"statistical in nature". The basis is an urban area served by a single bank, that
allocates credit to different investment projects. Some projects succeed but others
fail. When a project fails, the bank gains possession of an immobile and special-
ized asset. The value of the asset to the bank is its salvage value, that depends
on its value in its second best use. The second-best value is higher in large cities
because when a wide variety of firms exists there is a higher probability of a good
match between the initial firm's capital requirements and those of other firms.
Again, Helsley and Strange summarize the implications well:
Agglomeration economies arise because the expected second best use of
an immobile asset is worth more in a large city than in a small one. This
means that resource productivity rises with city size in two ways: used assets
are better matched and risk is reduced .... [C]ity size provides borrowers
with external collateral on loans. This external collateral is a public input in
urban capital markets: it is impossible to exclude borrowers from utilizing
the external collateral, and its use by one borrower does not limit its use by
others. The external collateral associated with greater city sizes resembles
the public inputs that Goldstein and Gronberg discuss. (Helsley and Strange,
1991: pp. 97-98)
Ch. 38: Agglomeration Economies and Urban PublicInfrastructure 1473
Another important implication of this paper is that large cities contain more types
of economic activities than small cities because "bad states of nature are less
costly" (p. 110).
Several studies (e.g., Abdel-Rahman, 1990b, 1996; Abdel-Rahman and Fujita,
1990; Dobkins, 1996; Mori, 1997) focus on the advantages that large-city diver-
sity provides. Abdel-Rahman (1990b) develops a model in which production of a
traded good depends on miscellaneous intermediate goods. The key result is that
the number of intermediate inputs produced in a city is an increasing function of
city size, which implies that larger cities produce a greater variety of goods. City
size is larger as labor requirements fall. If labor requirements vary with industrial
structure, city sizes can differ in equilibrium. Unlike most studies, that explain
the formation of specialized cities, Abdel-Rahman's (1990b) model produces a
system of cities with diversified industrial structures.l°
where Yij represents output for the ith firm in the jth city, and K, L, G and Z
represent capital, labor, public infrastructure and other inputs. We include public
infrastructure (G) as an input in the production function to show the importance
of including both measures of agglomeration and public infrastructure. As de-
scribed more fully in Section 4, studies typically include either agglomeration
measures or public capital but rarely do they include both.
The production function, f (e), is often but not always assumed to exhibit
constant returns to scale and may exhibit increasing returns to scale when public
infrastructure is included as an input. The shift factor, g(Sj), represents scale
economies, with g' > 0. Scale, Sj, is measured by either metropolitan employ-
ment or population to represent urbanization economies, or industry employment
10 Fujita and Krugman (1995) use a related approach to explain city existence. In their model, each worker
consumes a homogeneous agricultural good and a variety of differentiated manufacturing goods. Manufactur-
ing goods have internal scale economies, and are costly to ship. Workers prefer to live in large cities because
a greater number of manufacturing goods is available for consumption. The migration of workers to the city
lowers wages, attracting more firms. The result is a diversified city surrounded by an agricultural hinterland.
The model has not been extended to include a system of cities.
1474 R. W Eberts and D.P. McMillen
11 For example, if an agglomeration economy provides a cost advantage to locating in a city, firms will
enter the city until wages rise to the point of zero profits. The increased wages are a measure of the productive
advantages offered by the city.
Ch. 38: Agglomeration Economies and Urban PublicInfrastructure 1475
3.1. Functionalform of the production function
D
o
E
CCE8o
r:I
.2
Ib "B -a 5o" t;i=
B > >o R Z 1 . C- E'
O
z f7 -
. , I ~.,
CC .CCM~~~.
C
A ~C-Z8g
~~~~~~~~~~~~sl'2~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-o Zn Zn C
-. eNC-Co .
-NO - C-CCNC· C., Zn -
iE
5
'.
t O
o
I1 e, a ~~av; a D a
CCC 00
CC"
o Zi
n nC;O o C- a
~~~~~C
o-R
C-CC-C C . oCN
o t1 °5B
.. C C
i) a 0 °
5
N-n Znj F RC-. C-I
Zn' C- CC
- O 0
CCC C-?E
C~
- C-
-
)N
'8
o
o
CC Z
, '~
5 N*555 =-o E - ° =- C
Ez8
3
5 = eN E :,o8E8 U~~~~
2C--C.
>5
E C CC
3D
~o
2
C) 5C 5Z X 5C ; =< -
CCC
C eC C C
V, C C. C B 5aBC C
CC n,<_ N C-
C -o.~
ow
8 CC~ 2
C-, C<-C N<
ax C- o U C CS C - UN
-;n Zn Co C C
=- oC - R C C ': N- O
.
%8~ ~~~CCC " -C. CC ~ ~
D
CC-
N-
$.CCC
--C
~ C~O
n~a
N 8 Bo~ ~ ~9-~ l C
Cn
aC- CC~C~ZnC Zn~. ~
ZncO 3. 3E
6::
5
o I
I rr
II: Z. & 3CC Zn
I E I~ C_ I
z
Zn~~~ C
la 1e
;Q,
0
nO
5 _
II 1 < C n Zn
CI
L:CC~
O
C
C OF _
Ch. 38: Agglomeration Economies and Urban Public Infrastructure 1477
a = u
C
5 0
5
old 8~~~~~~~o 0
c "g
. - - s
- 5 0 , 8 8
E . C 5 *-
5 5 O0 Co
Ebt n o
* 5 5 Cc 0 a 3 5
o aE o
0 Da - .00
o 5o oC ~~ . 2 ..-. ~55 ~
a~ Xe5 a | 8 E C. 5 C¢ a .
5o~ E~ ,
R 5" 8 C ,:i a: :3
C a
a _
5 ~~~~8.- aC 'o O - -
5 0$.~ 0
' ~-,0 m'50 '5 m
0
5RE 0
0 ,0 S
2 o o 0 o 0 m
° c
5- 0
50 r E ...
C z 1 o 1- E
5 l 5 i~
j 0
U° 0 00
pP ~a 5 R X ~_R
_ R 2 3 .
a - a
0 a 3
- bt ,.2 "
- g*Q3Bi': 5 5 g 3 = 0
>
o ~Z0a o j a
3$
D o
Br oO
21 a
oa 0 oa.
0 06 ·o oi -2 a~'
~~ C a C O 0 0,
1478 R. W Eberts and D.P. McMillen
ceive compensation higher than they would have in the absence of the agglomer-
ation economies. As a result, private investment in a local labor market is subopti-
mal. Economic rents may dissipate as businesses expand or additional companies
and households move into an area in response to the economic rents. The degree
to which they fall depends on the marginal contribution of an additional business
to agglomeration economies versus diseconomies.
yj = g(Sj)L'jKj (3.2)
where yj, Lj, Kj denote output, labor and capital in the jth metropolitan area.
g(Sj) is the Hicks-neutral productivity, ao and are the output elasticities of
labor and capital, respectively. The returns-to-scale approach takes = a +
as the measure of agglomeration economies of metropolitan areas. The estimated
parameter is regressed against population or other scale measures to determine
whether, or not, there is a systematic relationship with and the measure of
scale. The other approach is to use the shift parameter g(Sj), expressed as a
15 Studies, such as Shefer (1973), Carlino (1978, 1982), Beeson (1983, 1987) and Calem and Carlino (1991),
find a positive relationship between returns to scale and city size. Mera (1973), Kawashima (1975), Sveikauskas
(1975), Segal (1976), Moomaw (1981), Nakamura (1985) and Henderson (1986), for example, show that city
size is positively correlated with a neutral shift in the production function.
1480 R. W. Eberts and D.P McMillen
output growth less the weighted contribution of input growth, and can be derived
from estimates of a production function (or cost function). Assuming that an
industry's production function takes the general form Y = F(X, T), TFP is
defined as l; - Yi rrXi, where the dot represents the time derivative. In this
expression, X denotes the growth rate of the private input, rT is the output share of
the ith private input, and T is time. Beeson (1987) decomposed this formulation
of TFP growth into its components of technical change and returns to scale. Each
component was then regressed on variables reflecting agglomeration economies.
The drawback of this approach, at least as executed by the two authors, is that
factor prices are not incorporated into the model. As a result, the models may
offer estimates of agglomeration economies that are biased upward.
Specifying the growth rate of total factor productivity and its components
offers an additional element of agglomeration economies. Sveikauskas (1975)
and Kaldor (1970) promote the importance of dynamic benefits of urban concen-
tration over the static advantages of specialization. Kaldor argues that agglom-
eration economies is more than large-scale production but rather the cumulative
advantage from the growth of an industry itself. These advantages result from
"the development of skills and know-how, the opportunities for easy communi-
cation of ideas and experience, the opportunity of every-increasing differentiation
of processes and specialization in human activities" (p. 340). In addition, as
stated by Calem and Carlino (1991), the use of pooled data allows the identi-
fication of technical change and returns to scale, without relying on the usual
control variables.
The urban area population is the standard measure of urban size in studies of
urbanization economies. 7 Studies incorporate this measure into the production
functions in one of two ways: (1) they include population directly in the pro-
duction function; or (2) they estimate the parameters of the production function
without population and then regress these estimated parameters on population
and other relevant variables (Carlino, 1979). In either case, population is typi-
17 Henderson (1986) argues that while the degree of urbanization economies may vary by industry, only the
size of the city, not its industry structure, affects the extent or level of scale effects relevant to firms in each
industry. Subsequent studies, such as Fogarty and Garofalo (1988), argue that intraurban spatial structure is
important in explaining agglomeration economies. They test the importance of spatial structure using Mills'
(1972) estimates of the manufacturing density gradient. Using pooled cross-section time series data for 13
metropolitan areas, they find that the decline of the manufacturing central density and flattening in the density
gradient reduced growth in manufacturing output by more than 1% per year.
1482 R. W Eberts and D.P. McMillen
p .
gi(Pj) = 1P
Factor prices have been shown to vary across cities of different sizes, which in
turn affect the demand for factors of production. Moomaw (1983) shows that as
city size increases equilibrium requires the percentage increase in productivity to
equal the percentage increase in wages multiplied by labor's share. Ignoring the
effect of city size on wages may overstate the productivity advantages of large
cities. Since Moomaw's critique appeared, most of the studies have incorporated
wages into the estimation equations by introducing a labor demand function or
by estimating factor share equations along with production functions. The labor
demand equations typically specify wages as a function of output, price of cap-
ital (when available), population, and regional or state dummy variables. Calem
and Carlino (1991) also include a labor supply function to capture the effect
of urban amenities on wages. Omitting the supply equation, they argue, would
likely bias the estimates, although they do not offer estimates comparing the two
specifications.
A more complete specification of including factor prices to model firm be-
havior is the use of the cost function. As discussed in the previous theoretical
section, few studies have used cost functions, although the theoretical models
of agglomeration economies are more in line with this specification. Henderson
(1986) estimates both a translog production function and a translog cost function
using data from Brazil, and finds similar results.
production function with private capital and labor as inputs and agglomeration
economies as measured by the Hicks-neutral technical change. Since estimates
of private capital are unavailable, Sveikauskas substitutes labor productivity for
the Hicks-neutral technical change. However, if capital intensity is correlated
with population (and uncorrelated with other independent variables), then the
estimates of agglomeration economies (as measured by the coefficient on the
population variable) are biased upward. In a later paper, Moomaw (1983) shows
that including a proxy for capital intensity reduces the population elasticity from
0.048 to 0.015.18 He explains that "this estimate implies that the largest MSA
has a more plausible 12 percent Hicks-neutral productivity advantage over the
smaller one" (p. 537). The 12% estimate is compared with the 50% productivity
advantage when capital intensity is not included.
Land is another important input in urban production activities that is rarely
included in the estimation of production functions or cost functions. Land prices
are difficult to obtain for industrial sites. Housing prices are more readily avail-
able for metropolitan areas and may provide a reliable proxy for industrial land
prices. However, housing prices may be correlated with wages and introduce
multicollinearity in the estimation. Labor market equilibrium conditions dictate
higher wages in areas in which agglomeration economies have increased land
prices. The omission of land may also bias the returns to scale estimate. The sum
of scale parameters in the industry production function that includes only capital
and labor necessarily do not correspond with the exact measure of internalized
localization economies (Nakamura, 1985: p. 109). Although considered impor-
tant by many authors, none of the papers surveyed in this chapter, except for
Mera (1973), included the quantity or price of land in the estimation equation,
primarily because of lack of data.
Public infrastructure is also typically omitted from the agglomeration studies,
as mentioned earlier. Public investment in highways and water treatment and
distribution and in human capital development is recognized as important to a fir-
m's productivity decisions, but difficult to measure. Infrastructure investment can
reduce agglomeration diseconomies due to traffic congestion, insufficient water
and sewer infrastructure, or shortage of qualified workers. Section 4 describes in
more detail the literature on the productivity effects of public infrastructure.
18 Moomaw (1983) estimates various production function specifications for two-digit industries and finds
that the effect of adding variables to the production function differs across industries.
Ch. 38: Agglomeration Economies and Urban PublicInfrastructure 1485
3.7. Wage and rent equations
4. Public infrastructure
19 See McGuire (1992) and Gramlich (1994) for reviews of the literature on public infrastructure and
productivity.
Ch. 38: Agglomeration Economies and Urban Public Infrastructure 1487
4.1. Infrastructure as a public input
the land input and T represents taxes. As discussed above, the cost function has
advantages to the production function.
4.2. Estimates
Eberts (1986) estimates the direct effect of public capital stock on manufactur-
ing output and the technical relationships between public capital and the other
production inputs. Public capital stock is estimated using the perpetual inventory
technique for each of 38 US metropolitan areas between 1958 and 1978. With this
method, capital is measured as the sum of the value of past investments adjusted
for depreciation and discard. Public capital stock includes highways, sewage
treatment facilities and water distribution facilities within the MSA. Eberts es-
timates a translog production function with value added as output. hours of pro-
duction and nonproduction workers as the labor input, and a value measure of
private manufacturing capital stock as private capital.
Eberts (1986) finds that public capital stock makes a positive and statistically
significant contribution to manufacturing output, supporting the concept of public
capital stock as an unpaid factor of production. Its output elasticity of 0.03 is
small relative to the magnitudes of the other inputs: 0.7 for labor and 0.3 for
private capital. It follows that the magnitude of the marginal product of public
capital is also small relative to the marginal product of private capital.2 1
Dalenberg (1987), using the same public capital stock as Eberts (1986), esti-
mates a cost function, with public capital stock as a quasifixed input, for 31 MSAs
for the years 1976 to 1978. He finds that a 10% increase in public capital stock
is associated with a 2% decrease in manufacturing costs. This shadow price is
close to the average imputed price of private capital, suggesting that the average
return to public capital is on par with that of private capital. Dalenberg (1987)
computes a shadow price for public capital for each of the 31 MSAs. Comparing
the shadow price of public capital with the imputed price of private capital (as
a proxy for the opportunity cost of investing) indicates that 26 of the 31 MSAs
have public capital stock at least as great as the desired level.
Deno (1988) estimates a profit function for 36 MSAs from 1970 to 1978,
using the same public capital stock as Eberts (1986) and Dalenberg (1987). To
take account of the shared nature of public capital, Deno (1988) multiplies the
public capital stocks by manufacturing's share of total metropolitan population.
He finds that output strongly responds to public capital: the output elasticities
of water, sewer and highway infrastructure are 0.08, 0.30 and 0.31, respectively.
21 With respect to technical relationships, Eberts (1986) and Dalenberg (1987) find that public capital and
private capital are complements.
Ch. 38: Agglomeration Economies and UrbanPublic Infrastructure 1489
Unlike those of Eberts and Dalenberg, Deno's estimates suggest a significant
underinvestment in public capital stock.
In general, these estimates of the effect of public capital stock on output are
smaller than those found at the state and national level. Some argue that the
more disaggregated studies cannot pick up the broad externalities generated by
the networks of public capital stock, such as the US interstate highway sys-
tem. However, another explanation argues that national time-series estimates and
state pooled cross-section time-series estimates are subject to estimation bias.
The national time series estimates are plagued by nonstationarity and spurious
correlation. Correcting for nonstationarity, Tatom (1991) finds that the effect of
public infrastructure on output is negative and statistically insignificant. State-
level estimates are criticized because of unexplained differences across states.
Holtz-Eakin (1992), using a state fixed-effects model, finds that the strong ef-
fect of infrastructure diminishes or disappears if the state-specific effects are
taken into account. McGuire (1992), on the other hand, finds a strong positive
and statistically significant effect of public capital stock on output under several
specifications, including a Cobb-Douglas production function with state fixed
and random effects.
Another criticism of state and national studies is the question of causality.
Does public infrastructure affect output, or does output affect public infrastruc-
ture? At the metropolitan level, Duffy-Deno and Eberts (1989) test the direction
of causation between infrastructure and output by positing two equations: a pro-
duction function and a demand equation for public capital. Their findings suggest
that the causation runs mostly from infrastructure capital to output growth. They
also find a smaller elasticity of output with respect to public capital when the
demand for public capital is taken into account.
Attention in Europe has also been drawn to the effect of public infrastructure
on output. Similar to the US, few studies have focused on metropolitan areas.2 2
Seitz (1993) is an exception. He estimates the effect of public infrastructure
investment on manufacturing cost for 85 (former West) German cities between
1980 and 1989. Public capital stock is estimated by applying the perpetual in-
ventory method to city-specific outlay data from 1969 to 1989, but using starting
values estimated from state-specific data. His cost elasticity estimate with re-
spect to public capital is -0.127. This estimate is smaller than that obtained by
22 Many more studies of the effect of public infrastructure on output have been conducted for European
countries, but these have focused on regional or national levels of aggregation. Berndt and Hansson (1991)
conclude that public infrastructure in Sweden reduces private productions costs. Using data from the UK,
Lynde and Richmond (1992) find an average elasticity of output with respect to public capital of 0.20, that is
within the range of many national-level studies for the US.
1490 R. W Eberts and D.P.McMillen
Dalenberg (1987) for the US, but is remarkably similar given different govern-
mental structures and national histories. In addition, Seitz (1983) includes city
population and taxes in the estimation, that Dalenberg (1987) did not include.
US metropolitan-level studies, or most of the other studies, have not incor-
porated taxes into the estimation. However, doing so may increase the estimated
effects of public infrastructure on output, particularly if local public infrastruc-
ture is financed through a distortionary tax. Jorgenson calculates that every $1
raised by taxes extracts $1.45 from the private sector, with the additional $0.45
resulting from the distortion of economic activities. Tax financing by local ju-
risdictions would probably cause greater distortion because of the mobility of
firms and households among jurisdictions. Therefore, the combined net effect of
infrastructure and taxes (when taxes are excluded) would be lower than the indi-
vidual effect of infrastructure on productivity if taxes were included separately
in the estimation.
The studies by Mera (1973), Moomaw (1983), Calem and Carlino (1991) and
Seitz (1993) are exceptions to the general tendency not to include measures of
both agglomeration and public infrastructure simultaneously in an analysis or
the productivity. Mera (1973) highlights the importance of public infrastructure
in achieving agglomeration economies and includes the level of employment as
a measure of localization economies. He contends that "the higher per capita
income in high-density areas can be explained by both savings in social over-
head capital cost and increased efficiency of inputs" (p. 318). Cities with more
investment in social overhead capital, defined by Mera as all capital improvement
with social implications, will have less diseconomies and greater net productivity
for a given city size. Therefore, including measures of public infrastructure in
the production (cost) function identifies the separate contributions of agglom-
eration economies and public capital stock to productivity advantages of larger
cities. Mera finds positive and statistically significant effects of both employment
and social overhead capital on the productivity of broad industries for Japanese
regions.
Moomaw's (1983) analysis includes a measure of transportation infrastructure
along with a population variable that measures net agglomerative effects. He
finds that for several two-digit industries, the transportation variable is positive
and statistically significant, indicating that public infrastructure, even when in-
cluded in an equation with population, has a positive effect on the productivity
Ch. 38: Agglomeration Economies and UrbanPublic Infrastructure 1491
of some US industries within metropolitan areas. Adding public infrastructure to
the equation reduces the magnitude of the elasticity of population only slightly.
Calem and Carlino (1991) also recognize that infrastructure and resource en-
dowment are likely to affect urban productivity. They acknowledge the difficulty
in measuring these factors, but they also recognize the potential bias of omit-
ting these factors. To circumvent the measurement problem, Calem and Carlino
estimate metropolitan-level production functions using a fixed-effects approach.
Entering cross-sectional dummy variables accounts for differences in the level of
the output, holding inputs constant, across US cities.
Seitz (1993) explicitly models and estimates the effect of both public in-
frastructure investment and agglomeration economies on private sector produc-
tivity within urban areas in Germany. He incorporates measures of public in-
frastructure and agglomeration economies within a translog specification of a
cost function. Public infrastructure, recorded on a per capita basis, is estimated
by applying the perpetual inventory method; agglomeration economies are mea-
sured by the total employment in each city. His estimates indicate significant but
decreasing agglomeration economies and a cost elasticity of public infrastructure
of 0.127.
5. Conclusions
The two strands of empirical research discussed in this chapter portray an urban
landscape in which the productivity of firms vary by city size and the level of
investment in public infrastructure. The considerable attention given to these
two topics in the last three decades has been generated by concerns about the
economic and social well-being of cities, particularly in the US, and by asking
what role government can take in promoting a more efficient spatial distribution
of economic activities. Studies in several countries find that manufacturing firms
are more productive in large cities than in smaller ones, and in cities with a larger
stock of public infrastructure. Studies covered in this chapter found positive ag-
glomeration and infrastructure effects in the US, Germany, Sweden, Japan and
Brazil. For the most part, the results hold up under various specifications of pro-
duction relationships and measures of agglomeration, and even when measures
of public infrastructure and agglomeration economies are entered simultaneously
in production functions (or cost functions).
One of the shortcomings of this literature, and a fruitful area for future re-
search, is the general failure to consider the effects of agglomeration economies
and public infrastructure simultaneously. As the theoretical papers clearly demon-
1492 R. W Eberts and D.P McMillen
strate, the two effects are closely related. Urban infrastructure is a shared input
that provides the means by which the close spatial proximity of economic ac-
tivities can lead to increased productivity for all parties. The dynamics of this
process is another area that deserves further attention. Only recently have studies
started to explore the length of time cities take to respond to economic shocks.
Moreover, the literature gives us little insight into whether, or not, the produc-
tivity advantages of cities of different sizes and infrastructure investment have
changed since these topics were first introduced.
References
Abdel-Rahman, H.M. (1990a), "Agglomeration economies, types, and sizes of cities", Journal of Urban
Economics 27:25-45.
Abdel-Rahman, H.M. (1990b), "Sharable inputs, product variety, and city sizes", Journal of Regional Science
20:359-374.
Abdel-Rahman, H.M. (1996), "When do cities specialize in production", Regional Science and Urban
Economics 26:1-22.
Abdel-Rahman, H.M. and M. Fujita (1990), "Product variety, Marshallian externalities and city size", Journal
of Regional Science 30:165-183.
Abdel-Rahman, H.M. and M. Fujita (1993), "Specialization and diversification in a system of cities", Journal
of Urban Economics 33:189-222.
Abdel-Rahman, H.M. and P. Wang (1995), "Toward a general-equilibrium theory of a core-periphery system
of cities", Regional Science and Urban Economics 25:529-546.
Abdel-Rahman, H.M. and P. Wang (1997), "Social welfare and income inequality in a system of cities", Journal
of Urban Economics 41:462-483.
Aberg, Y. (1973), "Regional productivity differences in Swedish manufacturing", Regional and Urban
Economics 3:131-156.
Anas, A. and I. Kim (1996), "General equilibrium models of polycentric urban land use with endogenous
congestion and job agglomeration", Journal of Urban Economics 40:232-256.
Appold, S.J. (1995), "Agglomeration, interorganizational networks, and competitive performance in the U.S.
metalworking sector", Economic Geography 71(1):27-54.
Aschauer, D.A. (1989), "Is public expenditure productive?", Journal of Monetary Economics 23:177-200.
Aydalot, P. and D. Keeble. (1988), High Technology Industry and Innovative Environments: The European
Experience (Routledge, London).
Beeson, PE. (1983), "Essays in regional productivity", unpublished Ph.D. thesis (Department of Economics,
University of Oregon).
Beeson, P.E. (1987), "Total factor productivity growth and agglomeration economies in manufacturing 1959-
1973", Journal of Regional Science 27:183-199.
Berndt, E. and B. Hansson (1991), "Measuring the contribution of public infrastructure capital in Sweden",
NBER Working Paper, No. 3842.
Blair, J.P. (1991), Urban and Regional Economics (Irwin, Boston).
Calem, P.S. and G.A. Carlino (1991), "Urban agglomeration economies in the presence of technical change",
Journal of Urban Economics 29:82-95.
Carlino, G.A. (1978), Economies of Scale in Manufacturing Location: Theory and Measurement (Martinus
Nijhoff, Leiden, The Netherlands).
Carlino, G.A. (1979), "Increasing returns to scale in metropolitan manufacturing", Journal of Regional Science
19:363-373.
Ch. 38: Agglomeration Economies and UrbanPublic Infrastructure 1493
Carlino, G.A. (1982), "Manufacturing agglomeration economies as returns to scale: A production function
approach, Papers of Regional Science Association 50:96-108.
Chinitz, B. (1961), "Contrasts in agglomeration: New York and Pittsburgh", American Economic Review
51:279-289.
Chipman, J.S. (1970), "External economies of scale and competitive equilibrium", Quarterly Journal of
Economics 84:347-385.
Choate, P. and S. Walter (1983), America in Ruins (Duke Press Paperbacks, Durham, NC).
Dalenberg, D. (1987), "Estimates of elasticities of substitution between public and private inputs in the
manufacturing sector of metropolitan areas", Ph.D. dissertation (University of Oregon).
Deno, K.T. (1988), "The effect of public capital on U.S. manufacturing activity: 1970 to 1978", Southern
Economic Journal 55:400-411.
Dobkins, L.H. (1996), "Location, innovation and trade: The role of localization and nation-based externalities",
Regional Science and Urban Economics 26:591-612.
Duffy-Deno, K. and R. Eberts (1989), "Public infrastructure and regional economic development: A simulta-
neous equations approach", Working Paper No. 8909 (Federal Reserve Bank of Cleveland).
Eberts, R.W. (1986), "Estimating the contribution of urban public infrastructure to regional economic growth",
Working Paper No. 8620 (Federal Reserve Bank of Cleveland).
Eberts, R.W. and J. Stone (1992), Wage and Employment Determination in Local Labor Markets (W.E. Upjohn
Institute for Employment Research Monograph).
Eberts, R.W., D. Dalenberg and C.S. Park (1986), "Public Infrastructure Data Development for NSF", mimeo
(University of Oregon).
Fogarty, M.S. and G.A. Garofalo (1978), "Environmental quality-income tradeoff functions with policy
applications", paper presented at the Southern Regional Science Association Meeting.
Fogarty, M.S. and G.A. Garofalo (1988), "Urban spatial structure and productivity growth in the manufacturing
sector of cities", Journal of Urban Economics 23:60-70.
Friedlaender, A.F. (1990), Comment: "How does public infrastructure affect regional economic performance?",
in: Is There a Shortfall In Public Capital Investment?, pp. 108-112.
Fujita, M. and P. Krugman (1995), "When is the economy monocentric?: von Thtinen and Chamberlin Unified",
Regional Science and Urban Economics 25:505-528.
Fuss, M. and D. McFadden (1979), Production Economics: A Dual Approach to Theory and Application
(North-Holland, Amsterdam).
Goldstein, G.S. and T.J. Gronberg (1984), "Economies of scope and economies of agglomeration", Journal of
Urban Economics 16:91-104.
Gramlich, E.M. (1994), "Infrastructure investment: A review essay", Journal of Economic Literature 32:1176-
1196.
Hanson, G.H. (1996), "Agglomeration, dispersion, and the pioneer firm", Journal of Urban Economics 39:255-
281.
Helsley, R.W. (1990), "Knowledge and production in the CBD", Journal of Urban Economics 28:391-403.
Helsley, R.W. and W.C. Strange (1990), "Matching and agglomeration economies in a system of cities",
Regional Science and Urban economics 20:189-212.
Helsley, R.W. and W.C. Strange (1991), "Agglomeration economies and urban capital markets", Journal of
Urban Economics 29:96-112.
Helsley, R.W. and A.M. Sullivan (1991), "Urban subcenter formation", Regional Science and Urban Eco-
nomics 21:255-275.
Henderson, J. (1974), "The size and type of cities", American Economic Review 64:640-656.
Henderson, J. (1982a), "The impact of government policies on urban concentration", Journal of Urban
Economics 12:280-303.
Henderson, J. (1982b), "Systems of cities in closed and open economies", Regional Science and Urban
Economics 12:325-350.
Henderson, J. (1983), "Industrial bases and city sizes", American Economic Review 73:164-168.
Henderson, J. (1985), Economic Theory and the Cities, 2nd ed. (Academic Press, New York).
Henderson, J. (1986), "The efficiency of resource usage and city size", Journal of Urban Economics 19:47-70.
Henderson, J. (1988), Urban Development (Oxford University Press, New York).
1494 R. W. Eberts and D.P. McMillen
Henderson, J.V. and A. Mitra (1996), "The new landscape: Developers and edge cities", Regional Science and
Urban Economics 26:613-643.
Holtz-Eakin, D. (1992), State-Specific Estimates of State and Local Government Capital (Syracuse University
Working Paper).
Hoover, E.M., Jr. (1937), Location Theory and the Shoe and Leather Industries (Harvard University Press,
Cambridge).
Jacobs, J. (1969), The Economy of Cities (Random House, New York).
Jovanovic, B. and R. Rob (1989), "The growth and diffusion of knowledge", Review of Economic Studies
56:569-582.
Kaldor, N. (1970), "The case for regional policies", Scottish Journal of Political Economy 17:337-347.
Kawashima, T. (1975), "Urban agglomeration economies in manufacturing industries", Papers of the Regional
Science Association 34:157-175.
Kim, S. (1989), "Labor specialization and the extent of the market", Journal of Political Economy 97:692-705.
Kim, S. (1990), "Labor heterogeneity, wage bargaining, and agglomeration economies", Journal of Urban
Economics 29:160-177.
Kim, S. (1991), "Heterogeneity of labor markets and city size in an open spatial economy", Regional Science
and Urban Economics 21:109-126.
Lynde, C. and J. Richmond (1992), "The role of public capital in production", The Review of Economics and
Statistics 74:37-44.
Marshall, A. (1890), Principles of Economics, 8th ed. (1977 reprint) (Macmillan Press, London).
McGuire, T.J. (1992), "Highways and macroeconomic productivity Phase Two", Report to Federal High-
way Administration, US Department of Transportation (Volpe national Transportation Systems Center, US
Department of Transportation).
Mera, K. (1973), "Regional production functions and social overhead capital:An analysis of the Japanese case",
Regional and Urban Economics 3:157-186.
Mills, E.S. (1967), "An aggregative model of resource allocation in a metropolitan area", American Economic
Review 57:197-210.
Mills, E.S. (1972), Studies in the Structure of the Urban Economy (Johns Hopkins University Press, Baltimore).
Mills, E.S. and B.W. Hamilton (1994), Urban Economics, 5th ed. (Harper Collins, New York).
Moomaw, R.L. (1980), "Optimal firm location and city size: Productivity vs. wages", Urban Institute Working
Paper 1447-1.
Moomaw, R.L. (1981), "Productivity and city size: A review of the evidence", Quarterly Journal of Economics
96:675-688.
Moomaw, R.L. (1983), "Is population scale a worthless surrogate for business agglomeration economies?",
Regional Science and Urban Economics 13:525-545.
Moomaw, R.L. (1985), "Firm location and city size: Reduced productivity advantages as a factor in the decline
of manufacturing in urban areas", Journal of Urban Economics 17:73-89.
Mori, T. (1997), "A modeling of megalopolis formation: the maturing of city systems", Journal of Urban
Economics 42:133-157.
Munnell, A., editor (1990), "Is there a shortfall in public capital investment?" Conference Series No. 34
(Federal Reserve Bank of Boston).
Nakamura, R. (1985), "Agglomeration economies in urban manufacturing industries: a case of Japanese cities",
Journal of Urban Economics 17:108-124.
National Council on Public Works Improvement (1988), Fragile Foundations: A Report on America's Public
Works, Final Report to the President and Congress.
Ohlin, B. (1935), Interregional and International Trade (Harvard University Press, Cambridge).
O'Sullivan, A.M. (1996), Urban Economics, 3rd ed. (Irwin, Boston).
Panzar, J.C. and R.D. Willig (1981), "Economies of scope", Pap. Proc., American Economic Association
71:268-272.
Palivos, T. and P. Wang (1996), "Spatial agglomeration and endogenous growth", Regional Science and Urban
Economics 26:645-669.
Rauch, J.E. (1993), "Productivity gains from geographic concentration of human capital: evidence from the
cities", Journal of Urban Economics 34:380-400.
Ch. 38: Agglomeration Economies and Urban PublicInfrastructure 1495
Richardson, H.W. (1973), The Economics of City Size (Lexington Books, Lexington).
Richardson, H.W. (1979), "Optimality in city size, systems of cities and urban policy", Regional Economics
(University of Illinois Press, Chicago).
Roback, J. (1982), "Wages, rents, and the quality of life", Journal of Political Economy 90.1257-1278.
Sasaki, K. and Se-Il Mun (1996), "A dynamic analysis of multiple-center formation in a city", Journal of Urban
Economics 40:257-278.
Segal, D. (1976), "Are there returns to scale in city size?", Review of Economics and Statistics 58:339-450
Seitz, H. (1993), "The impact of the provision of urban infrastructures on the manufacturing industry in cities",
paper presented at the 33rd European Congress of the Regional Science Association, Moscow, Russia.
Shefer, D. (1973), "Localization economies in SMSAs: A production function analysis", Journal of Regional
Science 13:55-64.
Smith, A. (1776), The Wealth of Nations (1976 reprint) (University of Chicago, Chicago).
Sullivan, A.M. (1983), "A general equilibrium model with external scale economies in production", Journal of
Urban Economics 13:235-255.
Sullivan, A.M. (1986), "A general equilibrium model with agglomerative economies and decentralized
employment", Journal of Urban Economics 20:55-74.
Sundquist, J.L. (1970), "Where shall they live?", The Public Interest 18:88-100
Sveikauskas, L.A. (1975), "The productivity of cities", Quarterly Journal of Economics 89:393-414.
Sveikauskas, L., J. Gowdy and M. Funk. (1988), "Urban productivity: city size or industry size", Journal of
Regional Science 28(2):185-202.
Tabuchi, T. (1986), "Urban agglomeration, capital augmenting technology, and labor market equilibrium",
Journal of Urban Economics 20:211-228.
Tatom, J. (1991), "Should government spending on capital goods be raised?", Review of the Federal Reserve
Bank of St. Louis 72:3-15.
Vining, D.R., Jr. and T. Kontuly (1977), "Increasing returns to city size in the fact of an impending decline in
the size of large cities: which is the bogus fact?" Environmental and Planning A 9:59-62.
von Bdventer, E. (1970), "Optimal spatial structure and regional development", Kyklos 23:903-924.
Walz, U. (1996), "Transport costs, intermediate goods, and localized growth", Regional Science and Urban
Economics 26:671-695.
Chapter39
GRAHAM R. CRAMPTON
Contents
1. Definitional and introductory issues 1501
1.1. Introduction 1501
1.2. Local labour market: definitions and operational problems 1502
2. Residential and workplace location 1504
2.1. Employment subcentres and the wasteful commuting controversy 1504
2.2. Simultaneous modelling of residential and workplace location 1504
3. Moves and quits, using micro and subaggregate data 1507
3.1. Empirical modelling of intraurban moves and/or quits in workplace and job-related
residential change 1507
4. The role of job search 1509
4.1. Labour market search mechanisms 1509
4.2. Some theoretical extensions to spatial labour market search models 1513
4.3. Racial differences in labour market search activity 1514
5. Implications of the changing structure of skill and status 1517
5.1. Variation in labour market behaviour with respect to skill, status and income 1517
5.2. Dual urban labour markets, the urban underclass, and the changing structure of labour
demand 1518
6. Impact of gender, family structure and demographic factors on
urban labour market behaviour 1521
6.1. Theoretical contributions 1521
6.2. Empirical research on gender, family structures and urban labour markets 1522
7. Spatial mismatch-residential segregation by race in cities, and
its implications for earnings and employment differences by race 1528
7.1. Original formulation 1528
7.2. More detailed research on racial segregation in urban residential models 1530
7.3. Reviews and descriptive evidence lying behind spatial mismatch studies 1531
7.4. Empirical research on the spatial mismatch 1533
7.5. Some recent research using microsamples 1538
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P. Cheshire
© 1999 Elsevier Science B. V All rights reserved.
1499
1500 G. R. Crampton
Abstract
The large research literature in urban labour market analysis is reviewed, with the
emphasis ranging from attempts to model aggregate simultaneous interactions
between residential and workplace location to more modem econometric work
researching individual labour market behaviour. The job search process is central
to the operation of the labour market, yet research is hampered by variable data
availability and the limited observability of the search mechanism. Individual
responses to major employer relocations have been recently studied, especially
the substitutability of the move or quit decision, and relationships to race and
gender.
The variation of commuting patterns by income and professional status has
also been analysed, and the functioning of the dispersed service-dominated mod-
em urban labour market raises challenging research issues including willingness
to search and commute over substantial areas, interacting with family circum-
stances and expected job security.
The continued growth in all developed economies of female labour force par-
ticipation and numbers of female-headed households have raised the importance
of urban labour market research focusing on gender, including the economic
understanding of patterns in the length of the female journey to work.
Study of the influence of racial segregation on outcomes in the urban labour
market has a longer history, with the "spatial mismatch hypothesis" having de-
veloped a large literature since the 1960s. With higher quality microdata and
modern computational power and econometric techniques, statistical research
has advanced considerably. Similar spatial relationships between race and labour
market and commuting outcomes are also intrinsic to the European urban labour
market, and have received particular attention from British, Dutch, Austrian and
Ch. 39: Urban Labour Markets 1501
Keywords: Dual labour market, gender, ghetto, high technology, job search, lo-
cal labour market area (LLMA), logit, mismatch, quit, reservation wage, skill,
unemployment duration, vacancy, wage gradient, youth unemployment
1.1. Introduction
ing of the relationships between family structure, or gender and urban labour
market outcomes (Section 6), into a more central part of urban economics.
The "spatial mismatch hypothesis" is one of the more long-established prob-
lem areas within urban labour market research (Section 7), and focused on the
labour market implications of residential segregation by race. In this area, the
seminal articles are now nearly 30 years old. The level of sophistication in the
econometric analysis of the mismatch hypothesis has grown substantially, and
European trends in ethnically diverse migration has sustained the interest in and
importance of this area. Regrettably, in this area as with most of this chapter we
found it difficult to extend the review much beyond the US and UK literature. A
limited amount of analysis of urban labour market conditions in various European
countries (e.g., France, Netherlands, Austria) has been published in accessible
English language sources, but this is a direction in which international academic
communication could be improved.
Overall, the sheer size of publicly available comprehensive data sources in
the US has given American researchers a considerable advantage, especially in
the development of microeconomic applied research. Standards of econometric
modelling have been set which relatively little of the European work has ap-
proached, till recently. Continued improvements of uniformly defined European
data sources, perhaps through Eurostat, may offer the potential to improve this
situation.
A number of the key issues in urban labour markets that will be discussed in
greater detail in the course of this chapter have arisen in connection with attempts
to define local labour markets operationally. These include the variation of search
areas and willingness to commute with the skill and economic status of workers,
or the possibility that local labour market areas (LLMAs) once defined may be
"dynamic". That is, they may be overtaken by macroeconomic events and rapidly
become less relevant and accurate for policy purposes.
An early awareness was found in Goldner (1955) of the likelihood that low
status workers may function with smaller LLMAs, stressing both the demand
side and search motives:
A more theoretical early paper by Termote (1978) extended the Losch model
of central place and urban hierarchy by including personal commuting costs.
Many features of the Lsch hierarchy remain, but the variation of wage rates
with urban centre size, required to compensate for commuting costs, would bring
about greater dispersion of urban centre size than in the basic model, and the
familiar neatly nested hexagonal pattern would not be stable.
European empirical research on local labour market operation has been hand-
icapped by the rather varied nature of the commuting data available. Austrian
data was used by Baumann et al. (1988). They addressed important issues such as
how to operationalise nondisjoint (i.e., overlapping) spatial labour markets, and
stressed the importance of estimating the parameters and regionalisation jointly.
An Israeli example came from Felsenstein (1994). This was a study of the
effect of the widespread Israeli practice of employers paying commuting costs,
either in the form of free bus provision, or liberal car expenses. Felsenstein found
that this had helped to bring about a surprisingly extensive spatial labour market
for low skill labour, strengthened by the low residentialmobility of Israeli labour.
His results were backed up by empirical estimation of the operation of spatial
labour markets for the Lod Airport and Haifa subcentres.
Most British work on LLMAs has been geared towards continued definition
or modification of the travel to work areas (TTWAs) used for the reporting of
unemployment rate statistics, and for the implementation of regional policy. A
weakness of these areas is perhaps the insistence on an exhaustive allocation of
all land areas, including nonurban land.
A minimum commuting self-containment level of 75% was introduced (Smart,
1974), and further work (Coombes and Openshaw, 1982) extended the defin-
itions to distinguish between high density multinodal metropolitan areas and
free-standing towns. In principle, the data-user should:
Other British work has focused on the variability over the business cycle of
LLMAs (Ball, 1980). This may be caused by a mixture of demand-side fac-
tors such as labour-hoarding by firms, who may shed in recession the shorter-
commute part of their labour force, with the supply motive of workers searching
1504 G. R. Crampton
Given the coverage of this part of the subject elsewhere in this volume (see
White, Chap. 36), we have omitted any review material for this section to avoid
duplication.
. The major stumbling block may not be model development but model
estimation and testing, since the spatially disaggregated panel microdata
best suited to this analysis will be difficult to find and expensive to develop.
(Simpson and Van Der Veen, 1992: p. 60)
The principal issues in the field of workplace and residential modelling have
been first, what plays the primary role in urban adjustment processes. That is, is
it possible to generalise about whether the employment distribution "drives" the
residential distribution, or vice versa? Second, the question remains of compen-
sation in the labour market for commuting distance, and the extent to which es-
timations of such compensation interact with the operation of the urban housing
market.
Ch. 39: Urban Labour Markets 1505
Early work on workplace and residential location emerged from the enormous
urban transportation studies of the 1960s. Kain (1962) summarised, in the light of
the (then new) monocentric urban model, broad patterns revealed by a large strat-
ified sample of Detroit households. However, despite its size, this was a single
cross-section dataset with no evidence on moves or other dynamic adjustment.
Beesley and Dalvi (1974) discussed an early version of the labour offer curve
from a fixed residence, with the acceptable wage rising with additional commut-
ing distance, and presented simple analysis of British Census data on the journey
to work by local authority. Possible interactions between car ownership, search,
and the commuting pattern were also mentioned.
Other simple empirical approaches included Steinnes (1977), who used Cen-
sus aggregates on population and broad employment categories for a pooled
sample of 15 large US MSAs, and claimed to find evidence that causality runs
from residence to employment, at least for certain sectors of employment, in-
cluding manufacturing. Gera and Kuhn (1980a, b) used Canadian Census data,
with tracts aggregated into zones. They found some suggestive evidence that
there may be a U-shaped relationship between a zone's Central Business District
(CBD) distance and the average commuting distance to that zone.
Some studies of zonal subaggregate data (e.g., Boarnet (1994)) included the
econometric modelling of spatial autocorrelation, and used maximum likelihood
techniques rather than ordinary least squares (OLS). Boarnet also found evidence
that it was more clearly the case that employment change was endogenous than
population change. This would seem to agree with Mills' (1986) finding that,
as between US central cities and suburbs, "jobs follow people" much more than
"people follow jobs".
Vickerman (1984) used British journey to work data for London and the outer
metropolitan local authorities and conducted OLS and two stage least squares
(2SLS) estimations with net incommuting as the dependent variable, and employ-
ment and migration endogenous. These estimates followed a thoughtful theoreti-
cal discussion of the problems of simultaneous modelling of intraurban moves of
residence or workplace or both, including considerations of various motives for
not moving (e.g., search and then conclude the net gain is negative, search and
then give up, do not search at all). From survey evidence on a separate sample
of about 1000 households, Vickerman concluded that, if anything, employment
factors were more likely to influence a residential move than vice versa, although,
... there is little immediate connection between the two and adjustments are
not made in an obviously fully optimising manner.
(Vickerman, 1984: p. 26)
1506 G. R. Crampton
Such detailed intraurban work was emerging at the same time as mainline
labour economics publications on applications of maximum likelihood logit tech-
niques to inter-MSA moves, such as Bartel (1979), which reiterated the important
role played by job tenure in reducing job move probabilities.
A series of studies of migration using British Census migration data from
1971 and concentrating on London emerged in the 1980s. Gordon and Vickerman
(1982) proposed a national 112-zone British migration model, with the London
metropolitan area modelled in greater detail than the rest of the country. Three
superimposed migration streams, with three separate distance decay parameters,
were included. These represented the complex mixture of short distance intra-
labour market moves, regional spatial choice, sometimes but not always with both
job and residence move, and long distance moves for job-related motives. Factor
analysis techniques were then used to identify motives for moves. However, the
explicit separation of the three migration streams was subject to the weakness that
the distance decay parameters were, in effect, imposed through "experimentation
and choice", and nonlinear methods to estimate them were unsuccessful.
Gordon and Lamont (1982) used this migration model as part of an ambitious
11-equation simultaneous model, with in- and outmigration under the three sepa-
rate streams discussed above, together with net commuting, employment growth,
unemployment, house building and owner-occupied house prices all treated as
endogenous. There are also 19 exogenous variables contributing to 2SLS and
3SLS estimations, making this one of the largest efforts at structural modelling of
an intraurban labour and housing market adjustment model. Policy implications
include:
... clear conclusions at least about the irrelevance of localised job creation
for the solution of pockets of high unemployment in the inner city.
(Gordon and Lamont, 1982: p. 259)
Zax (1994) also emphasised that job and residential change could be com-
plements in long-distance moves, but substitutes in shorter-distance intraregional
moves, so that pooling these data was inappropriate.
N- 3
s2/ (4.1)
B2 /3
V of course will vary cyclically for macroeconomic reasons. The model gen-
erated an implication of diminishing returns to search, whereas, many earlier
papers had just assumed this. This conclusion resulted from the less productive
nature of searching further away from the residence. The empirical estimations of
the power of search time in the formula giving the number of vacancies contacted
(two-thirds in the simple theory above) was then an issue of research interest,
with a value below unity implying diminishing returns to search. There was also
the implication that equiproportional increases in total vacancies and firms would
raise the number of vacancy contacts, implying certain "economies of labour
market density" in the biggest urban areas.
Barron and Gilley (1981) were able to estimate the Seater model using micro-
survey data. The "density of potential employers" was only crudely proxied by
a MSA dummy, and the exponent on search time came out at 0.45, confirming
diminishing returns to scale, although the spatial content as such is minimal.
In a separate econometric model, Barron and Gilley also presented maximum
likelihood estimates of a multinomial logit formulation with the employment
probability as the dependent variable.
An alternative empirical effort to model labour market search with a simple
spatial content was made by Chirinko (1982), using maximum likelihood estima-
tion of a Poisson distribution format on Current Population Survey data. This is
sensible given that the number of employer contacts is of course a nonnegative
integer. The decreasing returns to search are simply represented by a (negative)
quadratic term in total search time, and spatial density of potential employers
again simply proxied with a MSA dummy. This paper also made an early con-
tribution to the "different methods of search" debate (see Section 4.2). However,
Ch. 39: Urban Labour Markets 1511
one naturally looks for a spatial job search format extending beyond a MSA
dummy.
Simpson (1980) was perhaps the first substantial paper to apply search theory
ideas to the urban labour market and try to identify their outcome empirically,
extended later in Simpson (1992). The principal issue addressed here was the
relationship between skill or education level and the spatial extent of job search.
This can potentially explain the seemingly paradoxical tendency for higher in-
come workers to commute longer distances, even to nonCBD jobs (i.e., not just
for residential decentralisation reasons).
The data used in the Simpson (1980) study is a large sample of London house-
hold interviews. The problem is to confirm empirically whether, or not, for labour
market search reasons, higher income people are willing to commute on average
further, despite valuing their travel time more highly, and in a way in which
commuting distance is not separately compensated by lower property prices.
The London data also confirm the quantitative importance of job turnover (about
twice as frequent on average) relative to residential change, so that "choice of job
search area" from a predetermined residence will normally affect a significant
proportion of the labour force, whether searching unemployed or on-the-job.
The intensity of search (part of which is represented by greater search area) is
then thought to rise with skill and education level, despite the cost of search also
being higher. This is because of differences in productivity of search, formality
of search method, or the urgency of accepting an early offer.
The principal difficulty in econometrically modelling search mechanisms in
the urban labour market is the degree to which job search models and Alonso-
Muth type residential location models generate similar predictions with respect to
the variation of skill/education/income and commuting distance. That is, within
the search mechanism itself, a simultaneity exists of commuting distance from
a fixed residence generating demand for wage compensation, and higher wage
jobs themselves being associated with more extensive search. This is complicated
by the familiar monocentric residential location mechanism of compensation for
commuters by lower property prices at greater CBD distance.
Simpson's (1980) micro-London data, despite the impressive sample size,
had centre-job and centre-residence detail, without the journey-to-work trip it-
self. 2SLS techniques were applied to these linear distance variables, with skill
segmentation dummies trying to capture the suggested impacts of skill level on
commuting. These models were further developed in Simpson (1987, 1992).
More recent empirical work on the urban labour market search and job match-
ing mechanisms has focused on appropriate modelling of large microsamples.
O'Regan and Quigley (1993) were mainly concerned with the relative impor-
1512 G. R. Cramzpton
1 The printed coefficient in Table 6, p. 1555, is a typographical error, confirmed in correspondence with J.
Rouwendal.
Ch. 39: Urban Labour Markets 1513
and significant tendency was found for jobs with a higher level of education or
specific experience to involve wider recruitment areas, but larger firms also re-
cruited over a wider area. Previously unemployed applicants also were recruited
from significantly greater distances, more of a labour supply factor.
Van Ommeren et al. (1997) made use of a Dutch longitudinal microdata set
to study the commuting behaviour of individuals and relate it through maximum
likelihood estimations to their age, education, family status, and search-specific
factors such as the expected job offer arrival rate (an indicator of the relative
rareness of their skills) and whether they changed job the previous year (and
hence may be only partially adjusted). The theoretical prediction that people
who expect to receive more job offers will generally not have to accept a long
commute was strongly confirmed. The role of family structure was confirmed
by the finding that individuals with an employed working spouse also tended to
commute significantly further. Even with all the age, education, family status,
job change and search variables accounted for, males still tended to commute
significantly further.
In addition to these empirical efforts to gain insight into labour market search,
there have been a number of theoretical efforts. Much of this work, while de-
veloping search theory ideas, does so in a context in which commuting plays no
role, so that the relevance is more to an interurban than intraurban context.
For example, a series of papers by Rogerson and MacKinnon (1981, 1982),
and MacKinnon and Rogerson (1980), developed a Markov chain formulation
of the transmission of vacancies between locations, with a role for adaptive in-
terregional adjustment of "perceived vacancies" and an information transmission
mechanism proxied by lagged migration. Simulations endogenously represent
vacancy chains, and combined with exogenous private or public sector vacan-
cies produce cycles in migration flows. In Rogerson and MacKinnon (1981), the
emphasis was on the contrast between contracted versus speculative migration,
and Rogerson and MacKinnon (1982) developed policy prescriptions on spatially
selective advertising.
Another theoretical contribution to spatial labour market search modelling,
mainly oriented towards selective migration phenomena, came from Maier (1985,
1987). He suggested that many of the assumptions of standard search models
were empirically weak. For example, the knowledge of the wage distribution may
itself be accumulated through search. There is no reason why such models should
not be relevant to intraurban moves, as is Maier's discussion of the theoretical
1514 G. R. Crampton
A number of empirical studies have been done using microdata bases and at-
tempting to construct estimates of labour market search processes in which the
race of an individual affects either search behaviour or the outcome of search.
Rather than focusing directly on the outcome of ghettoisation, more general em-
pirical studies of search have race as simply one explanatory variable usually in
nonlinear hazard models.
The early work in this area of the literature was by Holzer (1986, 1987, 1988)
who formulated models of the choice of search method and level of reserva-
tion wage suitable for comparisons between racial subsamples. Holzer (1986)
used panel data on nonstudent youths and constructed models of the impact of
self-reported reservation wages on duration of unemployment and the wages
actually received. This carried certain implications over the degree to which
unemployment is actually "voluntary". It might be noted that whereas the theory
of the reservation wage is highly nonlinear, the estimates presented were of a
rather straightforward log-linear specification, although there is no evidence of
misspecification.
Interesting insights were gained about the comparative search behaviour of
blacks and whites. Among the control variables in regressions to explain unem-
ployment duration were simple spatial variables which may proxy local labour
demand (region and urban residence dummies) as well as personal supply side
factors such as schooling and even "presence of a library card". The main empiri-
cal finding was that the impact of a higher reservation wage on longer unemploy-
ment duration was found to be more severe for blacks than for whites, whereas
the impact of the reservation wage on the actual wage received was found to
be stronger for whites than for blacks. These results were striking evidence of
racial factors in hiring (with many personal characteristics controlled for). A
separate insight was that the severity with which the reservation wage increases
unemployment duration for blacks is more marked for the black unemployed
Ch. 39: Urban Labour Markets 1515
(i.e., those actively searching) than it is for the black nonemployed (i.e., those not
economically active at all, some of whom may be living off illicit activity).
The same sample was used by Holzer (1987) to study the racial detail of the
use of different search methods by white and black youth. The main objective
was to study the use of search using direct contacts compared to indirect meth-
ods using state agencies or newspapers. This paper also reviewed labour market
search theory in terms of the conditional probabilities, firstly of whether a search
technique is used, then whether an offer is received given the techniques used,
then whether the offer is accepted. The principal interracial difference related to
the rather low black use of informal (i.e., direct contact) search methods, and the
question of whether this indicated poor "networking", a low intensity of search,
or an acknowledgement that racial discrimination in hiring is less likely through
formalised hiring procedures such as agencies.
In Holzer (1988), a more detailed presentation of the same empirical study
included a theoretical model of job search. The theory generated comparative sta-
tics results showing the impact on choice of search method of the search method
cost, productivity, the level of nonwage income and the overall level of the wage
offer distribution. Simultaneous probit estimations were provided using a small
microsample, and detail on search methods used, as well as "employability"
characteristics of individuals and urban residence.
Although this work left one hoping for replicated studies with really large
data sets, it did shed light on questions crucial for the operations of urban labour
markets.
The same National Longitudinal Survey (NLS) Youth Cohort data source
with linked Census data, was used by Holzer, Ihlanfeldt and Sjoquist (1994).
The restriction to a youth sample has the effect of substantially removing the
endogeneity of the residential location decision, since youths living at home
normally have a residential location decided by parents. The principal objective
of this empirical approach was to attempt to identify the impact of MSA-wide
job decentralisation on the commuting distances of individuals, using a labour
market search model, including the effect of car ownership.
Only simple OLS estimates are offered based on merging micro and metro-
politan-wide evidence. The main findings were that black and inner city youths
1516 G. R. Crampton
do experience longer travel times to work, but this is partly the effect of lower
car ownership. It was also found that rapid job decentralisation by itself may not
cause longer commuting distances, but it does tend to result in lower employment
rates and longer unemployment duration among inner city blacks. This is clearly
closely related to the spatial mismatch debate.
A further contribution (Ihlanfeldt, 1997) was based on a detailed microstudy
of the degree to which blacks and whites accurately knew where the best job
opportunities were in the Atlanta metropolitan area. He found that the blacks
were worse informed, although information levels in general were poor. A logit
model of labour market knowledge suggested that central city residential location
contributed to making blacks badly informed, so that policies directed towards
improving the knowledge and efficiency of the labour market may be more cost-
effective than subsidised reverse commuting.
In terms of search activity, it was found that longer distance willingness to
commute (and implied longer distance search) did affect wage positively and
unemployment duration negatively. Of course, there were simultaneities with car
ownership, which results from higher wages and so through cheaper search costs
extends the area of search. The policy suggestion is that if public transport is
strikingly slower and more restricted in its routes than the private car, public
transport subsidy would not be an appropriate policy strategy.
The most detailed recent effort at incorporating spatial factors into a
Mortensen-type job search model was by Rogers (1997). Using a remarkable
spatially detailed microdata set of unemployment insurance claims in Pennsylva-
nia over 1980-1986, a logit model partly based on the work of Nickell (1979)
was formulated, with the probability of leaving unemployment affected by a
range of labour market, demographic and policy variables, but with an access
to employment index in addition.
ML estimation found that the access-weighted employment growth was in-
deed a significant variable in explaining the leaving-unemployment hazard.
The race dummy variable was included but found to be insignificant; it may be
partly collinear with a wide range of education and industry characteristics. The
implication appears to be that "access matters more than race".
Ch. 39: UrbanLabourMarkets 1517
5. Implications of the changing structure of skill and status
5.1. Variation in labour market behaviour with respect to skill, status and
income
Most applied work studying urban labour market behaviour of specific groups
has focused on questions of gender, family structure and race (covered below
in Sections 6 and 7). However, a small number of papers have tried to produce
useful generalisations about the operations of modem urban labour markets with
reference to the general skills and status distribution of workers. In more recent
work, much of the focus has been on the working of a mobile "high tech" labour
market in a single large urban area.
One of the earliest efforts, however, by Wheeler (1969), was a general study
of travel to work patterns by skill and status of workers commuting into the
Pittsburgh CBD. The fairly modest level of awareness of gender in commuting
at that time is revealed by the stress by Wheeler on the apparent paradox that
average commuting distance for women seemed to be fairly low (i.e., "distance
friction was high") despite their over-representation in higher status (white col-
lar) occupations. In contrast, the distance frictions for (mainly male) operatives
and labourers were fairly low. However, little attempt was made to incorporate in
detail the economics of residential location.
A more spatially specific study of high tech labour came from Scott (1992a),
as a detailed extension of a wider research programme to refine the understanding
of large, modern urban "industrial districts" (Scott, 1988). In the 1992a paper,
Scott conducted a questionnaire survey of engineers and scientists employed
at three Lockheed plants in the Los Angeles metropolitan area. There was a
strong overall sense of a large and complex agglomerated "local" labour mar-
ket, at metropolitan area scale. Job information was acquired partly by personal
contact through friends and relatives, partly through well-known interdependent
contacts with Southern Californian universities and colleges. Interindustry moves
within the high tech occupation range suggested agglomeration-specific skills.
One striking finding was the relative unimportance of distance to work in the
decision of whether to move residentially. The distributions of distances to work
in the final Lockheed job location were compared for residential movers and
nonmovers using a Kolmogorov-Smirnoff test and found to have no significant
difference. Although a modest sample size may weaken some of his conclusions,
an interesting picture was built up of a highly mobile and qualified type of urban
labour market evolution in which labour market agglomeration economies:
1518 G. R. Cramptoz
... reduce job search and recruitment costs by facilitating the retrieval of
local labor market information at low marginal cost.
(Scott, 1992a: p. 112)
5.2. Dual urban labourmarkets, the urban underclass, and the changing
structure of labour demand
The last decade has seen the development of a number of alternative approaches
to the study of labour markets from a more conceptual, political economy view-
point. Given the fact that these approaches do not generally make use of econo-
metric or other applied economic techniques, and usually develop their argu-
ments based on descriptive detail or recent historical evidence, we will restrict
ourselves to a brief outline. Several of these sources contribute alternative evo-
lutions of the urban labour market from the theory of the dual labour market,
much discussed in the 1970s. Piore and Sabel (1984) was one of the first popular
formulations of the flexible specialisation theory of industrial organisation. The
main emphasis in their work was on the industrial structure implications of com-
puterisation in production materials handling, and the greater flexibility in the
urban labour market it required. International comparisons of the US with four
alternative "faces" of flexible specialisation stressed the contrasts in centralised
Ch. 39: Urban Labour Markets 1519
... increasing polarisation of the labour market into low wage and high
wage sectors, (and) innovations in technology ...
(Wilson, 1992: p. 640)
be associated with them as a policy may not be a fair test of Affirmative Action
policies.
A useful review of the state of understanding of gender and urban labour
markets, in the light of rising female labour force participation (LFP) rates, was
provided by Madden and White (1980), discussed in Chapter 36 of this volume
by Michelle White. Their principal emphasis was on the implications of urban
spatial structure in continuing gender differences, or in the weakening over time
of familiar patterns of wage discrimination or job choice by gender. For example,
if women simply dislike long distance commuting more than men, the trend
towards higher female LFP in white collar service jobs might strengthen the
forces of gentrification. But women searching for jobs close to fixed suburban
residences will promote suburbanisation of jobs through providing a high quality
and possibly cheaper labour force.
Other work on family structure and urban labour market behaviour has ei-
ther been of an empirical nature, or has pursued a theoretical approach through
simulation, as in Hotchkiss and White (1993). They set up a model of spatial
allocation in a city with three household types, which are given separate Cobb-
Douglas utility functions and treated endogenously in an urban structure with a
CBD and two subcentres. Realistic patterns of desegregation of the spatial pattern
of household types are obtained, and stochastic wage distributions are assumed.
The detailed study of gender in urban labour markets has been closely associated
with the analysis of sex discrimination and other causes of wage differences
by gender in the economy as a whole. Sex discrimination can be assessed in
a similar way to racial discrimination (with which it may be mixed in complex
ways). Recently, more feminist viewpoints have focused on the manner in which
urban commuting patterns can help to generate or reinforce "occupational sex
segregation".
One of the earliest high quality papers on the subject of wage differentials by
gender in the economy as a whole was by Oaxaca (1973). This work only inciden-
tally included MSA size category as an explanatory variable in wage equations
attempting to separate the influences of occupational concentration from that of
different wages for equal work in explaining overall wage differentials. These
wage equations were estimated using a microsample from the 1967 Survey of
Economic Opportunity. Oaxaca concluded that the index of sex discrimination
derived from residuals of reduced form wage equations accounted for over half
the wage differentials by sex. But it is not unequal pay for equal jobs which is
Ch. 39: Urban Labour Markets 1523
the main ingredient, but rather the concentrations of women in low paid jobs
which produces the large overall wage differentials. Even with this detailed and
sophisticated formulation, the dynamic contribution of women's failure to invest
in human capital (or denial of the opportunity) is difficult to identify separately.
Another large US longitudinal microsample was used by Madden (1977b),
and the focus here was on the empirical estimation of local monopsony effects
in the urban labour market. These results comprised OLS estimations with the
worktrip time dependent, and produced suggestive results that married women
have a lower wage elasticity of trip time than men, indicating the possibility of
a more localised urban labour market for married women. However, difficulties
remained over the separate analysis of the participation decision and hours of
work in the labour supply decision.
Madden (1981) used yet another US microdata source, the Panel Survey of
Income Dynamics. Again, the empirical analysis of work trip length was the
principal econometric content, though the overall context was expressed in two
directions. Firstly, the study of the policy impact of job-creation programmes
requires care where labour markets are spatially constrained, as the author sug-
gests they are for married women. Secondly, the impact on urban structure of the
sustained rise in female LFP requires analysis.
The empirical evidence found by Madden (1981) suggested that work trip
lengths tend to vary positively with the wage and also with residential distance
from the CBD, but negatively with city size, controlling for household structure
and gender.
The study was perhaps weakened by the absence of information on job loca-
tion itself, so that no clear separation was made between the traditional monocen-
tric trade-off and wider search areas to high status jobs in the suburbs. However,
there is some good discussion of the important role of the expected turnover rate
of residence and job, in that residential location decisions are always implicitly
dynamic. But the unobserved nature of search and job or residence turnover
expectations remain as serious difficulties.
The question of the impact on urban structure of rising female activity rates
was further examined by Hekman (1980), who tried to "rescue" the Alonso model
from the criticisms of Wheaton (1977) by stressing the higher female LFP rates
found in lower income households. (Wheaton (ibid) had constructed simulations
of the trade-off model with realistic parameters indicating that there was little
difference in the slope of the bid rent curves for different groups). A more general
theory of the family labour market produces relationships between female LFP,
CBD distance and male earnings. The impact of the male wage on female LFP
1524 G. R. Crampton
They presented a three equation system, with wage, home value, and commute
time treated as endogenous, estimated by two stage least squares. The presence
of a "moved" dummy in the "commute time" equation seemed to confirm that
residential choice is based more heavily on the male's job location. Other fam-
ily structure variables, such as the presence of children, had expected results.
This was also one of the largest microsamples which made a success of esti-
mating a simultaneous system including firstly the "returns to commuting" (i.e.,
longer commutes causing a higher wage), and secondly the "cost of commuting",
(i.e., the higher wage causing higher time valuation and an incentive to reduce
commuting). The interaction of these opposite forces with the discrete "move"
decision, and family job priorities, goes to the heart of modern urban labour
market research, and deserves further work.
A different approach to the wage equation was taken by Madden and Chen
Chiu (1990), using the 5% Public Use Micro Sample (PUMS) of the 1980 US
Census. Limiting themselves to Detroit and Philadelphia, they estimated zone-
specific wage equations, separated by race and gender. The intention was to
identify and compare the intraurban wage variability of men and women. If
the wage variability for women were in fact greater, then constraints on female
commuting or residential mobility could be more damaging for them and could
contribute to male/female overall wage differentials. They found that conversely
there was evidence that for whites, the male coefficient of variation was higher,
though this was not true for blacks. There was however found to be a greater spa-
tial concentration of "top jobs" for women. The paper also included a theoretical
treatment of urban labour market behaviour for the two-earner household (with
commuting time as usual priced at full wage rate), with simulations for Detroit
and Philadelphia.
A number of papers from a more geographical point of view have adopted
a more explicitly feminist approach to urban commuting and labour market out-
comes [Hanson and Johnston (1985), Hanson and Pratt (1988), Johnston-
Anumonwo (1988), Nelson (1986)]. This research has used quite substantial
microsurvey data, but has limited the empirical analysis to analysis of variance.
Hanson and Johnston (1985) using Baltimore home interview data found that,
contrary to some preconceptions, gender commuting distance differences were
not greatly influenced by household responsibility or even part-time versus full-
time status. Shorter journey to work distances were best explained by women's
1526 G. R. Crampton
lower incomes and a greater use of the bus in commuting, which may also restrict
women to smaller job search areas. Perhaps studies such as this are most useful
in summarising evidence on "stylised facts", such as the stronger tendency for
working women to both live and work in the central city, for there to be more fe-
male headed households in the city, and for women to commute shorter distances
even within the same income class as men.
A similar approach was taken by Hanson and Pratt (1988) using special Cen-
sus runs for a specific town, Worcester Massachusetts. An attempt was made to
identify ". . . women's spatial entrapment" (p. 199), based on average journey to
work times for sub-aggregates of the labour force. In addition, full time and part
time workers were aggregated. However, there is a result of some interest. Male
average journey to work times were longer to a suburban job (by 9 minutes) and
from a suburban residence (by 6 minutes), compared to males living or working
in the central city, and compared to all female subgroups.
A closely related piece of work, using a small subsample of the same data was
done by Johnston-Anumonwo (1988). It likewise tried to identify the degree to
which
The sample was divided into "female" or "male-dominated" jobs, using na-
tional employment patterns. One interesting finding was that women in "female-
dominated" jobs had significantly shorter journey to work times, and were more
likely to work in suburban Worcester. In contrast, women from households with
no children were more likely to work in "male" or gender-integrated jobs. The
analysis ran quickly into small sample sizes when cross-tabulations were at-
tempted. It was one of very few examples of microsurvey evidence on a single
metropolitan area. A similar focus on a single metropolitan area, but the much
bigger one of San Francisco-Oakland was provided by Nelson (1986) in the
modern "political geography" style.
A more explicit discussion (from a male perspective) of the "patriarchal con-
straints hypothesis" was provided by Gordon et al. (1989a), using national US
microdata. They confirmed the finding that female worktrips were shorter than
male worktrips, even after standardising for a wide range of job and personal
characteristics. Thus, although the ranking of average journey to work distance
was very similar between men and women, women still had shorter average
Ch. 39: Urban Labour Markets 1527
journey to work within each occupational class. There was also a trend over
1977-1983 towards shorter journey to work trips, which the authors associated
with job dispersal.
This paper did not address the causes of male wage premia for a given oc-
cupation, but did suggest the likelihood that higher wage possibilities for each
occupation would induce wider search areas and consequently greater average
worktrip distance:
If the total wage and benefit package is consistently higher for males at the
occupational specific level, it may be more worthwhile for men to adopt
a spatially more extensive search area in seeking a suitable job than for
women.
(Gordon et al., 1989a: p. 508)
Almost all of the above research on gender and urban labour markets was
based on US data. Comparisons making use of the urban experience of Paris
was provided in a series of papers by Fagnani and colleagues (Fagnani, 1983,
1984, 1990; Brun and Fagnani, 1994). The 1984 descriptive survey of commuting
patterns in the Ile de France region focused on "... strategies to reconcile work
with family life". Women's jobs in the Ile de France are strongly concentrated in
the City of Paris, and the female LFP for the region is well above the national
average.
The survey showed that male/female differences in commuting distances are
only slight for the intraParis and reverse commute Paris-suburbs trips, but the
male average distance was greater for the conventional inwards radial worktrip.
In addition, the differences were even greater with respect to commuting times,
largely explained by women's greater use of public transit. The unequal access to
cars by gender was especially damaging for low income jobs in the outer suburbs.
Fagnani (1984) presented a descriptive data analysis for mature women in the Ile
de France. Again the work found that the single female heads of household were
much more likely to be in professions or top management and to reside in the
city of Paris:
... the higher a woman rises in the socio-occupational hierarchy, the more
she can offer herself the "luxury" of not abiding by social conventions.
(Fagnani, 1984: p. 145)
7.1. Originalformulation
Most of the derivative research which has since emerged developing the same
concept has either used more sophisticated data than was available to Kain, better
econometric techniques, or has tried to widen the concept of "access" beyond the
Ch. 39: Urban LabourMarkets 1529
cost of possible journeys to work. There were some fairly immediate follow-
ups to Kain's original work, which questioned the robustness of his findings
with respect to specification (Offner and Saks (1971)) or the measure of racial
segregation (Masters (1974, 1975)). The results were also dependent on the data
set used.
Kain and Meyer (1970) presented a clear policy-oriented analysis of the view
of the time, shared by the McCone Commission on the Watts riot, that
The view that poor physical access to suburban job concentrations, caused
partly by low car ownership and partly by the inevitably focused radial nature
of public transit, led Kain and Meyer to propose a range of flexible and innova-
tive service improvements to public transit, stressing small vehicle and demand-
activated transportation possibilities suitable to low density employment distrib-
utions. Inspired by this research and lobbying, the US Department of Transporta-
tion experimented with free-reverse-commuting bus trips, but the results were
poor and the experiments were ended in embarrassment.
It is relevant that older used cars are cheap in the US, and vehicle costs for
inner-city-suburban worktrips may normally be lower than bus fares if a modest
allowance is made for the value of time saved by reverse commuting (against the
main flow) by automobile. Mills and Lubuele (1997) also observed that in the US
labour market conditions of the mid-1990s (which admittedly were not typical)
there was excess demand for sales and clerical workers in both inner cities and
suburbs.
At the text book level of urban economics, incorporation of spatial mismatch
ideas was rapid. Kalachek (1973) stressed the likelihood that negative spatial
mismatch impacts were more severe amongst single persons, teenagers and mar-
ried women. Kalachek also anticipated later research by stressing the role played
by informal networks in disseminating information about potential jobs. The
policy choice at that time was seen as being between subsidising transport to the
suburban job concentrations, versus job creation programmes closer to low in-
come residential areas. He favoured the transport improvement approach, partly
because of flexibility (with bus infrastructure) and evenness of application.
1530 G. R. Crampton
Stimulated by the detailed study of "racial ghettos" which had developed as part
of both the study of urban housing markets and the spatial mismatch in urban
labour markets, a body of research developed on racial segregation. Some of
this work was largely theoretical, e.g., Yinger (1976), who took Bailey's (1959,
1966) original "border-minimising" model of the ghetto further, by showing that
residential racial instability could be a built-in and permanent feature of the ur-
ban housing market, if any subgroup of the population actually preferred racial
mixing. This meant that some form of "institutionalised discrimination" was
inevitable (e.g., through minimum lot zoning) if stability in racial composition
was to be the outcome.
Courant and Yinger (1977) reviewed theoretical models of ghettos and ex-
tended them by incorporating more realistic income distributions. They called for
better modelling of ghettos through simultaneous modelling of race and residen-
tial demand, incorporation of high search costs (both for housing and for labour),
and the possibility of collusive or institutionalised discrimination. The relatively
small numbers of highly concentrated racial ghettos, which tend to be quite stable
in location, are the stylised facts needing good theoretical explanations.
Straszheim (1980) was the first to provide an integrated model of housing
and labour market discrimination with respect to race including hiring policy
parameters such as a "black quota". The resulting nonlinear differential equa-
tion provided simulation results on how hiring practices interact with the spatial
pattern of job opportunities. The possible positive wage gradient effects was
studied empirically using data from San Francisco. Evidence was presented that
these effects are strongest for less educated blacks, who are also more segregated
residentially.
The explicit detailed study of segregation has not attracted so much atten-
tion recently. One exception came from Miller and Quigley (1990), who used
an entropy technique to study segregation by race and household type for San
Francisco, with entropy:
where i represents either household type or race subgroups and t tracts. A max-
imum of H would then be an equal distribution and H below this measures
the degree of segregation. Miller and Quigley found that segregation by race
was indeed greater than by household type, and both weakened slightly over
Ch. 39: Urban Labour Markets 1531
1970-1980. Another interesting finding was that the level of racial segregation
measured by these entropy indices in the suburbs was lower than in the central
city.
7.3. Reviews and descriptive evidence lying behind spatial mismatch studies
The research on the locational detail of the spatial mismatch process has been
formulated in the context of more general work on racial aspects of urban labour
markets. It is perhaps worth commenting on these first.
By the late 1970s in the US, the issue of black/white average wage conver-
gence was worthy of analysis. Smith and Welch (1979) used Current Population
Survey data to suggest factors which could be influencing this labour market
convergence. These included "cohort effects" as younger, better educated blacks
became more numerous, the effects of Affirmative Action Programmes, and long
term regional trends such as the urbanisation of the South and black migration
to the North. They concluded that the role of Affirmative Action was probably
small, and urbanisation within the South was a much more important factor in
furthering the job prospects of blacks.
One difficulty which emerged was that of distinguishing between "life cy-
cle" or career effects and the "cohort" effects, which operate as a labour force
"vintage" phenomenon. The life cycle effects could be properly tested only by
microlongitudinal data.
The particular life cycle stage which has received separate detailed research
attention concerns the problem of youth unemployment. It emerged in the early
1980s that a serious problem was developing with respect to youth joblessness
relative to prime age male unemployment. The paper by Wachter and Kim (1982),
and the detailed discussion of it by Gordon (1982) revolved around the effects
of "cohort overcrowding", the declining role of military service, the role of the
minimum wage or welfare payments on youth reservation wage rates. Gordon
suggested that switching occurred into full time school or the informal or illicit
sector, but it was difficult to distinguish the demand side from the pure supply
side effects of the reservation wage. Bureau of Labor Statistics data showed the
importance of youth unemployment "while in school" and the implied role of
school as a "substitute job" is somewhat deflating from a human capital point of
view.
By the mid 1980s, review material such as Kasarda (1985) on the position of
racial minorities in urban labour markets was suggesting a deterioration, that the
dynamics of urban change had:
1532 G. R. Crampton
... altered the capacity of America's older cities to offer entry-level jobs
and socially upgrade disadvantaged resident groups.
(Kasarda, 1985: p. 34)
The dynamics of urban change partly took the form of the dramatic growth
over 1965-1985 of the modem urban service sector at or near the CBD. Infor-
mational services grew faster still, especially in the largest cities, placing strong
emphasis on numeracy, literacy, and high motivation. The white population was
declining rapidly in the biggest cities. Kasarda's (1985) evidence from the C.P.S
tapes of 1969, 1977, 1982 showed that while average unemployment rates for all
educational groups had increased, the gaps between the better and worse edu-
cated had widened, and the same had happened for labour force nonparticipation
rates. The scale of economic marginalisation for minority groups, especially in
the large northern US cities, had become more severe. As suggested by Mills
and Lubuele (1997), wage inequality as such has increased by similar amounts
for blacks and whites over 1980-1997. A major reason for widening black wage
disparities between inner cities and suburbs is migration of higher wage blacks
to the suburbs.
Jencks and Mayer (1989) provided further discussion of the status of the
spatial mismatch hypothesis in the late 1980s. They acknowledged the evidence
of Kasarda on the widening of city-suburb differences in black labour market
outcomes, but they suggested that the role of selective city-suburb migration
was both important and difficult to estimate without good longitudinal data. In
addition, there were built-in simultaneities between individual job status and the
neighbourhood of residence, with each affecting the other. The earnings of blacks
living in the suburbs would be much affected by unmeasured heterogeneity so
that empirical measures of the role of commuting cost, which anyway might play
a modest role, may be statistically unreliable. Jencks and Mayer also stressed
the difficult mixture of demand and supply side factors in the spatial mismatch
process. The demand side factors involved the impact of racial discrimination
in hiring, either pro-black in the ghetto or pro-white in predominantly white
suburbs. Supply side factors included the role of search and information costs
and commuting costs from the ghetto.
A review of the status of the spatial mismatch hypothesis by Holzer (1991)
confirmed that the decentralisation of jobs and residents continued as did declines
in manufacturing employment. Although black suburbanisation was slowly ris-
ing, blacks were lagging Hispanic and Asian Americans in this respect. It was
also true that on average, employed blacks still had longer commute times than
employed whites and this might have some measurable relationship with the
Ch. 39: Urban Labour Markets 1533
poorer job access at neighbourhood level suffered by blacks. There seemed also
to be some evidence for a positive wage gradient for less-educated blacks com-
muting to the suburbs.
Holzer concluded that spatial mismatch effects could be identified for black
youth employment, but only if, firstly, microlevel Census data were used, and
secondly, separate estimates were produced for blacks and whites. If anything,
spatial mismatch was growing more relevant over time:
But uncertainty remains over the magnitude and existence of these effects.
(Holzer, 1991: p. 119)
It appeared clear that the severity of spatial mismatch phenomena varied be-
tween metropolitan areas, as one might expect discriminatory attitudes to do. The
developing empirical concentration on the youth labour market was welcome,
partly because their employment problems were more severe.
Perhaps Harrison was unduly pessimistic in feeling that policies to assist non-
whites in moving to the suburbs were ineffective; such surburbanization by better
off nonwhites was rapid during the 1980s (Mills, 1997).
Danziger and Weinstein (1976) took the estimation of spatial mismatch phe-
nomena further using a sample of 1970 Census Employment Service data from
three cities. They tried to resolve the issue of commuting cost differentials for
ghetto residents accessing suburban jobs. Using a reduced form wage equation,
the differential between the actual suburban wage and the imputed wage (with
characteristics controlled for) would be positive if commuting compensation was
paid. This was also compared with hypothetical cash and time commuting dif-
ferentials. Danziger and Weinstein concluded that the evidence was consistent
with labour market imperfections at wage rates which do not compensate for
commuting.
A further dimension was added by Myers and Phillips (1979) in stressing
the search and reservation wage mechanisms lying behind spatial mismatch. If
minority workers were searching from a suburban residence, search costs may
result in the reservation wage being higher and unemployment duration longer.
The original study estimated wage and job offer probabilities controlling for
many characteristic and labour market variables, including the duration of search.
Tentative evidence emerged of a positive suburban wage premium. But there was
evidence that black workers were more likely to be hired in suburban tracts with
large black populations. Raphael (1998) used 1980-1990 Census tract data for
the San Francisco Bay Area to explain the relationships between changing ac-
cessibility of employment and the employment rates of black and white youths.
This produced evidence that racial differences in youth employment rates were
significantly explained by changing job access.
The simultaneity problem intrinsic in modelling the relationship between per-
sonal characteristics, education, occupation and earnings was directly addressed
by Vrooman and Greenfield (1980), using a small microsample. A recursive
structural model was proposed and estimated for 6 sample subgroups. The
strength of this work was the explicit separation in an income formula of the
type:
Market Practices Survey. The system was estimated using two stage least squares,
supported by a Hausman specification test.
A microstudy of spatial mismatch using the percentage of blacks employed by
industrial firms, related to the distance from the principal black ghetto, was car-
ried out by Leonard (1987) and tested using Federal Contract Compliance data,
focusing on a comparison of Chicago and Los Angeles. A strong distance effect
was found for both Chicago and Los Angeles, but Leonard suggested that the
spatial mismatch phenomenon was more damaging in Chicago because of more
intense segregation there. Average commuting times were substantially lower in
Los Angeles. Zax (1990) provided another test of spatial mismatch phenomena
using evidence from a single large Detroit company which decentralised to Dear-
born. Clearly, after the move some black employees had moved to new suburban
mixed or ghetto neighbourhoods, added to those who commuted "down the rent
gradient". With the firm's suburbanisation, black commute times altered from
being shorter than those for whites (with given characteristics) to being longer.
The paper did not really address the simultaneous interaction with longer term
residential location decisions, indicated by the finding that black commute times
to the suburban workplace decreased with age and job tenure.
One of the most ingenious attempts to put some empirical detail on the costs
of spatial mismatch was carried out by Hughes and Madden (1991). Using micro
Census data for three cities, estimates were constructed of both rent and wage
gradients, with the intention of estimating the incidence of sub-optimality in-
duced by the residential patterns of both whites and blacks, for the fully employed
labour force. Hughes and Madden were able to offer estimates of sub-optimality
of the spatial distribution of jobs, given the residential pattern, or of residences
given the jobs, or of both unconstrained. Although significance testing was not
attempted, they found that the pattern of black residences given the jobs was
actually better than that for whites, though the pattern of black jobs, given the
residences, did impose suboptimalities. As a result they agreed with the modem
view that differences in access in terms of information may be more important
than commuting cost access.
The mid 1980s saw discussions of US spatial mismatch at a disaggregated
level being dominated by the sheer scale of black youth unemployment. An influ-
ential paper by Ellwood (1986) estimated that, if nonemployment was included
along with unemployment, about half of all black teenagers out of school were
"looking for work but can't find it". The spatial mismatch problem could be
most simply put as-would spatially rearranging jobs, with aggregate demand
fixed, have beneficial effects on black unemployment? Ellwood's study focused
on Chicago, and experimented with three alternative access measures for each
Ch. 39: Urban Labour Markets 1537
zone. A series of weighted OLS regressions found that access measures in fact
played a rather weak role in explaining the employment rate.
These results have been criticised (e.g., by Ihlanfeldt and Sjoquist, (1990))
for the absence of race subsampling, and the zone level subaggregates approach.
Ellwood felt however that he had supported the view that the impact of access on
wage rates was too small to play the dominant role claimed for it, and that
Finally, a series of papers using microsamples has attempted to estimate the role
of job access and race in the context of the much weaker US job market of the
1980s. Ihlanfeldt (1988) used 1980 Public Use Micro Sample (PUMS) data on
Atlanta ("one of the most segregated cities in the US") and found that white wage
premia were paid in some "labour shortage suburbs" in service sector jobs, but
that black wage premia were also paid in blue and white collar jobs in the outer
suburbs.
Ihlanfeldt and Sjoquist (1989) used 1978 microdata, with complementary
MSA Census data, to estimate annual earnings of central city residents net of
commuting costs. This is related to an index of decentralisation of low skill
jobs, and a number of human capital variables. They found a significant negative
effect of low skill job suburbanisation on the net wage for both black and white
central city men, but not for black women. This raised the issue of the degree
to which black women may substitute for teenagers in terms of labour demand.
They also found no support for the idea that the more segregated a city, the more
the adverse effect on the net wage. The paper also offered probit estimates of the
move probability, and found that low skill job suburbanisation was a significant
move-promoter for whites but not for blacks.
Ihlanfeldt and Sjoquist (1990) switched back to using microdata for a specific
city, Philadelphia. They tried to improve significantly on earlier studies of access
effects by using mean travel time for zones, and presenting logit estimates on
samples of youths or adults by race. The authors found the adverse impact of
mean journey to work time on employment probability, together with a number
of the personal and family characteristics variables and comparative runs for
Chicago and Los Angeles were offered. The issue remains that estimates of the
access effect are trying to separate out a labour supply mechanism, which itself
mixes commuting cost and information decay considerations, from racial dis-
crimination forces which are largely demand side. Single equation specifications
will always be open to this weakness.
The PUMS 1980 data was also used by Ihlanfeldt and Sjoquist (1991a) to
address three alternative explanations of black wage discrimination. Firstly, there
is the suggestion that white dominated areas have "better jobs", secondly the
conventional spatial mismatch mechanism of racial residential mix affecting the
hiring of blacks, and thirdly the possible tendency for black workers to be bi-
ased towards the public sector, the so-called "sheltered workplace" hypothesis.
Ihlanfeldt and Sjoquist used this data set of over 200,000 observations in 9 major
metropolitan areas. They used a multinomial logit technique for employment
Ch. 39: Urban Labour Markets 1539
by occupation category with the percentage black residents in the area as the
principal racial variable, in addition to education, experience and a range of
other dummy variables. A Wald test was used to test the significance of the
black residence variable on the occupational hiring pattern. The main finding was
that black males are more likely to get managerial or sales jobs, the higher the
percentage of black residents. The "sheltered workplace" hypothesis, however,
was not supported.
Further use of logit regression technique using the 1980 PUMS data was
provided in Ihlanfeldt and Sjoquist (1991b), who focused on teenagers living at
home in central cities. Average commuting times by private car were estimated
to low wage jobs and used as the principal index of job access. Other labour
market variables were included at an MSA level, such as indices of occupations
suitable for youths, and a ratio of numbers of low education white females to
youths to indicate competitiveness between these types of labour supply. These
additional variables gave the single equation specification a strong supply side
orientation. Ihlanfeldt and Sjoquist concluded that equalising access time to jobs
between blacks and whites would reduce the probability-of-employment gap by
about 19-23%.
Recent ideas on testing the importance of access for youth employment have
shifted towards attempts to quantify informational or networking elements of
access, rather than the spatial. O'Regan and Quigley (1991) also used the 1980
PUMS data and found that the importance of parent or sibling working could
itself represent both beneficial network and information effects, mixed with the
role of family values and culture. They also constructed isolation indices to
represent the level of social isolation within races, and these functioned well em-
pirically. They identified a large increase since 1970 in the spatial concentration
of US urban poverty, and the social and informational isolation associated with
this has identifiable labour market impacts. The programmes implied by such
findings, rather than the transportation or job creation orientation stressed in the
1970s, involve the difficult task of:
The paper then used a large sample of microdata from the 1970 PUMS to obtain
probit estimates in which the dichotomous unemployment variable was explained
by a combination of human capital, amenity, and purely economic variables.
In this equilibrium equation, the higher wage compensates for higher risk of
unemployment as do higher amenity levels. Despite the formidable size of the
data base, there is still a combination of equilibrium and dynamic mechanisms,
which are difficult to model effectively.
A study by Moffitt (1985) focused more closely on the dynamics of un-
employment using maximum likelihood methods with a sample of completed
unemployment spells over 1978-1993. The impacts of the exhaustion of unem-
ployment insurance on the "ending spell" hazard was successfully estimated, as
were racial differences. Kaplan-Meier statistics showed the "spike" in the hazard
over time especially at the unemployment insurance exhaustion point. Moffitt
also raised the issue of positive duration dependence of the job-finding hazard,
which interacts with the time pattern of unemployment insurance, also addressed
by Bailey (1994) with a Weibull hazard model, estimated using a sample of
over 12,000 aged 14-21. The concentration on youth was important because of
"job habit formation", and the adverse impacts of concentrated youth unemploy-
ment. Positive duration dependency for the "leave unemployment" probability
was found, and a range of 30 independent variables were used to try to min-
1542 G. R. Crampton
imise the bias caused by unobserved heterogeneity. These included factors such
as recent moves and mortgaged housing capital. Positive duration dependence
itself partly reflects "job shopping" by less risk averse young people, and the
exhaustion of savings. Finally, a British paper by Elliott and Theodossiou (1992)
focused on the distribution of the time structure of completed periods of unem-
ployment, for a sample of about 1000 adults in the city of Aberdeen. Lorenz
curves and Gini coefficients were calculated and revealed a clear minority who
suffered disproportionately from frequent spells.
The problem of the distributed lag structure in fluctuations in unemployment
between urban areas has yet to receive much detailed research attention. Gordon,
in two papers (1985a,b) suggested that there has in effect been a "modelling vac-
uum" in studying the complex patterns of lags arising in the spatial transmission
of employment and unemployment differentials. The recent slowdown in UK
migration as a result of higher unemployment rates has weakened interurban and
regional supply adjustments. In particular, the cyclical sensitivity of unemploy-
ment was highest in high unemployment regions. Through a very imperfectly
understood spatial adjustment, unemployment responds to adjacent interregional
change with a distributed lag, even if employment has no lag, or only a discrete
lag:
. . it is clear that rather more complex models are required, which incorpo-
rate explicitly the supply-side adjustment process.
(Gordon, 1985b: p. 855)
There have been examples in the US (e.g., the Massachusetts boom and bust)
where the adjustment process was inefficient, through excessive labour mobil-
ity and imperfect capital markets. In particular, the response of job creation to
movements in wages appears weak. In post-Single Market Europe, the efficiency
of labour mobility is much lower, so that unless there is more wage flexibility:
... shocks will have larger and longer lasting effects in relative unemploy-
ment in Europe.
(Blanchard and Katz, 1992: p. 56)
Interurban studies of wages geared towards the demand side have included the
influential work of Glaeser et al. (1992), which focused on mechanisms through
which knowledge spillovers can affect wage differentials. Three paradigms in-
cluded a Marshall-Arrow-Romer intraindustry spillover process, Porter-style
competitive innovation, and Jane Jacobs urbanisation externalities working
through diversification. A 31 year time series of employment and wage growth in
major US cities for the major industries in each MSA is used for OLS estimates of
city-industry observations. The sample is inevitably biased towards "mature" in-
dustries. Of the alternatives, they rejected the strong version of Marshall-
Arrow-Romer knowledge spillovers. If anything they lend support to a Jacobs-
type tendency for urbanisation externalities to promote industrial diversification.
The models however have a reduced form character, and:
The important development of urban quality of life measures and their detailed
empirical estimation is discussed in detail elsewhere in this volume by Gyourko,
Kahn and Tracy (Chapter 37). There are interactions between quality of life
analysis and urban labour market phenomena, both inter- and intraurban which
makes this a fertile research area receiving active interest in the 1980s and 1990s
in both Europe and the US.
Cheshire and colleagues (1986, 1990, 1995, 1996), in a series of innovative
papers extending urban quality of life analysis to the countries of the European
Union, designed a "problem index" (an inverse of quality of life) using discrim-
inant analysis based on a "training set" of European Functional Urban Regions
(FURs). This work is discussed in detail elsewhere in this volume, but it is worth
recalling that unemployment and in-migration rates are included in the urban
area attributes, a weighted average of which made up the problem index, which
as a result was geared much more towards the labour market components of
quality of life than the compensations operating through the housing market.
European-wide data on the housing market would be virtually impossible to
collect reliably, so that European replications of the US models discussed above
would be very difficult. The body of research put together by Cheshire and col-
leagues is, perhaps, most useful in identifying medium to long term patterns of
change in European urban welfare, using the problem index, and speculating on
the impact of urban policy by analysing problem score residuals, in addition to
the awareness of problems of growth as well as of decline.
Considering the established place this relationship has in urban economics, there
have been relatively few studies of the variation of wage rates by location within
urban areas. Moses' (1962) seminal theoretical paper developed the idea of the
wage rate for a standard unit of labour declining with CBD distance so as to pre-
serve equilibrium between workers commuting inwards to jobs at different radial
distances. More substantial suburban employment subcentres would modify this
(Ravallion (1979) and Madden (1985)).
The problem has been the difficulty of standardising occupations so that the
wage rate itself was well-defined. Supply side characteristic variables also play a
Ch. 39: Urban Labour Markets 1545
role, as does the difficulty over which measure of distance or accessibility to use
for an employer.
The first substantial empirical urban study of labour markets was the Rees
and Schultz (1970) monograph, based on detailed evidence for Chicago. Even
here, there was little spatial detail as such, but substantial insight on alternative
job search methods. Given its focus on the impact of gender on commuting
distance, the volume is as relevant to Sections 4 and 6 of this chapter as to the
wage gradient issue. It makes use of microdata to shed light on the commuting
distance/wage relationship, for a range of manual occupations which may be
simpler to study for wage gradients given the weaker role of salary structures.
A consistent positive effect of commuting distance on the wage rate is found,
though this is understood more as a compensation for commuters to relatively
inaccessible suburban subcentres, than a monocentric wage gradient.
In the Wachter (1972) study of the Boston, Massachusetts labour market, the
focus was more on the dynamics of change in the wage distribution over the
business cycle, for an individual urban labour market. This was most immedi-
ately in evidence in the narrowing of wage dispersion under boom conditions. A
suburban wage premium for female clerical workers was evident, though this lost
significance as the late 1960s boom developed. No evidence of wage premia was
found in other worker subgroups, though sample sizes may have been too small
for results to be robust.
Empirical work on New York City and Toronto was provided by Scott (1988).
He showed that production worker wage rates had become higher in the suburban
rings of these cities, though it is difficult to separate the role of compensation for
commuting from that of labour shortage at suburban employment subcentres.
Labour turnover rates were higher in New York City than in the suburban ring,
implying that job stability and seniority played a part in addition to spatial factors.
Eberts (1981) was one of the few who tested the wage gradient in a mono-
centric form. Eberts limited himself to five subgroups of Chicago area municipal
employees, estimating a reduced form with wage rates explained by CBD dis-
tance together with labour demand and supply characteristics interacting with
distance. The estimation technique used a Box-Cox transformation and maxi-
mum likelihood techniques. Eberts found a significant wage gradient for four of
the five labour groups, and a typical wage-distance elasticity was around 0.2-
0.3. The restriction to homogeneous municipal labour groups was useful, but
the wage gradient interacted with unionisation and the size of the municipal
government.
The theoretical treatment of the wage gradient in a model of optimal residen-
tial location choice was outlined by Madden (1985), who included hours of work,
1546 G. R. Crampton
... these results provide clear evidence of the existence of wage gradients.
(Madden, 1985: p. 300)
In effect, this work is also a contribution to the simultaneous modelling of
residential and workplace choice. One of Madden's strongest findings is that the
impact of wage change on work trip change is much greater for household heads
who job-change only, suggesting that job search and labour market flexibility
from a fixed residence may dominate dynamic wage gradient modelling.
Ihlanfeldt (1992) made use for three major urban areas (Philadelphia, Detroit,
and Boston) of the workplace and wage detail available in the 1980 PUMS, and
estimated wage gradient equations incorporating Box-Cox transformations. He
found that the predictions of the monocentric model were confirmed consistently
for white workers. However, as we have discussed elsewhere, black workers were
more likely to "reverse commute" from central city to suburb, and where job
concentrations were located in the suburbs, the monocentric model would pre-
dict wage gradients sloping the other way (positively with CBD distance). Only
patchy evidence was found for this, not surprising given the localised nature of
the job concentrations. This was the only paper which focused clearly on the
common ground between wage gradients and racial segregation in US cities.
Cheshire (1979) also provided a sharp and insightful discussion of the London
urban labour market, with special reference to the urban policy instruments under
discussion in Britain in the late 1970s, the development of which had not shown
much awareness of urban labour market processes. The role of a "trap" mecha-
nism for the poorest subgroups for example, could easily be oversimplified:
It was not the lowest socio-economic groups who might have been trapped
that suffered relatively the most unemployment in the GLC area, but the
higher groups.
(Cheshire, 1979: p. 37)
'Spatial' inequality may thus be just another and rather unimportant dimen-
sion of social inequality. It may consequently be best relieved by policies
aimed at greater social equality.
(Cheshire, 1979: p. 41)
10. Conclusions
One finishes a survey of this substantial part of the field of applied urban eco-
nomics by being impressed by some things but disappointed by others. The use of
modem computing power and econometric technique has been coupled (at least
in the US) with microdata samples of substantial size to make enormous advances
in the sophistication of empirical analysis of urban labour market phenomena. A
more limited amount of research of comparable quality has been carried out by
Dutch, British, or French researchers.
However despite these advances in analytical and computational power, the
estimated models are only as good as the data they analyse, and many labour mar-
ket phenomena of importance are simply not observable anyway. For example,
the activity of job search itself is largely nonobservable, merely its successful out-
come; likewise the internal "politics of the family" which affects the commuting,
move and quit behaviour of cohabiting partners. The vast complexity of labour
market behaviour once location is added to other relevant variables sometimes
defeats analysis. A plea for more high quality microlevel data becomes ever
more appropriate, as labour market problems are given more priority politically.
It would be comforting to think that Eurostat as well as the European national
governments might get to work on providing the kind of longitudinal microdata
which would encourage better understanding of the urban labour market in Eu-
rope, where residential move rates tend to be lower than in the US, and the role
of welfare benefit payments generally higher.
Among the trends, shared in varying measure by all countries, which emerged
over the time span covered by this survey are the tendency for greater earnings
inequality, partly as a result of the growth of a more advanced service sector. The
higher earning elite has become relatively richer, but often in parallel with shorter
job tenure and higher job aspirations by their domestic partners, both of which
then become incorporated in their labour market decisions. At least this part of
the labour market is normally well caught by data samples.
There has also been a trend towards further deterioration in the labour market
position of those youths and members of disadvantaged ethnic groups who have
not thrived in the state education system. In the US, suburbanization of blacks
has proceeded apace, but a large underclass has still been left behind in central
city areas where total formal employment has often been stagnant. The issue
of "nonemployment" as distinct from unemployment again raises the problem
for economists that important processes are taking place beyond the reach of
observed data. Cultural and social forces have always been somewhere on the
Ch. 39: Urban Labour Markets 1549
stage of the urban labour market, and may be progressively playing a greater role.
Within the reach of empirical research there are perhaps two major issues
which have continued to receive detailed attention, and will no doubt continue
to do so. The first is the question of the compensation for commuting distance,
its interaction with the extent and intensity of search for jobs, and the parallel
interaction with spatial variation of house prices. In practice, analysis of this has
been made more difficult by the additional roles of family structure and the career
aspirations of men and women, and the continuing question of spatial mismatch
and residential segregation of ethnic subgroups.
Even more difficult is the study of long term individual labour market be-
haviour, in which personal characteristics, earnings, education, and occupation
all affect and are affected by location in a metropolitan area. Good research of
these dynamic phenomena depend largely on the size and quality of the data
bases available. Some high quality research has emerged in the last 20 years,
but a great deal more in the way of empirical foundations for labour market and
welfare policies can be achieved especially in Europe.
References
Alperovich, G. (1993), "City size and the rate and duration of unemployment: Evidence from Israeli data",
Journal of Urban Economics 34:347-357.
Andrulis, J. (1982), "Intra-urban workplace and residential mobility under uncertainty", Journal of Urban
Economics 11:85-97.
Bailey, A.J. (1994), "Migration and unemployment duration among young adults", Papers in Regional Science
73:289-307.
Bailey, M.J. (1959), "Note on the economics of residential zoning and urban renewal", Land Economics
35:288-292.
Bailey, M.J. (1966), "Effects of race and other demographic factors on the values of single family homes",
Land Economics 42:215-220.
Ball, R.M. (1980), "The use and definition of travel-to-work areas in Great Britain: some problems", Regional
Studies 14:125-139.
Bane, M.J. and D.T. Ellwood (1994), Welfare Realities: From Rhetoric to Reform (Harvard University Press,
Cambridge, Massachusetts).
Barron. J.M. and 0. Gilley (1981), "Job search and vacancy contacts: Note", American Economic Review
71:747-752.
Bartel, A.P. (1979), "The migration decision: what role does job mobility play?", American Economic Review
69:775-786.
Bartik, T.J. (1994), "The effects of metropolitan job growth on the size distribution of family income", Journal
of Regional Science 34:483-502.
Baumann, J., M.M. Fischer and U. Schubert (1988), "A choice-theoretical labour market model: empirical tests
at the mesolevel", Environment and Planning A 20:1085-1102.
Beesley, M.E. and M.O. Dalvi (1974), "Spatial equilibrium and the journey to work", Journal of Transport
Economics and Policy 8:197-222.
Blanchard, 0. and L. Katz (1992), "Regional evolutions", Brookings Papers on Economic Activity 1:1-61.
1550 G. R. Crampton
Blomquist, G.C., M.C. Berger and J.P. Hoehn (1988), "New estimates of quality of life in urban areas",
American Economic Review 78:89-107.
Boarnet, M.G. (1994), "The monocentric model and employment location", Journal of Urban Economics
36:79-97.
Bound, J. and H.J. Holzer (1993), "Industrial shifts, skill level and the labor market for white and black males",
Review of Economics and Statistics 75:387-396.
Boyle, P. (1994), "Metropolitan out-migration in England and Wales, 1980-81", Urban Studies 31:1707-1722.
Brun, J. and J. Fagnani (1994), "Lifestyles and locational choices: a case study of middle-class couples living
in the Ile-de-France region", Urban Studies 31:921-934.
Burridge, P. and I.R. Gordon (1981), "Unemployment in the British metropolitan areas", Oxford Economic
Papers 33:274-297.
Cheshire, P.C. (1979), "Spatial labour markets; a critique of the inner area studies", Urban Studies 16:29-43.
Cheshire, P.C. (1990), "Explaining the recent performance of the European Community's major urban regions",
Urban Studies 27:311-333.
Cheshire, P.C. (1995), "A new phase of urban development in western Europe? The evidence for the 1980s",
Urban Studies 32:1045-1063.
Cheshire, P.C. and G. Carbonaro (1996), "Urban economic growth in Europe", Urban Studies 33:1111-1128.
Cheshire, P.C., G. Carbonaro and D.G. Hay (1986), "Problems of urban decline and growth in EEC countries:
or measuring degrees of elephantess", Urban Studies 23:131-149.
Chirinko, R. (1982), "An empirical investigation of the returns to job search", American Economic Review
72:498-501.
Cloutier, N.R. (1997), "Metropolitan income inequality during the 1980s: the impact of urban development,
industrial mix, and family structure", Journal of Regional Science 37:459-478.
Congdon, P. (1987), "The interdependence of geographical migration with job and housing mobility in
London", Regional Studies 22:81-93.
Coombes, M.G. and S. Openshaw (1982), "The use and definition of travel-to-work areas in Great Britain:
some comments", Regional Studies 16:141-149.
Courant, P. and J. Yinger (1977), "On models of racial prejudice and urban residential structure", Journal of
Urban Economics 4:272-291.
Crampton, G.R. (1990), "Commuting between local authorities in England and Wales: Econometric evidence
from the 1981 Census", Journal of Urban Economics 28:204-222.
Cropper, M.L. and A.S. Arriaga-Salinas (1980), "Inter-city wage differentials and the value of air quality",
Journal of Urban Economics 8:236-254.
Cropper, M.L. and P.L. Gordon (1991), "Wasteful commuting: a reexamination", Journal of Urban Economics
29:2-13.
Danziger, S. and M. Weinstein (1976), "Employment location and wage rates of poverty area residents",
Journal of Urban Economics 3:127-145.
Eberts, R.W. (1981). "An empirical investigation of intra-urban wage gradients", Journal of Urban Economics
10:50-60.
Ehrenberg, R.G. and G.S. Goldstein (1975), "A model of public sector wage determination", Journal of Urban
Economics 2:223-245.
Elliott, R.E. and I. Theodossiou (1992), "The inequality of unemployment experience in a local labour market",
Urban Studies 29:783-797.
Ellwood, D.T. (1986), "The spatial mismatch hypothesis: are there teenage jobs missing in the ghetto?", in:
R.B. Freeman and H.J. Holzer, eds., The Black Youth Employment Crisis (University of Chicago Press,
Chicago).
Evans, A.W. and R. Richardson (1981), "Urban unemployment: interpretation and additional evidence",
Scottish Journal of Political Economy 28:107-124.
Evers, G.H.M. and A. van der Veen (1985), "A simultaneous nonlinear model for labour migration and
commuting", Regional Studies 19:217-229.
Fagnani, J. (1983), "Women's commuting patterns in the Paris region", Tijdschrift voor economische en sociale
geografie 74:12-24.
Ch. 39: Urban LabourMarkets 1551
Fagnani, J. (1984), "Marital status and occupational structures in the Ile-de-France region", Urban Studies
21:139-148.
Fagnani, J. (1990), "City size and mothers' labour force participation", Tijdschrift voor economische en sociale
geografie 81:182-188.
Felsenstein, D. (1994), "Large high-technology firms and the spatial extension of metropolitan labour markets:
some evidence from Israel", Urban Studies 31:867-893.
Frank, R.H. (1978), "Family location constraints and the geographic distribution of female professionals",
Journal of Political Economy 86:117-130.
Freeman R.B. and H. Holzer (1986), The Black Youth Employment Crisis (University of Chicago Press,
Chicago).
Galster, G. (1987), "Residential segregation and interracial economic disparities: a simultaneous equations
approach", Journal of Urban Economics 21:21-44.
Gera, S. and P. Kuhn (1980a), "Job location and the journey to work: an empirical analysis", Journal of Socio-
Economic Planning Science 14:57-65.
Gera, S. and P. Kuhn (1980b), "An empirical model of residential location and the journey to work in a
metropolitan area", Journal of Socio-Economic Planning Science 14:67-77.
Glaeser, E.L., H.D. Kallal, J.A. Scheinkman and A. Schleifer (1992), "Growth in cities", Journal of Political
Economy 100:1126-1152.
Glaeser, E.L., J.A. Scheinkman and A. Schleifer (1995), "Economic growth in a cross-section of cities", Journal
of Monetary Economics 36:117-144.
Goldner, W. (1955), "Spatial and locational aspects of metropolitan labour markets", American Economic
Review 45:111-128.
Gordon, I.R. (1985a), "The cyclical sensitivity of regional employment and unemployment differentials",
Regional Studies 19:95-110.
Gordon, I.R. (1985b), "Distributed lags in local responses to fluctuations in unemployment", Environment and
Planning A 17:845-856.
Gordon, I.R. (1996), "Family structure, educational achievement and the inner city", Urban Studies 33:407-
423.
Gordon, I.R. and D. Lamont (1982), "A model of labour market interdependencies in the London region",
Environment and Planning A 14:237-264.
Gordon, I.R. and I. Molho (1995), "Duration dependence in migration behaviour: cumulative inertia versus
stochastic change", Environment and Planning A 27:1961-1975.
Gordon, I.R. and R.W. Vickerman (1982), "Opportunity, preference and constraint: an approach to the analysis
of metropolitan migration", Urban Studies 19:247-261.
Gordon, P., and H.W. Richardson (1996a), "Employment decentralization in US metropolitan areas: is Los
Angeles an outlier or the norm?", Environment and Planning A 28:1727-1744.
Gordon, P., and H.W. Richardson (1996b), "Beyond polycentricity: the dispersed metropolis, Los Angeles
1970-90", Journal of the American Planning Association 62:289-295.
Gordon, P., A. Kumar and H.W. Richardson (1989a), "Gender differences in metropolitan travel behaviour",
Regional Studies 23:499-510.
Gordon, P., A. Kumar and H.W. Richardson (1989b), "The spatial mismatch hypothesis: some new evidence",
Urban Studies 26:315-326.
Gordon, R.J. (1982), "Comment on: Time series changes in youth joblessness", in: R.B. Freeman and D.A.
Wise, eds., The Youth Labor Market Problem: Its Nature, Causes and Consequences (University of Chicago
Press, Chicago) pp. 189-198.
Gottfries, N. and B. McCormick (1995), "Discrimination and open unemployment in a segmented labour
market", European Economic Review 39:1-16.
Gottschalk, P. and R. Moffitt (1994), "The growth of earnings instability in the US labor market", Brookings
Papers on Economic Activity 2:217-272.
Green, A.E. and D.W. Owen (1990), "The development of a classification of travel-to-work areas", Progress in
Planning 34:1-92.
Gruver, G.W. and L.A. Zeager (1994), "Steady-state labor turnover as optimal household behaviour", Journal
of Regional Science 34:75-90.
t552 G. R. Crampton
Gyourko, J. and J. Tracy (1989), The importance of local fiscal conditions in analyzing local labor markets",
Journal of Political Economy 97:1208-1231.
Gyourko, J. and J. Tracy (1991), "The structure of local public finance and the quality of life", Journal of
Political Economy 99:774-806.
Hamnett, C. (1994), "Social polarisation in global cities-theory and evidence", Urban Studies 31:401-424.
Hanson, S. and I. Johnston (1985), "Gender differences in work-trip length: explanations and implications",
Urban Geography 3:193-219.
Hanson, S. and G. Pratt (1988), "Spatial dimensions of the gender division of labor in a local labor market",
Urban Geography 9:180-202.
Hanson, S. and G. Pratt (1992), "Dynamic dependencies-a geographic investigation of local labor markets",
Economic Geography 68:373-405.
Hanson, S. and G. Pratt (1995), Gender, Work and Place (London, New York: Routledge).
Harrison, B. (1972a), "Education and underemployment in the urban ghetto", American Economic Review
62:796-812.
Hanrrison, B. (1972b), "The intrametropolitan distribution of minority economic welfare", Journal of Regional
Science 12:23-43.
Harrison, B. (1974), Urban Economic Development: Suburbanization, Minority Opportunity, and the Condition
of the Central City (Urban Institute, Washington, DC).
Hanison, B. (1994), Lean and Mean: The Changing Landscape of Corporate Power in the Age of Flexibility
(Basic Books, New York).
Hekman, J. (1980), "Income, labor supply and urban residence", American Economic Review 70:805-811.
Herzog, H.W., A.M. Schlottman and D.L. Johnson (1986), "High technology jobs and worker mobility",
Journal of Regional Science 26:445-459.
Hochman, 0. and H. Ofek (1977), "The value of time in consumption and residential location in an urban
setting", American Economic Review 67:996-1003.
Holzer, H.J. (1986), "Reservation wages and their labor market effects for white and black male youth", Journal
of Human Resources 21:157-177.
Holzer, H.J. (1987). "Informal job search and black youth unemployment", American Economic Review
77:446-452.
Holzer, H.J. (1988), "Search method use by unemployed youth", Journal of Labor Economics 5:1-20.
Holzer, H.J. (1991), "The spatial mismatch hypothesis: what has the evidence shown?", Urban Studies 28:105-
122.
Holzer, H.J. (1994), "Black employment problems: new evidence, old questions", Journal of Policy Analysis
and Management 13:699-722.
Holzer, H.J. (1996), What Employers Want: Job Prospects for Less-Educated Workers (Russell Sage
Foundation, New York).
Holzer. H.J., K.R. Ihlanfeldt and D.L. Sjoquist (1994), "Work, Search and Travel Among White and Black
Youth", Journal of Urban Economics 35:320-345.
Hotchki:ss, D. and M.J. White (1993), "A simulation model of a decentralized metropolitan area with two-
worker, 'traditional', and female-headed households", Journal of Urban Economics 34:159-185.
Hughes, M.A. and J.F. Madden (1991), "Residential segregation and the economic status of black workers:
new evidence for an old debate", Journal of Urban Economics 29:28-49.
Hyclak, T. (1996), "Structural changes in labor demand and unemployment in local labor markets", Journal of
Regional Science 36:653-663.
lhlanfeldt, K.R. (1988), "Intrametropolitan variation in earnings and labor market discrimination: an economic
analysis of the Atlanta labor market", Southern Economic Journal 55:123-140.
Ihlanfeldt, K.R. (1992), "Intraurban wage gradients--evidence by race, gender, occupational class and sector",
Journal of Urban Economics 32:70-91.
Ihlanfeldt, K.R. (1997), "Information on the spatial distribution of job opportunities within metropolitan areas",
Journal of Urban Economics 41:218-242.
Ihlanfeldt, K.R. and D.L. Sjoquist (1989), "The impact of job decentralisation on the economic welfare of
central city blacks", Journal of Urban Economics 26:110-130.
Ch. 39: Urban Labour Markets 1553
Ihlanfeldt, K.R. and D.L. Sjoquist (1990), "Job accessibility and racial differences in youth employment rates",
American Economic Review 80:267-276.
Jhlanfeldt, K.R. and D.L. Sjoquist (1991a), "The role of space in determining the occupations of black and
white workers", Regional Science and Urban Economics 21:295-315.
IlHanfeldt, K.R. and D.L. Sjoquist (1991b), "The effect of job access on black youth unemployment: a cross
sectional analysis", Urban Studies 28:255-265.
Jencks, C. (1992), Rethinking Social Policy: Race, Poverty and the Underclass (Harvard University Press,
Cambridge, Massachusetts).
Jencks, C. (1994), The Homeless (Harvard University Press, Cambridge, Massachusetts).
Jencks, C. and S. Mayer (1989), "Residential segregation, job proximity and black job opportunities: the
empirical status of the spatial mismatch hypothesis", in: M. McGeary and L. Lynn, eds., Concentrated Urban
Poverty in America (National Academy Press, Washington, D.C.).
Jencks, C. and P.RE. Peterson (1991), The Urban Underclass (Brookings Institution, Washington, D.C.).
Johnston-Anumonwo, I. (1988), "The journey to work and occupational segregation", Urban Geography
9:138-154.
Kain, J.E (1962), "The journey to work as a determinant of residential location", Papers and Proceedings of
the Regional Science Association 9:137-160.
Kain, J.E (1968), "Housing segregation, negro employment, and metropolitan decentralization", Quarterly
Journal of Economics 82:175-197.
Kain, J.F. and J. Meyer (1970), "Transportation and poverty", The Public Interest, Winter.
Kalachek, E.D. (1973), "Ghetto dwellers, transportation and employment", in: D.W. Rasmussen and C.T.
Haworth, eds., The Modem City (Harper and Row, New York) pp. 72-80.
Kalachek, E.D. and F. Raines (1976), "The structure of wage differences among mature male workers", Journal
of Human Resources 11:484-506.
Kasarda, J.D. (1985), "Urban change and minority opportunities", in: PE. Peterson, ed., The New Urban
Reality (Brookings Institution, Washington, D.C.).
Kim, S. (1995), "Excess commuting for two worker households in the Los Angeles Metropolitan Area", Journal
of Urban Economics 38:166-182.
Layard, R. (1982), "Youth unemployment in Britain and the United States", in: R.B. Freeman and H.J. Holzer,
eds., The Black Youth Employment Crisis (University of Chicago Press, Chicago) pp. 499-531.
Leonard, J.S. (1986), "Comment on D.T. Ellwood's paper", in: R.B. Freeman and H.J. Holzer, eds., The Black
Youth Employment Crisis (University of Chicago Press, Chicago and London) pp. 185-190.
Leonard, J. (1987), "The interaction of residential segregation and employment discrimination", Journal of
Urban Economics 21:323-346.
Levy, E and R.J. Mumane (1992), "U.S. Earnings levels and earnings inequality: a review of recent trends and
proposed explanations", Journal of Economic Literature 30:1333-1381.
Lillydahl, J.H. and L.D. Singell (1985), "The spatial variation in unemployment and labour force participation
rates of male and female workers", Regional Studies 19:459-469.
Linneman, P. and P.E. Graves (1983), "Migration and job change: a multinomial logit approach", Journal of
Urban Economics 14:263-279.
MacKinnon, R.D. and P. Rogerson (1980), "Vacancy chains, information filters, and interregional migration",
Environment and Planning A 12:649-658.
Madden, J.F. (1977a), "A spatial theory of sex discrimination", Journal of Regional Science 17:369-380.
Madden, J.E (1977b), "An empirical analysis of the spatial elasticity of labor supply", Papers and Proceedings
of the Regional Science Association 39:157-171.
Madden, J.F. (1981), "Why women work closer to home", Urban Studies 18:181-194.
Madden, J.F. (1985), "Urban wage gradients: empirical evidence", Journal of Urban Economics 18:291-301.
Madden, J.E (1996), "Changes in the distribution of poverty across and within the US metropolitan areas
1979-1989", Urban Studies 33:1581-1600.
Madden, J.E and L. Chen Chiu (1990), "The wage effects of residential location and commuting constraints
on employed married women", Urban Studies 27:353-369.
Madden, J.E and W.J. Stull (1991), Work, Wages and Poverty-Income Distribution in Post-Industrial
Philadelphia (University of Pennsylvania Press, Philadelphia).
1554 G. R. Crampton
Madden, J.F. and M.J. White (1980), "Spatial implications of increases in the female labor force: a theoretical
and empirical synthesis", Land Economics 56:432-446.
Maier, G. (1985), "Cumulative causation and selectivity in labour market oriented migration caused by
imperfect information", Regional Studies 19:231-241.
Maier, G. (1987), "Job search and migration", in: M.M. Fischer and P. Nijkamp, eds., Regional Labour Markets
(North-Holland, Amsterdam) pp. 189-204.
Marston, S.T. (1985), "Two views of the geographic distribution of unemployment", Quarterly Journal of
Economics 100:57-79.
Massey, D. (1994), Spatial Divisions of Labour, 2nd edn. (Macmillan, London).
Masters, S.H. (1974), "A note on John Kain's 'Housing segregation, negro employment, and metropolitan
decentralization' ", Quarterly Journal of Economics 88:505-519.
Masters, S.H. (1975), Black-White Income Differentials: Empirical Studies and Policy Implications (Academic
Press, New York).
McCormick, B. (1986), "Employment opportunity, earnings, and the journey to work of minority workers in
Great Britain", Economic Journal 96:375-397.
Merriman, D. and D. Hellerstein (1994), "Compensation for commutes in the land and labor markets: some
evidence from the Tokyo metropolitan area", Journal of Regional Science 34:297-324.
Merriman, D., T. Ohkawara and T. Suzuki (1995), "Excess commuting in the Tokyo Metropolitan Area:
measurement and policy simulations", Urban Studies 32:69-85.
Metcalf, D. (1975), "Urban unemployment in England", Economic Journal 85:578-589.
Metcalf, D. and R. Richardson (1980), "Unemployment in London", in: A.W. Evans and D. Eversley, eds., The
Inner City: Employment and Industry (Heinemann, London) pp. 193-203.
Miller, V.P. and J.M. Quigley (1990), "Segregation by racial and demographic group: evidence from the San
Francisco Bay Area", Urban Studies 27:3-21.
Mills, E.S. (1986), "Metropolitan central city population and employment growth during the 1970s", in: M.H.
Peston and R.E. Quandt, eds., Prices, Competition and Equilibrium (Barnes and Noble, Lanham, Maryland).
Mills, E.S. (1992), "The measurement and determinants of suburbanization", Journal of Urban Economics
32:377-387.
Mills, E.S. and L.S. Lubuele (1997), "Inner cities", Journal of Economic Literature 35:727-756.
Moffitt, R. (1985), "Unemployment insurance and the distribution of unemployment spells", Journal of
Econometrics 28:85-101.
Molho, 1. (1995), "Spatial autocorrelation in British unemployment", Journal of Regional Science 35:641-658.
Moore, B. and P. Townroe (1990), Urban Labour Markets: Reviews of Research, Department of the
Environment (HMSO, London).
Mortensen, D. (1970), "Job search, the duration of unemployment and the Phillips curve", American Economic
Review 60:847-862.
Mortensen, D. (1986), "Job search and labor market analysis", in: O.C. Ashenfelter and R. Layard, eds.,
Handbook of Labor Economics (Elsevier, Amsterdam) pp. 849-919.
Mortensen, D. and C.A. Pissarides (1994), "Job creation and job destruction in the theory of unemployment",
Review of Economic Studies 61:397-416.
Moses, L. (1962), "Towards a theory of intra-urban wage differentials and their influence on travel patterns",
Papers and Proceedings of the Regional Science Association 9:53-63.
Myers, S.L. and K.E. Phillips (1979), "Housing segregation and black employment: another look at the ghetto
dispersal strategy", American Economic Review, Papers and Proceedings 69:298-302.
Nelson, K. (1986), "Female labor supply characteristics and the suburbanization of low-wage office work", in:
A. Scott and M. Storper, eds., Production, Work, Territory (Allen and Unwin, London) pp. 149-171.
Nickell, S.J. (1979), "Estimating the probability of leaving unemployment", Econometrica 47:249-266.
Oaxaca, R. (1973), "Male-female wage differentials in urban labor markets", International Economic Review
14:693-709.
Offer, P. and D.H. Saks (1971), "A note on John Kain's 'Housing segregation, negro employment, and
metropolitan decentralization", Quarterly Journal of Economics 85:147-160.
Oi, W. (1976), "Residential location and labor supply", Journal of Political Economy 84:S221-247.
Ch. 39: Urban Labour Markets 1555
O'Regan, K.M. (1993), "The effect of social networks and concentrated poverty on black and hispanic youth
unemployment", Annals of Regional Science 27:327-342.
O'Regan, K.M. and J.M. Quigley (1991), "Labor market access and labor market outcomes for urban youth",
Regional Science and Urban Economics 21:277-293.
O'Regan, K.M. and J.M. Quigley (1993), "Family networks and youth access to jobs", Journal of Urban
Economics 34:230-248.
O'Regan, K.M. and J.M. Quigley (1996), "Teenage employment and the spatial isolation of minority and
poverty households", Journal of Human Resources 31:692-702.
Peterson, P.E. (1992), "The urban underclass and the poverty paradox", Political Science Quarterly 106:617-
637.
Pickles, A. and P. Rogerson (1984), "Wage distribution and spatial references in competitive job search and
migration", Regional Studies 18:131-142.
Piore, M. and C.E Sabel (1984), The Second Industrial Divide (Basic Books, New York).
Price, R. and E.S. Mills (1985), "Race and residence in earnings determination", Journal of Urban Economics
17:1-18.
Quigley, J.M. and D.H. Weinberg (1977), "Intra-urban residential mobility: a review and synthesis", Interna-
tional Regional Science Review 2:41-66.
Raphael, S. (1998), "The spatial mismatch hypothesis and black youth joblessness: evidence from the San
Francisco Bay Area", Journal of Urban Economics 43:79-111.
Ravallion, M. (1979), "A note on intra-urban wage differentials", Urban Studies 16:213-215.
Rees, A.E. and G.P. Schultz (1970), Workers and wages in an urban labor market (University of Chicago Press,
Chicago).
Reid, C.E. (1985), "The effect of residential location on the wages of black women and white women", Journal
of Urban Economics 18:350-363.
Reza, A.M. (1978), "Geographical differences in earnings and unemployment rates", Review of Economics
and Statistics 60:201-208.
Rogers, C.L. (1994), "Job search and unemployment duration: implications for the spatial mismatch hypothe-
sis", Journal of Urban Economics 42:109-132.
Rogerson, P. and R.D. MacKinnon (1981), "A geographical model of job search, migration, and unemploy-
ment", Papers of the Regional Science Association 48:89-102.
Rogerson, P. and R.D. MacKinnon (1982), "Interregional migration models with source and interaction
information", Environment and Planning A 14:445-454.
Rosen, S. (1979), "Wage-based indexes of urban quality of life", in: P. Mieszkowski and M. Straszheim, eds.,
Current Issues in Urban Economics (Johns Hopkins Press, Baltimore) pp. 74-104.
Ross, S.L. (1998), "Racial differences in residential and job mobility: evidence concerning the spatial mismatch
hypothesis", Journal of Urban Economics 43:112-135.
Ross, S.L. and J. Yinger (1995), "Comparative static analysis of open urban models with a full labor market
and suburban employment", Regional Science and Urban Economics 25:575-605.
Rouwendal, J. (1994), "Search theory, spatial labour markets and commuting", Journal of Urban Economics
43:1-22.
Rouwendal, J. and P. Rietveld (1994), "Changes in commuting distances of Dutch households", Urban Studies
31:1545-1557.
Russo, G., P. Rietveld, P. Nijkamp and C. Gorter (1996), "Spatial aspects of recruitment behaviour of firms: an
empirical investigation", Environment and Planning A 28:1077-1094.
Sassen, S. (1991), The Global City-New York, London, Tokyo (Princeton University Press, Princeton, N.J.).
Schafer, R. (1978), "Metropolitan form and demographic change", Urban Studies 15:23-33.
Scott, A.J. (1988), Metropolis: From the Division of Labor to Urban Form (University of California Press,
Berkeley).
Scott, A.J. (1992a), "The spatial organisation of a local labor market: employment and residential patterns in a
cohort of engineering and scientific workers", Growth and Change 23:95-114.
Scott, A.J. (1992b), "Low wage workers in a high technology manufacturing complex", Urban Studies
29:1231-1246.
1556 G. R. Crampton
Scott, A.J. (1993), Technopolis: High Technology Industry and Regional Development in Southern California
(University of California Press, Los Angeles).
Scott, A.J. and M. Storper (1986), eds., Production, Work, Territory: The Geographical Anatomy of Industrial
Capitalism (Allen and Unwin, London and Winchester, Massachusetts).
Seater, J. (1979), "Job search and vacancy contacts", American Economic Review 69:411-419.
Siegel, J. (1975), "Intrametropolitan migration: a simultaneous model of employment and residential location
of white and black households", Journal of Urban Economics 2:29-47.
Simpson, W. (1980), "A simultaneous model of workplace and residential location incorporating job search",
Journal of Urban Economics 8:330-349.
Simpson, W. (1982), "Job search and the effect of urban structure on unemployment and married female
participation rates", Applied Economics 14:153-165.
Simpson, W. (1987), "Workplace location, residential location and urban commuting", Urban Studies 24:119-
128.
Simpson, W. (1990), "Starting even? Job mobility and the wage gap between young single males and females",
Applied Economics 22:723-737.
Simpson, W. (1992), Urban Structure and the Labour Market (Clarendon Press, Oxford).
Simpson, W. and A. van der Veen (1992), "The economics of commuting and the urban labour market", Journal
of Economic Surveys 6:45-62.
Singell, L.D. and J.H. Lillydahl (1986), "An empirical analysis of the commute to work patterns of males and
females in two-earner households", Urban Studies 23:119-129.
Sirmans, CF. (1977), "City size and unemployment: some new elements", Urban Studies 14:99-101.
Smart, M.W. (1974), "Labour market areas: uses and definition", Progress in Planning 2:239-383.
Smith, J. and E. Welch (1979), "Race differences in earnings: a survey and new evidence", in: P. Mieszkowski
and M. Straszheim, eds., Current Issues in Urban Economics (Johns Hopkins Press, Baltimore) pp. 40-73.
Steinnes, D.N. (1977), "Causality and intra-urban location", Journal of Urban Economics 4:67-79.
Steinnes, D.N. and W.D. Fisher (1974), "An econometric model of intraurban location", Journal of Regional
Science 14:65-80.
Storper, M. and A.J. Scott (1992), Pathways to Industrialization and Regional Development (Routledge,
London).
Stover, M.E. and C.L. Leven (1992), "Methodological issues in the determination of quality of life in urban
areas", Urban Studies 29:737-754.
Straszheim, M. (1980), "Discrimination and the spatial characteristics of the urban labor market for black
workers", Journal of Urban Economics 7:119-140.
Termote, M. (1978), "Migration and commuting in Lsch central place system", in: R. Funck and J.B. Parr,
eds., The Analysis of Regional Structure: Essays in Honour of August Lsch (Pion, London).
Van Ommeren, J., P. Rietveld and P. Nijkamp (1996), "Residence and workplace location: a bivariate duration
model approach", Geographical Analysis 28:315-329.
Van Ommeren, J., P. Rietveld and P. Nijkamp (1997), "Commuting: in search of jobs and residences", Journal
of Urban Economics 42:402-421.
Veen, A. van der, and W. Simpson (1991), "Compensation for commuting in the Netherlands: Do skilled
workers receive more?" Paper presented at the European Congress of the Regional Science Association,
Lisbon.
Vickerman, R.W. (1984), "Urban and regional change, migration and commuting-the dynamics of workplace,
residence and transport choice", Urban Studies 21:15-29.
Vipond, J. (1974), "City size and unemployment", Urban Studies 11:39-46.
Vipond, J. (1980), "Intra-urban unemployment differentials in Sydney, 1971", Urban Studies 17:131-138.
Vipond, J. (1984), "The intra-urban unemployment gradient: the influence of location on unemployment",
Urban Studies 21:377-388.
Vrooman, J. and S. Greenfield (1980), "Are blacks making it in the suburbs? Some new evidence on intra-
metropolitan spatial segmentation", Journal of Urban Economics 7:155-167.
Wachter, M.L. (1972), "Wage determination in a local labor market: a case study of the Boston labor market",
Journal of Human Resources 7:87-103
Ch. 39: Urban Labour Markets 1557
Wachter, M. and C. Kim (1982), "The time series changes in youth joblessness", in: R.B. Freeman and D.A.
Wise, eds., The Youth Labor Market Problem: Its Nature, Causes and Consequences (University of Chicago
Press, Chicago and London), pp. 155-185.
Wales, T. (1978), "Labour supply and commuting time", Journal of Econometrics 8:215-226.
Weinberg, D. (1979), "The determinants of intra-urban household mobility", Regional Science and Urban
Economics 9:219-246.
Wheaton, W.C. (1977), "Income and urban residence: an analysis of consumer demand for location", American
Economic Review 67:620-631.
Wheeler, J.O. (1969), "Some effects of occupational status on work trips", Journal of Regional Science 9:69-
78.
White, M.J. (1976), "Firm suburbanization and urban subcenters", Journal of Urban Economics 3:323-343.
White, M.J. (1977), "A model of residential location choice and commuting by men and women workers",
Journal of Regional Science 17:41-52.
White, M.J. (1988), "Location choice and commuting behaviour in cities with decentralized employment",
Journal of Urban Economics 24:129-152.
White, M.J. (1990), "Commuting and congestion: a simulation model of a decentralized metropolitan area",
Journal of the American Real Estate and Urban Economics Association 18:335-368.
Wilson, W.J. (1987), The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy (University
of Chicago Press, Chicago).
Wilson, W.J. (1992), "Another look at 'The Truly Disadvantaged", Political Science Quarterly 106:639-656.
Wilson, W.J. (1997), "When work disappears", Political Science Quarterly 111:567-595.
Yinger, J. (1976), "Racial prejudice and racial residential segregation in an urban model", Journal of Urban
Economics 3:383-406.
Zax, J.S. (1990), "Race and commutes", Journal of Urban Economics 28:336-348.
Zax, J.S. (1991a), "Compensation for commutes in labor and housing markets", Journal of Urban Economics
30:192-207.
Zax, J.S. (1991b), "The substitution between moves and quits", Economic Journal 101:1510-1521.
Zax, J.S. (1994), "When is a move a migration?", Regional Science and Urban Economics 24:341-360.
Zax, J.S. and J.E Kain (1991), "Commutes, quits and moves", Journal of Urban Economics 29:153-165.
Zax, J.S. and J.E Kain (1996), "Moving to the suburbs: do relocating companies leave their black employees
behind?", Journal of Labor Economics 14:472-504.
Chapter40
Contents
1. Introduction 1560
2. The basic market relationships 1564
2.1. What is housing? 1564
2.2. The demand for housing 1566
2.3. Housing supply 1569
2.3.1. The supply of new housing 1569
2.3.2. Investment in the existing stock 1572
2.4. Housing markets 1573
2.4.1. Aggregate models of the housing market 1573
2.4.2. Urban housing models 1575
3. Urban housing policy 1577
3.1. Market versus administrative provision 1579
3.2. Housing and labour markets 1581
3.3. The impact of local taxation 1581
3.4. Rent control and its impact on the provision of rented housing 1582
4. Conclusions 1583
References 1585
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P. Cheshire
© 1999 Elsevier Science B. V All rights reserved.
1559
1560 C.M.E. Whitehead
Abstract
1. Introduction
housing markets, particularly with respect to durability and filtering. The conclu-
sion was however quite negative arguing that, because there was no well defined
alternative approach able to include imperfect competition thus sophistication
had little to offer policy makers. The survey of the empirical literature by Olsen
(1988) again stressed the relative extent of work on demand rather than supply,
as well as the potential for empirical work made possible by the large scale data
sets generated for the Housing Allowance Experiment and the Annual Housing
Survey. Other chapters, notably that by Kain, examined the value of the computer
simulation urban housing market models developed by the Urban Institute and
the NBER (De Leeuw and Struyk, 1975; Struyk, 1976; Ingram, 1977; Bradbury
et al., 1977). The availability of these large data sets has been an important driver
in research significantly affecting the ways in which understanding of American
housing markets have developed.
A major review of the housing economic literature, concentrating more on
housing market models and less on both urban structures and large scale empir-
ical analysis, was published at much the same time (Smith et al., 1988). This
still found it necessary to set out general empirical information about the Amer-
ican market (without reference to any other country) before detailing the state
of existing knowledge with respect to the treatment of particular characteristics,
notably durability, heterogeneity, spatial fixity and slow and costly adjustment.
The growing body of material on the impact of inflation and the economic cy-
cle on housing markets was also discussed. Finally, the role of government was
acknowledged and a wide range of positive economic analyses of particular pro-
grammes, notably those relating to zoning, rent regulations and the impact of
taxation were included.
The conclusions of this review were remarkably similar to those of Quigley
a decade earlier: that housing is a complex product, whose special characteris-
tics need to be incorporated into formal models, but that the resulting range of
attributes and relationships cannot then all be dealt with within a single frame-
work. The result was seen to be a growing range of models addressing specific
elements of the housing market, almost all within a traditional comparative static
framework, and almost all concentrating on positive analysis. The bibliography
included was vast and, although the importance of institutional realism was noted
in the text, wholly North American in its orientation.
Published before the review, but not included there, was what is now argued
to be a seminal work by Poterba (1984). This innovative contribution placed
housing as an asset into a general equilibrium framework, where the emphasis
is on the equalisation of risk adjusted rates of return. The immediate problem
addressed was that of the impact of tax subsidies on demand for housing and
Ch. 40: Urban Housing Markets: Theory and Policy 1563
but beyond this there are a variety of different and partial approaches. Another
is that the vast majority of work undertaken in the field sees government as one
actor modifying specific market relationships rather than, as is the case in many
European countries, as a pervasive influence on the way in which the housing
system operates. This in turn places the emphasis on comparing outcomes across
localities, while in the European context far greater importance is placed on
analysing the impact of national policies on the operation of the housing system
as a whole rather than on how these policies impact at the local level (Maclennan,
1982; Meen, 1989, 1996).
therefore the rationale for government intervention in these markets (Hills, 1991;
Whitehead, 1993).
The analysis of demand started with the measurement of income and thereafter
price elasticities. In both cases the story is one where initial estimates were
seen to generate unrealistically high and low estimates respectively. More recent
analyses suggest that these early elasticity measures were inadequately founded
in theory and suffered from technical deficiencies in estimation. As a result these
approaches have now been continually refined so that they generate results that
are more in line with general expectations. The treatment of the economic deter-
minants of household formation and their impact on the demand for housing has
been much less formally examined (Smith et al., 1984).
Income elasticities are particularly important with respect to both urban struc-
ture and housing policy. The starting point is the work by Reid (1962) in which,
using Friedman's permanent income hypothesis and aggregate data, she found
that the income elasticity of demand might be nearer two than the usually as-
sumed value of unity or below. Starting with Lee (1964, 1968) much of the work
thereafter has been concentrated on refining both specification and testing, in
particular to take account of the problem that what is observable is expenditure
rather than price and quantity separately (de Leeuw, 1971 and 1976). Household
level studies have tended to result in far lower income elasticities. In particular
Polinsky (1977) showed how separating out transitory income, interpreted as
variance within aggregate estimates, allowed a better specification of permanent
income while Polinsky and Ellwood (1979) took account of spatial variations in
measured prices to improve the estimates of both price and income elasticities.
An equally important question has been whether and how income elasticities
vary with level of income. Ihlanfeldt (1982) suggested elasticities were higher
for higher income groups, while Hansen et al. (1996) using Lorenz curves shows,
income elasticity to be less than unity at all levels. Finally, problems associated
with the equilibrium assumptions have also been a matter of concern, with some
arguing that only recent moves should be included (Maclennan, 1982; Hanushek
and Quigley, 1982), while others suggest little difference between movers and
nonmovers (Goodman and Kawai, 1984).
Initial UK analyses tended to stress the importance of tenure, for instance
Byatt et al. (1973), which suggested that the elasticity was higher for owner-
occupation than for the private rented sector. A second area of concern in the
UK context was the impact of financial constraints on the individual house-
Ch. 40: Urban Housing Markets: Theory and Policy 1567
hold's capacity to adjust housing expenditure to desired levels. King (1980), for
instance, showed that, while measured income elasticities were still inelastic,
without constraints income elasticity would be much closer to unity. Later work
by Meen (1996), where the emphasis is on improved estimation of long-run
supply, generates income elasticities as high as 1.25.
In the comparative context the detailed work undertaken at the World Bank
(Mayo, 1981; Mayo and Malpezzi, 1985; Jimenez and Keane, 1984) tends to
suggest that income elasticities are reasonably constant across countries, that
there are significant differences between those for rental and owned housing and
that governmental and other constraints are important in limiting adjustment.
Estimates of price elasticities presented similar, closely related, problems.
Early unit cost based measures tended to generate estimates ranging from -0.3
to -0.7 in the United States (De Leeuw, 1971; Maisel et al., 1971). Estimates
based on appraisal data suggested the upper end of the range (Polinsky, 1977;
Polinsky and Elwood, 1979). Using grouped data tended to increase both income
and price elasticities in comparison to individual observations but still gener-
ated estimates of a similar order (Rosenthal, 1989). Excluding tenure choice
is agreed to bias both price and income elasticities upwards (Gillingham and
Hagemann, 1983). These and other authors have also argued that elasticities vary
with income (Friedman and Weinberg, 1981), while Harrington (1989) suggested
that taking a lifetime intertemporal approach generates a wider range of price
parameters.
An important question from the point of the view of the operation of housing
markets is the extent to which income and price elasticities, as well as cross
elasticities of demand, vary across different attributes. Early studies based on
hedonic approaches suggested significant differences in both price and income
elasticities between different attributes (Kain and Quigley, 1975; King, 1976,
1977; Witte et al., 1979; Follain and Jimenez, 1985). These studies use a range of
approaches, for instance, Awan et al. (1982) and Parsons (1986) which estimated
an Almost Ideal Demand System developed by Deaton and Muellbauer (1980)
across a range of attributes, Ball and Kirwan (1975, 1977) which took account of
space and structure and Boehm (1982), which examined the demand for dwelling
size and quality in relation to tenure choice. All these studies tend to suggest that
income and price elasticities vary considerably across the ranges of attributes that
have been identified, notably with respect to space, structure and environment as
do cross elasticities of demand with respect to substitute and complementary
attributes.
Later work has evaluated price and income elasticities within wider analysis
of urban areas. In the USA these differential estimates are crucial to the analysis
1568 C.M.E. Whitehead
of segmented markets (Rothenburg et al., 1991). In the UK, Cheshire and Shep-
pard (1995, 1998) suggest particularly high income elasticities of demand, as
well as larger price elasticities because of the capacity to adjust between at-
tributes within an analysis of the impact of land use planning on demand and
house prices.
An area of particular concern with respect to demand is that of tenure choice.
The American literature tends to treat the decision as dependent simply upon
economic and demographic variables with particular emphasis on price differen-
tials between tenure (Kain and Quigley, 1972; Struyk, 1976; McDonald, 1979;
Megbolugbe and Linneman, 1993). Later developments have concentrated on
the determination of the user cost of housing, taking account of both local and
national taxation policies as well as inflationary expectations within the context
of life cycle models (Kearl, 1979; Hendershott, 1980; Follain, 1982; Wheaton,
1985; Goodman, 1988). Similar studies have been undertaken in Australia where
the inclusion of user costs have significantly improved goodness-of-fit (Bourassa,
1995). There has been growing emphasis on portfolio aspects of tenure choice
(Haurin and Lee, 1989; Follain and Ling, 1988; Haurin, 1991; Pollakowski and
Wachter, 1990; Meyer and Wiend, 1996) and on the positive linkages between de-
cisions about quantity and tenure (Gillingham and Hagemann, 1983; Henderson
and Ionnides, 1983, 1986, 1989; Borsch-Supen and Pollakowski, 1990).
In the European context, far more of the emphasis has been on constraints
on tenure choice arising from financial restrictions and rent regulation as well as
the relationship between the market and administratively allocated social hous-
ing. King (1980) for instance, stressed the nature of the joint decision between
quantity and tenure in a system where there was rationing in both of the two
main tenures. Ermisch et al. (1996) using data for 1988 and 1989 when financial
rationing for owner-occupation had almost disappeared, found aggregated price
elasticities ranging from -0.35 to 0.44.
As models have become more comprehensive both income and price elastici-
ties tend to be generated as an element of the analysis, rather than being the object
of central concern (Rothenberg et al., 1991; DiPasquale and Wheaton, 1994; Pain
and Westerway, 1996). The implied values generally appear relatively similar to
those already discussed. The main conclusion must be that on the demand side
both income and price elasticities are generally less than one but vary between
tenures, income and demographic groups as well as between different housing
attributes.
Ch. 40: Urban Housing Markets: Theory and Policy 1569
2.3. Housing supply
One of the most important attributes of housing is the fact that new supply forms
only a very small proportion of the total supply of housing at any given time
and that the overall supply of housing is modified not just by new building but
also by improvement and conversion of the existing stock on the one hand and
depreciation of that stock on the other. Equally new housing supply is made up of
both land and structure, which means that the analysis of housing supply must be
linked both to the operation of land markets and the organisation of the construc-
tion industry. It is perhaps not surprising therefore that, for both technical and
data reasons, models of supply appear to have made relatively little progress as
compared to those dealing with demand (Quigley, 1979; Olsen in Mills, 1987).
adjustment approach of the type first put forward by Muth (1969) and similar to
that employed in the UK literature. Applying this suggests both slow adjustment
and a long run price elasticity of supply of around 1.2 in the USA. The work
by Tsoukis and Westaway (1994), which built heavily on Topel and Rosen's ap-
proach when examining the UK data, suggested a much lower long-run elasticity
of housing starts with respect to a permanent price increase, of around unity.
This model also finds no evidence that turnover in the housing market influ-
ences starts. Again the findings are somewhat different to the US where it has
been suggested that quantity signals in the form of vacancies play a relevant
part in determining the construction decision (Arnott in Mills, 1987) but that,
anyway, new construction reacts slowly to both market prices and vacancy rates
(Wheaton, 1990).
There is now growing interest in the UK in examining spatial interactions with
respect to both housing starts and house prices. Meen (1996) in particular has
clarified the conditions under which it is possible to consider national markets
independently from spatial structure, based on the standard aggregation condi-
tions of whether markets are homogenous, dependant and convergent. While he
suggests that the evidence points to the need to treat housing markets as a set of
inter-related local or regional markets, using a Poterba style model for the English
regions, he found that housing starts equations were remarkably similar across
regions and hence construction could be characterised as a national market.
Another important issue in the context of new construction is the treatment
of the impact of land markets and the planning system on the provision of new
housing. In the US there is a long tradition of analysis of the impact of zoning in
land and housing markets (e.g., Grieson and White, 1981; Mark and Goldberg,
1986) detailed in Chapter 42. The model by Mayo and Sheppard (1991, 1996)
examines the effect of stochastic development controls and suggests that there
may be situations in which increases in demand could impact more on the price
of vacant land than on developed land, resulting in a reduction in current supply.
In the UK context Cheshire and Sheppard (1989) examined the impact of land use
controls on supply and access to housing to two areas one with tight controls, the
other less constrained. Bramley (1993) estimated a cross-section model across
local authorities suggesting that land availability at the district level impacts
significantly on housing output. Monk et al. (1996) examined the processes by
which the planning system modifies supply and price over the economic cycle
to reduce effective output levels in a range of areas with different levels of con-
straint. Later work by Pryce (forthcoming) compares responsiveness in boom
and slump and finds evidence of a backward bending supply curve in periods of
boom. The research estimates average price elasticities of supply of 0.58 in the
1572 C.M.E. Whitehead
boom and 1.03 in the slump as well as considerable variation between districts.
All the evidence so far available therefore suggests that, even where supply of
new construction is very elastic in the medium term, the attributes of the land
market are extremely important in determining both the rate of response and the
overall supply elasticity.
So far we have discussed specific elements of the housing market and of urban
housing markets in particular. In so doing it has often been the case that extreme
assumptions have had to be made about the way in which the market actually
operates, for instance hypothesising that the conditions for long-run competitive
equilibrium exist in either a single market or a set of sub-markets.
Questions about whether housing markets can be effectively analysed in this
way lie at the forefront of the evaluation of empirical analysis and their rele-
vance for policy determination. Again there is a tension between incorporating
adjustment processes and introducing other, perhaps more realistic, assumptions
about the attributes of housing. Much of the analysis of adjustment has occurred
within aggregate models of the housing market while that introducing additional
attributes of housing occurs in the context of urban analysis.
head, 1971, 1974). Yet clearly results at the aggregate level impact significantly
on urban housing markets, notably in terms of the incentive to invest in and
maintain the quality of the housing stock. They are also important with respect
to the speed of adjustment, the impact of macro variables on housing decisions
and the robustness of equilibrium assumptions.
The possibility of a collapse in real house prices as a result of demographic
changes has generated new interest in aggregate housing market models in the
US. Mankiw and Weil (1989, 1992) were the first to put forward a model suggest-
ing that lower rates of new household formation and an ageing population would
result in a 40% fall in real house prices. Their results have been contested by,
among others, Englehardt and Poterba (1990), Hamilton (1991), Holland (1991)
Hendershott (1991) and Woodward (1991). DiPasquale and Wheaton (1994) used
the question as a starting point for the reevaluation of this type of approach,
generating a stock flow model with price and output adjustment. The strong con-
clusion was that prices in the owner-occupied market adjust slowly, in contrast to
the general assumptions of instantaneous market clearing, with the implication
that housing markets operate in ways that are far different from other financial
assets.
Aggregate models incorporating adjustment mechanisms have been funda-
mental to the developments in housing market analyses in the UK, notably in the
1990s by Meen (1990, 1993, 1996b), Milne (1991) and Muellbauer and Mur-
phy (1994, 1997), although the usual starting point here has been the impact of
finance market liberalisation as well as fundamental demographic and income
relationships.
A rather different approach at the aggregate level has been that of evaluating
the efficiency of housing in an asset market. The starting point here was the article
by Poterba (1984) which postulates a rational expectations approach to housing
in an efficient asset market. A range of studies of the North American evidence
in Hendershott and Hu (1981), Case and Schiller (1989, 1990), Mankiw and Weil
(1991), Poterba (1991) and, in the Canadian context, Clayton (1996, 1997), all
suggest that the arbitrage conditions do not hold at a given point in time, so
that excess returns can be maintained for long periods. This does not necessarily
negate the basic Poterba model, as long as market behaviour converges to the
basic equation over both time and space (Meen, 1996). However, this may be
an heroic assumption. Housing market efficiency has also been discussed in the
context of Tokyo (Ito and Hirono, 1993) and Britain (Evans, 1995).
Nontraditional models of housing market disequilibrium concentrate on the
extent to which house prices can be regarded as the sole determinant of supply
and the speed of the supply adjustment (Fair, 1972; Eubank and Sirmans, 1979;
Ch. 40: UrbanHousing Markets: Theory and Policy 1575
Rosen and Smith, 1983; Hendershott and Haurin, 1988). All tend to suggest that
other factors, notably the vacancy rate, sales rates, financial market and macro-
economic variables can be significant (Arnott, 1989; Kaserman et al., 1989).
None suggest that instantaneous adjustment is a reasonable assumption.
In this context there has been considerable interest in the question of transac-
tions costs, notably with respect to demand. Chinloy (1980) for instance incor-
porated a range of implicit and explicit costs many of them related to financing,
Goodman (1995) examined the differences in these costs between owners and
renters while Read (1991) looked at national rates of vacancies. A range of
studies has shown that these costs affect mobility as much, if not more than, the
benefits of these moves (Hanuschek and Quigley, 1978; Weinberg, Friedman and
Mayo, 1981). In the UK context, Muellbauer and Murphy (1996) have suggested
that these can impact on market behaviour to generate important non-linearities
in house prices.
The approach to analysing housing policy is very different in America and Eu-
rope. In the American literature the government and particularly local govern-
ment, is seen as one actor that modifies private investment decisions through
taxation, subsidy and quantitative constraints. The impacts of these policies are
then traced through market behaviour. Few question whether the mainstream the-
ory of competitive equilibrium discussed in this chapter is appropriate to policy
1578 C.M.E. Whitehead
analysis. Arnott (in Mills, 1987) is unusual when he argues that "the highest
priority item on the research agenda in the economic theory of housing markets
is the development of imperfectly competitive and non-competitive models of the
housing market" (p. 985). Most of the emphasis in later research has in fact gone
the other way, improving the theoretical base but in the context of competitive
equilibrium rather than market failure and welfare analysis.
Second, the American literature emphasises local market analysis both be-
cause of limited federal intervention in housing and the fact that policies are
mediated and initiated by local government in many different ways. The Experi-
mental Housing Allowance Program undertaken in the 1970s is a good example
in that it was intentionally implemented in only a small number of areas allowing
relatively easy comparison with similar areas where no subsidy was available
(Bradbury and Downs, 1982; Friedman and Weinberg, 1981 and 1982). The Ex-
perimental Housing Allowance Program generated important empirical results
with strong policy implications. On the demand side the experiment included a
range of different subsidies to selected households in six cities for a three year
period. The outcome was far lower price elasticities of demand than had been pre-
dicted but relatively large price elasticities of supply ensuring that the majority of
benefits went to consumers. The supply experiment which lasted for ten years and
was available to all eligible landlords included a minimum quality requirement
which effectively raised standards at relatively low cost (Rydell, 1976; Rydell et
al., 1982; Lowry, 1983).
Equally rent control, dwelling-based taxation, housing assistance to low in-
come families and zoning and planning controls are all fundamentally local poli-
cies which can thus be evaluated by comparison across areas (Apgar, 1990;
Malpezzi and Green, 1996; Galster, 1997). In this context, Galster, for instance,
has argued that the demand side approach to subsidy will normally be supe-
rior to supply side housing assistance because of the capacity of sub-markets
to respond effectively (Galster, 1997). Anas and Arnott (1994) generate similar
conclusions within a dynamic-equilibrium simulation model of Chicago. Critics
of this approach from a European angle, argue both that the range of objections
is too narrowly defined and that the outcome depends more on the specifics of
the subsidy (Yates and Whitehead, 1998).
A third important element in the American approach to policy analysis is the
emphasis placed on the costs of government intervention, the extent of admin-
istrative failure and the evaluation of alternative policies in terms of these costs
(Weicher, 1979; Nesslein, 1982, 1988; Apgar, 1990; Galster, 1997).
In Europe housing has been seen far more as a fundamental part of national
social policy throughout much of the century. This impacts on what is seen as
Ch. 40: UrbanHousing Markets: Theory and Policy 1579
relevant theory, as well as on the approach to policy analysis. In particular the
starting point tends to be that of market imperfection, the relevance of wel-
fare economics and the rationale of government intervention (Maclennan, 1982;
Whitehead, 1984; Turner et al., 1977).
Second, the emphasis on housing as part of welfare and social policy means
that there has been large scale intervention at the national level, including con-
siderable emphasis on direct provision of housing as well as large scale supply
subsidies and controls on the pricing and financing of housing (Maclennan and
Williams, 1990a, b; Boelhouwer and van der Heijden, 1992, 1993; Turner and
Whitehead, 1993). These policies are generally national in character and, while
often mediated by local government, apply across all jurisdictions. It is not nor-
mally politically acceptable to apply policies to specific areas as was done in
the US with the Housing Allowance Experiment. Instead policies apply either
across the country as a whole or to relevant categories of households, providers
or localities. Policy analysis tends either to be undertaken at the national level
or through case studies of the impact of particular programmes on local housing
conditions rather than the operation of the market. Moreover because of the scale
and type of intervention much of the analysis is tenure rather than market driven
(Freeman et al., 1996; Boelhouwer and van der Heijden, 1993).
It is impossible in a review chapter of this kind to give more than a flavour
of the types of policy issues that are relevant to the economics of urban housing
markets, particularly as the approaches are so different between countries. Here
we touch on four areas of current debate: market versus administrative provision;
the relationship between housing and labour markets; the impact of local taxation
on house prices; and the impact of rent control on investment in rental housing.
The impact of zoning and land use planning on housing provision and prices is
another issue of particular importance, but is discussed in detail in Chapter 42.
Two areas of concern have not been included: the relationship between racial
segregation and housing markets, which is seen as of particular importance in
the US context and the impact of housing tax expenditures on housing demand
which is seen as of more national the urban interest. Both are well covered in
the literature (Yinger, 1979; Smith et al., 1988; Wood, 1988; Lundqvist, 1988;
Englund et al., 1995).
An important aspect of the social provision of housing is that the output is allo-
cated to households by administrative rather than market mechanisms. This has
impacts on household behaviour and on the nature of local housing and urban
systems. The question which has been much discussed in the UK literature has
been the impact of administrative allocation on mobility and ultimately employ-
ment. Hughes & McCormick (1981, 1985, 1987, 1990) have examined migration
behaviour, comparing outcomes with the United States and with the UK, between
council tenants and equivalent owner-occupiers. They show that council house
tenants tend to migrate less often that those in the market sector and when they do
move it is likely to be for housing rather than labour market reasons. More gen-
erally administrative allocation, rent controls and subsidies to owner-occupation
are all shown adversely to affect the operation of the labour market (Bover et
al., 1988). Earlier work by, for instance, Gordon and Lamont (1982), tended to
point to tenure as an important determinant of differential unemployment at the
local level. In the 1990s, much of the discussion has turned to tenure polarisation,
social exclusion and urban regeneration rather than to interaction with the labour
market per se.
The impact of local taxation on house prices and community satisfaction has
been an area of considerable interest in both North America and Europe. In the
American literature much of the discussion has concentrated on Tiebout (1966)
style analysis, evaluating the impact of the mix of expenditure and taxes on de-
mand and price (Oates, 1969; Yinger, 1982; Wheaton, 1993). A survey of these
approaches can be found in Yinger et al. (1988), which also includes estimates
of the impact of tax reassessment, suggesting capitalisation in house prices of
around 15-25%. Mieszkowski and Zodrow (1989) examines the impact of local
expenditure and shows it to have an important impact on property values.
The UK, with its local taxation changes from property taxes to person-based
taxes to a mixture of the two provides an interesting case study. Work for the
Department of the Environment suggested on the basis of relative demand and
supply elasticities subject to a range of constraints, uncertainties and transaction
costs, that the impact of the change from property to person-based taxation would
be of the order of 5% of total house prices (Department of the Environment,
1988). Hughes (1989a, b), using a similar approach, suggested a much larger
impact.
1582 C.M.E. Whitehead
3.4. Rent control and its impact on the provision of rented housing
The impact of rent controls on the operation of the housing market has been the
subject of more theoretic and empirical analysis on both sides of the Atlantic
than almost any other housing topic. Many countries in Europe introduced rent
controls during the 1914-1918 war with constraints continuing in varying de-
grees throughout the century (Arnott, 1988). In the US certain cities, notably
New York, have had long experience of controls as well as periods of decon-
trol (Gyourko and Linneman, 1990; Ault and Saba, 1990). There have been
detailed studies in countries around the world (e.g., Cheung, 1979 for Hong
Kong; Malpezzi, 1986 for Cairo).
Early work concentrated on setting out the analytic framework in which to
analyse the impact of rent control particularly on housing as an investment but
also on prices in other markets (Frankena, 1975; Olsen, 1969; Cheung, 1975;
Fallis and Smith, 1984; Kempner and Fine, 1989) as well as on empirical analyses
of simple relationships (e.g., Olsen, 1972; Stegman, 1982). Much of this work has
been criticised as being based on too simple models and on casual empiricism
(Olsen, 1988).
One area where there has been considerable development both of theoretical
models and empirical analysis is that of the impact of rent control on maintenance
investment (Frankena, 1975; Dildine et al., 1974; Olsen, 1969, 1988; Albon and
Stafford, 1990; Kutty, 1995, 1996). These have tended to suggest that there will
be a deleterious effect on both quality and the life of the building but that better
formulated policies can limit or even reverse these effects. Empirical studies
have generated more results although the majority suggest reduced investment
and increased dereliction (Cheung, 1975; Gyorko and Linneman, 1990). What
they do suggest very strongly is that the specifics of controls are important in
determining the extent and nature of landlord response.
In Europe and elsewhere much of the discussion on rent control has been
subsumed into a wider discussion of the incentives and disincentives associated
with the provision of privately rented housing (van der Heijden and Boelhouwer,
1996; Yates, 1996; Crook and Kemp, 1996) and of the attributes of private rented
accommodation for landlords, tenants and investment (Whitehead, 1996). Again
Ch. 40: Urban Housing Markets: Theory and Policy 1583
the analyses tend to be based on implicit models and to provide a general de-
scription of trends, incentives and constraints, rather than tests of more formal
hypotheses.
4. Conclusions
are researched. The areas where the least progress has occurred are the far more
intractable problems of how effectively to address questions of market failure
and to model and evaluate the interaction between government and the housing
market.
References
Albon, R.P and D.C. Stafford (1990), "Rents control and housing maintenance", Urban Studies 27:233-240.
Alonso, W. (1964), Location and Land Use (Harvard University Press, Cambridge, MA).
Anas, A. (1978), "Dynamics of urban residential growth", Journal of Urban Economics 5:66-87.
Anas, A. and R.J. Arnott (1994), "The Chicago prototype housing market model with tenure choice and its
policy implication", Journal of Housing Research 5:23-90.
Apgar, W.C. (1990), "Which housing policy is best?", Housing Policy Debate 1:1-32.
Arcelus, F and A.H. Meltzer (1973), "The markets for housing and housing services", Journal of Money Credit
and Banking 1:78-99.
Amott, R. (1980), "A simple urban growth model with durable housing", Regional Science and Urban
Economics 10:53-76.
Arnott, R. (1988), "Rent control: the international experience", Journal of Real Estate Finance and Economics
1:3.
Arnott, R. (1989), "Housing vacancies, thin markets and idiosyncratic tastes", Journal of Real Estate Finance
and Economics 2:5-30.
Arnott, R., R. Davidson and D. Pines (1983), "Housing quality maintenance and rehabilitation", Review of
Economic Studies 50:467-494.
Amott, R., R. Davidson and D. Pines (1986), "The spatial aspects of housing quality, quantity and
maintenance", Journal of Urban Economics 19:190-217.
Ault, R. and R. Saba (1990), "The economic effects of long-term rent control: the case of New York City",
Journal of Real Estate Finance and Economics 3:25-41.
Awan, K., J.C. Odling-Smee and C.M.E. Whitehead (1982), "Housing attributes and the demand for private
rented housing", Economica 49:183-200.
Ball, M. (1981), "The development of capitalism in housing provision", International Journal of Urban and
Regional Research 5:145-177.
Ball, M. (1983), Housing Policy and Economic Power (Methuen, London and New York).
Ball, M. and R. Kirwa (1975), The Economics of an Urban Housing Market, Bristol Area Study (Centre for
Environmental Studies, London).
Ball, M. and R. Kirwan (1977), "Accessibility and supply constraints in an urban housing market", Urban
Studies 14:11-32.
Boehm, T. (1982), "A hierarchical model of housing choice", Urban Studies 19:17-31.
Boelhower, P. and H. van der Heijden (1992), Housing Systems in Europe (Delft University Press, Delft).
Boelhower, P. and H. van der Heijden (1993), "Housing policy and housing finances in seven European coun-
tries", in: B. Turner and C.M.E. Whitehead, eds., Housing Finance in the 1990s, SB56 (The National Institute
for Building Research, Givle).
Bi6rsch-Supan, A. and H. Pollakowski (1990), "Estimating housing consumption adjustments from panel data",
Journal of Urban Economics 27:131-150.
Bourassa, S.C. (1995), "A model of housing tenure choice in Australia", Journal of Urban Economics, 37:161-
175.
Bover, O., J. Muellbauer and A. Murphy (1988), "Housing, wages and UK labour markets", Oxford Bulletin
of Economics and Statistics 51:97-136.
Bradbury, K. and A. Downs, eds. (1982), Do Housing Allowances Work? (Brookings Institution, Washington,
DC).
1586 C.M.E. Whitehead
Bradbury, K., R. Engle, 0. Irvine and J. Rothenberg (1977), "Simulation estimates of supply and demand for
housing location in a multizoned metropolitan area", in: G.K. Ingram ed., Residential Location and Urban
Housing Markets, pp. 51-86 (Lippincott, Ballinger for National Bureau of Economic Research, Cambridge,
MA).
Braid, R. (1981), The short run comparative statistics of a rental housing market", Journal of Urban Economics
10:280-310.
Bramley, G. (1993), "The impact of land use planning and tax subsidies on the supply and price of housing in
Britain", Urban Studies 30:5-30.
Bmeckner, J. (1981), "A dynamic model of housing production", Journal of Urban Economics 10: 1-14.
Bucldey. R.M. (1982), "A simple theory of the UK housing sector", Urban Studies 19(3):303-312.
Buckley, R. and J. Ermisch (1982), "Government policy and house prices in the United Kingdom: an
econometric analysis", Oxford Bulletin of Economics and Statistics 44:273-304.
Byatt, I.C.R., P. Holmes and D.W. Laidlaw (1973), "Income and the demand for housing: some evidence for
Great Britain", in: Parkinson, ed., Essays in Modem Economics (Longman, London).
Case, K.E. and R. Schiller (1989) "The efficiency of the market for single family homes", American Economic
Review 79:125-137.
Case, K.E. and R.J. Schiller (1990), "Forecasting prices and excess returns in the housing market", Journal of
the American Real Estate and Urban Economics Association 18.
Cheshire, P. and S. Sheppard (1989), "British planning policy and access to housing", Urban Studies 26:469-
485.
Cheshire, P. and S. Sheppard (1995), "On the price of land and the value of amenity", Econometrica 62:247-
267
Cheshire, P. and S. Sheppard (1998), "Estimating the demand for housing, land and neighbourhood character-
istics", Oxford Bulletin of Economics and Statistics 60(3):357-382.
Cheung, S.N.S. (1975), "Roofs or stars: the stated intent and actual effects of a rents ordinance", Economic
Inquiry 13:1-21.
Cheung, S.N.S. (1979), "Rent control and housing reconstruction:the post-war experience of pre-war premises
in Hong Kong", Journal of Law and Economics 22(1):27-53.
Chinloy, P. (1980), "An empirical model of the market for resale homes", Journal of Urban Economics 7:279-
292.
Clayton, J. (1996), "Rational expectations, market fundamentals and housing price volatility", Real Estate
Economics 4:441-470.
Clayton, J. (1997), "Are house price cycles driven by irrational expectations?", Journal of Real Estate Finance
and Economics 14:341-364.
Crook, A.D.H. and P.A Kemp (1996), "The revival of private rented housing in Britain", Housing Studies
1:51-68.
Davies, G. (1974), "An econometric analysis of residential amenity", Urban Studies 11:217-225.
Davies, O. and A. Whinston (1961), "The economics of urban renewal", Law and Contemporary Problems
26:105-117.
Deaton, A. and J. Muellbauer (1980), "An almost ideal demand system", American Economics Review 70:312-
326.
de Leeuw, F. (197 1), "The demand for housing: a review of cross-section evidence", Review of Economics and
Statistics 53:1-10.
de Leeuw, F. (1976), "The demand for housing: a review of the cross-sectional evidence", Review of Economics
and Statistics 53:1-10.
de Leeuw, F. and N.E Ekanem (1971), "The supply of rental housing", American Economic Review 61: 806-
817.
de Leeuw, F. and R. Struyk (1975), The Web of Urban Housing (Urban Institute, Washington).
Dicks, M. (1990), "A simple model of the housing market", Bank of England Discussion Paper, No. 49.
Dildine, L.L., F.A. Massey and R.E. Worth (1974), "Dynamic model of private incentives to housing
maintenance", Southern Economic Journal 40:631-639.
DiPasquale, D. and W.C. Wheaton (1992), "The cost of capital, tax reform and the future of the rental housing
market", Journal of Urban Economics 31:337-359.
Ch. 40: Urban Housing Markets: Theory and Policy 1587
DiPasquale, D. and W.C. Wheaton (1994), "Housing market dynamics and the future of house prices", Journal
of Urban Economics 35:1-27.
Edell, M. (1972), "Filtering in the private housing market", in: M. Edel and J. Rothenberg, eds., Reading in
Urban Economics, pp. 204-215 (Macmillan, New York).
Engelhardt, G.V. and J.M. Poterba (1990), "House prices and demographic change: Canadian evidence",
Regional Science and urban Economics 21:539-546.
Englund, P., P.H. Hendershott and B. Turner (1995), The Tax Reform and the Housing Market (National
Institute of Economic Research Economic Council, Stockholm).
Ericsson, N.R. and D.E Hendry (1985), "Conditional econometric modelling: an application to new house
prices in the UK", in A.C. Atkinson and S.E. Feinberg, eds., A Celebration of Statistics (Springer-Verlag,
Berlin).
Ermisch, J., J. Findlay, and K. Gibb (1996), "The price elasticity of demand for housing in Britain: issues of
sample selection", Journal of Housing Economics 5:64-86.
Eubank, A.A. and C.F. Sirmans (1979), "The price adjustment mechanism for rental housing in the United
States", Quarterly Journal of Economics 93(1):163-168.
Evans, A.W. (1973), The Economics of Residential Location (Macmillan, London).
Evans, A.W. (1995), "The property market:ninety per cent efficient", Urban Studies 32:5-29.
Fair, R.C. (1972), "Disequilibrium in housing models", Journal of Finance 27:207-221.
Fallis, G. and L.B. Smith (1984), "Uncontrolled prices in a controlled market: the case of rent controls",
American Economic Review 74:193-200.
Follain, J.R. (1979), "The price elasticity of the long-run supply of new housing", Land Economics 55:190-
199.
Follain, J.R. (1982), "Does inflation affect real behaviour?: the case of housing", Southern Economic Journal
48:570-582.
Follain, J.R. and E. Jimenez (1985), "Estimating the demand for housing characteristics: a survey and critique",
Regional Science and Urban Economics 15:77-107.
Follain, J.R. and D.C. Ling (1988), "Another look at tenure choice, inflation and taxes", Journal of the American
Real Estate and Urban Economics Association 16(3):207-229.
Frankena, M. (1975), "Alternative models of rent control", Urban Studies 12:303-308.
Freeman, A., A. Holmans and C.M.E. Whitehead (1996), Is the UK Different? International Comparisons of
Tenure Expenditure and Subsidy (Council of Mortgage Lenders, London).
Friedman, J. and D.H. Weinberg (1981), "The demand for rental housing:evidence from the housing allowance
demand experiment", Journal of Urban Economics 9:311-331.
Friedman, J. and D.A. Weinberg (1982), The Economics of Housing Vouchers (Academic Press, New York).
Fujita, M. (1976), "Spatial patterns of optimal growth", Journal of Urban Economics 3:193-208.
Gabriel, S.A. and FE. Nothaff (1988), "Rental housing markets and the natural vacancy rate", Journal of the
American Real Estate and Urban Economics Association 16:343-353.
Galster, G. (1981), "A neighbourhood interaction model of housing maintenance and quality changes by owner-
occupants", Regional Science Perspectives 11:29-48.
Galster, G. (1987), Homeowners and Neighbourhood Reinvestment (Duke University Press, Durham, NC).
Galster, G. (1997), "Comparing demand side and supply-side housing policies: market and spatial perspec-
tives", Housing Studies 12:561-577.
Gillingham, R. and R. Hagemann (1983), "Cross-sectional estimation of a simultaneous model of tenure choice
and housing services demand", Journal of Urban Economics 14:16-39.
Giussani, B. and G. Hadjimatheou (1990), "House prices: an econometrics model of the UK", The APEX
Centre Economics Discussion Paper No. 1.
Giussani, B. and G. Hadjimatheou (1991), "Modelling regional house prices in the UK", Regional Science
70:201-219.
Goodman, A.C. (1981), "Housing submarkets within urban areas:definitions and evidence", Journal of
Residential Science 21:175-185.
Goodman, A.C. (1988), "An econometric model of housing price, permanent income, tenure choice and
housing demand", Journal of Urban Economics 23(1):327-353.
1588 C.M.E. Whitehead
Goodman, A.C. (1995), "A dynamic equilibrium model of housing demand and mobility with transactions
costs", Journal of Housing Ecotiomics 4:307-327.
Goodman, A.C. and M. Kawai (1984), "Replicative evidence on the demand for owner-occupied and rental
housing", Southern Economic Journal 50:1036-1037.
Gordon, I.R. and D. Lamont (1982), "A model of labour market interdependencies in the London region",
Environment and Planning A 14:237-264.
Gosling, J., G. Keogh and M. Stabler (1993), "House extensions and housing market adjustment: a case study
of Wokingham", Urban Studies 30:1561-1576.
Grebler, L. and S.J. Maisel (1963), "Determinants of residential construction: a review of present knowledge",
in: D. Suits et al., eds., Impacts of Monetary Policy (Prentice Hall, Englewood Cliffs).
Grieson, R.E. and J.R. White (1981), "The effects of zoning on structure and land markets", Journal of Urban
Economics 10:27-85.
Grigsby, W. (1963), Housing Markets and Public Policy (University of Pennsylvania Press, Philadelphia).
Gyourko, J. and P. Linneman (1990), "Equity and efficiency aspects of rent control: an empirical study of New
York City", Journal of Urban Economics 26:54-74.
Hadjimatheou, G. (1976), Housing and Mortgage Markets (Saxon House, Farnborough).
Hamilton, B.W. (1991), "The baby boom, the baby bust and the housing market: a second look", Regional
Science and Urban Economics 21:547-552.
Hancock, K. (1993), "Can't pay, won't pay: or economic principles of affordability", Urban Studies 30(1): 127-
145.
Hancock, K. and D. Maclennan (1990), Housing Price Monitoring Systems and Housing Planning in Scotland
(Central Research Unit, Scottish Office, Edinburgh).
Hansen, J.L., J.P. Formby and W.J. Smith (1996), "The income elasticity of demand for housing: evidence
from concentration curves", Journal of Urban Economics 39:173-192.
Hanushek, E. and J. Quigley (1978), "An explicit model of intra-metropolitan mobility", Land Economy
54:411-429.
Hanushek, E. and J. Quigley (1979), "The dynamics of the housing market: a stock adjustment model of
housing consumption", Journal of Urban Economics 6:90-111.
Hanushek, E. and J. Quigley (1980), "What is the price elasticity of renewal?", Review of Economics and
Statistics 62:449-454.
Hanushek, E. and J. Quigley (1982), "The determinants of housing demand", in: J. Vernon Henderson, ed.,
Research in Urban Economics, vol. 2, pp. 221-242 (JAI Press, Greenwich, Corin).
Harrington, D.E. (1989), "An intertemporal model of housing demand: implications for the price elasticity",
Journal of Urban Economics 25(2):230-246.
Hrsman, B. and J. Quigley, eds., (1991), Housing Markets and Housing Institutions: An International
Comparison (Kluwer Academic, Boston, Dordrecht, London).
Haurin, D.R. (1991), "Income variability, home-ownership and housing demand", Journal of Housing
Economics 1(1):60-74.
Haurin, D.R. and K. Lee (1989), "A structural model of the demand for owner-occupied housing", Journal of
Urban Economics 26:348-360.
Hendershott, P. (1980), "Real user costs and the demand for single-family housing", Brookings Papers on
Economic Activity, vol. 2, pp. 401-444 (Washington DC).
Hendershott, P.H. (1991), "Are real estate prices likely to decline by 47%?", Regional Science and Urban
Economics 21:553-563.
Hendershott, PH. and D.R. Haurin (1988), "Adjustments in the real estate market", Journal of the American
Real Estate and Urban Economics Association 16:
Hendershott, P. and S.C. Hu (1981), "Inflation and extraordinary returns on owner-occupied housing: some
implications of our capital allocation and productivity growth", Journal of Macroeconomics 3:177-203.
Henderson, J.V. (1985), Economic Theory and the Cities (Academic Press, Orlando).
Henderson, J.V. and Y.M. loannmides (1983), "A model of housing tenure choice", American Economic Review
73:98-113.
Henderson, J.V. and Y.M. Ioannides (1986), "Tenure choice and the demand for housing", Economica 53:231-
246.
Ch. 40: Urban Housing Markets: Theory and Policy 1589
Henderson, J.V. and Y.M. Ioannides (1989), "Dynamic aspects of consumer decisions in housing markets",
Journal of Urban Economics 26:212-230.
Hendry, D.F. (1984), "Econometric modelling of house prices in the UK", in: D.F Hendry and K.E Wallis,
eds., Econometrics and Quantitative Economics (Basil Blackwell, Oxford).
Hills, J. (1991), Unravelling Housing Finance (Oxford University Press, Oxford).
Hills, J., F Hubert, H. Tomann and C.M.E. Whitehead (1990), "Shifting substitution for bricks and mortar to
people", Housing Studies 5:147-167.
Holland, S.A. (1991), "The baby boom and the housing market: another look at the evidence", Regional
Science and Urban Economics 21:565-571.
Holmans, A.E. (1995), Housing Demand and Need in England 1991-2011 (Joseph Rowntree Foundation,
York).
Hoyt, H. (1939), Structure and Growth of Residential Neighbourhoods in American Cities (Federal Adminis-
tration, Washington DC).
Hughes, G.A. (1989a), "The poll tax and the housing market", The Economic Review September: 10-13.
Hughes, G.A. (1989b), "The switch from domestic rates to the community charge in Scotland", Fiscal Studies
10:1-12.
Hughes, G.A. and B. McCormick (1981), "Do council house policies reduce mobility between regions?",
Economic Journal 91:919-937.
Hughes, G.A. and B. McCormick (1985), "Migration intentions in the UK: which households want to migrate
and which succeed", Economic Journal 95:113-123.
Hughes, G.A. and B. McCormick (1987), "Housing markets, unemployment and labour market flexibility in
the UK", European Economic Review 31:615-645.
Hughes, G.A. and B. McCormick (1990), "Housing and labour market mobility", in J. Ermisch, ed., Housing
and the National Economy (National Institute of Economic and Social Research, London).
Igarashi, M. (1991), "The rent vacancy relationship in the rental housing market", Journal of Housing
Economics 1.
Ihlanfeldt, K.R. (1982), "Property tax incidence on owner-occupied housing: evidence from the Annual
Housing Survey", National Tax Journal 35:89-97.
Ingram, G.K., ed. (1977), Residential Location and Urban Housing Markets (Ballinger, Cambridge Mass).
Ingram, G.K. and J.F. Kain (1973), "A simple model of production and the abandonment problem", American
Real Estate and Urban Economics Association Journal 1:79-106.
Ingram, G.K., J.E Kain and J.R. Ginn (1972), The Rental Prototype of the NBER Urban Simulation Model
(National Bureau of Economic Research, New York).
Ito, T. and K. Hirono (1993), "Efficiency of the Tokyo housing market", NBER Working Paper No. 4382.
Jimenez, E. and D. Keane (1984), "Housing consumption and permanent income in developing countries:
estimates from panel data in El Salvador", Journal of Urban Economics 15:172-194.
Kain J.E and J.M. Quigley (1970), "Measuring the value of housing quality", Journal of the American
Statistical Association June:532-547.
Kain, J. and J. Quigley (1972), "Housing market discrimination, home-ownership and savings behaviour",
American Economic Review 62:263-277.
Kain, J.E and J.M. Quigley (1975), Housing Markets and Racial Discrimination (National Bureau of Economic
Research, New York).
Kaserman, D.L., J.L. Trimble and R.C. Johnson (1989), "Equilibrium in a negotiated market: evidence from
housing", Journal of Urban Economics 26(1):30-42.
Kearl, J.R. (1979), "Inflation, mortgages and housing", Journal of Political Economy 87(5):115-1138.
Kempner, J.L. and M.E. Fine (1989), "The high cost of rent control", Journal of Real Estate Finance and
Economics.
King, A.T. (1976), "The demand for housing", in: N. Terleckyj, ed., Household Production and Consumption
(National Bureau of Economic Research, New York).
King, A.T. (1977), "The demand for housing: a Lancaster approach", Southern Economic Journal 43:1077-
1087.
King, M.A. (1980), "An econometric model of tenure choice and the demand for housing as a joint decision",
Journal of Public Economics 14:137-159.
1590 C.M.E. Whitehead
Kirwan, R. and D. Martin (1972), The Economics of Urban Residential Renewal and Improvement, Working
Paper CES WP77 (Centre for Environmental Studies, London).
Kiyotaki, N. and J.E. Moore (1997), "Credit cycles", Journal of Political Economy 105:211-248.
Kutty, N.K. (1995), "A dynamic model of landlord reinvestment behaviour", Journal of Urban Economics
37:212-237.
Kutty, N.K. (1996), "The impact of rent control in housing maintenance: a dynamic analysis incorporating
European and North American rent regulations", Housing Studies 11:69-88.
Lancaster, K.J. (1966), "A new approach to consumer theory", Journal of Political Economy 74:132-157.
Lancaster, K. (1979), Variety, Equity and Efficiency (Basil Blackwell, Oxford).
Lansing, J.B., C.W. Clifton and J.N. Morgan (1969). New Homes and Poor People: a Study of Chains of Moves
(Survey Research Centre, Ann Arbor).
Lee, T.H. (1964), "The stock demand elasticities of non-farm housing", Review of Economics and Statistics
49:82-89.
Lee, T.H. (1968), "Housing and permanent income", Review of Economics and Statistics 50:480-490.
Leven, C., J. Little, H. Nourse and R. Read (1976), Neighbourhood Change (Praeger, New York).
Linneman, P. (1985), "An economic analysis of the home-ownership decision", Journal of Urban Economics
17:230-246.
Lowry, I.S. (1960), "Filtering and housing standards: a conceptual analysis", Land Economics 36:362-370.
Lowry, I.S. (1983), Experimenting with Housing Allowances: The Final Report of the Housing Assistance
Supply Experiment (Oelgechlager, Gunn and Hain, Cambridge, MA).
Lundqvist, L.J. (1988), Housing Policy and Tenures in Sweden, the Quest for Neutrality (Gower/Avebury,
Aldershot).
Maclennan. D. (1977), "Some thoughts on the nature and purpose of house price studies", Urban Studies
14:39-71.
Maclennan, D. (1982), Housing Economics (Longman, Harlow).
Maclennan, D. and R. Williams, eds. (1990a), Affordable Housing in Britain and America (Joseph Rowntree
Foundation, York).
Maclennan, D. and R. Williams, eds. (1990b), Affordable Housing in Europe (Joseph Rowntree Foundation,
York).
Maclennan, D., M. Munro and G. Wood (1987), "Housing choice and the structure of the housing submarket",
in: B. Turner, J. Kemeny and L. Lundquist, eds., Between State and Market Housing in the Post Industrial
Era, pp. 26-51 (Almquist and Wicksell International, Stockholm).
Maclennan, D. and Y. Tu (1996), "The microeconomics of local housing market structure", Housing Studies
11:387-406.
Maisel, S.J., J.B. Burnham and J.S. Austin (1971), "The demand for housing: a comment", Review of
Economics and Statistics 53:410-415.
Malpezzi, S.J. (1986), Rent Control and Housing Market Equilibrium: Theory and Evidence from Cairo,
Doctoral Dissertation (George Washington University).
Malpezzi, S.J. and R. Green (1996), "What has happened at the bottom of the US housing market?", Urban
Studies 33:1807-1820.
Mankjw, N.G. and D.N. Weil (1989), "The baby boom, the baby bust and the housing market", Regional
Science and Urban Economics 19:235-258.
Mankiw, N.G. and D.N. Weil (1992), "The baby boom, the baby bust and the housing market", Regional
Science and Urban Economics 21:573-579.
Margolis, S. (1982), "Depreciation of housing: an empirical consideration of the filtering hypothesis", Review
of Economics and Statistics 64:90-96.
Mark, J.H. and M.A. Goldberg (1986), "A study of the impacts of zoning on housing values over time", Journal
of Urban Economics 20: 257-273.
Mayo, S.K. (1981), "Theory and estimation in the economics of housing demand", Journal of Urban Economics
10:95-116.
Mayo, S.K. and S. Malpezzi (1985), Housing Demand in Developing Countries, WP-733 (The World Bank,
Washington DC).
Ch. 40: Urban Housing Markets: Theory and Policy 1591
Mayo, S.K. and S. Sheppard (1991), "Housing supply and the effects of stochastic development control",
Discussion Paper in Economics, Oberlin.
Mayo, S.K. and S. Sheppard (1996), "Housing supply under rapid economic growth and varying regulatory
stringency: an international comparison", Journal of Housing Economics 5:274-289.
McDonald, J.E (1979), Economic Analysis of an Urban Housing Market (Academic Press, New York).
McDonald, J.E (1981), "Capital-land substitution in urban housing: a survey of empirical estimates", Journal
of Urban Economics 9:190-211.
McLeod, B. and J. Ellis (1982), "Housing consumption over the family life cycle: an empirical analysis",
Urban Studies 19:177-185.
Meen, G. (1989), "The ending of mortgage rationing and its effects on the housing market: a simulation study",
Urban Studies 26:240-253.
Meen, G. (1990), "The removal of mortgage market constraints and the implications for econometric
modelling", Oxford Bulletin of Economics and Statistics 52.1:1-24.
Meen, G. (1993), The Treatment of House Prices in Macroeconometric Models (Department of the Environ-
ment, London).
Meen, G. (1996a), "Spatial aggregation, spatial dependence and predictability in the UK housing market",
Housing Studies 11:345-372.
Meen, G. (1996b), "Ten propositions in UK housing macroeconomics: an overview of the 1980s and early
1990s", Urban Studies 33:425-444.
Meen, G. (1998), "Modelling sustainable home-ownership: demographics or economics", Urban Studies
35(11):1919-1934.
Meen, G. and M. Andres (1998), Modeling Regional House Prices, A Review of the Literature (Department of
Environment, Transport &the Regions, London).
Megbolugbe, I. and P. Linneman (1993), "Home ownership", Urban Studies 30:659-682.
Mendelsohn, R. (1977), "Empirical evidence on home improvements", Journal of Urban Economics 4:459-
468.
Meyer, N. (1981), "Rehabilitation decisions in rental housing", Journal of Urban Economics 10:76-94.
Meyer, N. (1985), "The impacts of lending, race and ownership on rental housing rehabilitation", Journal of
Urban Economics 17:349-374.
Meyer, R. and K. Wiend (1996), "Risk and return to housing, tenure choice and the value of housing in an asset
pricing context", Real Estate Economics 24:113-132.
Mieskowski, P. and G.R. Zodrow (1989), "Taxation and the Tiebout model", Journal of Economic Literature
27:1098-1146.
Mills, E.S. (1967), "An aggregative model of resource allocation in a metropolitan area", American Economic
Review 57:197-210.
Mills, E.S. ed. (1987), Handbook of Regional and Urban Economics (North Holland, Amsterdam).
Milne, A. (1991), Incomes, Demography and UK House Prices, LBS Centre for Economic Forecasting,
Discussion Paper 39-90.
Monk, S., B.J. Pearce and C.M.E. Whitehead (1996), "Land use planning, land supply and house prices",
Environment and Planning A 28:495-511.
Muellbauer, J. and A. Murphy (1994), "Explaining regional house prices in the UK", mimeo (Universities of
Oxford and Durham).
Muellbauer, J. and A. Murphy (1997), "Booms and busts in the UK housing market", Economic Journal
107:1701-1727.
Muth, R.E (1960), "The demand for non-farm housing", in: A. Harberger, ed., The Demand for Durable Goods,
pp. 29-96 (University of Chicago Press, Chicago).
Muth, R.F. (1969), Cities and Housing (University of Chicago Press, Chicago).
Muth, R.F. (1973), "A vintage model of the housing stock", Paper of the Regional Science Association 30:141-
156.
Muth, R.E (1988), "Housing market dynamics", Regional Science and Urban Economics 18:345-356.
Needleman, L. (1969), The Economics of Housing (Staples Press, London).
Nellis, J.G. and J.A. Longbottom (1981), "An empirical analysis of the determination of house prices in the
UK", Urban Studies 18:9-22.
1592 C.M.E. Whitehead
Nesslein, T.S. (1982), "The Swedish housing model: an assessment", Urban Studies 19:235-246.
Nesslein, T.S. (1988), "Housing: the market versus the welfare state model revisited". Urban Studies 25:95-
108.
Oates, W.E. (1969), 'The effects of property taxes and local public spending on property values: an empirical
study of tax capitalisation and the Tiebout hypothesis", Journal of Political Economy 77:957-971.
Ohls, J. (1975), "Public policy towards low income housing and filtering in housing markets", Journal of Urban
Economics 2:144-171.
Olsen, E.O. (1969), "A competitive theory of the housing market", American Economic Journal Part 1,59:612-
621.
Olsen, E.O. (1972), "An econometric analysis of rent control", Journal of Political Economy 80:1081-1100.
Olsen, E.O. (1988), "What do economists know about the effect of rent control on housing maintenance?",
Journal of Real Estate Finance and Economics 295-307.
Ozanne, L. (1996), "Do supply curves slope up? the empirical relevance of the critique of product economics",
Cambridge Journal of Economics 20:749-762.
Ozanne, L. and R. Struyk (1976), Housing from the Existing Stock (Urban Institute, Washington DC).
Ozanne, L. and R. Struyk (1978), "The price elasticity of supply of housing services", in L. Bourne and J.R.
Hitchcock, eds., Urban Housing Markets: Recent Directives in Research and Policy (University of Toronto
Press, Toronto).
Pain, N. and P. Westerway (1996), Modelling Structural Change in the UK Housing Market (National Institute
of Economics and Social Research, London).
Parsons, G.R. (1986), "An almost ideal demand system for housing attributes", Southern Economic Journal
53:347-363.
Peterson, W., C. Pratten and J. Tatch (1998), An Economic Model of the Demand and Need for Social Housing
(Department of Environment, Transport and the Regions, London).
Polinsky, A.M. (1977), "The demand for housing: a study in specification and grouping", Econometrica
45:447-461.
Polinsky, A.M. and D.T. Ellwood (1979), "An empirical reconciliation of micro and grouped estimates of the
demand for housing", Review of Economics and Statistics 61:199-205.
Pollakowsli, H.O. and S.M. Wachter (1990), "The effects of land use constraints on house prices", Land
Economics 66:315-324.
Potepan, M.J. (1989), "Interest rates, income and home improvement decisions", Journal of Urban Economics
25:282-294.
Poterba, J.M. (1984), "Tax subsidies to owner-occupied housing: and asset market approach", Quarterly
Journal of Economics 99:729-752.
Poterba, J.M. (1991), "House price dynamics: the role of tax policy and demography", Brookings Papers in
Economic Activity 2:143-203.
Pozdena, R.J. (1988), The Modem Economics of Housing: a Guide to Theory and Policy (Quorum Books,
Greenwood Press).
Pryce, G. (in press) "Constructive elasticities and land availability", Urban Studies.
Quigley, J. (1979), "What have we learned about urban housing markets?", in: P. Mieszkowski and M.
Straszheim, eds., Current Issues in Urban Economics, pp. 391-429 (John's Hopkins, Baltimore).
Rabianski, J. and J. Spitzer (1978), "A two-sector cross-section econometric model of the 1970 housing stock",
Growth and Change 9:2-8.
Read, C. (1991), "An equilibrium model of advertising and natural vacancies in a spatially differentiated
housing market", Journal of Housing Economics 1.
Reece, B.F. (1988), "The price adjustment process for rental housing: some further evidence", Journal of the
American Real Estate and Urban Economics Association 16:411-418.
Reid, M. (1962), Housing and Income (University of Chicago Press, Chicago).
Renaud, B. (1994), "Book review", Journal of Housing Economics 3:156-163.
Rosen, K.T. (1974), "Hedonic prices and implicit markets: product differentiation in pure competition", Journal
of Political Economy 82:34-55.
Rosen, K.T. and L.B. Smith (1983), "The price adjustment process for rental housing and the natural vacancy
rate", American Economic Review 73:779-786.
Ch. 40: UrbanHousing Markets: Theory and Policy 1593
Rosenthal, L. (1989), "Income and price elasticities of demand for owner-occupied housing in the UK:
evidence pooled cross-sectional and time series data", Applied Economics 21:761-775.
Rothenberg, J. (1967), The Economics of Urban Renewal (Brookings Institute, Washington).
Rothenberg, J. (1983), "Housing investment, housing consumption and tenure choice", in: R. Grieson, ed., The
Urban Economy and Housing (Lexington Books, Lexington).
Rothenberg, J., G.C. Galster, R.V. Butler and J. Pitkin (1991), The Maze of Urban Housing Markets (University
of Chicago Press, Chicago).
Rouwendal, J. (1989), Choice and Allocation Models for the Housing Market (Kluwer Academic Publishers,
Dordrecht).
Rydell, C.P. (1976), "Measuring the supply response to housing allowances", Papers of the Regional Science
Association, 37:31-57.
Rydell, C.P., K. Neels and C.L. Barnett (1982), Price Effects of a Housing Allowance Program R-2720HUD
(Rand Corporation, Santa Monica).
Seek, N. (1983), "Adjusting housing consumption: improve or move", Urban Studies 20:455-469.
Segal, D. (1979), The Economics of Neighbourhood (Academic Press, New York).
Shea, J. (1993), "Do supply curves slope up?", Quarterly Journal of Economics 108(1):1-32.
Smith, J.W. (1989), "A theoretical analysis of the supply of housing", Journal of Urban Economics 26:174-188.
Smith, L.B. et al. (1984), "The demand for housing, headship rates and household formation: an international
comparison", Urban Studies 21:407-414.
Smith, L.B., K.T. Rosen and G. Fallis (1988), "Recent developments in economic models of housing markets",
Journal of Economic Literature 26:29-64.
Stahl, K. (1985), Microeconomic Models of Housing Markets (Springer-Verlag, Berlin).
Stegman, M. (1972), Housing Investment in the Inner City: The Dynamics of Decay (MIT Press, Cambridge).
Stegman, M.A. (1982), The Dynamics of Rental Housing in New York City (Rutgers University, New Jersey).
Stemlieb, G. and B. Burchall (1973), Residential Abandonment: The Tenement Landlord Revisited (Rutgers
University Press, New Brunswick, NJ).
Stover, M. (1986), "The price elasticity of supply of single family detached housing", Journal of Urban
Economics 20:331-340.
Straszheim, M. (1975), An Econometric Analysis of the Urban Housing Market (National Bureau of Economic
Research, New York).
Struyk, R.J. (1976), Urban Home-ownership (DC Heath, Lexington).
Struyk, R.J. and M. Bendick (1981), Housing Vouchers for the Poor: Lessons from a National Experiment (The
Urban Institute Press, Washington DC).
Sweeney, J.L. (1974a), "A commodity hierarchy model of the rental housing market", Journal of Urban
Economics 1:288-323.
Sweeney, J.L. (1974b), "Quality commodity hierarchies and housing markets", Econometrica: 147-167.
Taub, R., G.D. Taylor and D. Durban (1984), Paths of Neighbourhood Change (Chicago University Press,
Chicago).
Tiebout, C.M. (1966), "A pure theory of local expenditures", Journal of Political Economy 64:416-424.
Tompkinson, P. (1979), "A model of housebuilder's supply behaviour", Applied Economics 11:195-210.
Topel, R. and S. Rosen (1988), "Housing investment in the United States", Journal of Political Economy
96:718-740.
Tsoukis, C. and P. Westaway (1994), "A forward looking model of housing construction in the UK", Economic
Modelling 11:266-279.
Tu, Y. (1997), "The local housing sub-market structure and its properties", Urban Studies 34:337-353.
Turner, B., J. Kemeny and L.J. Lundqvist (1977), Between State and Market: Housing in the Post-Industrial
Era (Almqvist and Wiksell International, Gbteborg).
Turner, B. and C.M.E. Whitehead (1993), Housing Finance in the 1990s, SB56 (The National Institute for
Building Research, Gavle).
Turner, B., C.M.E. Whitehead and J. Jacobsson (1996), "Comparative housing finance", in: Bostadspolitik
2000 Statens offentliga utredningar 1996: 156 (SOU, Stockholm).
van der Heijden, H. and P. Boelhouwer (1996), "The private rented sector in Western Europe: developments
since the Second World War and prospects for the future", Housing Studies 11(1): 13-34.
1594 C.M.E. Whitehead
Vandell, K. (1981), "The effects of racial composition in neighbourhood succession", Urban Studies 18:315-
333.
Watson, C.J. (1973), Household Movement in West Central Scotland, Occasional Paper No 26 (Centre for
Urban and Regional Studies, Birmingham).
Weicher, J. (1979), "Urban housing policy", in: P. Mieszkowski and M. Stroszheim, eds., Current Issues in
Urban Economics, pp. 469-508 (John's Hopkins, Baltimore).
Weicher, J. and T. Thibodeau (1988), '"Filtering and housing markets", Journal of Urban Economics 23:21-40.
Weinberg, D.H., J. Friedman, and S.K. Mayo (1981), "Intraurban residential mobility: the role of transaction
costs, market imperfections and disequilibrium", Journal of Urban Economics 9:332-348.
Wheaton, W.C. (1979), "Monocentric models of urban land use: contributions and critiques", in: P.
Mieszkowski and M. Straszheim, eds., Current Issues in Urban Economics, pp. 107-129 (Jolhn's Hopkins
University Press, Baltimore).
Wheaton, W.C. (1985), "Life-cycle theory, inflation and the demand for housing", Journal of Urban Economics
18:161-179.
Wheaton, W.C. (1990), "Vacancy, search and prices in a housing market matching model", Journal of Political
Economy 98:1270-1292.
Wheaton, W.C. (1993), "Land capitalisation, Tiebout mobility and the role of zoning regulations", Journal of
Urban Economics 34.
White, M.J. (1977), "On cumulative urban growth and urban density functions", Journal of Urban Economics
4:104-112.
White, M.J. (1986), "Property taxes and urban housing abandonment", Journal of Urban Economics 20:312-
330.
White, M.J. and L. White (1977), "The tax subsidy to owner-occupied housing: who benefits?", Journal of
Public Economics 7:111-126.
Whitehead, C.M.E. (1971), "A model of the UK housing market", Bulletin of Oxford University Institute of
Economics and Statistics 33:245-266.
Whitehead, C.M.E. (1974), The UK Housing Market: An Econometric Model (Saxon House, Farnborough).
Whitehead, C.M.E. (1984), "Privatisation and housing", in: J. Le Grand and R. Robinson, eds., Privatisation
and the Welfare State (Allen and Unwin, London).
Whitehead, C.M.E. (1991), "From need to affordability", Urban Studies 8:871-886.
Whitehead, C.M.E. (1993), "Privatising housing: an assessment of UK experience", Housing Policy Debate
4:101-139.
Whitehead, C.M.E. (1996), "Private renting in the 1990s", Housing Studies 11:7-12.
Whitehead, C.M.E. and J. Odling-Smee (1975), "Long run disequilibrium in the housing market", Urban
Studies 12:315-318.
Witte, A., H. Sumko and H. Erikson (1979), "An Estimate of a structural hedonic price model of the housing
market: an application of Rosen's theory of implicit markets", Econometric 47:1151-1174.
Wood, G. (1988), "Housing tax expenditures in OECD countries: economic impacts and prospects for reform",
Policy and Politics 16.
Wood, G. (1991), Taxation and Housing, National Housing Strategy Background Paper No. 5 (AGPS,
Canberra).
Woodward, S.E. (1991), "Economists' prejudices: why the Mankiw-Weil story is not credible", Regional
Science and Urban Economics 21:531-537.
Yates, J. (1996), "Towards a reassessment of the private rental market", Housing Studies 11(1):35-50.
Yates, J. and C.M.E. Whitehead (1998), "In defence of greater agnoticism: a response to Galster's 'Comparing
demand side and supply side housing policies' ", Housing Studies 13(3):415-424.
Yinger, J. (1979), "Prejudice and discrimination in the urban housing market", in P. Mieszkowski and M.
Straszheim, eds., Current Issues in Urban Economics (John's Hopkins, Baltimore).
Yinger, J. (1982), "Capitalization and the theory of local public finance", Journal of Political Economy 90:917-
943.
Yinger, J., H. Bloom, A. Borsch-Supan and H. Ladd (1988), Property Taxes and House Values: the Theory and
Estimation of Intra-Jurisdictional Property Tax Capitalisation (Academic Press, San Diego, CA).
Chapter41
STEPHEN SHEPPARD*
Contents
1. Introduction 1596
2. Theoretical foundation 1598
2.1. Implicit markets 1598
2.2. Hedonic price functions 1599
2.3. Structure of demand and welfare evaluation 1603
2.3.1. Properties of attribute demand 1604
2.3.2. Welfare analysis 1610
2.4. Hedonic price indices 1613
3. Implementation 1613
3.1. Estimation of hedonic price functions 1613
3.1.1. Parametric approaches 1614
3.1.2. Nonparametric approaches 1619
3.2. Estimation of demand for housing attributes 1619
3.2.1. Identification and endogeneity 1621
3.2.2. Specification 1630
3.2.3. Comparison with discrete choice approaches 1631
4. Conclusion 1631
References 1632
* I would like to thank Joseph Gyourko, Paul Chesire and Ed Mills for helpful comments on an earlier
draft of this chapter.
Handbook ofRegional and Urban Economics. Edited by E.S. Mills and P. Cheshire
( 1999 Elsevier Science B. V All rights reserved.
1595
1596 S. Sheppard
Abstract
This chapter examines the hedonic analysis of housing markets. These techniques
have been widely applied in studies of the demand for housing attributes and
environmental amenities. The chapter discusses the theoretical foundation of
hedonic analysis, the use of hedonic estimates of demand for welfare analysis,
the empirical difficulties that arise in such studies, and some of the methods for
overcoming these difficulties.
1. Introduction
Imagine, for a moment, that you are a private investigator or market researcher
studying the demand for food. You have a particular disadvantage, however, in
that you have been banned from entering the local grocer. You have found a place
outside where you can sit and photograph shoppers as they approach the checkout
counter, and from these photographs you can pretty much tell what foods each
customer has purchased (although some items may be obscured in the shopping
basket) and the total cost of all items combined. By bribing a contact at the local
bank, you are able to find out each shopper's income. That is all the information
you have. From this, can you infer the demand for eggs? Can you determine how
much households would be willing to pay to remove sugar import quotas?
Such a difficult assignment is essentially what is undertaken in the hedonic
analysis of housing markets. We have no direct observations of attribute prices.
We have somewhat imprecise observations of what attributes are purchased. We
have reasonably good observations of what is spent for the entire bundle, and how
much income the household has. From this we attempt to answer some important
questions.
Housing and residential construction are of central importance for determina-
tion of both the level of welfare in society and the level of aggregate economic
activity. In many economies a residence represents the most valuable single asset
owned by most individuals, and a very large share of total household wealth.
Ch. 41: Hedonic Analysis of Housing Markets 1597
In all economies the share of income spent on housing represents a very large
fraction of total expenditure. It is thus to be expected that economists would
devote considerable effort towards understanding the structure of the demand for
housing and equilibria in these markets.
The first difficulty that arises in this effort stems from the obvious heterogene-
ity of the product. To attempt a comparison between the price of houses in China
and Chile, or between housing prices in 1996 and 1946 requires us to address
the similarity and appropriateness of making such a comparison. We observe a
low-income central city resident who pays $500 per month for housing, while
the more affluent suburban family in the same city has a monthly expenditure of
$1500. Can we conclude that the central city resident is lucky and faces a price of
housing lower than in the suburbs? More subtly, suppose we observe that these
residences are really very different. Are the different choices made by the two
households attributable to different incomes, different preferences, nonmarket
constraints on choice, or differences in the effective prices of the attributes that
characterize each house? To answer such questions, we have no choice but to
undertake hedonic analysis of the market.
The important work of Griliches (1961, 1971) did much to introduce hedonic
analysis and techniques for dealing with commodity heterogeneity to a wider
audience of economists. Griliches and many others have rightly referred to the
work of Court (1939) as an early pioneer in the application of these techniques, as
well as the first to apply the term "hedonic" to analysis of prices and demand for
the individual sources of pleasure-the attributes which combine to characterize
heterogeneous commodities. A decade before the appearance of Court's work,
however, the study of Waugh (1929) appears to be the first to provide a systematic
analysis of the impact of "quality" on the price of a commodity. Waugh charac-
terizes quality using a variety of observable attributes, and estimates the implicit
price of each of these attributes. Not only is the study the first to estimate what we
now call a "hedonic price function", it is an important early application of multi-
variate statistical techniques to economics. It nicely illustrates the way in which
hedonic analysis was, and is, an important source of econometric innovation for
the profession.
This chapter reviews several aspects of the hedonic analysis of housing mar-
kets, focusing broadly on the theoretical foundation of such analysis followed by
the practical implementation. Within each of these areas, we consider the issues
which affect the determination and estimation of implicit prices for housing at-
tributes, as well as the difficult problems which confront the use of these prices
to estimate the demand for characteristics.
1598 S. Sheppard
This reflects a basic dichotomy which is also present in the literature. There
is a large body of work that focuses more or less exclusively on inference of the
implicit prices of housing and environmental characteristics. Such information
is of considerable value in the construction of price indices which take proper
account of changes in the quality of the goods produced, and also in estimating
or forecasting values of real estate assets. A second branch of the literature uses
these implicit prices to estimate the structure of household demand for housing
attributes, and to evaluate the welfare consequences of changes in attribute prices
or environmental quality.
There have been several excellent surveys of all or important parts of this
literature. The papers by Mayo (1981), Follain and Jimenez (1985), Palmquist
(1991) and others provide particularly useful accounts of the progress made in
this area, and one might "review" this area of urban economics by simply refer-
ring the reader to these. The discussion below, in addition to highlighting some
of the more recent contributions in this area, provides an alternative view of some
of these developments.
2. Theoretical foundation
Before discussing the actual estimation and use of hedonic models of hous-
ing markets, we begin with a discussion of some of the theoretical constructs
that are the foundation of this approach. We begin with some preliminary re-
marks concerning "implicit markets" before turning to discussion of hedonic
price functions and the structure of demand for housing attributes.
The notion of implicit markets denotes the process of production, exchange, and
consumption of commodities that are primarily (perhaps exclusively) traded in
"bundles". The explicit market, with observed prices and transactions, is for the
bundles themselves. Such a market, however, might be thought of as constituting
several implicit markets for the components of the bundles themselves. This is
of particular importance when the bundles are not homogeneous, but vary due to
the varying amounts of different components that they contain.
There are at least two possible perspectives to take on implicit markets, that
differ more in their emphasis and orientation than in the final theory whose
development is thus motivated. On the one hand, we might regard the demand
for all goods-even those that appear to be homogeneous-as not being based
Ch. 41: Hedonic Analysis ofHousing Markets 1599
on the goods themselves but on the characteristics they embody. The household
purchases these goods and uses them as a type of 'input', transforming them into
utility, the level of which depends on the quantity of characteristics embodied
in the goods purchased. This approach, motivated at least in part by the work
of Lancaster (1966), tends to place particular emphasis on household produc-
tion and the properties of household demand for the (sometimes unobservable)
characteristics.
An alternative view emphasizes the idea that some goods are usefully com-
bined and thought of as being traded in a single "market", but they are quite
heterogeneous-automobiles, workers and houses are all reasonable examples.
Such markets are incapable of being analyzed with the usual economic mod-
els because they are not characterized or even approximated by a single price,
but rather by a range of prices that depend on the quality of the commodity
or the characteristics it contains. The hedonic approach attacks this difficulty
by asserting that these goods, while globally heterogeneous, are composed of
aggregates of (more or less) homogeneous parts, and while the aggregate bundle
may not have a common price, the component attributes do (or at least have
a common price structure). The hedonic approach provides a methodology for
identifying the structure of prices of the component attributes (estimation of the
hedonic price function). Analysis of demand can then proceed using these prices,
estimating a demand system in which the attributes are treated as goods. This
involves an implicit assumption that a variety of aggregate bundles are available
in the market so that consumers can choose any bundle of attributes they wish,
being constrained only by their incomes and the price of the resulting bundle.
If we let Z represent a vector of quantities of these attributes that differenti-
ate the commodities, the hedonic approach starts by recognizing that first, each
consumer may consume a different commodity, in the sense that each consumer
may consume a unit of the good with a different amount of Z embodied. Second,
each consumer may pay a different price for the good, so that a range of marginal
prices may exist that, in general, depend on the quantity of Z.
Hedonic price functions are estimated for two primary reasons: (1) for use in con-
struction of overall price indices that account for changes in the quality of goods
produced; and (2) as an input in the analysis of consumer demand for attributes of
heterogeneous goods. To understand the appropriate estimation techniques and
problems, and to interpret the results, we must begin with an understanding of
1600 S. Sheppard
how a market for heterogeneous goods can be expected to function, and what
types of equilibria we can expect to observe.
We begin with a recapitulation of the basic theory of hedonic markets. Con-
sumers are assumed to derive utility from consumption of a commodity that
embodies a vector Z of J different characteristics, plus consumption of a com-
posite good Y. They have fixed income M and face a price function P(Z) that
gives the price of the heterogeneous good (that we refer to as housing) as a
function of the embodied characteristics Z. The preferences of the household
are represented by the utility function
The derivative of the bid rent function apl/azi gives the rate at which the house-
hold would be willing to change expenditure on a house as characteristic i in-
creases, while holding the utility level constant. This is the inverse of the com-
pensated demand curve.
The household chooses a house with characteristics Z, and consumption of
composite commodity Y, to solve:
There are a variety of producers, and their distribution is described by the proba-
bility density g(y). Each of these producers is assumed to take the price function
P (Z) as given and solve,
Pi = CiVi (8a)
P(Z) = CN. (8b)
Thus, each producer equates the marginal cost of each characteristic to its hedo-
nic price, and builds houses until the marginal cost of building another house (of
1602 S. Sheppard
type Z) is equal to the value of the house P(Z). There will typically be a large
number of housing producers active in the market, with sellers of existing hous-
ing consisting of a special type of producer with N = 1 and a cost function C(.)
that is determined by the costs and technology of house repair and remodeling.
Equilibrium in this market for heterogeneous goods requires a hedonic price
function P(Z) that equates supply and demand for every type of house Z. This
equilibrium is conventionally represented as a locus of tangencies between a
series of marginal cost curves, Ci, and the derivatives fBi of the bid rent curves.
In general, as pointed out in Epple (1987) inter alia, the equilibrium price func-
tion depends on the distributions f (a, M) and g(y). This dependence has also
been discussed in the context of labor markets in Tinbergen (1959) and Sattinger
(1980).
This dependency is clear when one considers extreme examples. Suppose that
there are a variety of consumers but only one type of producer, so that g(y) is
degenerate. In this case, the equilibrium will have only a single Ci for each char-
acteristic, and all Bi will be tangent to it. An equilibrium hedonic price function
would be the cost function C. In general, then, the equilibrium hedonic price
function depends on the distribution of consumers and producers. It is possible
to endogenize the distribution g using a zero profit condition to generate "en-
try" and "exit" of house producers who behave as described above, increasing
or decreasing the density of each producer type y until a type of "long run"
equilibrium condition is met.
This provides a theoretically coherent foundation for explaining the relation-
ship between the price of a house (or other heterogeneous good) and the charac-
teristics it possesses. It describes the actions of market participants, and provides
equilibrium conditions in which these actions combine to determine the hedonic
price relationship. Furthermore, it does so in a way that underscores the potential
usefulness of observing this relationship. Apart from the difficulties in actually
carrying out the estimation required for this "observation", there are still some
gaps that have been addressed by researchers or provide opportunities for future
research.
The gaps that seem of greatest relevance in the present context relate to as-
pects that set the market for houses apart from the markets for other hetero-
geneous commodities, such as breakfast cereals or automobiles. Some of these
aspects are shared with a few other markets, but in any event are not incorporated
into the standard formulations of hedonic models of housing markets. Consider
the following features of housing markets:
* Housing markets involve search: It is costly to collect information concern-
ing the characteristics embodied in a particular structure, and as a result
Ch. 41: Hedonic Analysis of Housing Markets 1603
there is uncertainty about the exact nature of the hedonic price function. A
consumer purchases a house by examining a sequence of structures, even-
tually beginning the process of making offers on properties (possibly con-
tinuing to sample) and continuing until the expected increase in utility from
continued search is less than the cost. In this sense, the housing market is
similar to other "matching" type markets such as the labor market.
* Housing markets are intrinsically spatial:Houses involve varying quantities
of land, and possess particular locations. Many empirical studies do not even
include location as a characteristic of the house, let alone take advantage of
the special nature of this characteristic. The special nature arises when all
(or most) relevant local amenities have been accounted for, so that the only
factor that distinguishes one location from another is the transport costs.
In this case, the bid rent curve for locations with greater accessibility is
determined by the transport cost function.
* Housing markets involve both new and existing homes: Perhaps the most
significant difference between housing and other heterogeneous goods mar-
kets is the importance of the sale of existing (previously produced) goods.
The share of house sales accounted for by new construction is relatively
small in almost all housing markets. Consumers can substitute between
them, and seek out the type, whether new or old, that maximizes utility.
While this poses no problem for the consumer side of the market, how
should this factor be incorporated into the supply side of the market?
In addition to the observations made above, we can also question whether, or
not, the "hedonic approach" itself has any testable implications. Are there ob-
servations that could be made concerning the implicit price of housing attributes
that would refute some or all of the theory outlined above? One issue concerns
the convexity of the hedonic price function. This in turn translates into the con-
vexity of the consumer's budget set, that would, if violated, make problematic
the first-order characterization of consumer choice outlined in Eq. (2.4) above.
In a restrictive setting (all attributes and commodities perfectly divisible) Jones
(1988) shows that in equilibrium, hedonic price functions must be convex. He
also shows that in this setting they must be linearly homogeneous. In practice,
most estimated hedonic price functions satisfy convexity (see Anderson, 1985),
but a great many fail to satisfy the homogeneity restriction.
ties. Such understanding is essential for predicting the response to changes in the
housing market and for providing welfare estimates of the costs and benefits
associated with such changes. In this section, we explore some of the properties
that can be expected of such demand, and the way in which these properties
inform and constrain the application of hedonic analysis. In this context, we begin
with a consideration of the extent to which "demand functions" can be derived
within the hedonic framework, and which properties these relationships might be
expected to exhibit. We then discuss the way in which these relationships might
be applied to evaluate the welfare consequences of changes that affect the prices
of attributes or the levels of environmental amenities.
We could then proceed to estimate this in the form of a share equation with an
additive error term,
Pi · qi(P1, , P, M, ) . (2.10)
Wi = MM
+8 (2.10)
There are two important problems with this strategy: the first concerns the esti-
mation technique and lack of stochastic independence between the "variables"
Pi and the error term E. This issue has received a great deal of attention and is
discussed more completely in Section 3.2 below. The second problem concerns
the appropriateness of this approach when the budget set faced by the household
is potentially nonlinear due to the nonlinearities that may be present in P (Z).
1606 S. Sheppard
160 S Separ
M1
Mo
M1
Mo
"0
8
o
E
0
(..
Y = M - P(Z). (2.11)
where we assume that the hedonic price function is convex so that the house-
hold's budget set will be convex. As long as the hedonic price function is indepen-
dent of household income (which will in general be true if sellers of houses are
price-takers) the budget lines themselves are "quasilinear"-the slope of the bud-
get line is -P'(Z) and independent of consumption of the composite good. Thus,
the distance (A - C) = (B - D) = P(Z), and M - M = M - Mo = (C - D).
A household with income MI, facing hedonic price P(Z) and making choice
C exhibits the same behavior as a household with income M facing an attribute
price that is constant and equal to aPlaZ. As discussed in Section 2.2, the first-
order conditions for household choice imply a tangency between the indifference
curve, the true budget line MI, and the linearized budget line M at the point C.
Provided that preferences are strictly convex, demand will be a function, and the
linearized demand will be the same as the hedonic demand.'
I A discrepency clearly arises if preferences are not strictly convex. While a point such as E in Fig. I
would not be an element of hedonic demand, it could be an element along with point C in linearized demand
Ch. 41: Hedonic Analysis of Housing Markets 1607
Thus, if we want to estimate the structure of household preferences by esti-
mating household demand for attributes, we can take as data the hedonic prices
Pi = aP(Z)/aZi, and the actual household choices, but we must use income
level M M - P (Z) + EJ-I Pi Zi. Based on this data, the structure of demand
can be estimated.
Such demand is what Murray (1983) refers to as "mythical" demand, stressing
that the resulting estimated demands are those that would obtain if the consumer
faced linearized prices. While the demand functions might be "mythical", there
is nothing mythical about the household preferences that can be inferred from
the estimated demands. If a parametric form for demand functions is chosen
to facilitate integrability, then we can determine the parameters of a utility or
expenditure function directly from the estimated demand systems. For example,
if a generalized CES (as in Quigley (1982) or Follain and Jimenez (1983)) or
almost ideal demand system (as in Cheshire and Sheppard (1998) or Parsons
(1986)) is applied, then the parameters of the utility or expenditure function are
directly obtainable from estimates of the demand.
It is important to qualify the interpretation of the estimated demands and pref-
erences by recognizing that households subject to nonlinear budget constraints do
not exhibit the same behavior as households subject to linear constraints, even if
(locally) the linear and nonlinear budgets are equivalent. In this sense, the "linear
approximation" demands are not so much mythical as erroneous, with the error
depending on the nonlinearity of the hedonic prices and the preferences of the
household.
Conventional demands derived from linear budget constraints exhibit a variety
of standard properties that have proven useful in applied demand analysis. These
can be used either as the basis of tests of the model, or potentially as a source of
"nonsample" information to improve the quality of estimates when, for example,
sample data exhibit colinearity. Which of these properties are preserved in the
linearized demands?
The study by Turnbull (1994) derives some properties of overall housing de-
mand in a context that explicitly considers the endogenous location choices that
determine the spatial structure of housing consumption. This paper (along with
related studies such as Blackley and Follain (1987), DeSalvo (1985) or Turnbull
(1993)) establish clearly that demand for housing is affected in important ways
by endogenous location choice, and that such demand does not possess the same
properties as conventional demand. In particular, the "law of demand" may be
if the indifference curve had a flat portion that was colinear with the linearised budget constraint M1. While
of minor theoretical interest, this is of minimal relevance in applications since estimation of demand functions
already requires the assumption of strictly convex preferences.
1608 S. Sheppard
violated for housing even if housing is not an inferior good. Turnbull derives
sufficient conditions to rule this out, and shows that with convex equilibrium rent
functions this is less likely to occur near the urban periphery. The fact that it can
occur at all might make one despair at using these properties to test or inform
housing demand estimates. While the analyst might seek comfort in the fact that
the analysis focuses on location choice, the difficulty is much more general.
The problem arises because of the endogeneity of prices, and the potential for
a household to respond to a pattern of price increases by not only altering the
amount of space they consume, but also their location (and, hence, the price of
space and their effective income). In a nonspatial setting, an almost identical
problem arises, in which an increase in the general price of an attribute can
generate not only a change in the consumption of that attribute, but also a recon-
figuration of the entire dwelling and neighborhood. In general, this will change
the marginal price of the attribute under consideration, and the final adjustment
in consumption of the attribute may be quite unexpected.
The most complete treatment of this problem (surprisingly rarely cited in
the hedonic literature) has been provided by Blomquist (1989), who investigates
the general properties exhibited by demands that are obtained by maximizing a
strictly quasiconcave differentiable utility function subject to a general nonlinear
budget constraint M > P (Z, 0), where 0 is a vector of parameters of the hedonic
price function. 2 Blomquist assumes that P is strictly increasing in Z and differ-
entiable, and is such that a unique differentiable solution exists to the household
optimization problem. Clearly, it is sufficient for this if P is convex. Let Zi (M, 0,
a) represent the hedonic demand for housing attribute i, that is the ith component
of the solution to the problem,
max u(Z, Y, a) subject to M > P(Z, 0) + Y. (2.12)
z,Y
For a fixed (M*, 0*), we have well defined household choice given by Zi (M*,
0*, a), and can define the linearized demand ZL(M, p) as the ith component of
the solution to the problem,
max u(Z, Y, a) subject to M > p'Z + Y, (2.13)
,Y
where
aP(Zi(M*, 0*, a), 0*)
Pi = and M= pi Zi(M*, 0,a).
i=1
2 Blomquist (1989) does not refer to P as a "hedonic price" function. The focus of his paper is oriented
towards complex tax schedules or wage functions.
Ch. 41: Hedonic Analysis of Housing Markets 1609
Let S represent the Slutsky matrix of substitution terms derived from the lin-
earized demands ZL, and let H P represent the matrix of second derivatives of
the hedonic price function with typical element given by, 3
2
Hip '=a P(Zi(M*, 0*, a), a*)
HP: (2.14)
Blomquist (1989) shows that there is a well-defined link between the income
effects of the true hedonic demand and the income effect of the associated lin-
earized demand. The relation is given by,
az - aZL .
Z
= (I -S . H) (2.15)
aM aM
This elegant formulation illustrates how the curvature of the true budget con-
straint (captured via HP), and the curvature of preferences (captured via S),
interact to produce a complex relation between the estimated income effects of
the linearized demand and the income effects that can be expected of a utility
maximizing household. A similar expression allows evaluation of the impacts of
a change in the parameters 0 that determine the hedonic price function. Define
a vector HoP = (a 2P)/(aZiaOj), whose ith element gives the impact on the
hedonic price of attribute i of a change in parameter Oj. Taking
aP aZ L
A =S HP- A, (2.16)
az
= (I- S. HP)-A. (2.17)
aOj
Again, we see the possibility of expressing the actual change in household at-
tribute demand as a function of the structure of linearized demand (via S and the
"income effect" component of A), the curvature of the budget constraint, and the
way in which the parameter in question affects the budget constraint.
This analysis demonstrates that there is nothing problematic or incorrect in
estimating linearized demands using the derivatives of the hedonic price function
as prices and using the modified income M defined in Eq. (2.13). Not only is
3 Clearly this matrix comprises the first J rows and columns of the hessian matrix of the hedonic price
function.
1610 S. Sheppard
M1
C
C
C
C
E
C
C
utility to determine utility levels U1 and U0. We might then try to estimate the
welfare cost using this expenditure function, and the estimate would be E(Pz,
U 1) - E(Pz, U) = E - F in Fig. 2, where the hedonic price of the attribute Pz
is held constant. What relation does this estimate have to the correct magnitude
M 1 - M?
As indicated in Fig. 2, the welfare loss estimated from the conventional ex-
penditure function is usually greater than the true welfare cost, and the error
is indicated by . This error results from the convexity of the hedonic budget
constraint and the failure of the usual technique using traditional expenditure
functions to account for the endogeneity of price. This is related to the ob-
servation noted by Ohsfeldt and Smith (1990) that nonlinearity of the budget
generates greater substitution between housing attributes and nonhousing goods
(the composite commodity).
One circumstance in which the traditional approach gives an exact evaluation
is if there is "zero income effect" in demand for the attribute. In such an unlikely
case, the utility maximizing choice for a household facing the reduced income
level Mo would involve no change in the quantity of the attribute, and the change
in total expenditures would equal the equivalent change in income.
1612 S. Sheppard
The nonlinearity of budgets also has implications for forecasting the response
of house attribute demand to a change in household income. Again referring
to Fig. 2, the estimated demand structure would lead us to expect an increase
in attribute demand from Zo to Z1 to be generated by an increase in income of
E- F. In fact, demand will be somewhat more income elastic, with this change in
attribute demand being generated by the smaller increase in income of M 1 - Mo .
Clearly, the accuracy of approximation obtained using the conventional "linear
budget" constructs to analyze household choice in a hedonic framework depends
on the structure of preferences and the curvature of the hedonic price function.
Similar problems arise when we seek to evaluate the impact of a change in the
hedonic price function. Of course, there are a wide variety of possible changes,
but suppose we consider an "increase" in the hedonic price in which, at every
level of attribute consumption, the hedonic price increases so that the budget line
is "steeper". Such a situation is illustrated in Fig. 3, where the budget line pro-
duced by the hedonic price function is rotated downwards by the price increase.
The household responds by changing from optimal choice B to choice C. It is
reasonable to seek to measure the variation in income that is equivalent to this
price change.
Figure 3 indicates that if the hedonic attribute prices were kept constant but
income were reduced by M 1 - Mo, then the effect on welfare would be the same
as the increase in attribute price. Using a conventional measure of the welfare
effect based on the expenditure function would give E - F as the change in
income that is equivalent to the price increase. As indicated, this overestimates
the actual welfare effect by £. The accuracy of the approximation will depend on
the nonlinearity of the hedonic price function and the structure of preferences.
Thus, a basis exists for using estimated hedonic prices and modified income
levels to estimate the structure of demand. A variety of econometric difficulties
confront such estimation, and these are discussed in Section 3.2. To be most
useful, a parametric form could be chosen that is explicitly derived from an
underlying expenditure or utility function, but in this respect hedonic analysis
of house attribute demand is no different from other types of demand analysis.
Bartik (1988) provided a comprehensive guide to accurate measurement of
benefits from changes in house attributes or environmental amenities using he-
donic price models. The methods discussed implicitly provide a way to estimate
the error E in the linearized welfare measure noted above. Although difficult to
implement, it is possible to obtain an exact welfare measure. In practice, many
studies communicate the approximate welfare measure, but it is important to note
that this is at best an approximation.
Ch. 41: Hedonic Analysis of Housing Markets 1613
E
T
C
C
C
E
C
3. Implementation
data. It is perhaps surprising that one of the first applications of multiple re-
gression should be a problem that is in fact very difficult. This may account
for the continued attention that the problem draws from econometricians and
applied economists. Estimation of hedonic prices confronts the economist with
a rich sampling of the standard difficulties that arise in estimation using cross-
section data. These include choices of the proper parametric specification-both
of functional form and of variables to be included-coping with colinearity and
ill-conditioned data, potential heteroscedastic and nonnormal errors, regressors
subject to measurement error, and maximum likelihood estimation of relation-
ships that are nonlinear.
Confronted with these difficulties, a variety of approaches are possible. We
review here some approaches that appear to be most useful or promising, along
with the traditional ones. Choosing an estimation methodology cannot be done
independent of consideration of the particular data sources available and the ob-
jectives of the analysis. For example, the "best" approach to use will generally
depend on whether the estimated hedonic price function is to be used to infer
the implicit prices of attributes, or simply to forecast or appraise the value of
individual properties.
We first discuss the traditional parametric approaches in which a specific
functional form is chosen, and the parameters that define it are estimated. This is
followed by noting the alternative nonparametric or semiparametric approaches
that make inferences about the implicit prices of attributes without imposing a
priori a functional relation between the total price of the house and the quantities
of attributes which characterize it.
3.1.1. Parametricapproaches
From the early work of Waugh (1929), Court (1939) and Griliches (1961) to
the most recent studies, the standard approach to estimation of the hedonic price
function has been to adopt some form of parametric approach. This means that
the analyst must choose a functional form whose actual values are determined by
a finite number of parameters. The estimation then proceeds by selecting those
parameter values resulting in a hedonic price function that gives the "best fit" to
the data.
This procedure has evolved in several ways, as allowed by the constraints
of computation technology, the availability of data, and the understanding of
the nature of the statistical problems involved. Studies such as those of Ridker
and Henning (1967), Kain and Quigley (1970), Harrison and Rubinfeld (1978),
Goodman (1978) and Linneman (1981), tended to rely on linear or logarithmic
parametric forms that performed reasonably well and were computationally fea-
Ch. 41: Hedonic Analysis of Housing Markets 1615
sible to estimate. In the 1980s, beginning with the work of Linneman (1980),
hedonic studies began to use more flexible functional forms obtained by apply-
ing the Box-Cox transformation, either to housing prices or nondichotomous
attribute quantities. This transformation uses a single parameter X to transform
the variable x as follows:
by how well they fit the data, but by how accurately they estimated true marginal
bids. When all valuable attributes were observed, and without measurement error,
the model that was linear in Box-Cox variables performed the best, while the
quadratic Box-Cox performed the worst.
When some important attributes were omitted or observed imprecisely, the
simpler forms such as linear Box-Cox and logarithmic forms were the most
accurate. They attribute this, at least in part, to the fact that in quadratic forms,
each hedonic price depends on more coefficients than in the more nearly linear
cases.
Related to the issue of the parametric form of the hedonic price function are
issues of which variables to include and the way in which they should enter.
Consider, for example, the role of lot size or land area. Examination of any
standard presentation of urban economic theory reveals the central role played by
models of land value determination. One might reasonably assert that a central
principal of urban economics is that the price of land will vary with location, and
this varying land price is what produces the varying types and intensity of land
use in cities.
Given this observation, it is surprising how many hedonic models lack either
a variable for land area, or a variable that explicitly identifies the location of the
structure. 5 Some studies, such as the interesting analysis of Laakso (1997), in-
clude location explicitly and land implicitly via a separate variable that measures
the density of structures. Very few models explicitly incorporate a land value
function that depends on location. Two exceptions are Jackson et al. (1984) and
Cheshire and Sheppard (1995). Both of these studies are able to obtain reasonable
estimates of the value of land in residential use, and to reject the hypothesis that
this value is constant over locations. Recent research by Colwell and Munneke
(1997), who studied prices of vacant land, also show that land prices are most
definitely not constant over locations.
While it is commonplace to justify adoption of a flexible form for estimation
of a hedonic price function by observing that "theory places few restrictions"
on the form, there are at least some restrictions. As observed above, urban eco-
nomic theory suggests that the form should include land values that depend on
location. Furthermore, Jones (1988) show that in equilibrium the hedonic price
functions should be convex. This guarantees convexity of consumer budget sets
and, with convexity of preferences, establishes continuity properties of house-
hold attribute demand. Although Anderson (1985) develops and applies a test
of hedonic function convexity, it is again surprising how few subsequent studies
have applied this simple test. At this point the study of Colwell and Munneke
(1997) is worth mentioning again because in examination of prices of vacant
land they obtain concavity of the price function for land. Whether, or not, this
holds more generally for land as an attribute in housing markets has not been
widely investigated.
A third way in which theory might suggest appropriate restrictions on the
parametric structure of the hedonic price function estimated concerns public
goods or environmental amenities. Parsons (1990) argues for weighting local
public goods and neighborhood amenities by lot size, based on land market equi-
librium and the fact that consumption of these amenities is limited by available
land for residential consumption. While his argument depends on an assumption
that consumption of these attributes is nonexclusive (and therefore may not be
appropriate for some local attributes such as schools or parks), for a great many
environmental amenities his analysis suggests a reasonable theoretical constraint
on the parametric specification for hedonic models.
Specification. Many of the issues raised above might be viewed under the general
heading of "model specification" for the hedonic price function. Over the past
decade, this topic has generated considerable interest amongst econometricians.
Several studies have considered techniques for testing or evaluating alternative
specifications for the hedonic price function. Among these are Butler (1982),
Milon et al. (1984), Dubin and Sung (1990), Burgess and Harmon (1991) and
Craig et al. (1991).
This last paper is interesting because it introduces a new test that is par-
ticularly appropriate in the setting of hedonic estimation. They apply the test
to housing market data to test for the systematic nonlinearities in the relation-
ship between structure price and attribute quantity. In addition to testing for the
appropriateness of nonlinearities in the functional form, the test can also pro-
vide an indication of important omitted variables. Interestingly (and a propos
the comments above about the need to include land area that varies with loca-
tion), their analysis identified an inability to "discover a specification for lot size
which ... passes" the test they derive. Whether a specification that did not impose
spatially invariant land value would pass was not tested.
Colinearity and error structure Two econometric problems that seem intrinsic
to estimation of hedonic price functions are colinearity or ill-conditioned data
and lack of stochastic independence between observations. The first problem has
received considerable attention in the literature. Only recently has the second
begun to be addressed.
1618 S. Sheppard
account for the spatial autocorrelation are seen to be warranted, and the results
suggest that models that account for these types of spatial structure are likely to
produce more reliable results.
3.1.2. Non-parametricapproaches
An alternative to specification of a parametric functional form for the hedonic
price function is to adopt a non- or semiparametric approach to estimation, that
attempts to infer attribute prices directly from the data without benefit of an as-
sumed functional relationship. The difficulty that arises in the application of these
techniques is the extremely large amounts of data required (formally, the slow
rate at which these estimation techniques converge to the true value as sample
size increases).
Feasible estimates can be obtained, however, by using approaches that con-
sider truly nonparametric combinations of simple parametric forms. Papers that
have investigated this approach include Knight et al. (1993), Pace (1993, 1995)
and Anglin and Gencay (1995). These approaches turn out to be much more
robust to specification and measurement error than many parametric estimates,
although they can be considerably more computationally complex to
estimate.
The analysis of Pace (1995) suggests in particular that semiparametric esti-
mates of hedonic price functions seem to suffer fewer incorrectly signed
parameters-a common problem with colinear data. Even more interesting, the
study of Anglin and Gencay (1995) compares a parametric estimator with a semi-
parametric one. The parametric model was easily rejected in tests that compare it
to the semiparametric. Furthermore, this was not for a poorly specified parametric
model, but one that passed a variety of standard tests of model specification.
This suggests that considerable gains in accuracy may be available by utilizing
semiparametric techniques for estimating hedonic price functions.
Estimation of the demand for any heterogeneous good such as housing presents
a variety of difficult problems, some of which are rarely encountered in other
economic contexts. It is perhaps a peculiar attribute of the literature concerning
estimation of demand for housing attributes that, during the past decade, much
of it has been obsessed with clarifying the problem of endogeneity of price.
Partly, this stems from a peculiarity in the development of the literature, in
which the problem of estimating the demand for housing attributes was originally
1620 S. Sheppard
This perspective (that at least implicitly lies at the foundation of a very large
literature) holds that since we have observations of individual behavior, along
with the prices and incomes that have produced that behavior, we can estimate
the structural equations of household demand as long as we are willing to assume
that a common preference ordering determines the decisions of all households.
All that is required are data in which there is linearly independent variation in
income and hedonic prices, and the estimates will be readily obtained.
The first objection to this perspective is the simple one that the hedonic prices
are not known with certainty, but must be estimated. The stochastic nature of the
relation between attribute quantities and house price will impart some uncertainty
to the estimated hedonic prices. To this, one might reply that demand estimation
is a classic "errors in variables" problem, and that as long as the hedonic price
function has been correctly specified, the parameters will be consistently esti-
mated and for large samples we need not worry. Of course, for finite samples
any hypothesis testing we might wish to do with our demand estimates will be
troubled by variance estimates that are biased downwards, but in principle, these
could be corrected by using the covariance matrix for parameters of the hedonic
price function to obtain an estimate of the covariance structure for the hedonic
prices. This approach, however, cannot by itself compensate for violations of the
Gauss-Markov conditions, and produces consistent estimates only if the errors
in hedonic prices (and "linearized income") are independent of the error terms in
the demand functions. 'This is precisely the core of the problem.
The problem is easily seen in Fig. 5. In the upper portion of the figure, a
hedonic price function P(Z) is shown with a household. bid function for the
attribute Z that is tangent at point B so that the household would demand Z1
units of the attribute given this price structure. 7 The object of hedonic analysis is
to obtain an estimate of the actual demand or marginal benefit function drawn as
a solid line in the lower portion of Fig. 5. For example, we might assume additive
error e and estimate
Z = o+ i Pz + c, (3.2)
using the observed household consumption and the estimated hedonic price func-
tion to determine the hedonic prices Pz. In the case illustrated, the parameter P1
is negative, and we try to obtain an estimate ,BI. When the error £ is near zero, we
observe the household consuming Z1 and infer from the hedonic price function
that the price which produces this behavior is the slope of P (Z) at point B.
7 Alternatively, the hedonic price function could be used to construct a budget line, and the bid function
would be an indifference curve tangent to the budget line at B.
Ch. 41: Hedonic Analysis of Housing Markets 1623
7
$
7
Zo Z1 Z2
Fig. 5. Biased estimates from price endogeneity.
For strictly convex hedonic price functions, the correlation p is positive and we
obtain inconsistent demand estimates whose slopes are biased upwards. For more
complex (and realistic) demands with several attributes, and linearized income
which is also correlated with E, the formula for the bias is more complex but the
basic intuition remains unchanged: the correlation between estimated hedonic
prices and errors in measured demand behavior leads to inconsistent estimates
of the structure of demand.
There are a variety of possible responses to this problem, but most of the
recent literature on estimation of housing demand begins with this basic obser-
vation. Thus, the analyses of Blomquist and Worley (1982), Brown and Rosen
(1982), Murray (1983), Diamond and Smith (1985), Ohsfeldt and Smith (1985),
Bartik (1987b), Epple (1987) and McConnell and Phipps (1987) all share a com-
mon initial theme: with individual household data, there is no structural simul-
taneity; the difficulty arises because the endogeneity of prices gives rise to cor-
relation between the random error in the model and the "independent variables".
This results in inconsistent estimates.
The required response to such a problem is to devise consistent, "instrumental
variables" estimates of those variables which appear on the righthand side of the
structural equations. Thus, for example, in Eq. (3.2) we need to identify variables
that are uncorrelated with s which can be used to provide a consistent estimate of
Pz. This estimate is then used in subsequent stages of the procedure to estimate
attribute demand. More generally, we identify instruments that permit estimation
of all of the hedonic attribute prices and linearized income.
Three basic approaches exist to obtain such instruments. First, we may be able
to find or construct other variables that are independent of the errors, but suffi-
ciently correlated with hedonic prices to provide admissible instruments. Second,
we might take advantage of nonlinearities that exist in actual hedonic price and
marginal benefit relations to identify the models (essentially using transforma-
tions of the variables as instruments). Third, we might use other variables that
occur in the structural equations to obtain a set of instruments for consistently
estimating hedonic prices.
The third method is familiar from its use in solving the identification prob-
lem and estimating models in which there is true simultaneity. Its application
is stressed as a potential solution in several of the papers cited above, and is
illustrated in the righthand portion of Fig. 6. In the upper righthand quadrant
of the figure, we see two budget curves associated with hedonic price functions
from two different markets. 8 In one market, the observed household has income
8 We know these are two different markets because the budget lines are not vertically parallel.
Ch. 41: Hedonic Analysis of Housing Markets 1625
Y Y
M1 M1
Mo
Mo
can choose among several urban areas to live, the urban area itself may well be
endogenous and not serve as a valid instrument.
From a practical perspective the difficulties are even larger. There are almost
no individual level datasets that are comparable across a large number of urban
areas and available to the public. There are none9 in Europe or Asia. In the US,
there are two potential data sources: the American Housing Survey and property
market data available from local tax authorities or through real estate listing ser-
vices. The first of these is available at very low cost and includes a large number
of variables for about 30 metropolitan areas. The difficulty is that it does not
include location. This makes it impossible to estimate a "land value" component
in the hedonic price function, although there are some accessibility variables
that provide information on the journey to work and mode of transport used.
The second data source contains much better information concerning location
(since property address is included), but typically contains no detail concerning
the occupants. In particular, it does not contain household income, nor does it
include any data on the age or structure of the occupants. It might be possible
(given sufficient resources) to survey a sample of properties to determine values
for these other important variables.
Most of the "multiple market" studies done have used the American Hous-
ing Survey. ° These studies were severely constrained by the nature of the data
available. None of these included location, despite the theoretical importance of
location in determining the value of land and the relatively large proportion of
residential construction costs accounted for by land. None included more than
eight attributes, with such obviously important factors as local school quality
or accessibility to parks and open space never appearing in any study. These
limitations arise because of the limited data available, and might be corrected in
the future. At present, however, it is rather difficult to understand the attraction of
seeking to correct one sort of bias (arising from endogenous prices) by introduc-
ing bias of another sort (severe specification error by failing to include important
attributes). This is particularly true when other approaches exist that might solve
the problem.
The second approach mentioned above relies on imposition of mostly
untestable restrictions of nonlinearity on both the hedonic price function and the
structure of demand. This approach is illustrated in the lefthand side of Fig. 6.
In the top left quadrant, we see two budget sets drawn from the same hedonic
market but generated by the two income levels Mo and Mi. At the lower income
tribute is not included in the hedonic price function, it will be impossible to create
instruments that are not correlated with the error. The analysis of Bartik (1987a)
makes this argument most forcefully, arguing that unless unobserved variability
in preferences is assumed not to exist, it will be impossible to construct valid in-
struments for consistent estimation of household attribute demand. His argument
proceeds by examples, however, and is restricted implicitly to an assumption
that the source of instruments are other variables that appear in the structural
equations of the model.
Are such variables the only source of instruments to which we might turn?
Surely that depends on the stochastic structure of the application, but in principle
one need not restrict attention to only these variables. It is helpful to keep in mind
that our problem is not one of a truly simultaneous equation system. Consistent
estimation of the hedonic demand for J attributes requires that we obtain con-
sistent estimates of the J hedonic prices (and hence also linearized income) that
define the budget set of the household. For this we need J + 1 or more instruments
that are uncorrelated with the error in household attribute demand, but which are
not so weak as to give extremely imprecise (even if consistent) estimates of the
actual hedonic prices.
Murray (1983) makes a variety of interesting suggestions concerning possible
instruments, and recently Cheshire and Sheppard (1998), have pursued the idea
of using the average attribute prices paid by "similar" households as instruments.
There are numerous dimensions along which one might define similarity, and
they consider two: (1) taking those households that occupy the locations that
are nearest to the household, and taking those households that have chosen to
consume similar houses (including both the attributes and location). For the lat-
ter concept of similarity, they construct an index that measures the Euclidean
distance in characteristics space weighted by consumption shares. For both geo-
graphic distance and characteristics space distance, they experimented using as
instruments the prices paid by the two "nearest" households in their sample.
Whether, or not, such an approach is likely to provide a valid set of in-
struments is, as noted above, dependent on the stochastic specification of the
model. l If these errors come from simple measurement error on the quantities of
attributes, then consideration of prices paid by similar households might provide
good instruments as long as the measurement errors were uncorrelated between
observations. Whether, or not, this is true will naturally depend on the attribute
and the nature of the market, but fortunately it is relatively straightforward to
Zi = Bo + k Pk +' M M M+ E. (3.4)
k=1
I T
plimrT I'V Et = 0, (3.5)
t=l
where the st error terms have common variance or,. A test statistic for this hy-
pothesis is
T = 7 ( )- Wf (3.6)
where ' is the vector of residuals from the two-stage least square estimate ob-
tained using JI,as instruments. A remarkable feature of this statistic is that it is
equal to the product of the sample size T and R2 ,the coefficient of determination
obtained from regressing the two-stage least squares residuals e on the set of
instruments AI. The statistic ~T is distributed X 2 with K - J degrees of freedom
under the hypothesis of admissibility.
This is completely intuitive: we have as an estimate of the (unobserved) error
the residuals . Validity of the instruments P requires that they be independent
of these errors. If the variables 'l are able to explain the variation in , then it
seems unlikely that WIand e will be independent.
Using this test, Cheshire and Sheppard (1998) found that the combined char-
acteristics distance instruments are admissible for use in estimating the hedonic
demand for land area and open space amenities. They found that the geographic
distance instruments are admissible for all of the characteristics for which they
had data. The geographic distance based instruments were somewhat weaker,
although both provided demand system estimates with acceptable (in sample)
accuracy.
1630 S. Sheppard
3.2.2. Specification
In estimating household hedonic demand or marginal bid functions, there are
three basic approaches that have appeared in the literature. First, and probably
least reliable, is to simply estimate some reasonably flexible nonlinear demand
that depends on (linearized) household income and attribute hedonic prices. The
difficulty with this is the lack of connection between the estimated demand and a
preference ordering which presumably generates it.12
A preferable approach is to estimate a demand structure derived from an ex-
plicit utility or expenditure function. For example, the studies by Quigley (1982)
and Kanemoto and Nakamura (1986) undertake this as a possible solution to
the price endogeneity problem discussed above. Without regard to whether, or
not, their technique is the best solution to price endogeneity, they do certainly
obtain estimable demand functions derived from an explicit preference ordering.
An alternative is to use a demand system derived from a flexible expenditure
function. The "almost ideal demand system" presented in Deaton and Muellbauer
(1980) is used in Parsons (1986) and Cheshire and Sheppard (1998).
12 As noted, for example, by McConnell and Phipps (1987) in discussing the example used by Brown and
Rosen (1982).
Ch. 41: Hedonic Analysis of Housing Markets 1631
While some techniques have been proposed for estimating compensating and
equivalent income variations directly from hedonic price functions (see Horowitz
(1984)), general welfare analysis will typically require more detailed knowl-
edge of the household expenditure functions. This is most directly obtained by
estimation of a completely specified demand system.
4. Conclusion
that they correspond to the restrictions implicit in the theory of urban housing
markets. Estimation must confront inadequate data, and make use of whatever
information sources are available.
Use of these hedonic prices to estimate the structure of demand brings more
difficulties, many of which have been only poorly understood. In addition to the
usual problems of model specification and measurement error, the nonlinearity
in household budgets implies endogenous determination of attribute prices. A
variety of strategies for meeting the challenge of endogeneity have been pre-
sented, ranging from use of multiple market data to construction of alternative
"nonstructural" instrumental variables for prices and income. Whatever approach
is adopted, it is clear that accurate estimation requires some explicit acknowl-
edgment of the endogeneity, although the actual quantitative significance of the
problem may in some cases be modest.
References
Anderson, J.E. (1985), "On testing the convexity of hedonic price functions", Journal of Urban Economics
18:334-37.
Anglin, P.M. and R. Gencay (1995), "Semiparametric estimation of a hedonic price function", Discussion
paper, University of Windsor, University of Windsor Working Paper.
Bartik, T.J. (1987a), "Estimating hedonic demand parameters with single market data: the problems caused by
unobserved tastes", Review of Economics and Statistics 69:178-80.
Bartik, TJ. (1987b), "The estimation of demand parameters in hedonic price models", Journal of Political
Economy 95:81-88.
Bartik, T.J. (1988), "Measuring the benefits of amenity improvements in hedonic price models", Land
Economics 64.:72-83.
Belsley, D.A., K.E. Belsley and R.E. Welsch (1980), Regression Diagnostics, Identifying Influential Data and
Sources of Collinearity (Wiley, New York).
Blackley, D.M. and J.R. Follain (1987), "Tests of locational equilibrium in the standard urban model", Land
Economics 63:46-61.
Blackley, P., J. Follain and J. Ondrich (1984), "Box-Cox estimation of hedonic models: how serious is the
iterative OLS variance bias?", Review of Economics and Statistics 66:348-353.
Blomquist, G. and L. Worley (1982), "Specifying the demand for housing characteristics: the exogeneity
issue", in: D.B. Diamond and G.S. Tolley, eds., The Economics of Urban Amenities, Chap. 4, pp. 89-102
(Academic Press, New York).
Blomquist, N.S. (1989), "Comparative statics for utility maximization models with nonlinear budget con-
straints", International Economic Review 30:275 296.
Brown, J.N. and H.S. Rosen (1982), "On the estimation of structural hedonic price models", Econometrica
50:765-768.
Burgess, J.F. and O.R. Harmon (1991), "Specification tests in hedonic models", Journal of Real Estate Finance
and Economics 4:375-393.
Butler, R.V. (1982), "The specification of hedonic indexes for urban housing", Land Economics 58:96-108.
Can, A. (1992), "Specification and estimation of hedonic housing price models", Regional Science and Urban
Economics 22:453-474.
Cassel, E. and R. Mendelsohn (1985), "The choice of functional forms for hedonic price equations: comment",
Journal of Urban Economics, 18:135-142.
Ch. 41: Hedonic Analysis of Housing Markets 1633
Cheshire, P. and S. Sheppard (1995), "On the price of land and the value of amenities", Economica 62:247-267.
Cheshire, P. and S. Sheppard (1998), "Estimating the demand for housing, land, and neighbourhood
characteristics", Oxford Bulletin of Economics and Statistics 60:357-382.
Colwell, P.RE and H.J. Munneke (1997), "The structure of urban land prices", Journal of Urban Economics
41:321-336.
Court, A.T. (1939), "Hedonic price indexes with automotive examples", in: The Dynamics of Automobile
Demand (General Motors, New York).
Craig, S.G., J.E. Kohlhase and D.H. Papell (1991), "Chaos theory and microeconomics: an application to
model specification and hedonic estimation", Review of Economics and Statistics 73:208-215.
Cropper, M.L., L.B. Deck and K.E. McConnell (1988), "On the choice of functional form for hedonic price
functions", Review of Economics and Statistics 70:668-675.
Cropper, M.L., L. Deck, N. Kishor and K.E. McConnell (1993), "Valuing product attributes using single market
data: a comparison of hedonic and discrete choice approaches", Review of Economics and Statistics 75:225-
232.
Deaton, A. and J. Muellbauer (1980), "An almost ideal demand system", American Economic Review 70:312-
326.
DeSalvo, J.S. (1985), "A model of urban household behavior with leisure choice", Journal of Regional Science
25:159-174.
Diamond, D.B. Jr. and B.A. Smith (1985), "Simultaneity in the market for housing characteristics", Journal of
Urban Economics 17:280-292.
Dubin, R.A. (1988), "Estimation of regression coefficients in the presence of spatially autocorrelated error
terms", Review of Economics and Statistics 70:466-474.
Dubin, R.A. (1992), "Spatial autocorrelation and neighborhood quality", Regional Science and Urban
Economics 22:433-452.
Dubin, R.A. and C.H. Sung (1990), "Specification of hedonic regressions: non-nested tests on measures of
neighborhood quality", Journal of Urban Economics 27:97-110.
Ellickson, B. (1981), "An alternative test of the hedonic theory of housing markets", Journal of Urban
Economics 9:56-79.
Epple, D. (1987), "Hedonic prices and implicit markets: estimating demand and supply functions for
differentiated products", Journal of Political Economy 95:59-80.
Feenstra, R.C. (1995), "Exact hedonic price indexes", Review of Economics and Statistics 77:634-653.
Follain, J. and E. Jimenez (1983), "The demand for housing characteristics in developing countries",
Discussion Paper WUDD 43 (The World Bank Water Supply and Urban Development Department).
Follain, J.R. and E. Jimenez (1985), "Estimating the demand for housing characteristics: a survey and critique",
Regional Science and Urban Economics 15:77-107.
Freeman, A.M. (1979), "Hedonic prices, property values and measuring environmental benefits: a survey of
the issues", Scandinavian Journal of Economics 81:154-173.
Gilley, O.W. and R.K. Pace (1995), "Improving hedonic estimation with an inequality restricted estimator",
The Review of Economics and Statistics 77:609-621.
Goodman, A.C. (1978), "Hedonic prices, price indices and housing markets", Journal of Urban Economics
5:471-484.
Gourieroux, C. and A. Monfort (1995), Statistics and Econometric Models, Vol. 2 (Cambridge University
Press, Cambridge).
Griliches, Z. (1961), "Hedonic price indexes for automobiles: an econometric analysis of quality change", The
Price Statistics of the Federal Government, No. 73.
Griliches, Z. (1971), Price Indexes and Quality Change: Studies in New Methods of Measurement (Harvard
University Press, Cambridge, MA).
Halvorsen, R. and H.O. Pollakowski (1981), "Choice of functional form for hedonic price equations", Journal
of Urban Economics 10:37-49.
Harrison, D.J. and D.L. Rubinfeld (1978), "Hedonic housing prices and the demand for clean air", Journal of
Environmental Economics and Management 5:81-102.
Horowitz, J.L. (1984), "Estimating compensating and equivalent income variations from hedonic price
models", Economics Letters 14:303-308.
1634 S. Sheppard
Horowitz, J.L. (1987), Identification and stochastic specification in Rosen's hedonic price model", Journal of
Urban Economics 22:165-173.
Jackson, J.R., R.C. Johnson and D.L. Kaserman (1984), "The measurement of land prices and the elasticity of
substitution in housing production", Journal of Urban Economics 16:1-12.
Jones, L.E. (1988), "The characteristics model, hedonic prices, and the clientele effect", Journal of Political
Economy 96:551-567.
Kain, J.F. and J.M. Quigley (1970), "Measuring the value of housing quality", Journal of the American
Statistical Association 65:532-548.
Kanemoto, Y. and R. Nakamura (1986), "A new approach to the estimation of structural equations in hedonic
models", Journal of Urban Economics 19:218-233.
Knight, J.R., R.C. Hill and C.F. Sinnans (1993), "Estimation of hedonic housing price models using nonsample
information: a Monte Carlo study", Journal of Urban Economics 34:319-346.
Laakso, S. (1997), Urban Housing Prices and the Demand for Housing Characteristics (ETLA, The Research
Institute of the Finnish Economy, Helsinki).
Lancaster, K. (1966), "A new approach to consumer theory", Journal of Political Economy 74:132-157.
Lerman, S.R. and C.R. Kern (1983), "Hedonic theory, bid rents, and willingness-to-pay: some extensions of
Ellickson's results", Journal of Urban Economics 13:358-363.
Linneman, P. (1980), "Some empirical results on the nature of the hedonic price function for the urban housing
market", Journal of Urban Economics 8:47-68.
Linneman, P. (1981), "The demand for residence site characteristics", Journal of Urban Economics 9:129-148.
Mayo, S.K. (1981), "Theory and estimation in the economics of housing demand", Journal of Urban Economics
10:95-116.
McConnell, K.E. and T.T. Phipps (1987), "Identification of preference parameters in hedonic models: consumer
demands with nonlinear budgets", Journal of Urban Economics 22(1):35-52.
McFadden, D. (1977), "Modeling the choice of residential location", Discussion Paper 477 (Cowles Founda-
tion).
Meese, R.A, and N.E. Wallace (1997), "The construction of residential housing price indices: a comparison
of repeat-sales, hedonic-regression and hybrid approaches", Journal of Real Estate Finance and Economics
14:51-73.
Milon, J.W., J. Gressel and D. Mulkey (1984), "Hedonic amenity valuation and functional form specification",
Land Economics 60:378-387.
Murray, M.P. (1983), "Mythical demands and mythical supplies for proper estimation of Rosen's hedonic price
model", Journal of Urban Economics 14:326-337.
Ohsfeldt, R.L. and B.A. Smith (1985), "Estimating the demand for heterogeneous goods", Review of
Economics and Statistics 61:165-171.
Ohsfeldt, R.L. and B.A. Smith (1990), "Calculating elasticities from structural parameters in implicit markets",
Journal of Urban Economics 27:212-221.
Pace, R.K. (1993), "Nonparametric methods with applications to hedonic models", Journal of Real Estate
Finance and Economics 7:185-204.
Pace, R.K. (1995), "Parametric, semiparametric, and nonparametric estimation of characteristic values within
mass assessment and hedonic pricing models", Journal of Real Estate Finance and Economics 11:195-217.
Palmquist, R.B. (1991), "Hedonic methods", in: J.B. Braden and C.D. Kolstad, eds., Measuring the Demand
for Environmental Quality, pp.77-120 (Elsevier Science, Amsterdam).
Parsons, G.R. (1986), "An almost ideal demand system for housing attributes", Southern Economic Journal
53:347-363.
Parsons, G.R. (1990), "Hedonic prices and public goods: an argument for weighting locational attributes in
hedonic regressions by lot size", Journal of Urban Economics 27:308-321.
Quigley, J.M. (1982), "Nonlinear budget constraints and consumer demand: an application to public programs
for residential housing", Journal of Urban Economics 12:177-201.
Ridker, R.G. and J.A. Henning (1967), "The determinants of residential property values with special reference
to air pollution", Review of Economics and Statistics 49:246-257.
Rosen, S. (1974), "Hedonic prices and implicit markets: product differentiation in pure competition", Journal
of Political Economy 82:34-55.
Ch. 41: Hedonic Analysis of Housing Markets 1635
Sattinger, M. (1980), Capital and the Distribution of Labor Earnings (North-Holland, Amsterdam).
Spitzer, J.J. (1982), "A primer on Box-Cox estimation", Review of Economics and Statistics 64:307-313.
Tinbergen, J. (1959), "On the theory of income distribution", in: L.M.K.L.H. Klaasen and H.J. Witteveen, eds.,
Selected Papers of Jan Tinbergen (North-Holland, Amsterdam).
Turnbull, G.K. (1993), "The substitution theorem in urban consumer theory", Journal of Urban Economics
33:331-343.
Turnbull, G.K. (1994), "Housing demand properties in the monocentric market form", Regional Science and
Urban Economics 24:253-263.
Waugh, EV. (1929), Quality as a Determinant of Vegetable Prices (Columbia University Press, New York).
Chapter42
ALAN W. EVANS*
Contents
1. Introduction 1638
2. Zoning 1641
3. Conservation areas and historic districts 1647
4. Growth controls 1649
5. Planning gain and impact fees 1653
6. Growth controls matter 1655
7. The supply of land 1657
8. Conclusions 1665
References 1667
* I am indebted to the Leverhulme Trust for financial support in the completion of this survey.
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P. Cheshire
© 1999 Elsevier Science B. V All rights reserved.
1637
1638 A. W Evans
Abstract
Two kinds of government intervention in the land market are considered. The
first is the control of development, of which the most studied form is zoning, but
we also consider the designation of conservation areas and the effects of growth
controls. Growth control may raise the price of land, allowing an infrastructure
charge to be made or growth may be limited by charging impact fees. The second
kind of intervention aims to increase the supply of land, either by direct action, as
in the Netherlands, through compulsory purchase or eminent domain, or through
the reallocation of land ownership as in land readjustment schemes.
1. Introduction
point of view, it is obvious (even though Abercrombie does not provide the in-
formation) that the constraint on expansion had been followed by a continuing
excess demand for housing and hence rising prices, because the price of housing
in the area had risen enough to make the cost of excavation worth bearing and
the risk of demolition worth taking.
This question of the risk of incurring a penalty if a regulation is disobeyed is
important. In the Elizabethan case, there was clearly a risk of discovery which
was significant, but not very high, since a temporary building was worth con-
structing and presumably had a reasonable, if short, anticipated life. If it had
been known that discovery would be immediate and certain, no such houses
would have been built. If the regulation was not enforced at all then a properly
constructed permanent house would have been built. So the same regulation may
have different effects at different times and under differing enforcement regimes.
Many of the studies to which we refer report on the research carried out in the
US within an approximately homogeneous legal system, but even here Fischel
(1990) suggested that some knowledge of the enforcement regime is important in
gaining an understanding of its economic effects. Within this system it has been
suggested by Wallace (1988) and McMillen and McDonald (1991, 1993), that
zoning follows the market, i.e., the pattern of land use that occurs after a system
of land use regulation and planning is imposed may be exactly what would have
occurred without such a system, only in a more ordered fashion, the land use
regulatory system therefore providing a kind of imprimatur. This will happen,
it is argued, because political pressure will tend to ensure that land values are
maximised, if only to maximise local government revenues from property taxes.
In other legal systems different principles may apply. The British system can
be placed at one extreme, where land use regulation is enforced (and enforce-
able) against the pressure of market forces, where the fiscal system gives no
encouragement to local authorities to maxinise revenue from property taxes,
and where deliberate flouting of the rules by would-be developers is likely to
result in financial penalties and demolition. Near the other extreme is the Italian
system, where land use controls exist but may often be ignored or flouted, and
where amnesties in 1985 and 1994 allowed illegally constructed buildings to be
legalised on the payment of a fine (Ave, 1996: p. 49f). Since the fines were not set
too high to ensure that people paid them, the result was that controls appeared to
be, to say the least, highly flexible. It may be noted that environmental pressure
groups strongly opposed such amnesties on the grounds that each one that was
announced reduced the degree to which people were willing to abide by the ex-
isting physical controls. Moreover, the amnesty/fines system tends to transform a
system of physical controls into a fiscal system, since the result is a system where
Ch. 42: The Land Market and Government Intervention 1641
nonconforming uses are, in effect, allowed provided a payment is made, with a
consequent loss of the possibility of fine tuning that can occur with a physical
planning system (but with greater benefit to the Treasury).
2. Zoning
There have been two recent surveys of the empirical literature on the economic
effects of municipal zoning in the US. One, more recent and explicitly on this
subject is by Pogodzinski and Sass (1991), and the other, on the related topic
of growth controls is by Fischel (1990) and summarised in Fischel (1989). In
the circumstances, this topic does not need to be dealt with at length, since this
would mean covering the same ground as these authors, and their work can be
readily assessed by those wishing to study the topic in more depth. The aim of
these topics in this chapter is therefore to summarise their findings and, where
possible, draw attention to non-North American literature, since the two surveys
limit themselves to US and Canadian work.
In general, researchers have found it difficult to measure the extent of the
external diseconomies or externalities caused by nonconforming land uses, at
least in part, it has been argued, because the reallocation of land uses caused
by zoning would have been intended to reduce the extent and impact of these
externalities and because it is difficult to find an urban area where there is not
a zoning system in place. In this respect the position in Houston, Texas, is of
interest. In Houston, there are districts with zoning ordinances, districts with
systems of covenants, and districts without either. Speyrer (1989) found that
property values were lower in the latter, and that buyers were prepared to pay a
substantial premium for properties covered by either covenants or zoning. Thus,
zoning must achieve something, from the local property owners' point of view,
even though, as Speyrer points out, "premiums paid for land use restrictions do
not imply such restrictions may be optimal ... gains may be small compared with
the losses incurred by excluded groups" (Speyrer, 1989: p. 128). An alternative
interpretation of the position was put forward by McDonald (1995) after the
voters of Houston had narrowly voted in 1993 against a proposed zoning plan. In
general, low income areas voted against zoning while middle income areas voted
in favour. McDonald (1995) suggests that this was primarily because zoning was
perceived, by both groups, as a means of keeping out the (low income) "riffraff".
An alternative interpretation might be that the negative externalities minimised
by zoning are given less weight by low income families, i.e., the presence of a
1642 A. W. Evans
urban area affects the price of land and housing there. It is hypothesised that
the smaller the number of jurisdictions the more likely it is that they can, and
will, act monopolistically. Hamilton (1978) proposed and tested this hypothesis
using the 1960 Census data for census tracts in 13 urban areas in the northeast-
ern US. Fischel (1980) argued that Hamilton's measure of zoning power (MSA
jurisdictions/MSA population, or a dummy if number of jurisdictions less than
four) was misleading, first, because the cities with few jurisdictions were all in
New England, and second, because the number of jurisdictions in a Standard
Metropolitian Statistical Area was misleading since the data should relate to the
Urbanised Area (UA) so that the UA jurisdictions/UA population would be the
correct measure. Using these measures, and the 1970 Census data, Fischel found
the monopoly zoning hypothesis to be "not proven".
Two later papers discuss determinants of interurban house prices (Ozanne and
Thibodeau, 1983), and interurban land prices (Manning, 1988), and include mea-
sures of monopoly zoning power as independent variables. Data for the first paper
were from the 1974-1976 Annual Housing Survey and data for the second were
from the 1980 Census. In both these studies, the variable is highly significant but
both use MSA jurisdictions despite the strictures in Fischel's paper, to which nei-
ther refers. We do not know, therefore, whether the significance of the measure of
monopoly zoning in their later work is because they are using an incorrect mea-
sure, or using later data and the situation has changed over time, or their equations
are more carefully specified. Manning (1988), for example, includes a measure of
the high "real income tail" to take account of the way in which taxes and inflation
may interact to reduce the cost of housing for high income households, and this
was highly significant. But this variable was itself introduced by Gottlieb (1965),
only to be neglected until Manning's (1988) apparent rediscovery of it. A later
paper by Thorson (1996) goes some way towards settling the question in that he
takes account of the points made by Fischel (1980), and carefully specifies his
equations taking account of the specifications used by others such as Manning
(1988), and shows that "zoning power is associated with higher housing prices
for both 1980 and 1990", arguing that inflation and population changes since
1970 account for the increase in the importance of zoning power since that date.
It would seem better if there were an attempt at consistency, taking cognisance
of the argument in the econometrics of time series studies that later research
should be able to encompass the earlier results in order to demonstrate the su-
periority of the later formulation. So, for example, with respect to house price
studies, the model formulated by Giussani and Hadjimatheou (1990) is shown
to encompass the model set out by Hendry (1984). Because researchers using
urban data are using cross-sectional analyses involving different cities and dates,
Ch. 42: The Land Market and Government Intervention 1647
it is impossible to follow this principle literally, but the paper by Thorson (1996)
shows what should be done in the case of interurban property price studies. With
intraurban house and land price studies, a check-list of what might be expected
in a hedonic study of house prices could be used. One effort in this direction
is by Andersson, who explicitly considers the relationship between what others
have already found and what he would therefore expect to find in his study of the
Swedish cities of Malmo and Lund (Andersson, 1994), and of the condominium
market in Singapore (Andersson, 1997).
One form of zoning that should be considered on its own terms is the designation
of an area as worthy of conservation and preservation, with as little as possible
alteration, either because of its historic interest or its architectural or environ-
mental interest. From an economic point of view, it is not immediately clear
what the effect should be on the value of the properties within the designated
area. In the first place, one would expect that the value of properties in the area
would be higher than elsewhere because of the characteristics which make the
area worthy of designation. Nevertheless, this enhanced value might be small and
not statistically observable, particularly if the characteristics of the area are not
valued by those who might be expected to live there normally. Furthermore, since
designation usually restricts the ability of owners to alter properties in ways they
might wish, but which would be regarded as altering the character of the area, this
feature would tend to reduce the value of properties. On the other hand, designa-
tion ensures that other owners cannot alter the character of their properties and
so ensures that the favourable environment is preserved when it might otherwise
change. Because of designation, it follows that a purchase of a property in the
area becomes slightly less risky because the favourable environment, for which
a premium has to be paid, becomes more secure and less likely to be degraded.
With respect to the designation of individual buildings, we would expect their
value to fall since the owner's ability to alter the building is restricted, but there is
no compensating guarantee that the sulrounding environment will not be altered
for the worse. Hence, Asabere et al. (1994) found that local designation of build-
ings in Philadelphia was "associated with a 24% price decrease". On the other
hand, as we would expect, empirical evidence with respect to the designation of
conservation areas is mixed. In Baltimore, Maryland, for example, Ford (1989)
found that designation of an area as a historic district resulted in prices being
higher than one would expect, while, on the other hand, Asabere et al. (1989)
1648 A. W Evans
future and hold their land as a kind of option, and, not least, they may speculate
as to the price currently available and the price that they might obtain later. It is
this feature of the land market, the inelasticity of the supply of land for urban
development, that is an important feature of Bramley's results. What he found
was that a substantial release of land as available for development would have
little sustained impact on house prices-after a few years these prices would only
have fallen by some 5%, and the output of housing would only have increased by
some 10%.
Now, it may be that not too much reliance should be placed on Bramley's
(1993a, b) model when it is used for the simulation of changes that are supposed
to occur over substantial periods of time, at least because it is based on cross-
sectional data (see Evans, 1996; Pryce, 1997). Nevertheless, Bramley's findings
are of interest. In his simulation, the increase in supply is supposed to occur
because the local authorities designate a greater amount of land as available
to be developed, and this is represented as an alteration in the variable "struc-
ture plan provision for housing". But in his empirical analysis Bramley finds
an extremely weak relationship between this "structure plan provision" and the
output of housing. One might have expected the coefficient to be close to one if
there was a close relationship between the allocation of land for development by
planning authorities and the development of this land by developers. In fact, the
coefficient is much less than one. Most of the effect of any increase in structure
plan provision for new housing is therefore dissipated in what Bramley calls an
"implementation gap". Land may be declared to be available for development
by planning authorities but this does not mean that landowners will sell it im-
mediately for development or that developers will immediately buy it. Each will
respond in their own time, rationally seeking to maximise benefits within their
own time horizon.
The evidence relating to the economic effects of the designation of urban
growth boundaries (UGBs) in the state of Oregon is relevant to this argument.
Studies by Knaap (1985) of the Portland UGB and by Nelson (1985) of the
Salem UGB both found that four years after designation of the UGBs the price
of nonurban land within a UGB was higher than that outside the UGB. Since
the UGBs were designed and intended to meet the demand for urban land for
20 years, this evidence indicates that even with a presumed 16-year supply of
development land available within a UGB, together with a presumption that at
the end of the 16 years the UGB would be relaxed, there was still a significant
price effect.
Taken together, the British and Oregon evidence demonstrates that a con-
straint on the availability of land for development need not be completely binding
Ch. 42: The Land Market and Government Intervention 1653
As we have noted, while there has been some discussion and considerable eco-
nomic research to discover whether, or not, the imposition of growth controls
in and around US cities has resulted in increased property values, in the UK the
difference between the value of a site without permission to develop and its value
with permission to develop has become so great over the years, certainly in south-
em England, that it could not be ignored. In particular, it could not be ignored by
the local planning authorities who were granting planning permission, and, in the
process, giving the landowners substantial increases in value. The observation by
the local authorities that planning permission could be worth a considerable sum
of money has therefore led, over the past 25 years or so, to what have come to
be called "planning gain" agreements. Under such an agreement, the developer
agrees to certain conditions which both reduce the value of the permission and
transfer some of the benefits to the local authority. Initially, these agreements
mainly involved the designation of part of the site for some public use, e.g.,
an open space or a road improvement, later they included the construction of
a building on the site that might be handed over to the local authority, e.g. a
community centre or housing for low income households. By the 1980s they had
widened in scope to include payments for facilities elsewhere in the area that
might have only a tenuous relationship with the proposed development, e.g., part
of the cost of a swimming pool in a nearby town.
It is interesting to note that, as the British system has moved from a system
of wholly physical controls to one involving fiscal elements, so the same move-
ment has occurred in the US as local authorities have buttressed exclusionary
zoning by the levying of so-called impact fees, fees which are represented as
covering the cost to the city of the additional population or development. The
shift has occurred for different reasons in the two countries, however. In Britain,
the local authorities have recognised, in a manner any economist would under-
stand and applaud, that there is no point in giving anything away free if you
can obtain something you want in exchange. In the US, on the other hand, it
would appear that the wish is to prevent further development in an area and to
1654 A. W. Evans
use fiscal methods when physical controls have been perceived as being less than
fully successful. So Gyourko (1991) reports evidence from the Wharton Urban
Decentralisation Project that "among the cities which impose impact fees, there
is a statistically significant positive correlation between the amount of the fee
and the rating of its exclusionary strength [and the] amount of the fee is sig-
nificantly negatively correlated with the respondents' ratings of the exclusionary
strength of other growth control devices including ordinances and building permit
restrictions. Thus, it appears that some cities may have turned to fees after other
restrictive policies did not work as well as desired" (p. 244).
Because planning gains in Britain are subject to ad hoc negotiation between
the local planning authority and the developer, and because they vary in level,
type and cost, from development to development, little empirical research has
been carried out into their magnitude and effect. It does appear, however, that,
as one might expect, the cost of planning gain is borne by the landowner in
reduced land values, rather than being borne by the developer or passed on to
the purchaser of the property. To exemplify this, one should note that in the late
1980s it became standard practice for a developer to purchase an option to buy
a piece of land and then seek planning permission. One of the conditions of the
option would often be that the cost of planning gain, that had to be conceded in
order to increase the probability of obtaining permission, would result in a lower
price being paid to the landowner for the land in the event of permission being
granted and the option to buy being exercised.
In the US and Canada, on the other hand, because within any local authority
the level of the impact fee is set and standardised, some empirical work has been
carried out to ascertain the incidence of the fees. Singell and Lillydahl (1990)
report a study of house prices in Loveland, Colorado, over a period including
July 1984 when fees to developers were increased by $1182 over and above the
level of the traditional water and sewage fees. They found that the price of both
new and old houses increased after July 1984, indeed by substantially more than
the amount of this impact fee.
In a later study of the effects of changes in impact fees in three Toronto sub-
urbs, Skaburski and Qabeer (1992) also found that property prices rose by more
than the amount of the impact fees. They found that "each dollar ... increase in
the fee results in a $1.88 increase in lot prices when the growth rate is zero" but
that the extent of forward shifting is reduced by growth so that, "when the region
is growing at the study period's average rate of 2.33 per cent a year, each dollar
increase in development impact fees leads to a $1.23 increase in lot prices" which
they estimate corresponds to the impact fee plus the administrative cost (p. 663).
Ch. 42: The Land Market and Government Intervention 1655
Since land was available for development in the areas studied, the finding
that the amount of the impact fee was passed on to the buyers of new properties
is theoretically plausible, as is Singell and Lillydahl's (1990) finding that the
increased price of new properties resulted in higher prices for older houses. If the
supply of land is elastic, one would expect the fee and any increase in the fee to
be passed on to buyers.
In other contexts, however, the impact fee may not be passed on at all, but
borne by the landowner. Skaburkis (1990) points out that in a housing boom
such as occurred in Vancouver in the early 1980s when prices doubled within
two years (and then halved), the impact fee is unlikely to be passed on to the buy-
ers because the price of housing is demand determined. In these circumstances
the response is likely to be a reduction in the prices offered by developers to
landowners. Similarly, in Britain, in southern England at least, since the price of
land is determined by demand because the supply is fixed by the local planning
authorities, as opposed to land being freely available at a price set by the owners
of agricultural land, any form of impact fee is likely to reduce the price of land
for development, rather than passed on in the form of higher house prices.
have deliberately set out to restrict the growth and movement of firms (Evans,
1992), it would also seem difficult to deny. We have already noted that, in any
event, Cheshire and Sheppard (1997) estimate the static costs of containment
in southern England as equivalent to a 10% tax on incomes. The oddity is that
because macroeconomists have little interest in town planning, planning controls
are rarely cited by economists as one of the causes of the slow rate of growth of
the British economy. The interesting question lies in explaining why one of the
fastest growing economies in the world, Korea, should share stringent planning
controls with one of the slowest growing economies in the world.
One difference between the two economies lies in the field of housing finance.
As house prices have risen in the UK it has become evident that homeowners
have saved less because they have perceived that their wealth has risen. The
sophisticated system for financing the purchase of housing allowed people to
borrow almost all the cost of the property so that households purchased housing
early in life and the rented sector declined. On the other hand, the less sophis-
ticated Korean system for financing house purchase did not permit households
to purchase a home until they had saved a very high proportion of the cost. So
in Korea, households were encouraged to save a high proportion of their income
in order to get on the house price escalator, while in Britain, households, once
on the house price escalator, were discouraged from saving by the rising value
of their home, because of inflation and land use controls. Between, say, 1970
and 1990, it was quite possible for the value of someone's home in Britain to
increase by more than the total amount received (net of taxes, etc.) from their
employers so that any saving out of income would have affected their net wealth
very little. On the other hand, on the credit side, the housing finance system in
Britain has resulted in loans on the security of housing being a significant source
of finance for new entrepreneurs, particularly in southern England where house
prices increased the most (Black et al., 1996).
But, as is increasingly realised, the market for land-or to be exact, the market
for urban land-is not a perfectly competitive market. Assuming that it is may
result in investigators reaching false conclusions, and, in the context of this chap-
ter, failing to realise the reasons for various forms of government intervention in
the land market which would seem to be intended to ease rather than to restrict
the supply of land for some use. The land market is a highly imperfect market-
the good being sold, a piece of real estate, is not a homogeneous product, there
are relatively few buyers and sellers involved in trading pieces of real estate
that might be regarded as homogeneous products, and the buyers and sellers do
not have full information about the alternatives available, particularly alternative
future states (Evans, 1995).
Real estate has a particular characteristic that distinguishes it from other kinds
of asset, indeed it could be described as its defining characteristic. It is fixed
in location. This is easily forgotten by economists schooled in the neoclassical
tradition in which all factors of production can be treated identically. It is not,
of course, necessary to go as far as the classical economists and construct a
theory of land rent and land pricing that is entirely different from that relating
to capital and labour, but it is necessary to recognise this intrinsic difference.
It is, after all, not forgotten by the owners and buyers of land. From it follow
two other features of real property-the implications of contiguity or relative
location for land as a production good, which we will look at shortly, and its
peculiar status as a consumption good. In many places and at many times, real
property has been regarded as a better kind of asset to own than other, movable,
assets. This seems to be implicit in the use of the terms "real" estate, "real"
property. Real estate cannot be stolen like gold, jewellery or money. Nor can
it be concealed. It follows that in a stable society subject to the rule of law,
land may be preferred as an asset, particularly when its ownership may confer
wealth and power. (In an unstable society, of course, portable valuables may be
preferred to land and property.) Many nineteenth century novels could be cited in
support of this view, but perhaps Trollope's Framley Parsonage may serve as an
example. At a less exalted level, and closer to the present day, one could cite the
British Conservative government's policies of the 1980s that encouraged people
to become homeowners because it was felt that this would lead them to become
responsible citizens (and, also of course, it was hoped, become more likely to
vote Conservative).
Whether this judgement was politically correct or not, certainly it is evidence
that land is, in some way different. The economic implication is that people
become attached to their property and their home and may be reluctant to sell
at what would be presumed to be the market price (Dynarski, 1986). This at-
Ch. 42: The Land Market and Government Intervention 1659
ment in a zone on the periphery of a city during a particular period with respect
to the land in agricultural use (Evans, 1983), where the upward sloping supply
curve represents the supply of land for sale for development.
Some land would normally be sold to other fanners, because, say, the owners
wished to retire from farming and so were willing to sell at the ruling agricultural
market price (or less). A demand for land for urban development would result
in the price of land for development being bid up above the agricultural use
price. The premium will encourage some other landowners to sell who otherwise
would not. Thus, more land will be sold for development than would be sold for
agricultural use and it will be sold at the higher price necessary to encourage
landowners to sell.
Nevertheless, in the period, it is probable that only some of the land that
might be sold for development actually will be sold. The rest will not be sold
but will remain in agricultural use, because the price offered is not enough to
induce the owners to sell. Furthermore, there is no reason why the pieces of
land sold for development should be contiguous. As a result, in the absence of
government intervention, there will be urban sprawl and scattered development
as noncontiguous pieces of land are developed and intervening pieces are not. It
should be noted that scattered development is not necessarily nonoptimal from
an economic point of view. Scattered development does not completely fix the
pattern of development for the future and gives flexibility. Scattered development
does appear to be nonoptimal, however, and governments have intervened to try
to control and minimise its extent.
One reaction has been to define green belts, urban growth boundaries or some
other physical control to limit the extent of sprawl by limiting the area of land
available for development. The results of this kind of limitation on the availability
of land have been discussed earlier in this chapter. The price of the land that is
available for development is raised. The land that might have been developed
further from the urban area is now unavailable and the demand for land closer to
the city is therefore higher. The price of land for development rises as does the
quantity of land developed during the period in the area where it is permitted.
An alternative government reaction is that of the governments of Sweden and
the Netherlands. In both cases it has been government policy for many years
to buy up agricultural land that it was anticipated would be developed in the
foreseeable future. Therefore, as urban growth occurs, the land adjacent to the
city on which building should take place to maintain an orderly pattern of devel-
opment is land that is already in public ownership. Land for development is, in
effect, supplied to developers "off the shelf" at a market price that allows only
for the cost of acquiring the land at the agricultural use price and the cost of any
Ch. 42: The Land Market and Government Intervention 1661
infrastructure. The process is discussed in detail by Duncan (1985) in the case of
Sweden, and by Needham (1992) in the case of the Netherlands.
The paper by Needham (1992) is of particular interest in that he gives in-
formation on the cost of providing infrastructure (pp. 679f), and therefore sheds
light both on the operation of the policy and also, along the way, on the possi-
ble different meanings of "the price of land". Land in agricultural use (towards
the end of the 1980s) was sold at between DFL2 and DFL5 per square metre,
depending on soil properties. When municipalities acquired rural land for urban
development they paid (at the beginning of the 1980s) about DFL10 per square
metre. Thus, some compensation was paid to the seller, in part because the seller
may often have been otherwise unwilling to sell-Needham suggests that DFL10
was close to the price that would be paid if powers of compulsory purchase (i.e.
eminent domain) were to be exercised. The evidence suggests (but this time for
the 1970s) that the cost of servicing this land, i.e., the infrastructure cost, was
about DFL40 per square metre. Serviced land was sold to developers for housing
at a price of about DFL100 per square metre (1970s figures again), but it would
appear that the municipalities could in fact only sell on about half of the land that
they bought, because the rest was used for roads and other community uses. The
figures, even allowing for the differences in the periods for which the data on
prices were obtained, suggest that municipalities did not try to make a profit but
merely covered their costs as they sold land on to developers. Thus, land could
be bought for DFL10 and sold for DFL100 and either of these prices would be
"the price of land". The information is an explanation of the considerable range
in quoted market prices, and a warning to researchers to take special care in using
statistics on "the price of land".
Even taking the small compensation element into account, since the mu-
nicipalities pass on the land at cost price, it is clear that the price of land for
development is significantly lower than it would have been if green belts, or some
other form of physical growth control, had been imposed.
It should also be noted that this form of intervention does not lead to windfall
gains for developers who appear to be buying land from the municipalities at a
lower price than they would otherwise have to pay. It is the greater availability
of land for development which allows the land price to be lower, and if the de-
velopers supply housing in a competitive market the result is that their costs are
lower and, hence, the price of housing should be lower. Certainly the evidence
indicates that the cost of land is a much smaller proportion of the cost of housing
in these two countries than it is elsewhere. For example, Duncan (1985: p. 322f)
notes that in the 1970s the plot prices as a proportion of selling price of housing
1662 A. W. Eans
ranged between 15 and 27% in England and Wales, whereas they ranged between
1 and 4% in Sweden.
As mentioned earlier, there is a second feature of the land market which fol-
lows from the characteristic of land that it is fixed in location, and that is that
the location of one piece of land relative to another can be important. So one
might be able to advertise for capital to build a railway, widen a road, construct
an office park or develop an industrial or housing estate and accept any funds
subscribed--one might also be able to advertise for labour and accept any with
the necessary skills that might apply-but one cannot advertise for land in the
same way. The pieces of land necessary to build a railway would have to form a
continuous strip, there is only one piece of land that could be used to widen the
road, the pieces of land for the office park or the industrial or housing estate
have, at the least, to be contiguous. This feature of the land market may be
relatively unimportant if large areas of land are owned or controlled by single
landowners, whether individuals, companies or trusts, but it becomes of great
importance if the ownership of land is fragmented. Now, it is possible that a
would-be developer may wait and painstakingly acquire all the various pieces
of land necessary to carry out a development at the right location of the scale
needed to achieve all the possible economies. The earlier analysis suggests that
some sites will come on the market, that others may follow in due course, but that
it may be necessary to pay higher prices to some owners in order to induce them
to sell, given either their attachment to the property or their expectations about its
future value. Some developers do this. Nevertheless, it is well known that if some
owners discover that theirs is the key property necessary to complete the scheme
then they may attempt to exercise their apparent monopoly position by holding
out for an excessive price-"price gouging" as it is called in the US. Moreover,
as Eckart (1985) has shown, even if the developer were to negotiate with all the
owners at the same time, very great fragmentation of ownership may result in a
higher price being asked in total as each small owner operates a strategy that their
site is such a small proportion of the whole that they might as well ask for a high
price. Because of this, some developers may operate using a number of "front"
companies with each one apparently buying separately so that it is not apparent
that there is a single overall buyer until the whole site has been acquired. This
is not possible in many cases. The local authority that wishes to widen a road
has to acquire a particular site and both it and the site's owner know this. The
owner has a monopoly and could use it, and it is for this reason that governments
have powers of compulsory purchase or eminent domain which allow them to
force the supply of land for a particular use where, for one reason or another, the
public interest is felt to be involved.
Ch. 42: The Land Market and Government Intervention 1663
Empirical (as opposed to theoretical) research into the economics of eminent
domain is scarce. Indeed, a search of the Social Science CitationsIndex suggests
that there may be only one published paper, a study by Munch (1976) of eminent
domain in the acquisition of land for urban renewal in three areas of Chicago
between 1962 and 1970. She found the price paid was, on average, nearly 40%
greater than the predicted market value, thus implying some recognition that sell-
ers' reserve prices were likely to be higher than the market value. Some sellers
sold voluntarily, and others were subject to the legal process of eminent domain
with the price being settled by the court. In the former case, prices were 45%
higher than the predicted market value and in the latter case, they were only 25%
higher. Munch argued that since the prices for compulsory sales were settled in
the courts there was scope for influencing the price paid through legal argument.
Since the amount and quality of legal assistance that sellers might use would
depend on the total amount at stake, small owners would be at a disadvantage
relative to large owners. This implied, Munch said, that the prices obtained by the
sellers of large properties would be higher than the prices obtained by the sellers
of small properties and this is what she found. Nevertheless, it seems remarkable
that the differences were as great as they appeared to be. She concluded that,
under eminent domain, "high-valued properties systematically receive more than
market value and low-valued properties receive less than market value" (Munch,
1976: p. 495). While one might accept that small sellers who went to court did
badly, it seems odd that they should do worse than might be expected from an
open market sale as well as from a voluntary sale to the government authority.
Although the property market is an imperfect one, this would indicate both a lack
of information among a group of sellers who might have been expected to be in
communication and that their legal avisers were also giving faulty advice (since
one would have expected that with better information and advice small sellers
would not have gone to court).
As already stated, there is a scarcity of empirical research on this topic, and a
number of questions need to be answered. Even within the US setting the ques-
tion arises as to whether, or not, Munch's (1976) results are replicable. There are
also further questions as to the operation of compulsory purchase under different
legal jurisdictions. To what extent are her results location specific? What appears
to be the same form of intervention may take different forms in different places.
So in Italy (Ave, 1996: pp. 69-77) and Taiwan, the price paid for land purchased
compulsorily is substantially lower than the market price. On the other hand, in
the UK the price paid for residential properties is greater than the market price
to include an element of compensation for disturbance. What is meant by com-
pulsory purchase is important because its threat is often present when apparently
1664 A. W. Evas
voluntary agreements are made for the sale of land under schemes arranged by
governments, whether central or local. So, in the discussion above of the system
of advance purchase operated in the Netherlands, it was indicated that the price
at which land was sold to the local governments was close to the price at which it
would be compulsorily purchased. Thus, the compulsory purchase price, plus (or
minus) any differences in legal costs, provides an upper bound to the price that
might be agreed in any "voluntary" agreement.
The threat of legal coercion is also present in the schemes for the readjustment
of land ownership which exist in some countries such as Germany and Taiwan.
Under such schemes, an agreement is reached with and by the landowners in
an area that some land will be allocated for roads and other infrastructure in an
area, as laid out in a proposed plan. If this amounts to, say, one-third of the land
area, then each landowner is allocated a share of the remaining land equal to two-
thirds of the area of their original land holding, but in a single unit (if the holding
was previously split as it might be in Taiwan), and of a shape, size and location
suitable for development, for example, rectangular and bordering a planned road.
Each owner is then in a position to sell his or her landholding for development. (A
description of the process in Germany, where it is called replotting or Umnlegung
is given by Dieterich et al., 1993.)
Such land readjustment schemes are unknown in the UK or the US, possibly
because in both countries, but for different historical reasons, agricultural land
holdings are relatively large so that it is unnecessary to put together small peas-
ant landholdings to obtain sites of a reasonable size for urban development. In
countries where social, constitutional or agricultural factors have encouraged the
splitting up of the landholdings amongst numerous owners, powers to encourage
land reallocation are more necessary.
Once again, however, there will be differences between countries so that ap-
parently similar schemes will differ in their operation. As noted earlier, the pres-
sure on landowners to reach an agreement may be backed by government powers
to force an agreement if one is not reached voluntarily. But in Germany land
may be compulsorily purchased while in Taiwan a scheme of land reallocation
may be imposed on the landowners. On the other hand, the land readjustment
schemes adopted in Japan are wholly voluntary, apart from anything else because
of strong public opposition to land expropriation. "Between 1955 and 1975 the
Japan Housing and Urban Development Corporation exercised the power only six
times" (Hebbert and Nakai, 1988: p. 42). Land readjustment, however accounts
for over 20% of urban land supply (Hebbert and Nakai, 1988: p. 25), both in
urban promotion areas and in urban control areas. In the latter, schemes are
used by landowners to get around the presumption against urban development
Ch. 42: The Land Market and Government Intervention 1665
outside the major cities. If a scheme for land readjustment is agreed within an
urban control area then it has to be permitted to proceed (Hebbert and Nakai,
1988: p. 60), the landowners then selling off some land to cover the costs of
the infrastructure and the creation of the scheme, and reallocating the remainder
among themselves, which they may either choose to sell at that time or continue
to hold on to as an asset that it is hoped will appreciate in value. This practice
encourages urban sprawl since only a few sites may be sold for development at
the time that the ownership of the land is reallocated. The rest will be farmed or
left vacant, with a few sites sold each year as and when the owners need to realise
capital.
Land readjustment schemes are used to put together sites for urban develop-
ment which were previously used for agriculture. In all countries, similar prob-
lems of site assembly occur within urban areas where ownership is fragmented
and a reordering of the pattern of land use and ownership would be desirable.
For this reason, powers of compulsory purchase or eminent domain are available
to government or state authorised activities such as utilities. In most developed
economies, such powers were used to force through schemes of urban renewal
or slum clearance in the period after World War II, particularly in cities that had
suffered extensive bomb damage. In Britain, in the 1970s and 1980s these powers
have been used to promote the redevelopment of the commercial centres of cities
by the local authority, using its powers of compulsory purchase, in partnership
with a private developer, building and operating a new shopping or commercial
centre. As we have already noted, however, there has been little or no economic
research into either the effects, efficiency or equitability of these schemes.
8. Conclusions
This survey has been limited to looking at research into government intervention
in the land market in the market economies. Of course land was allocated ac-
cording to government directives and plans in the former communist countries,
but this was not really intervention in the market because no market existed.
Recently, of course, the situation has changed as, in the former communist coun-
tries of eastern Europe, governments have tried to ensure the creation of markets
where none existed before, and in China where, within an overtly communist
state, some kind of simulacrum of a land market is being created, and where
the apparent price of land must be determined prior to the market coming into
existence. Previously the view had been that, in effect, land had no value, with the
result that the pattern of land use came to differ significantly from the pattern in
1666 A. W. Evans
the market economies. Bertaud and Renaud (1997) demonstrate that Moscow has
a "perverse" population density gradient because households are concentrated at
the periphery, and that low intensity industrial uses occupy prime central areas.
The process of adjustment to differentiated land prices is likely to be lengthy. As
Li (1997) notes, "government intervention in the operation of the land market
can leave behind an inflexible system that cannot react to the market system
immediately, even after a market mechanism is created" (Li, 1997: p. 165). It is
almost certainly too early to survey research into this topic and, anyway, to do so
would lengthen this survey excessively.
Of the conclusions which can be drawn from this survey of government inter-
vention in the land markets of the market economies, the main one must surely be
that there is a disproportion in the research effort allocated to the various topics
discussed. The major part has been devoted to the analysis of the economics of
zoning and growth controls in cities in the US. In part, of course, this is inevitable
given that there are more academic economists in the US than elsewhere, but the
disproportion is such that there must be other factors. One such is surely the rela-
tive ease with which data on house prices can be obtained and analysed, coupled
with the view, which we noted, that hedonic price methods seem to have been
regarded as almost the only acceptable way in which the economics of zoning
could be researched. So there are almost no studies of zoning using contingent
valuation methods, despite their continued use in the study of the economics of
the rural environment. Is it also significant that there is only one empirical study,
using any method, of the economics of eminent domain and that this, which was
certainly not the last word on the subject, was published 20 years ago? It would
appear to be a topic that could equally well be researched. Presumably, however,
the collection of sufficient accurate data to generate tenable conclusions would
be costly and time consuming, and the absence of other research deters rather
than encourages exploration of the topic.
I am aware, therefore, that a different survey could have been written, one
which updated the previously published surveys of the economics of zoning by
Pogodzinski and Sass (1991) and of growth controls by Fischel (1990), and that
such a survey, in covering 95% or more of the published research would have
been regarded as normal and acceptable. Instead the existence of these recent
surveys has allowed the coverage of other topics in more detail, indicating in this
way more the areas where research might be carried out than those where it has
been carried out.
Because the forms of government intervention, and the operation of land
markets vary from country to country, and even within countries, this survey is
certainly incomplete. I only hope that this incompleteness will encourage others
Ch. 42: The Land Market and Government Intervention 1667
References
Abercrombie, P. (1945), Greater London Plan 1944 (HMSO, London).
Adair, A.S., J.N. Berry and W.S. McGreal (1991), "Land availability, housing demand and the property
market", Journal of Property Research 8:59-67.
Andersson, D.E. (1994), Households and accessibility: an empirical study of households' valuation of acces-
sibility to one or more concentrations of employment or services, University of Reading Discussion Papers
in Urban and Regional Economics, No. 97.
Andersson, D.E. (1997), Hedonic Prices and Center Accessibility: Conceptual Foundations and an Empirical
Hedonic Study of the Market for Condominium Housing in Singapore (Department of Infrastructure and
Planning, Royal Institute of Technology, Stockholm).
Asabere, P.K., G. Hachey and S. Grubaugh (1989), "Architecture, historic zoning, and the value of homes",
Journal of Real Estate Finance and Economics 2:181-193.
Asabere, PK., F.E. Huffman and S. Mehdian (1994), "The adverse impacts of local historic designation: the
case of small apartment buildings in Philadelphia", Journal of Real Estate Finance and Economics 8:225-
234.
Ave, G. (1996), Urban Land and Property Markets in Italy (UCL Press, London).
Bates, L.J. and R.E. Santerre (1994), "The determinants of restrictive residential zoning: some empirical
findings", Journal of Regional Science 34:253-263.
Bertaud, A. and B. Renaud (1997), "Socialist cities without land markets", Journal of Urban Economics
41:137-151.
Black, J., D. De Meza and D. Jeffreys (1996), "House prices, the supply of collateral, and the enterprise
economy", Economic Journal 106:60-75.
Bramley, G. (1993a), "The impact of land use planning and tax subsidies on the supply and price of housing in
Britain", Urban Studies 30:5-30.
Bramley, G. (1993b), "Land use planning and the housing market in Britain: the impact on housebuilding and
house prices", Environment and Planning A 25:1021-1051.
Cheshire, P.C. and S. Sheppard (1989), "British planning policy and access to housing: some empirical
estimates", Urban Studies 26:469-485.
Cheshire, P.C. and S. Sheppard (1995), "On the price of land and the value of amenities", Economica 62:247-
267.
Cheshire, P.C. and S. Sheppard (1997), Welfare economics of land use regulation, Research Papers in En-
vironmental and Spatial Analysis, No. 42 (Department of Geography, London School of Economics,
London).
Crecine, J.P., O.A. Davis and J.E. Jackson (1967), "Urban property markets: some empirical results and their
implications for municipal zoning", Journal of Law and Economics 10:79-100.
Dieterich, H., E. Dransfeld and W. Voss (1993), Urban Land and Property Markets in Germany (UCL Press,
London).
Duncan, S. (1985), "Land policy in Sweden: separating ownership from development", in: S. Barrett and P.
Healey, eds., Land Policy: Problems and Alternatives, Chapter 15 pp. 308-344 (Gower, Aldershot).
Dynarski, M. (1986), "Residential attachment and housing demand," Urban Studies 23:11-20.
Eckart, W. (1985), "On the land assembly problem", Journal of Urban Economics 18:364-378.
Evans, A.W. (1983), "The determination of the price of land", Urban Studies 20:119-139.
Evans, A.W. (1988), No Room! No Room! (Institute of Economic Affairs, London).
Evans, A.W. (1992), Town Planning and the Supply of Housing. The State of the Economy, 1992 (Institute of
Economic Affairs, London).
Evans, A.W. (1995), "The property market: ninety per cent efficient?", Urban Studies 32:5-29.
1668 A. W. Evans
Evans, A.W. (1996), "The impact of land use planning and tax subsidies on the supply and price of housing in
Britain: a comment", Urban Studies 33:581-585.
Fischel, W.A. (1980), "Zoning and the exercise of monopoly power: a reevaluation". Journal of Urban
Economics 8:283-293.
Fischel, W.A. (1989), "What do economists know about growth controls: a research review", in: D.J. Brower,
D.R. Godschalk and D.R. Porter. eds., Understanding Growth Management (The Urban Land Institute,
Washington. DC).
Fischel, WA. (1990), Do Growth Controls Matter? (Lincoln Institute of Land Policy. Cambridge, Mass.).
Ford, D.A. (1989) "The effect of historic district zoning on single-family home prices", Journal of the American
Real Estate and Urban Economics Association 17:353-362.
Gale, D.E. (1991), "The impacts of historic district designation: planning and policy implications", Journal of
the American Planning Association 57:325-340.
GB Commission on the Third London Airport (1970), Papers and proceedings, Vol. VII (HMSO. London).
Giussani, B. and G. Hadjimatheou (1990), "Econometric modelling of UK house prices"', Cyprus Journal of
Economics 3:36-57.
Gottlieb, M. (1965), "Influences on value in urban land markets", Journal of Regional Science 6:1-16.
Gyourko, J. (1991), "Impact fees, exclusionary zoning, and the density of new development", Journal of Urban
Economics 30:242-256.
Hamilton, B.W. (1978), "Zoning and the exercise of monopoly power", Journal of Urban Economics 5:116-
130.
Hebbert, M. and N. Nakai (1988), How Tokyo grows: land development and planning on the metropolitan
fringe. Suntory-Toyota International Centre for Economics and Related Disciplines. Occasional Paper No
11 (London School of Economics).
Hendry, D.F (1984), "Econometric modelling of house prices in the United Kingdom", in: D.F. Hendry and K.
Willis, eds., Econometrics and Quantitative Economics (Blackwell, Oxford).
Ihlanfeldt, K. and T.P. Boehm (1987), "Government intervention in the housing market: an empirical test of
the externalities rationale", Journal of Urban Economics 22:276-290.
Joseph Rowntree Foundation (1994), Inquiry into Planning for Housing (Joseph Rowntree Foundation, York).
Kim, K-H. (1993), "Housing policies, affordability, and government policy: Korea", Journal of Real Estate
Finance and Economics 6:55-71.
Kim, K-H. (1994), Controlled developments and densification: the case of Seoul, Korea. Discussion Paper,
Dept of Economics (Sogang University, CPO Box 1142, Korea).
Knaap, G-J. (1985), "The price effects of urban growth boundaries in metropolitan Portland, Oregon", Land
Economics 61:26-35.
Lakatos, I. (1970), "Falsification and the methodology of scientific research programmes". in: I. Lakatos and
A. Musgrave, eds., Criticism and the Growth of Knowledge (Cambridge University Press, Cambridge).
Lenon, M., S.K. Chattopadhyay and D.R. Heffley (1996). "Zoning and fiscal interdependencies", Journal of
Real Estate Finance and Economics 12:221-234.
Li, L-H. (1997), "Privatization of the urban land market in Shanghai", Journal of Real Estate Literature 5:161-
170.
Li, M.M. and H.J. Brown (1980), "Micro-neighbourhood externalities and hedonic housing prices", Land
Economics 56:125-141.
Malpezzi, S. and S.K. Mayo (1987), "User cost and housing tenure in developing countries". Journal of
Development Economics 25:197-220.
Malpezzi, S. and S.K. Mayo (1997), "Getting housing incentives right: a case study of the effects of regulation,
taxes, and subsidies on housing supply in Malaysia", Land Economics 73:372-391.
Manning, C.A. (1988), "The determinants of intercity home building site price differences", Land Economics
64:1-14.
McDonald, J.F. (1995), "Houston remains unzoned", Land Economics 71:137-140.
McMillen, D.P. and J.F. McDonald (1991), "Urban land value functions with endogenous zoning". Journal of
Urban Economics 29:14-27.
McMillen, D.P. and J.F. McDonald (1993), "Could zoning have increased land values in Chicago?" Journal of
Urban Economics 33:167-188.
Ch. 42: The Land Market and Government Intervention 1669
Mirrlees, J.A. (1972), "The optimum town", Swedish Journal of Economics 74:114-135.
Monk, S., B.J. Pearce and C.M.E. Whitehead (1996), "Land-use planning, land supply and house prices",
Environment and Planning A 28:495-511.
Munch, P. (1976), "An economic analysis of eminent domain", Journal of Political Economy 84:473-497.
Needham, B. (1992), "A theory of land prices when land is supplied publicly: the case of the Netherlands",
Urban Studies 29(5):669-686.
Nelson, A.C. (1985), "Demand, segmentation and timing effects of an urban containment program on urban
fringe land values", Urban Studies 22:435-443.
Neutze, G.M. (1987), "The supply of land for a particular use", Urban Studies 24:379-388.
Oizumi, E. (1994), "Property finance in Japan: expansion and collapse of the bubble economy", Environment
and Planning A 26:199-214.
Ozanne, L. and T. Thibodeau (1983), "Explaining metropolitan housing price differences", Journal of Urban
Economics 13:51-66.
Pogodzinski, J.M. and T.R. Sass (1991), "Measuring the effects of municipal zoning regulations: a survey",
Urban Studies 28(4):597-621.
Pryce, G. (1997) Construction Elasticities and Land Availability (Centre for Housing Research and Urban
Studies, University of Glasgow).
Schaeffer, P.V. and C.A. Millerick (1991), "The impact of historic district designation on property values: an
empirical study", Economic Development Quarterly 5:301-312.
Singell, L.D. and J.H. Lillydahl (1990), "An empirical examination of the effect of impact fees on the housing
market", Land Economics 66:82-92.
Skaburskis, A. (1990), "The burden of development impact fees", Land Development Studies 7:173-185.
Skaburskis, A. and M. Qabeer (1992), "An empirical estimation of the price effects of development impact
fees", Urban Studies 29:653-667.
Son, J-Y., and K-H. Kim (1998), "Analysis of urban land shortages: the case of Korean cities", Journal of
Urban Economics 43:362-384.
Speyrer, J.F. (1989), "The effect of land-use restrictions on market values of single-family houses in Houston",
Journal of Real Estate Finance and Economics 2:107-113.
Takeuchi, K., K. Nishimura, D-K. Lee and H. Ikegeuchi (1994), "Land prices and Japanese city planning:
evaluating the effects of land use control", in: B. Koppel and D. Young Kim, eds., Land Policy Problems in
East Asia-Toward New Choices (East-West Center, and Korea Research Institute for Human Settlements:
Kyounggi-Do, Korea, Honolulu).
Thorson, J.A. (1996), "An examination of the monopoly zoning hypothesis", Land Economics 72:43-55.
Titman, S. (1985), "Urban land price under uncertainty", American Economic Review 75:505-514.
Wallace, N.E. (1988), "The market effects of zoning undeveloped land: does zoning follow the market?",
Journal of Urban Economics 23:307-326.
Willis, K.G. and M.C. Whitby (1985), "The value of green belt land", Journal of Rural Studies 1:147-162.
Chapter43
CHARLES M. BECKER*
ANDREW R. MORRISON
Inter-AmericanDevelopment Bank
Contents
1. Patterns of urbanization in developing countries 1675
1.1. Urbanization patterns and data issues 1675
1.2. The great fertility deceleration and components of urban growth 1678
1.3. Survey themes 1681
2. Models of city growth and empirical estimates 1682
2.1. Urban labor markets: empirical microfoundations 1682
2.1.1. Inequality in urban labor markets: segmentation and discrimination 1682
2.1.2. Education and labor market outcomes 1687
2.1.3. Economic crisis, adjustment and urban labor markets 1690
2.2. Aggregate partial equilibrium migration models 1691
2.3. Analytical general equilibrium models 1699
2.4. Computable general equilibrium (CGE) models 1703
2.5. Demo-economic approaches 1707
2.6. Migration of women and families 1712
3. Impacts of public policy on urban growth 1718
3.1. Structural adjustment 1718
3.1.1. The structure of structural adjustment 1718
3.1.2. Impacts of SAPs on urbanization 1719
3.2. The urban bias thesis 1721
3.3. Special problems of (ex)-socialist countries 1723
3.3.1. Urbanization in the former USSR 1723
3.3.2. Urbanization in the People's Republic of China 1727
* We are deeply grateful to Omar Bello, Christopher Grewe, Ramon Key, Stephen Malpezzi and Edwin
Mills, all of whom gave extensive and valuable comments on earlier versions of this manuscript. All errors,
omissions and misinterpretations, however, are our own.
1673
1674 C.M. Becker and A.R. Morrison
Abstract
The past half-century has witnessed a dramatic change in the way in which peo-
ple live. Fifty years ago, only a small proportion of the less developed world lived
in cities, and world poverty was overwhelmingly rural. In 1950, less than one-
fifth of the population of the "third world" was urban; in the next five years or
so, a majority of developing countries' populations will be urban. This dramatic
social change has captured the attention of development economists and, to a
lesser degree, urban economists. This chapter examines what has been learned in
a variety of areas. Section 1 discusses the stylized patterns of urbanization in the
developing world, while Section 2 turns to models of third world city growth and
their empirical estimates, discussing partial equilibrium models, general equi-
librium models, economy-wide computable general equilibrium (CGE) models,
demographic-economic perspectives, and household migration modeling. Sec-
tion 3 considers the impact of government policies on urbanization. Particular
attention is devoted to structural adjustment policies, urban biases in public ex-
penditures, and issues unique to (ex)-socialist economies. Section 4 examines
structural impediments to urban development, including labor and land mar-
kets, transportation issues, public finance and social infrastructure concerns, and
urban spatial structure. The final section looks at the macroeconomic impacts
Ch. 43: Urbanization in Transforming Economies 1675
Broad trends are shown in Table 1. The urban populations of developing countries
grew at about 4.7% annually during the 1950s and at about 4.0% during the
1960s. What happened thereafter is a matter of some debate. Estimates through
the early 1990s suggest that urban growth rates accelerated somewhat in the
1970s, perhaps to 4.4%, and remained quite high in the 1980s. But the most
recent estimates by international agencies indicate much more moderate urban
population growth, both in the 1980s, and, more surprisingly, in the 1970s (only
3.6%, according to World Bank (1995).
The extent of disagreement is disconcerting, especially to those inclined to
regard as inviolately factual anything printed by authoritative international agen-
1 This survey should be regarded as a complement to Lucas' (1997) survey of internal migration in
developing countries for the Handbook of Populationand Family Economics.
1676 C.M. Becker and A.R. Morrison
Table 1
Patterns of urbanization, 1950-1995 (all figures are in percentages)
Region 1950 1960 1970 1975 1980 1990 1995e 1995 2015e
Source: United Nations (1980: Tables 5, 6); final two columns: United Nations (1996: Tables 1-2).
cies. These agencies, however, have to use census data, and these are collected
only every decade or so, typically around the beginning of a new decade. Given
that processing often takes three years or more, it is fairly natural that large
revisions to world statistics occur at mid-decade.
Data on the proportion of populations living in urban areas tend to be less
volatile. In part, this is simply because it is a stock rather than a flow measure.
But it also reflects the fact that much of the error in predictions of city growth is
caused by erroneous estimates of population growth rates, rather than mistaken
assessments of the geographical distribution of population. On the other hand,
the overall level of urbanization in developing countries is even more sensitive to
Ch. 43: Urbanizationin Transforming Economies 1677
definition of "developing", since wealthier "developing" countries are far more
urban than their poorer counterparts.
Estimates also have been muddied in recent years by the inclusion of data
from a considerable number of countries which had not reported (or not reported
accurately) earlier. Most of these are from the current or former socialist bloc, and
it should be clear that, even now, many of these official statistics are implausible. 2
In these cases, historical "pass" systems and other forms of population control
have begun to break down, but censal data continue to reflect official locations of
residence.
Ironically, then, we may know less about world urbanization today than we
did 15 years ago. But two key patterns are clear: (1) the developing world contin-
ues to urbanize, but it is doing so at a slower pace than in the 1950s and 1960s;
(2) excluding China and east Asia, urban growth in the early 1990s is slower
than in the 1970s; it is also slower than in the 1980s for all regions except for
east Asia. 3 Thus, Preston's (1979: p. 196) claim that "The rate of change in the
proportion urban in developing countries is not exceptionally rapid by historical
standards; rather it is the growth rates of urban population that represent an un-
precedented phenomenon" is somewhat less valid today than it was two decades
ago. Urban growth rates remain very high, but they have slowed significantly in
many regions.
A final note on forecasts. Earlier forecasts of urbanization appear to have
been on the high side, especially in Africa, south Asia and China. The Preston
Report (United Nations 1980; Preston, 1979) had 42.5% of Africans living in
cities by 1995 (revised down to an implied 37.3% in a UN 1991 re-estimate cited
in Lucas, 1997); more recent estimates by the United Nations (1996) place the
actual figure at 34.4%. Forecasts for south Asia and India suffer from similar
problems. For "less developed regions" as a whole, the United Nations (1980)
forecasted an increase in urbanization from 1980 to 1995 by 13%, to 43.5%
2 It stretches credulity that the urban population of a country like Vietnam is only growing at 2.7 to 3.1%
(with background total population growth of 1.6 to 2.1%), given the obvious mushrooming of major cities
and the very low initial urbanization rate. China's official and actual urbanization levels may be similarly
distorted, especially if one considers that most of the official urban growth was "achieved mainly through the
transformation of villages into new-style cities and towns, rather than through migration of residents from rural
to established urban areas" (Young and Deng, 1997: p. 5).
3 If one accepts the veracity of official Chinese data. Zhu (1996) suggests that recent growth may be far
larger than official figures suggest, as official figure miss much or all of the "floating population". Individuals
are registered as "agricultural" or "nonagricultural;" perhaps surprisingly, Zhu's (1996: p. 11) empirical results
from a microsurvey are that people from agricultural households are more likely to migrate to urban areas. The
interpretation offered is that such people face more limited opportunities at home, and therefore must migrate
to enhance earnings. The finding by Young and Deng (1997) that political regime changes did not appear to
directly affect urban growth in China also merits notice.
1678 C.M. Becker and A.R. Morrison
urban. The United Nations (1996) revised estimates for 1995 indicated a rise
of only 7.1%-a dramatic revision.
There are clear demographic explanations for these large prediction errors. In
times of considerable behavioral change across cohorts, such as during the past
decade, simple, highly aggregated, a behavioral demographic accounting models
are subject to huge forecasting errors. Behavioral economic models and more
sophisticated demographic frameworks perform better, and one might expect a
combined approach to do better still.
One of the primary reasons for overestimating urban growth rates has been the re-
cent decline in world fertility rates. Previous studies indicate that, for developing
countries as a whole, somewhat less than half of the increase in urban population
is due to migration from rural areas (see United Nations, 1996, 1980; Preston,
1979; Gregory, 1986; Ledent, 1982a); the remainder is due to natural population
increase in urban areas. In Preston's survey, 61% of developing countries' urban
population growth was due to natural increase, and this share was expected to
rise with time. The United Nations (1996) study of urbanization compares 30
countries between the 1960s and 1970s and 19 countries between the 1970s and
1980s; it shows migration's contribution to urban growth declining over time for
most countries.
If one is willing to make the assumption that crude birth rates are identical
in urban and rural areas, then it is possible to provide a simple decomposition
of urban growth. 4 This decomposition appears in Table 2, which provides esti-
mates of the urban net inmigration rate (net change in rural-urban migrants as
a fraction of base urban population) both for regional aggregates, as well as for
a small number of Asian, Latin American and African countries which are at
the forefront of the economic and/or demographic transitions. Compared with
1970-1980, the period 1980-1993 was one of decreasing urban net inmigration
in Africa, Latin America and south Asia, but of rising inmigration in east Asia.
Among the east Asian "tigers", however, the (unweighted average) inmigration
rate declined, albeit not to a very low rate.
Population growth is a different matter. Growth rates have been stable in south
Asia and Africa (where they actually rose in the 1980s relative to the 1970s).
In contrast, population growth rates have declined by one-third or more in east
Asia and Latin America from the 1970s to the early 1990s; it is clear that large
4 These UN data reported above are based on separate estimates of urban and rural fertility, and such data
are available in consistent form only for a limited number of countries and years.
Ch. 43: Urbanization in Transforming Economies 1679
10 "v I I I
q 1rvlr
I
e
I
O
NI1NN _
~oo _ ci
0
C)
ll- - ll lz
C)
aC
O. N . 99
-.- 9 0
N N ,q-m
o
C C
ci 7 o o s Iq -m -a 1 Iz e v, n vI
C) I- 7 > V) 90 V) 0 " rj Nr I i I' Cr CN ©
C)
u
0
00 I
9 CO
0 S w
E
C)l
-q
._ nnel ti Is
Ai N iO
c9 qo N
o -
PI ~n \o
ej a9
oi
ns
l ie
N Nh
i - oi
0
L)
Ce
-0 i
a
N0 CCN
w
C
C oO NNCC N" I
io N N0 N NC
CC C 01-
rC jO N Or NI
o O
o
o
00
ESD 3
.
0
0-
0
o N o - Co C
Cs
55 E
"0
CO C C Nim o N O N I
ci
o
8I
0 0`
-c
ci
CoNt CNOriN NO ! oo - "
Cci CO ~oN W ON O
C))
o
ci
" i n
Zt i ~n m N' N _ P\ "r
-0 W
:j
io 0
v
C)
It lz 'ot ll mNti N Nm v,.,
00 _ ri d ri O K 6 O 6 &K· M I
0
, a
S i
uz =i o H\
20 0
e I66 me> w e zo <6
' 0 6 5 i 6 o• Ir
7 fi
C c
CES
c 0e~r ON
ON
3
1 pXi c:
7s.·· ·
i4 C O V1I
:e b On u
C)o
o
0,
r
S~~~
-0 ,~B
CC 5
S ~ C
CC
C, C
C
>--
i t C ) *
C , a
CO CCC ~~~o
~~~~·5 CO CC
rm~~~ CCC
)CC
C~
,
C)5o~
O•O
5
C 0
-C·
CC
50
0
<CCC-
S oN ~
-j E)O
C O ZN g
OZ g~ y), C)CJ5,;
1680 C.M. Becker and A.R. Morrison
declines will continue in the coming years, given that total fertility rates are now
at or below replacement in east Asia.
Consequently, declining fertility is currently the major source of decelerating
urban population growth in east Asia and Latin America, while declining urban
inmigration rates are the larger source of slowing urbanization in Africa and (to
the extent that changes are taking place) south Asia. The east Asian pattern shows
clearly both in China and in the "tigers": during the current decade, migration is
likely to account for two-thirds or more of total urban growth. In Latin America
and south Asia, much lower migration rates and much higher population growth
rates prevail, so that natural population growth will be the main source of urban
population growth in both regions. This is especially true in the most advanced
Latin economies, which have urbanization levels comparable to those in the most
developed countries.
The emerging African pattern may be that of its relatively prosperous and
stable nations, which appear to be entering a fertility transition. Migration and
natural population growth appear to be roughly comparable in importance, al-
though the role of migration is decreasing. Moreover, while still quite high
(6.0%), annual urban population growth in prosperous Africa has declined con-
siderably from the 1970s.
There is good reason to believe that declining population growth will con-
tinue. Declines in population growth rates have been modest thus far. World
Bank (1995) reports that for all low income countries, the weighted average
annual growth rate was 2.1% in the 1970s, 2.0% from 1980-1983, and is pro-
jected to be 1.8% for 1993-2000), but should accelerate for two reasons. First,
demographic momentum has largely maintained population growth thus far: for
low income countries, the weighted crude birth rate declined from 39 births
per thousand people in 1970 to 29 births per thousand in 1993 (the decline is
only from 45 to 40 if China and India are excluded). But the total fertility rate,
defined as the expected number of children a woman will bear, applying current
cohort-specific birth rates, fell from 5.9 in 1970 to 3.6 in 1993. A rising share
in the total population of women of child-bearing ages during the past 25 years
thus maintained the crude birth rate, but in the coming decades, the impact of
declining fertility will not be so offset by the momentum created by past high
fertility.
Second, the mortality component of the "demographic transition" has been
largely completed. The decline in the crude birth rate by 11 births per thousand
was largely offset by a fall in the crude death rate from 14 to 10 per thousand
(and from 19 to 13 for low income countries other than China and India, more
than offsetting the birth rate decline). These declines reflect health care improve-
Ch. 43: Urbanizationin TransformingEconomies 1681
ments which raise life expectancies and limit deaths, especially from contagious
diseases. However, as populations age, increasing mortality from chronic disease
begins to compensate (especially as many of the "easy gains" already have been
realized), so that further crude death rate declines are likely to be more moderate.
In short, the developing world overall is moving into a more mature stage of
the demographic transition, with sharply declining population growth rates. Thus,
while the contribution of migration to urban growth eventually must decline, the
combined effects of the demographic transition and rapid economic growth in
much of the transforming world's cities virtually ensure that migration's role
will be large in the coming decades.
There is little doubt that the nature of urbanization is in the midst of major
change. Education and infrastructure projects of the preceding three decades
have led to much greater integration of economies today than ever before. Thus,
the stylized facts are changing-but differently in different places, to such a
degree that it is a terrible mischaracterization to write of a single "third" or
"developing" world. Hereafter, we therefore refer to economies other than in
advanced capitalist nations as transforming-fornearly all are undergoing major
change. We further distinguish them in two ways: whether they have integrated
or segmented labor and other factor markets, and, affirming the distinction made
more than two decades ago by Chenery and Syrquin (1975), whether coun-
tries are large, small and open, or small and inward-looking. The essay obvi-
ously will not make constant reference to which of the six categories a model
or empirical finding is appropriate, but will make these distinctions as neces-
sary. Clearly, other delineations are possible as well; among these, high and low
economic growth regimes might be the most critical. Prolonged rapid growth
seems strongly associated with growing economic integration, however, and for
purposes of analyzing human mobility we prefer to emphasize the latter feature.
A second theme which appears throughout the chapter concerns the need to
develop a new set of "stylized facts", for the traditional ones are surely dated.
The explosion of microhousehold and firm datasets gives us information of al-
most unimaginable detail. In the coming pages, we argue that much of the new
information is inconsistent with the "informed impressions" which guide model
builders and policymakers alike. At the same time, the gap between microstudies
and aggregate models appears to be growing, and we see efforts at reconciliation
to be of critical importance. A related theme is that economic models of urban-
1682 C.M. Becker and A.R. Morrison
ization need to connect with models of demographic change, especially given the
dramatic transitions underway in many societies.
Finally, we believe that economic models of urban growth are correct to re-
main focused on the role of migration. As noted, there are good reasons to believe
that it will continue to be the dominant source of growth in China and east Asia,
and will play an important role everywhere except for Latin America. 5 Second,
economic analysis focuses on marginal decisions, and it seems plausible spatial
living decisions react more immediately to economic incentives than do fertility
decisions in transforming economies.
This section discusses three topics related to the functioning of urban labor mar-
kets: inequality in urban labor markets, the relationship between educational
attainment and labor force outcomes, and the performance of urban labor markets
during economic adjustment programs. These aspects of urban labor markets
have important implications for urbanization in terms of incentives for potential
migrants and the efficiency of production processes in urban areas.
estimates for Latin America range from a low of 19% of the economically active
in urban areas in Uruguay to a high of 39% in Peru (Infante and Klein, 1991: p.
126). Informal labor markets are also important sources of employment in Africa
and in some Asian countries. 7
Labor markets are segmented because workers in one sector (the protected
or formal sector) earn higher wages than identical workers in the other sector
(the unprotected or informal sector). Neoclassical economic theory suggests that
such a wage gap should be competed away by informal workers moving to the
formal sector, but the theory of segmented labor markets argues that mobility
is extremely limited between the sectors. Note that the mere existence of wage
gaps between formal and informal sectors is not sufficient evidence to prove
segmentation; mobility must be limited as well.
One of the most common-and commonly misused-techniques for identi-
fying segmented labor markets is to estimate wage equations for a sample of
workers in each labor market segment. A typical set of wage equations would be:
where W is the wage rate, the subscripts f and I represent formal and infor-
mal sectors, respectively, X is a vector of human capital characteristics, a is a
constant, 3 is a vector of coefficients that captures the return to human capital,
and E is a random error term. The conventional interpretation is that if Pf > i
for some or all of the estimated coefficients then labor market segmentation is
present, since returns to human capital in the formal sector exceed those in the
informal sector.
This is an incorrect interpretation for two reasons. The first problem with this
approach is that there may be a selection process not captured in the regressions.
Gindling (1991) corrects selection in his investigation of segmentation in San
Jose, Costa Rica, by estimating multinomial logit selection equations that allo-
cate workers to the private formal, public formal and informal sectors. 8 He finds
that many of the coefficients in the selection equations, including education, sex,
age, head of household and marital status, are statistically significant predictors
7 Sethuraman (1981) provides an early but very comprehensive estimate of the share of the urban labor
force working in the informal sector for a long list of cities. His estimates for the late 1970s are worth listing
here: Abidjan (31%), Lagos (50%), Kumasi (60-70%), Nairobi (44%), Calcutta (40-50%), Ahmedabad (47%),
Jakarta (45%), Colombo (19%), Singapore (23%), Sao Paulo (43%), Rio de Janeiro (24%), Bogota (43%),
Guayaquil (48%), Mexico City (27%) and Caracas (40%).
8 Gindling is not the only researcher to note that the conventional approach is subject to selectivity bias.
Gronau (1973) and Magnac (1991) also offer intelligent discussions of this issue.
1684 C.M. Becker and A.R. Morrison
There are, however, several articles which examine the applicability of efficiency wage theory to rural Indian
labor markets (see Rodgers, 1975; Bliss and Stem, 1978, Part II).
12 As Birdsall and Sabot (1991) note, "tradition" may confine women to working in family businesses or
household production, or segregate castes by occupation or industry. Tradition may be viewed as a societal taste
for discrimination. Statistical discrimination occurs when employers generalize a certain group's "average"
behavior to all members of the group, e.g., a higher likelihood of leaving the labor force is ascribed to an
individual woman, since women in general have higher rates of voluntary separation from the labor force.
13 The standard methodological references are Blinder (1973), Oaxaca (1973) and Brown et al. (1980).
14 The samples are from urban labor markets, with the exception of Nicaragua. In the case of Nicaragua,
the sample is taken from the national labor market.
1686 C.M. Becker and A.R. Morrison
neurial ability will choose informal sector work, and compares predicted earnings
in informal sector versus public and private formal sector jobs. Informal sector
work actually provided an income premium relative to the private formal sector
both in 1985 and 1990; the public sector appeared to offer a protected premium
in 1985, but by 1990 after a period of pronounced depression, its workers were at
a strong disadvantage. Yamada's work, with its emphasis on sectoral selection as
well as thoughtful sectoral definitions, provides powerful evidence against crude
segmentation. On the other hand, it is difficult to claim that no labor market
imperfections exist in transforming economies, or that all of the returns to distin-
guishing features reflect differing capacities: rather, the nature of imperfections
is complex, and it is not evident that their crude depiction in modeling is a second
best solution.
Funkhouser (1997) examines labor market segmentation in Guatemala, and
also corrects for self-selection-which greatly increases the formal/informal sec-
tor difference in returns to education (12 versus 6%, respectively). Much of the
earnings gaps in Guatemala can be tied to gender and indigenous/nonindigenous
differences; broadly speaking, these are more important than differences among
(and are generally maintained within) the formal private, formal government, and
informal sectors.
15 Return figures are for the "latest year"; for most countries this is during the 1980s.
1688 C.M. Becker and A.R. Morrison
16 Such stringent data requirements are unlikely to be met in many transforming economies, though.
17 They term the comparison a "natural experiment" because the countries are structurally very similar, but
pursued strikingly different educational policies upon obtaining independence. Kenya embarked upon a rapid
expansion of secondary education, including allowing private secondary schools to open. Tanzania, on the
other hand. stringently limited access to secondary schooling, using a manpower planning approach to avoid
excess supply of secondary school leavers.
Ch. 43: Urbanization in Transforming Economies 1689
Another contribution of Knight and Sabot's work is a correction of the wage
equation estimates for cohort effects. The conventional measurement of the re-
turn to primary education is 17%, but the return available to current school
leavers is only 12%. This significant difference is largely due to the "filtering
down" of secondary school graduates, i.e., the most attractive jobs formerly filled
by primary school graduates are now filled by secondary school graduates
(Knight and Sabot, 1990: p. 42). The rate of return to secondary school, on the
other hand, is not affected by correcting for cohort effects; it remains 13%.
A more general problem with Mincerian wage equations concerns their in-
terpretation, though these problems are mitigated with thoughtful econometric
techniques (such as correction for sectoral selection bias) and good datasets.
Nevertheless, it takes something of a leap of faith to ascribe the full effect of
human capital variables to human capital effects, especially when selectivity,
screening, credentialism and cohort effects are not accounted for. An example
of the sort of careful work which does inspire confidence is provided by Lam and
Schoeni's (1993) work on Brazil. Returns to schooling decline by one-fourth to
one-third when family background measures are included, a finding consistent
with Behrman and Wolfe's (1984c) study of 500 sister pairs from Nicaragua.
Lam and Schoeni's most memorable finding, though, is that the schooling of a
man's father-in-law has a larger effect than own father's schooling, which they
take as an indication of assertive mating.
The other important effect of educational expansion on urban labor markets is
its impact on the distribution of income. Two potentially opposing effects must be
considered. The first is the compression effect. Education expansion compresses
the premium on education by: (1) reducing wages in occupations that employ
educated individuals, and (2) causing more educated workers to filter down into
jobs formerly occupied by less-educated individuals, reducing more educated
individuals' average wages in relation to the less-educated (Knight and Sabot,
1990: p. 28). The composition effect, by changing the educational composition
of the labor force, may either increase or decrease wage inequality.18 Educational
expansion is likely to increase wage inequality only if the composition effect in-
creases inequality and if this effect dominates the compression effect. In practice,
this is likely to happen only in the very early stages of economic development.
Educational expansion not only increases the productivity of labor, but also is
likely to make labor market outcomes more equitable.
18 If the initial number of educated workers is small, an increase in their number is likely to increase wage
inequality. After some critical point is passed, however, further increases in the number of educated workers
will decrease wage inequality.
1690 C.M. Becker aund A.R. Morrison
19 Suructural adjustment programs are designed to restore external balance and, in the long run, achieve pos-
itive economic growth. Structural adjustment programs, at least in theory, differ from stabilization programs.
which were designed to reestablish external balance in the short run via demand compression (Conway. 1994:
p. 267).
20 The urban unemployment rate in Latin America rose from 6.9% (1980) to 10.9% (1985) (Portes, 1989: p.
27). Is this evidence that real wage declines were not large enough? It is difficult to determine. Certainly labor
should not bear the entire cost of adjustment; in addition, as mentioned above, a too severe decline in wages
will constrict aggregate demand and reduce output.
Ch. 43: Urbanizationin TransformingEconomies 1691
2.2. Aggregate partialequilibrium migrationmodels
22 The surplus labor versus neoclassical rural sector distinction is not terribly important. Bhatia (1979)
extends HT to a rural surplus labor economy. His findings (1979: p. 409) are:
Result 1 An increase in expected urban wage will lead to outmigration from agriculture whether work hours
on the farm are fixed or flexible ...
Result 2 ... for a given increase in rural wage, flexible hours in farming will cause larger outmigration than in
the HT model ...
Result 4 . . for a given, small increase in expected urban wage, the presence of surplus labour will lead to a
larger outmigration than in Result 2.
23 This paper does not deal with the econometric problems encountered in estimating aggregate migration
functions, but there are many. Among the more obvious, but rarely recognized, is that interregional net migra-
tion coefficients must sum to zero, as pointed out by Greenwood and Hunt (1984). Data problems abound as
well.
1694 C.M. Becker and A.R. Morrison
but also finds that origin poverty incidence is a deterrent to migration. The HT
variable-a measure of the expected income differential-also has the expected
sign, though it is only significant at the 10% level.
24 In an important paper, Jamal and Weeks (1988) document the vanishing" formal sector premium over
rural (and IFS) activities; further evidence of gap narrowing is summarized in Becker et al. (1994).
Ch. 43: Urbanizationin TransformingEconomies 1695
among labor economists, who were challenged to generate a wage gap within a
neoclassical framework.
Among the responses to HT were those which emphasized educational level
as a screening device. Barnum and Sabot (1977) examine Tanzanian migration
from 1955-1970 and find rural-urban male migration rates for those with some
secondary schooling (Standard 5-8; the differential is even greater for those with
high school education) to be about six times greater than those with only primary
schooling, and 20 times as great as for those without any schooling. By the late
1960s, the likelihood of finding a job within four months had fallen considerably
for all groups, but those with secondary education still had dramatic advantages
in terms of finding a job and also received much higher wages. Using aggregated
data, the HT expected wage term is highly significant, but so is education.
For Kenya (the country which originally inspired the HT model), Collier and
Bigsten (1981), using microdata from the late 1970s, find evidence of wage
flexibility, substantial increases in screening in response to rising educational
standards, and falling unemployment rates among those least likely to obtain for-
mal sector jobs. In a study of Venezuelan interstate migration, Levy and Wadycki
(1974: p. 387) find that those with secondary education are far "more respon-
sive than the uneducated to variables which represent the costs and benefits of
migration".
Sabot (1979) and his associates synthesize these results into a "queuing"
model of migration. As growth of formal sector jobs with wages above what
can be earned in informal activities is outstripped by growth in the supply of
secondary school leavers, educated workers "filter down" and take increasingly
less skill-using jobs. Potential migrants with less education are crowded out,
and migrate only to take advantage of competitive, unskilled formal or informal
sector jobs. Collier (1979) generalizes this framework to a model for Tanzania in
which the perceived probability of finding ajob depends on anticipated duration,
and in which rural labor supply conditions matter. In effect, his model accommo-
dates both the education-based arguments and HT, but he concludes that shifting
supply forces and changing employment probability perceptions are the main
forces driving inmigration.
This paradigm suggests that stable equilibria will occur at low rates of open
unemployment. Migration will exceed new job opportunities mainly when in-
formation is unclear-that is, when rapid economic changes are taking place.
Lengthy job searches will not persist, though, nor will random turnover give new
hope to those without the requisite qualifications.
1696 C.M. Becker and A.R. Morrison
Ya = - tx(P, y, t, ra),
where t is the commuting cost per km, P is city population, and y and Ya are
urban and rural incomes, respectively. This equation can be inverted and solved
for equilibrium city population P*. A standard stock adjustment framework thus
gives:
aP* P**
* aP*
P* =f[P*(Y, y, , rai - P0] > 0> (2.3)
ay ara a a ' t 't
where P0 is initial population. In contrast to other aggregate models, Brueck-
ner emphasizes rural land rents and urban commuting costs. In his formulation,
nominal incomes are exogenous in both sectors, though it should be noted that
equilibrium requires y > Ya, since urban residents bear additional commuting
costs. To estimate the model, a Cobb-Douglas agricultural production function
implies that r, is proportionate to output per hectare, while an international public
transit dataset exists for t. Unfortunately, neither of these variables are significant
in the cross-country regressions: urban/rural per capita income and lagged pop-
ulation dominate the regressions. But measurement of the rent and commuting
cost variables is difficult, and they may well be important.
for those not presently employed for wages. as well as an iterative procedure
to correct for simultaneity of hours worked and proportion of households with
a migrant away in urban areas. Bigsten's findings are striking, and somewhat at
variance with those from Asia and Latin America. Neither rural land pressure nor
local economic potential appear to influence migration: potential urban income
appears to be the overwhelmingly dominant force.
Other stylized facts from microstudies include the following. The incentives
generated by the household and community structures appear to matter enor-
mously. Outmigration often rises with education, but not invariably: in Ecuador,
the least educated young women tended to migrate, presumably to enter domestic
service and other jobs at the low end of the urban service sector (Bilsborrow et
al., 1987). Land holdings tend to have mixed effects as well. As with education
and origin income, greater family land ownership is associated with relaxed mi-
gration (and job search) financing constraints, but also with a higher opportunity
cost. Amenities associated with rural infrastructure have mixed effects as well:
their absence represents a "push" factor, but poorer infrastructure is associated
with greater remoteness, and hence greater difficulties in arranging migration,
especially if it is likely to be circular. Migration also tends to diminish with
rural job opportunities and marriage, and rises with the household's effective
labor supply (Bilsborrow et al., 1987). As Findley (1987) emphasizes, household
and community variables are strongly interactive, so that linear or even loglin-
ear specifications without interaction terms are likely to miss a great deal. Her
work on the Philippines suggests that outmigration will rise with agricultural
commercialization-except for those communities where social and economic
infrastructure is weak. She also finds (Findlay, 1987: p. 185) that in "communities
with little previous migration, the lower- and upper-class families are more likely
to migrate than the middle-class families, but in communities with much previous
migration, this differential is reversed".
In short, the microstudies typically offer little in the way of clear implications
for aggregate studies, much less policy measures. They do tend to find that in-
dividuals and households respond to incentives: it is the incentives which vary
enormously across contexts. The glib interpretation of the vast micromigration
literature (and consistent with a few studies, such as Bigsten, 1996) is that the
myriad variables and contexts represent much random noise that can be ignored
at aggregate levels-but this statement is one of faith rather than an empirical
assessment, and, as we indicate below, may well be wrong.
Ch. 43: Urbanizationin TransformingEconomies 1699
2.3. Analytical general equilibrium models
HT models spurred several attempts to generate high and low wage sectors within
a competitive environment. Three themes merit pursuit here: efforts to endoge-
nize wages, work which extends the HT model to include product markets and
international trade, and general equilibrium models based on monocentric city
models.
Endogenous wages: labor turnover, efficient search, and efficiency wages. The
endogenous wage literature in turn has several branches. By far the most influ-
ential is Stiglitz's (1974) seminal piece (also written in Nairobi) which advances
a "labor turnover" model. In this paradigm, an urban modem sector's total labor
costs depend both on the wage bill and on the costs T of training new workers,
since the existing force quits at rate q. This quit rate in turn depends on the
firm's wage relative to the expected urban wage, its wage relative to prevailing
rural earnings, and to the urban unemployment rate U-and is decreasing in each
argument. Thus, the firm's wage bill may be written as
Firms will take U and the general urban wage rate parametrically, and then
choose w, to minimize w*. The simple rule which follows is that "marginal
savings in turnover costs must be equal to the extra wage costs" (Stiglitz, 1974:
p. 199).
Assuming that modem sector firms can buy loyalty and that quitting is costly
to the firm, then firms will offer a wage above Wr (but not Ewu in equilibrium,
since firms are identical), and in consequence a natural urban wage premium will
occur-leading to HT-like unemployment, though under a more sophisticated
turnover regime than the random turnover assumed by Harris and Todaro. The
scope for unemployment is also more limited, since its presence directly inhibits
quits, which in turn means fewer new jobs, and hence less immigration.
Beyond the turnover equation and an intersectoral wage ratio
other firms. On the other hand, it is too low because they consider neither the
endogeneity of wr nor the effect of their actions on U. Stiglitz asserts that the
first effect is likely to dominate, though the matter is uncertain theoretically. In
any event, the socially optimal urban wage will still exceed the rural wage.
The labor turnover perspective moves attention from institutional forces which
cause high wages, and instead focuses on labor market characteristics. If an
economy is highly dualistic in the sense that modern sector production involves
substantial accumulation of job-specific skills which cannot be acquired outside
the firm, then w,, will be high. It seems plausible that quits will be less sensitive
to market conditions during periods of rapid economic growth, while high profits
in those periods may also raise the opportunity costs of quits to the firm in terms
of foregone profits (captured in T). During stagnant periods, on the other hand,
urban unemployment might even fall as q approaches a lower bound, especially
if Wr is low and falling.
Turnover rates in transforming economies are often very low, especially in
African countries for which the models were first formulated. East Africa in 1970
may have had turnover rates comparable to those in the US (Collier and Lal,
1980: p. 42), but more recent evidence (for example, Knight and Sabot, 1990:
p. 171) suggests that these rates have fallen greatly, presumably in response to
economic difficulties and the rapid increase in educated labor.
A second reason for endogenous intersectoral wage gaps is "expense pref-
erence" by managers. The expense preference model (Williamson, 1964) posits
that the payment of higher wages to workers allows managers to reduce their
effort level, by reducing the need for monitoring, oversight and training. This
model has not yet been applied to urban labor markets of transforming economies.
Costly turnover and expense preference are not the only factors which may
impel firms to pay wages above workers' reservation prices. Bliss and Stem
(1978) argue that firms may pay additional wages to ensure that their workers
are physically capable of working-that is, they pay a "nutrition premium" (see
also Deolalikar (1988)). More broadly, Stiglitz (1976) proposes an "efficiency
wage" hypothesis, which claims that a worker's productivity rises with his or her
wages. This effect may stem both from greater capacity (some of the additional
income goes to more and better food, some may go to more reliable transport
to work, and some may go to education) and effort. Recently, Bencivenga and
Smith (1997) provide a model in which a formal sector wage generates under-
employment in a three-sector, two-labor type model. Put simply, formal sector
firms are compelled to offer both high and low productivity workers the same
wage, and cannot (perfectly) monitor performance. Migrants can either search
for formal sector jobs or work in the informal sector, where high quality workers
Ch. 43: Urbanization in Transforming Economies 1701
are more productive. An equilibrium which attracts some good workers must
also be one with some informal sector underemployment in order to compensate
for adverse selection and somewhat discourage low quality migrant workers. In
all cases, HT-like unemployment can emerge as a result of endogenous wage-
setting-as will rural-urban migration of those willing to endure a considerable
spell of unemployment in order to gain access to a high-wage job.
Endogenous wages: rent-seeking models. Knight and Sabot (1988) recognize that
public enterprises, which have dominated many African modem sectors for the
past 25 years, are under pressure to provide employment regardless of needs. But
one should not regard this as limited to Africa: evidence that large enterprises are
pressured into employing an above optimal number of workers in order to limit
unemployment can be cited from Turkey (Knight and Sabot, 1988), the former
Soviet Union, China, Latin America, Korea and Indonesia 2 6 -indeed, it is hard
to list transforming countries where such pressure does or did not exist. 27 In a
small simulation model, Knight and Sabot (1988) show that an overly expansion-
ary government employment policy will actually exacerbate HT unemployment,
though under different labor supply conditions this finding could be reversed.
The next question is to ask what will happen if wages rather than public
employment become endogenous. Calvo (1978) modifies HT by assuming union
power in the formal sector, although only within the context of a small, open
economy (fixed product prices). In a similar vein, Becker and Morrison (1993)
provide a simple three-sector model of a stereotypical African economy which is
neoclassical in the urban traditional and agricultural sectors, and in which urban
modem (UM) employers act as rational profit maximizers. However, organized
labor is able to capture a share of UM value added; this share depends on govem-
ment's share in total output, the unemployment rate, the degree of protection from
foreign competition, the government budget deficit as a share of GDP (indicating
the public sector's financial flexibility), the size of foreign capital inflows (on the
assumption that foreign firms are more susceptible to political pressure), and the
balance of trade (as an indicator of the health of the domestic UM sector, since
access to imports is a key determinant of production and profitability).
26 Lumbantoruan (1997) analyzes frontier cost functions for state-owned and private industrial enterprises
(mainly textile firms) in Indonesia, and finds that there is no significant difference in efficiency levels: most
firms are highly inefficient regardless of ownership. This finding can be attributed to the fact that few firms
are truly private in the sense of being independent of government or immune to its pressures; for a variety of
political reasons, government has also encouraged firms to build capacity (and employ) way beyond market
needs, so that optimal production scale is rarely reached.
27 Privatization programs underway in many countries are changing this panorama.
1702 C.M. Becker and A.R. Morrison
Migration models: dynamics, product markets and trade. Bennett and Phelps
(1983) design a neoclassical two-sector, small, open economy model in which
households have money (or other liquid assets). Their model contains goods
markets' equilibria, fixed exchange rates but an explicit trade balance statement,
labor-leisure choices, exogenous urban labor demand due to an above-equilibrium
wage, and urban wage dependence on the cost of living. With an implicit gov-
ernment budget constraint, their framework provides an exceptionally lucid pre-
sentation of the impact of short-run macropolicies on a developing economy,
and in particular on rural-urban migration. Key comparative static results include
the finding that urban public works employment will have a total employment
impact less than one, and possibly less than zero. Export taxes on the rural good
affect the urban sector through a falling domestic price level, but of course hurt
the rural economy. A rise in tariffs on urban goods raises output prices but not
workers' living costs (consisting of rural food), so protectionism is bound to
expand cities. Devaluation raises prices of both urban and rural goods, but most
plausibly the net effect is to increase urban labor demand. Devaluation is also
less costly than the other measures considered (tariffs, rural export taxes, urban
Ch. 43: Urbanizationin TransformingEconomies 1703
employment subsidies and public works programs) in terms of rural jobs lost per
urban job created.
Becker and Morrison (1988) employ a model of an African economy that is
similar in many respects, building on the basic HT fixed urban modern sector
wage hypotheses. Unlike Bennett and Phelps (1983), they have neither money
nor an explicit labor supply choice; they do, however, build in a balance of
payments constraint, an endogenous exchange rate, and goods markets. Becker
and Morrison (1988) further include a traditional urban sector T, along with agri-
culture A and the urban modern UM sectors, and consider the consequences of
imported goods used as intermediate inputs rather than simply being competitors
with UM outputs. They also permit T goods to be used as intermediates by UM.
The model predicts an uncertain effect on migration from raising the UM wage
for standard Harris-Todaro reasons. Government demands are urbanizing, as are
foreign capital inflows. Trade effects are similar to those in Bennett and Phelps
(1983), except that protectionism plays an ambiguous role, since it both raises
UM costs and diminishes foreign competition.
The Becker-Morrison formulation is useful in part for explicitly laying out
the erstwhile implicit assumptions undertaken in single equation migration or
urban growth regressions. These equations can be generated as a reduced form,
but only from highly restrictive models. Becker and Morrison estimate the model
for a cross section of sub-Saharan African countries and find a very tenuous link
between urban population growth and UM employment growth. This result is
disturbing for all models which posit industrialization and modem sector growth
as the primary engine of urbanization. Indeed, while Becker and Morrison (1988)
do find that growth of government, GDP and the rate of urban capital formation
matter in expected ways, rural push factors also turn out to be very important.
During the past 25 years, a vast range of CGE models of developing economies
has been produced. Most, but not all, lack spatial content. Of those with spatial
content, some contain urban and rural locations for different sectors, but do not
seriously model spatially related forces. CGE models with detailed spatial struc-
ture appear to be limited to Kelley and Williamson (1984), itself an outgrowth
of Kelley et al. (1973), and Kelley and Williamson's successors, Becker et al.
(1992). The general structure of CGE models is well known, and need not be
repeated here. We instead focus on what makes the Kelley and Williamson and
Becker et al. models unique and the insights they present into the urbanization
process.
1704 C.M. Becker and A.R. Morrison
28 Only Mohan's (1985) smaller CGE model and Rogers (1984) produce urbanization rates close to the low
rates forecast by Becker et al. (1992). See Becker and Morrison (1993: p. 70).
1706 C.M. Becker and A.R. Morrison
shifts (which filter through to the Indian economy via Armington equations) con-
tinue to play important roles in Becker et al. (1992), they are of far less relative
importance than in the Kelley and Williamson (1984), RDC model. Factors cen-
tral to India's low urban growth rate include its low manufacturing TFP growth
rate, low rates of foreign capital inflows, rising imported raw materials prices,
and low rate of overall TFP growth. Modest savings rates do not help, but are
clearly exacerbated by low foreign investment rates and low capital productivity.
About one-third of the slowdown in urban growth following the early 1960s can
be attributed to unfavorable environmental conditions (including TFP change);
the rest is due to the large and undynamic agricultural sector.
The Becker et al. (1992) simulations contain several important messages for
students of urbanization. India appears to have become increasingly sensitive
to exogenous shocks over time, mainly because its structure is shifting toward
activities (urban manufacturing and modern services) capable of responding to
opportunities. Migration restriction has very small GDP impacts in the Indian
context, but does have the anticipated neoclassical distributional effects, and as a
whole is highly inegalitarian.
GDP is a poor measure of welfare, especially in fragmented and distorted
economies. Growth of government is unambiguously urbanizing and harmful to
the poor: governments consume skill-intensive goods and, depending on how
their growth is funded, may reduce capital formation. Sophisticated policies of
all types have unanticipated consequences which may swamp the direct effects.
More critically, the underlying economic environment crucially determines the
impacts of various policies: impact multipliers in 1981 India often differ strik-
ingly from those for 1960 India, and both differ even more from those in the
Kelley and Williamson small, open economy. Perhaps most importantly, though,
there is a clear policy direction for India-like economies: welfare of the poor
is most enhanced, as Mellor (1976) has long asserted, if agriculture and labor-
intensive employment creation policies are followed.2 9 Its urbanization impacts
are ambiguous, but its welfare consequences are not.
A final model worth mentioning is Adelman and Robinson (1978a, b). Para-
meterized for 1968 and historically simulated over 1968-1972 for Korea, it is
superficially similar in structure to Kelley and Williamson (1984). In practice
it differs greatly because of the greater degree of disaggregation (four firm size
groups in each of 29 sectors; 15 household categories; and 6 skill levels). With
a migration statement based on urban/rural wage gaps and competitive product
29 Job creation policies need not be activist; they may be as simple-and as effective--as a benign policy
environment which does not articificially lower the price of capital.
Ch. 43: Urbanizationin TransformingEconomies 1707
and factor price determination elsewhere, the model offers distributional insights
unobtainable from Kelley and Williamson (1984) or Becker et al. (1992).
A key finding of Adelman and Robinson concerns the distributional effect
of rural-urban migration. Had migration not taken place, the simulation 9 years
out (Adelman and Robinson, 1977) would have yielded a Gini coefficient of
0.598 and a bottom quintile income share of 1.2%-far worse than the baseline
simulation values of 0.457 and 5.0%, respectively. As they write (Adelman and
Robinson, 1988b: p. 21), "rural-urban migration is beneficial to the poor and
detrimental to the rich. It improves the distribution of income, increases overall
productivity of the economy, and reduces poverty". As in the other CGE mod-
els, migration restriction counterfactuals have little immediate impact on GDP,
but are highly inegalitarian. On the whole, though, migration unambiguously
raises GDP in Adelman and Robinson, since urban activities enjoy a considerable
productivity advantage over rural pursuits.
Virtually all recorded migration patterns show strong patterns by age, and fre-
quently by gender. Figure 1 provides unweighted average net urban inmigration
rates for 10 Asian, 13 African and 14 Latin American countries. 3 0 Inmigration
rates peak strongly for both genders, especially for African men and women. But
for all regions, the immigration rates for 15-29-year-olds are far larger than for
other age groups. Rural outmigration rates show a similar peaking pattern.
Figures 2 and 3 show urban inmigration and rural outmigration rates for one
country, Korea, at roughly 5-year intervals from 1960-1985. Korea exhibits a
peculiar double spiking for men at ages 15-19 and 25-29; the intervening lull re-
flects nearly universal, lengthy military service. For both sexes, the pattern is one
of exceptionally high migration, with urban inmigration rates gradually diminish-
ing for a given age group over time following the 1966-1970 peak. In contrast,
rural outmigration rates do not appear to be systematically diminishing: on the
contrary, for males aged 5-19 and 40-69 they peak during 1980-1985; and for all
females aged over 5, except those 25-29 and 75+, the highest recorded rate was
during 1980-1985. This difference between in- and outmigration rates reflects
the rapid urbanization of the country, coupled with high and rising outmigration
rates.
30 We are indebted to Christopher Grewe for generating the figures in this chapter, as well as for calculating
the migration rates which underlie them. In order to avoid making fairly arbitrary assumptions on region-
specific birth rate patterns, migration rates for those born during intercensal periods (age 0-9) were not
computed. Computation methodology by residual inference is discussed in Becker and Grewe (1996).
[708 C.M. Becker and A.R. Morrison
1(1-14 99=1
:
4ral J apan
Jal, 5
Korean
Femalle
7 Ne
tua
Japan 5
oe
aMale
Nep
7a Urban
1r
Q
I2
Age Cohort 6 74
Year
74
These findings should be taken as a first rather than last word, but a key point
emerges: different forces matter to different groups.
The lack of attention to demographic forces in understanding urbanization
derives neither from its omission in classic works (e.g., Sjaastad, 1962), nor from
lack of emphasis by economic historians or demographers. Nor, for that matter,
have past North-HollandHandbook chapters ignored the topic, since Williamson
(1988a) emphasizes migrant age selectivity. Williamson (1988b) emphasizes the
"young adult bias" of England's cities during the industrial revolution, and their
parallel relatively low dependency ratios. Furthermore, the age distribution of
migrants from the countryside to English cities was overwhelmingly of young
working age people: by the 1980s, two-thirds of all inmigrants were aged 15-
29. This age selectivity enabled urban labor force participation rates to rise, and
actually curbed total rural inmigration, both immediately, and in the longer run
via the inmigration of adults of child-bearing years.
Demographic forces are also important determinants of natural rates of popu-
lation in urban areas. Williamson (1988b) finds that in the 1970s for developing
economies as a whole, 29.4% of urban and 26.0% of rural populations were in
the 20-39 year age group, respectively. These differing age structures imply that,
even if urban totalfertility rates are considerably below rural TFRs, urban crude
birth ratesmay differ little from rural CBRs. In short, it is clear that age selectivity
plays a key role in urbanization patterns, both by affecting relative sizes of the
population at greatest risk to migrate and by compensating for rural-urban TFR
differentials.
The "new economics of labor migration". Rural-urban migration thus far has
been presented largely as an implicitly permanent event undertaken by full-time
workers, accompanied perhaps by their families. We know, however, that much
migration takes place for quite different reasons, and that the relevant decision-
making units may be large, extended families rather than individuals. By now a
wealth of literature on the migration decision-making process at the microlevel
has emerged, and it can only be cursorily summarized here (as noted above, the
Lucas (1997) survey treats this topic in far greater detail, as does Massey's (1990)
survey). The economist perhaps most responsible for the "new economics of
labor migration" is Stark (see Stark (1991) for a collection of major papers, Stark
and Bloom (1985) for an overview, and Katz and Stark (1986b) for a formal
statement of rational low-earnings migration due to risk aversion), though the
new paradigm may be traced to Stiglitz (1969).
Ch. 43: Urbanization in Transforming Economies 1713
Finally, Lucas and Stark have emphasized that households are most likely to
fund those who have already demonstrated altruistic behavior beforehand, as this
reduces enforcement costs. Conversely, children are more likely to send large
remittances to families which have demonstrated altruistic behavior in the past.
Such mutual altruism reduces contract enforcement costs (and perceived outcome
variance)-in a sense, it defines a family.
31 Lauby and Stark (1988) conjecture that migrants have poorer access to information and network capital
needed to successfully start one's own business.
1716 C.M. Becker and A.R. Morrison
women's migration in the coming decades will be quite different from that of the
past half-century.
To the extent that the different components of SAPs have distinct impacts on
urbanization, we should not expect the effect of SAPs to be identical across
countries. Third, while SAP policies themselves are important, the outcomes
produced by SAPs may be more crucial determinants of urbanization. Thus, the
decline in wages produced by SAPs (see Section 2.1.3) is likely to have a more
important impact on urbanization than either trade reform or the reform of public
enterprises, though these policies will themselves affect wages. Finally, and quite
soberingly, there is almost no empirical literature which examines the impact of
SAPs on urbanization rates. The reasons are simple: the research would of ne-
cessity need to be cross-sectional in nature, and within this structure it is difficult
to: (1) develop cross-country measures of the strictness of SAP programs; (2)
to control for other factors that shape urbanization rates; and (3) interpret the
impact of SAPs, given that dependable urbanization rates only can be calculated
over intercensal intervals, which typically last much longer than a SAP.
34 Counterfactual exercises conducted within the framework of CGE models also imply that SAPs will slow
urbanization, since they show that closed economies tend to have slower urbanization rates than more open
economies (see Section 2.4).
35 Much of the impact of SAPs on urbanization occurs because of SAPs impact on urban labor markets; for
a discussion of this impact, see Section 2.1.3.
36 As Portes (1989: pp. 33-34) notes, the implantation of an export-oriented growth model in Latin Amer-
ica has been accompanied by growth of industries outside of the region's large cities, such as commercial
agriculture, forestry, mining and product assembly.
37 This is true for most transforming economies. For those producing primarily manufactured exports,
however, export incentives will favor increased urbanization.
1720 C.M. Becker and A.R. Morrison
of the capital city, fiscal reform--to the extent it is successful-will make the
capital a relatively less attractive place to locate for firms and consumers. If, on
the other hand, the government's taxing agency has national reach, fiscal reform
will have little spatial impact.
The reform of public enterprises is often a key element of budget reform,
since the closure or sale of money-losing state-owned enterprises is one step in
reducing budget deficits. The closure of public firms results in immediate em-
ployment losses, and privatization is often preceded by reductions in employment
to make the firms more attractive to potential buyers. These firms are located
predominantly in the largest urban areas, so the employment losses serve to
make these areas less attractive places to live.3 8 Other elements of budget reform
include reduction of subsidies for food and public services; these actions also
have their principal impact in the largest cities. Agriculturalpolicy and exchange
rate reform are both designed to increase production of tradable goods, the first
specifically in the agricultural sector and the second economy-wide. Given the
structure of production in most transforming economies, these reforms likely
will slow urbanization. 3 9
It is difficult to gauge the impact of the decrease in investment (associated with
SAPs) on urbanization. To the extent that the investment is public investment and
there is urban bias in investment expenditure (see Section 3.2), the decrease in
investment spending may slow urbanization. 4 0 The effect of a decrease in private
investment on urbanization depends on the spatial distribution of private invest-
ment. Since data on the spatial distribution of private investment in transforming
economies are rare, it is impossible to make any generalizations.
The improvement in the current account associated with SAPs is clearly ur-
banizing. Most rural areas are net producers of foreign exchange, while urban
areas are net users of foreign exchange.41 The improvement in the current account
loosens the import constraint on urban areas, both for producers and consumers.
38 It is important to note that the eventual impact on employment and incomes is larger than the initial
impact, since there will be multiplier effects on the production of other firms. Proponents of privatization
would argue that in the long run private ownership will lead to increased output and employment in privatized
firms.
39 Note that agro-processing is often an urban industry. In the African context, growth of agro-processing is
likely to stimulate the growth of secondary cities (Becker et al., 1994).
40 There is ample evidence that as public budgets are squeezed, capital expenditure is reduced to a greater
extent than current expenditure (Greenaway and Morrissey, 1993: p. 252).
41 The magnitude of the gap between rural and urban areas presumably has declined with the demise of
import substitution industrialization, though.
Ch. 43: Urbanizationin TransformingEconomies 1721
3.2. The urban bias thesis
As the name implies, the urban bias thesis (Lipton, 1976) describes ways in
which the government favors producers and consumers in urban areas at the ex-
pense of rural residents. According to the urban bias thesis, government policies
typically have been biased against agriculture in three ways: (1) macroeconomic
policy (mainly trade and price policies) distorted economic signals, raising non-
agricultural value-added above what it would have been at world prices; (2) gov-
ernments have allocated investment funds in urban infrastructure projects without
regard for the higher rates of return available in nonurban areas; and (3) public
employment has been expanded in urban areas-especially in capital cities-to
a degree not justified by any conceivable efficiency criteria (Becker et al., 1994:
p. 94).
Lipton (1993; p. 233), originator of the urban bias thesis, has stressed the
need to distinguish between two types of urban-biased policies: price policies
and expenditure policies. This distinction is crucial, since price distortions-
in the form of below-market prices for agricultural output, and above-market
prices for urban-produced goods-have been greatly reduced in recent years,
largely as a result of structural adjustment programs. But the mere reduction of
price distortions does not imply the elimination of urban bias: the government
may react by spending more money on goods and services produced in urban
areas, leaving urban bias intact (Braverman and Kanbur, 1987: p. 1180). Indeed,
Lipton (1993: p. 248) argues that price distortions are not the main source of
urban bias. He suggests that expenditure biases are exacerbated by an economic
structure in which multiplier effects-backward and forward linkages-are far
stronger in urban than rural areas, while price biases are compounded by rural
mark-ups due to poor transport, and limited competition among wholesalers and
retailers. In a sense, rural neglect rather than specific policies lies at the heart of
his indictment. 4 2 Nor can there be any doubt that health care and education are
considerably better in urban than rural areas nearly everywhere; there is also con-
siderable evidence that urban children are better fed than their rural counterparts
(see World Resources Institute, 1996: p. 10).
42 This claim, however, does not have clear empirical support. Lipton's thesis implies that urbanization
could easily stunt rural development, a proposition at variance both with neoclassical spatial economics and
Loschian central place theory. Shukla (1996: Chap. 10) relates rural nonfarm employment across districts in
Maharashtra to local urbanization, and finds a fairly strong positive connection, even controlling for measures
of rural infrastructure, capital, credit access, value of crop production, and Government rural development
expenditures. It is difficult to read her work without getting the sense that at least secondary cities spark rather
than stifle rural development opportunities.
1722 C.M. Becker and A.R. Morrison
But how can price and expenditure bias be measured? The measure of price
bias employed in some empirical studies is the ratio of local to world prices
for farm products. This measure has three serious problems. First, it does not
account for any price bias favoring urban-produced products. Second, as Timmer
(1993) points out, domestic prices for commodities tend to change less-and
less frequently-than do international prices; thus, much of the variation in this
ratio results from changes in international market conditions (including variation
in exchange rates), not from conscious governmental policy decisions. Third, the
mere existence of price controls on agricultural output does not necessarily lower
agricultural incomes, since these controls can often be side-stepped by farmers
who sell their produce in black markets (Bates, 1993: p. 227). Furthermore, as
Braverman and Kanbur (1987) illustrate, general equilibrium effects of price dis-
tortions may completely reverse the partial equilibrium effects. Expenditure bias
is no easier to measure, since without a well-specified social welfare function
there is no clear reason to prefer one spatial distribution of expenditure over
another on welfare grounds. Indeed, given certain conditions, urban bias may be
the socially optimal policy. 43
Given these severe measurement problems, the major contribution of this
literature is not to produce a precise measure of bias; rather, it is to focus at-
tention on the fact that macroeconomic and sectoral policies shape urbanization
patterns. In a distortion- and externality-free economy, urbanization must be
welfare-improving, since it reflects the utility-maximizing behavior of rational
economic agents. In such an environment, the optimal city size-much discussed
in the urban economics literature-is whatever size results from the maximizing
behavior of individual agents. Urban-biased policies introduced by governments
probably have led to more rapid urban growth than would otherwise have been
the case. 44
"Overurbanization" per se is not the problem, though; the problem is urban-
biased policies which respect neither static nor dynamic comparative advantage,
and consequently imply efficiency losses for countries that pursue them. Indeed,
43 Arnott and Gersovitz (1986), for example, build a model in which urban-biased expenditure is socially
optimal. because some public goods (e.g., education. medical care and electricity) are more costly to provide
in rural than in urban areas. Note, however, that different models would give quite different results. A model
which included congestion and pollution costs in urban areas, for example, would generate an optimal policy
rule involving rural expenditure bias.
44 Although, as Williamson (1988: p. 442) points out, no studies have documented empirically the link
between urban-biased policies and faster urbanization. Malpezzi's (1990) efforts to link urban growth to a
measure of price distortion found virtually no correlation, a finding consistent with Becker and Morrison
(1988). In addition, urban-biased policies need not always attract more migrants to the city. Policies that
artificially raise urban firms' profitability by subsidizing capital may result in less labor demand, not more.
Ch. 43: Urbanization in Transforming Economies 1723
perhaps the healthiest approach to the topic is the political economy rational
choice approach taken by Bates (1993) and, though couched in terms of social
class competition, Lipton (1993). They see the underlying problem as that of
unequal political power, especially in nondemocratic, nonpluralistic polities. In
this setting, the transition from urban-biased policies characteristic of developing
societies to rural-biased policies prevalent in advanced capitalist economies can
be explained by shifting political power (and coalitions), as well as declining
costs in meeting infrastructure and other needs of rural communities.
A large proportion of people inhabiting the world's poorer countries live in coun-
tries which experienced prolonged periods of communist rule. According to Pop-
ulation Reference Bureau (1997), some 1.733 billion people live in current or
former communist states; this number accounts for 37% of the 4.666 billion
inhabitants of the "less developed" world. While the economic aspects of com-
munism are in various stages of abandonment, the legacy continues to exert a
profound impact.
Urbanization has been a quite different experience under communism, which
has also given rise to urban patterns and problems dissimilar to those in
economies even with mild market orientations. The critical distinction is not
that urban growth has been abnormally slow, although it has been in Vietnam
and China until recently, but rather the manner in which it has been formed and
channeled. The communist structure has also greatly influenced post-communist
patterns. This section will examine first the former Soviet Union and then turn
to China; for reasons of space, eastern Europe, Indochina and Cuba will not be
discussed.
et al., 1997: p. 43); in much of the former USSR, crude death rates also exceed
crude birth rates. While immigration to Russia from other republics of the former
USSR has roughly maintained its population thus far, Andreev et al.'s projected
population declines in the next 30 years range from 2 to 32 million, depending
on the demographic scenario employed. Becker and Hemley's (1998) simulations
based on demo-economic forecasts give little reason to adopt the less pessimistic
scenario, barring a miraculous economic recovery.
Nor are these projected population declines likely to be evenly distributed.
Andreev et al.'s "medium" projections are for an overall 1995-2025 population
decline in Russia of 10.9%. While not broken down by rural and urban areas,
the largest declines are for the heavily urbanized Northwest (-20.6%), Central
(-22.7%), Volga-Vyatsk (-15.4%) and Ural (-13.5%) regions. Although TFRs
in cities like Moscow and St. Peterburg are very low (1.01 for St. Petersburg: see
Safarova (1996)), it appears likely that inmigration spurred by their economic
recovery will ensure population growth. 4 5 Rather, the greatest population losses
are likely in: (1) northern and far east industrial cities (ASW (1997: p. 27) report
that 15% of the population left Magadan oblast in 1994 alone); (2) defense cities
such as Novosibirsk, St. Petersburg and Ekaterinburg, where demand for core
products has vanished; and (3) heavy industrial "company towns", that produced
goods uneconomically or for which demand has collapsed.
Ethnic-based migration began in a significant way during the mid-1980s,
when the first decentralization of Soviet power was accompanied by growing
regional nationalism. It accelerated following the break-up of the Soviet Union,
and was especially great in the republics on the periphery of the Soviet Union.
Unsurprisingly, outmigration was greatest from the heavily ethnically Russian
large cities of these republics. In Kazakstan (United Nations Development Pro-
gramme, 1996: Chap. 3), ethnic Slavs, Germans and Tartars make up 90% of
emigrants.
Nor is this emigration unique to countries like Kazakstan and Uzbekistan,
where strongly nationalistic policies are being imposed. The capital of the
Kyrgyz Republic, Bishkek, saw its population decline from 640 thousand in 1992
to 594 thousand in 1995, a loss of 7.3% of the total population (and 14% of
the Russian and 63% of the German populations, respectively: see Natskomstat,
1995).
45 The process will nonetheless be traumatic. Safarova's (1996) simulations for St. Petersburg for the period
1995-2005 indicate a 5.2% population decline for St. Petersburg under an optimistic scenario. But this decline
occurs almost entirely among children under 15 (indeed, the elderly population continues to grow), who are
projected to fall by 32%, and hence from 19% to 13% of the population.
Ch. 43: Urbanization in Transforming Economies 1725
Kazakstan emigration, 1993-1995.
1993 1994 1995 1996
At the same time, deteriorating urban living conditions and job opportunities
appear to have given rise to net urban outmigration to rural areas (Willikens
and Scherbov, 1995: p. 217). This pattern is clearly noticeable during the period
initially following the collapse of the Soviet Union; whether it has continued in
more recent years as urban food supplies have improved remains to be seen.
Soviet housing policies were dictated by the principal that investments should
go first to industry and defense, leading to appalling crowding and low quality
throughout the Soviet period. Populist efforts to construct more mass housing
began under Khrushchev and were followed by the construction of sturdier but
depressingly homogeneous and seemingly endless apartment complexes on the
outskirts of large Soviet cities during the Brezhnev era. The construction spurt
at the end of the Gorbachev era was marked by deteriorating standards as the
economy weakened.
Because housing was in chronically short supply and was tied to one's job,
labor mobility was severely constrained throughout the Soviet period. So was
urban fertility: based on cross-oblast regressions, Becker and Hemley (1998)
have estimated that an increase in 15 m2 of living space (a one-third increase
in mean urban apartment size) would lead to an extra birth per household.
Failure to allocate land via a market mechanism also led to growth of Soviet
cities in concentric rings, with vast amounts of unused land throughout the city. A
very low proportion of built-up area is used for residential land. In contrast, vast
amounts of even the largest cities are used for industrial sites (31.5% of Moscow,
as compared with 5% of Paris) and railroad lines. In large part due to the absence
of an efficient land market, obsolete industrial areas rather than residential areas
dominate land use just beyond the city center. Because housing construction is
determined by enterprise wealth and success in obtaining space, housing and
land consumption per person do not decrease exponentially with distance fi-om
the city center. Residential densities do tend to fall, but not consistently, with
distance from the center (Bertaud and Renaud, 1994; Bertaud and Kolko, 1994).
Such patterns obviously add to infrastructure costs, especially in a society
where urban heating and hot water is centrally generated and then piped to apart-
ment complexes throughout the city. Excellent but very costly metro lines in
many large cities have enabled population dispersal, but further growth will re-
quire further large infrastructure investments.
There are indications that market forces are beginning to have an impact in
the post-Soviet era. Maddala et al. (1993) examined Moscow land price gradi-
ents, and found that they were becoming increasingly steep over time. With the
privatization of housing stock, emergence of significant private housing construc-
tion, and the growing disassociation of housing from employment, normal rent
Ch. 43: Urbanization in Transforming Economies 1727
gradients seem likely to emerge in the former Soviet Union. On the other hand,
mechanisms for land transfer remain seriously stunted, as do efforts to recover
the costs of social infrastructure provided by local governments.46
There is certainly evidence of rapidly appreciating property values in real
terms in desirable areas of major cities (for example, see Kaganova and Berezin,
1993). Indeed, it seems likely that these prices exceed long run equilibrium val-
ues, since increases in the housing stock remain constrained, albeit now more by
the tiny new levels of housing construction and existing shortages rather than by
prohibition of housing sale and transfer.
46 As of 1989, only 2.5% of Soviet household incomes were devoted to apartment rents or "communal ser-
vices" (Renaud, 1993: p. 15). Comparable figures elsewhere in eastern Europe ranged from 1.0% in Romania
to 7.1% in Hungary.
1728 C.M. Becker and A.R. Morrison
quite clear, though, as is the fact that urbanization has been forcefully suppressed
in China. 4 7
The fascinating finding by Young and Deng (1997), however, is that the po-
litical regime did not matter directly. Rather, estimating a disequilibrium model
of supply and demand for urban labor, they find that food supplies have been
the critical determinant of urban growth (and these, of course, have been greatly
influenced by political policies). They find as well that agricultural sector con-
straints rather than industrial demand have been the driving force in population
movements throughout the communist period; the real achievement of the Deng
era, in that light, has been to create agricultural surpluses sufficient to enable
rapid urban and industrial growth.
The emphasis in recent years on rural industrialization and spatial decen-
tralization (following, and at times inconsistent with another policy of favoring
coastal cities) has resulted in the surging growth of new urban areas.4 8 Migration
patterns have been affected in curious ways as well. Zhu (1996) examines rural
outmigration at the household level, and most of the results for China are unex-
ceptional. 4 9 But migration is actually more likely for people who are registered
as agricultural workers; even though they cannot hope for an urban grain card,
they also have even more relatively constrained opportunities at home. Goldstein
et al. (1991) find that migration to Shanghai (based on the 1984 Shanghai Float-
ing Population Study), arguably the most desirable location in China, is mainly
47 The most definitive account of the roles of reclassification and rural-urban migration in affecting Chinese
urban population growth is given by Goldstein (1990). The 1987 1% National Survey reported an increase in
China's urban population share from 20.6% in 1982 to 46.6% in 1987 according to Goldstein (1990), and from
20.8 to 36.9% according to Yi and Vaupel (1989); the 1990 Census, however, reported an urban share of only
26.4% (see Chang (1994)). Despite the presence of a large unregistered population, this smaller figure seems
more plausible (World Bank 1997 estimate for 1995 is 30%). The 1984 reclassification led to an increase in the
"agricultural" population of cities from 33% in 1982 to 50% in 1987, indicating that many of the new urbanites
were in fact peasants-although, of course, many people designated as "agricultural" were engaged in urban
services and manufacturing. Nonetheless, as Goldstein terms it, there undoubtedly has been a "ruralization"
of China's cities. An extreme case is that of "the newly created city of Zibo, in Shandong Province ...Zibo's
1987 population of 2.4 million was 66% agricultural .... These 2.4 million are all counted as part of China's
1987 urban population defined by residence within city boundaries". (Goldstein, 1990: p. 677). Ma (cited in
Ma and Fan, 1994: p. 1640) estimates that boundary enlargement accounted for 49.5% of 1982-1987 urban
population growth. and migration accounted for 46.3%--but that urban population grew by only 46 million,
a figure vastly different from the increases reported by Goldstein of 116 million in cities and 181 million in
towns for the same period.
48 Perhaps even more than Soviet planners, Chinese planners have been taken by notions of growth poles
and leading regions. Since the Deng era, special economic zones have also been in favor; in effect, this gives a
rationale for favored treatment to coastal regions.
49 Zhu (1996) finds that the likelihood of migration is greater for men; it rises with education, number of
relatives in the city, family size and per capita income, land ownership (an asset measurement), and the level of
,illage socioeconomic development; it rises and then falls with age it declines with the household dependency
ratio, marriage, the number of people engaged in fanning, and the extent of village industrialization.
Ch. 43: Urbanization in Transforming Economies 1729
50 Restricted migration, young workers confined largely to members of the "pendulum population", and
low fertility rates imply that populations of cities like Shanghai (where TFRs have ranged from 1.1-1.3 for the
past 20 years) will age very rapidly in the years to come. Chen et al. (1997) forecast that by 2025, one-third of
Shanhai's population will be over 60 years of age.
51 Both these workers and migrants to Jiangsu's towns appear to be motivated overwhelmingly by work
prospects. Ma and Fan's regression analysis based on a survey of 190 towns in Jiangsu Province indicates
that land scarcity is an important push factor, while employment prospects in township and village enterprises
(TVEs) are the major attraction - especially as this employment does not require state approval. The combi-
nation of greater stringency in state control of large cities and explosive TVE employment growth has led to
more rapid growth of China's towns and smaller cities than its largest cities in recent years, even discounting
definitional changes.
52 Oi (1993) reports a mean urban:rural income ratio of 2.42 for China, well above many other east Asian
countries. Beyond wages, she also reports that subsidies account for 39% of Chinese urban income (Oi, 1993:
p. 144).
1730 C.M. Becker and A.R. Morrison
greatly in recent years, causing recent urbanization and urban growth rates to be
understated.53
But even given this understatement, the limited extent of rural-urban mi-
gration since the 1978 reforms until the 1990s remains striking. The 1987 1%
survey, which asked migration questions for the first time, found that only 30.5
million people moved residences in the preceding five years. implying that an
annual permanent migration rate of only 0.6% (Ma and Fan, 1994: p. 1630)-
and, in their study of Jiangsu Province, they find that only 48% of these migrants
are from villages to towns or cities). The 1990 national census also collected
migration information, as did the 1988 Two Per Thousand Fertility Survey, that
covered the floating population as well. This latter survey enabled Liang and
White (1996) to infer interprovincial migration rates as far back as 1950, and
confirms low long distance mobility (the 1965-1984 average gross migration
rate was 0.3%). This rate rose to 0.43% in the latter half of the 1980s, while
1985-1988 gross inmigration rates for economically surging destinations such
as Sichuan (9.7%), and Shandong (8.0%) were much higher. 54 Liang and White
also confirm the virtually universal pattern that young adults have the highest
propensity to move: the distinguishing feature of China has not been migration
motive, but rather its government's ability to suppress much of the drive. Figures
from the 1995 1% survey, however, indicate an acceleration in rural-urban (city
and town) migration: the data from Zhang et al. (1997) imply an annual net
inmigration rate of 1.45% for 1990-1995. Chan and Yang (1997) however, argue
that the floating population has probably risen from about 50 million in 1990 to
90 million in 1995, implying a gross urban inmigration rate of about 2.4%, and a
net inmigration rate of about 1%.55
53 Recent migrants, primarily of the floating variety, also differ from base populations (and probably ealier
migration streams) in terms of gender, especially to towns. Historically, male workers were allowed to move
place of employment, but could not take their families out of rural areas. This restriction was relaxed with
respect to towns in 1984; the 1982-1987 sex ratio (100*females/males) of town migrants was only 81, as
compared with a sex ratio of 107 for migrants to cities, and a base town population sex ratio of 116 (Goldstein.
1990).
54 However, the 1987 1%eNational Survey (which defined residents as those who reported having lived in
a location for more than six months, or who had official registration) indicates that 80% of migrants move
intraprovincially (Goldstein, 1990). This inclusive survey finds that 1982-1987 net migration to cities was
5.8 million, and to towns was 7.2 million. These imply annualized inmigration rates of 0.8% for cities and
2.3% for towns, respectively, using base populations (which one recalls are vastly different than terminal
populations because of definitional changes). These data suggest that migration to towns has been an important
phenomenon (and has probably increased in the 1990s). Migration from cities to the countryside is rare, since
it would entail potential loss of urban registration. itself a valuable asset.
55 One should not discount the possibility that these large estimates are nonetheless vast underestimates.
Rawski and Mead (1998) discount census figures and official data on agricultural labor force, and in place
construct estimates of farm employment based on crop area and estimated labor inputs. by province. Their
Ch. 43: Urbanization in Transfonning Economies 1731
Some of China's urban problems are similar to the Soviet ones; others are
not. Like the Soviet Union, much Chinese urban land was not allocated by any
market or efficiency-based principles, although a much larger proportion of new
growth involves costly land acquisition in China. As in the Soviet Union, China
starved its cities of infrastructure and, until recently, such infrastructure as was
constructed was planned and implemented by the central government. This policy
greatly constrained urban growth. In the past two decades, however, significant
decentralization of power has enabled local governments to undertake infrastruc-
ture projects of their own. 56 Their response to decentralization of authority and
responsibility has become gradually more effective. Initially, much land was
handed out virtually freely (for a detailed study of Shenzhen, see Zhu, 1994); un-
surprisingly, the consequences were local government revenue shortages, excess
demand for land, and low utilization rates of new allocations. Indeed, in Shen-
zhen, industrial land "hoarded" exceeded industrial land developments around
1990; obviously, incentives for corruption in such a system were enormous. Need
for local government revenues and aversion to giving land to foreign investors,
however, today appear to be generating price-based land allocation systems in at
least part of China.
Oi (1993) argues that relative rural living standards have risen with economic
liberalization, especially at the outset of the reforms, as they were felt first in the
countryside. This rural growth does not signify movement away from an "urban
bias", however-indeed, she does not perceive a deliberate bias toward cities or
countryside, but rather a need by the Chinese Communist Party to maintain a
prosperous and complacent urban population. Rather, the "bottom-up" revenue-
sharing system and abandonment of the compulsory unified grain procurement
system (because by 1983-1984, record harvests depressed market prices below
the state price) generated an environment of local autonomy and relaxation of
rural controls. Pressed for revenues, local governments pushed the development
of "township and village enterprises" (TVEs) rather than agricultural expansion,
however, leading to exceptionally rapid growth of industries in rural China. Thus,
despite decollectivization, net income from grain production appears to have
figures. and similar Chinese ones, suggest that 75-100 million fewer people are employed in agriculture than
the 330 million reported for 1993. Official data at the same time appear to markedly undercount employ-
ment in construction, transport, and trade, both rural and urban. Moreover, some "rural" areas may well have
sufficiently high densities of nonagricultural activities that they would be more accurately classified as urban.
56 As Wang (1994) details, local governments have responded by requiring developers to provide "rent-in-
kind" as they build new factories or housing. The ultimate incidence of such quasi-taxes is hard to determine,
since government officials grant rent-generating privileges to developers at the same time, and since many
construction projects are contracted on a cost-plus basis. One clear consequence is that it provides local officials
with a powerful reason to encourage local growth. since existing enterprises and households cannot be similarly
taxed.
1732 C.M. Becker and A.R. Morrison
fallen during 1984-1988 by perhaps 15% (Oi, 1993: p. 135), while by 1988 the
majority of ruraloutput was industrial rather than agricultural, and 24 percent of
the rural labor force was employed in TVEs. To reiterate, it is clear that "rural"
and "urban" designations in the Chinese contexts have quite different meanings
than elsewhere.
Chinese cities are not as residentially decentralized as post-Soviet ones, partly
because of lower industrialization and greater land pressure, but mainly because
commuting methods (mainly bicycles) necessitate higher densities and proxim-
ities to jobs. Consequently, as Bertaud and Kolko (1994) note, Chinese cities
have structures similar to those in market economies, though for quite different
reasons.5 7
A distinct feature of Chinese development, continued in the reform era, has
been the relative autarky of its provinces. Due partly to poor interregional in-
frastructure, partly in reaction to national pricing policies which sought to achieve
"primitive socialist accumulation" (in Preobrazhenkii's words) by extracting sur-
pluses from the producers of primary products, partly in reaction to self-
sufficiency drives, and partly because of trade and factor movement restrictions
(Cai, 1997), China's regions and cities are far less integrated than one would
expect."8 In consequence, cities are also less specialized (Henderson, 1988a): in
contrast, Soviet cities were probably overly specialized.
As China continues to develop rapidly, it seems likely that its overall urbaniza-
tion and urban structure will increasingly resemble those of other poor but rapidly
industrializing market economies. Economic growth will not remove all of the
existing inefficiencies, but it will make their consequences less pronounced.
respect to land. This section focuses strictly on the problems associated with land
use imperfections: housing issues are the subject of Malpezzi's work (Chapter 44,
this volume).
To start with, governments' roles in urban land markets tend to be terribly
confused. Justifiable rationale for government intervention are to correct market
failures, to compensate for externalities, or to achieve welfare redistribution (for
a nice statement, see Dowell and Clarke (1991)). But while there is much need
on all three counts, government interventions are rarely welfare enhancing: on
the contrary, there is substantial evidence that in much of the world, they make
all three problems worse.
Even in the absence of government intervention, urban land markets suf-
fer from severe problems. Perhaps most critical is an absence of well-defined
property rights for households and businesses. In many transforming economies,
multiple property right systems coexist simultaneously in an ill-defined man-
ner. Formal (Western) property registration is often restricted to upper income
housing and formal sector enterprises; it may coexist with traditional, indigenous
systems (sometimes codified legally), and almost certainly much housing and
small businesses will be governed by extra- or semi-legal arrangements which,
nonetheless, have fairly widespread acceptance.
Somewhat surprisingly, the economist's obvious answer to the problem of ill-
defined property rights (register everything quickly and clearly) is not always
appropriate. Formal land registration is expensive, especially if done precisely,
and may be perceived by many as a mechanism for alienating them from their
land and dwellings. Management design strategists suggest that adding more
legal authorization to existing semi-official institutions for resolving disputes and
authorizing these bodies to define titles and property rights to units when owners
wish to sell will have more relevance and be more readily accepted.
A thorough survey of urban land policies is provided in Farvacque and McAus-
lan (1992). They detail the painfully difficult journey toward legal property status
which ends up effectively excluding all those without considerable resources
and patience. They also discuss several reasons for unclear land titling and the
absence of straightforward policies (beyond self-interested supply restrictions
by powerful officials). Skilled labor constraints surely matter: legal systems in
many countries are nearly paralyzed by excess demand, and at the same time
government capacity is often severely limited. Centralized management in much
of the developing world adds to skill constraints by requiring the involvement of
many layers in making land use (or other) decisions. Registration and develop-
ment authorization thereby becomes needlessly time consuming and civil servant
labor-intensive. Complex and contradictory laws subject to sporadic enforcement
1734 C.M. Becker anud A.R. Morrison
add to difficulties, as do weak, small-scale capital markets. Unless one has a great
deal to gain by formal authorization and has considerable resources to wield,
the consequence is likely to be failure to obtain the required authorization-
necessitating recourse to informal means of acquisition. But manufacturing and
construction enterprises thus constrained are also unlikely to grow beyond a mod-
est size, thereby restricting economic growth of potentially dynamic firms. The
powerful and well-connected, on the other hand, will be tempted to seek "spe-
cial" authorizations and concessions, thus circumventing the impractical formal
process.
Poorly defined property rights and ill-designed government interventions are
astonishingly costly. Poorly definedproperty rights ensure that lower value struc-
tures are constructed, thereby reducing densities, which in turn increase total ur-
ban space requirements, consequently raising unit infrastructure costs, increas-
ing commuting costs and diverting land from other valued uses. This problem
is especially acute in Africa (where up to one third of urban households engage
in some farming: see Becker et al., 1993), but is pervasive elsewhere as well.5 9
Just how great the costs of misallocated land and uncertain property rights are
is difficult to determine, and a comprehensive study remains to be done. Dowall
and Clarke (1991: p. 2) cite a World Bank estimate that land use and housing
restrictions cost Malaysia about 3% of its GDP annually, and Malaysia is hardly
the worst offender-nor are all of the costs enumerated in this section included
in that study. Data from Kampala, Uganda indicate that in 1992, some 56% of
total city land was devoted to agriculture (Maxwell, 1995); moreover, since only
20% of urban farmers owned the land they tended, input utilization and crop
choice were suboptimal. Moreover, because of the dependence of the poor on
urban farming for food supplements (and in Kampala there is a strong linkage be-
tween urban farming and nutritional status of children in lower income classes),
efficiency-oriented measures may have disastrous social consequences.
As Aina's (1992) study of Lagos finds, the exact implications of a given mix
of land tenure laws tend to be complicated. Nigeria sought to merge colonial and
customary legal practices in the Land Use Act of 1978, which, not surprisingly,
was virtually inaccessible to the poor. Its practical consequences seem to be to
59 Rakodi's (1988) study of Zambia reports similar figures. and also that in some squatter areas, roughly
75% of squatter households have plots. While the income from household plots only amounted to 10-15% of
food budgets, these savings were important for the poor. Maxwell (1995) in his study of Kampala, Uganda,
found that 35% of households engaged in agricultural production. As in Zambia, urban farming in Uganda
is overwhelmingly done by women in order to enhance food security: only 2.5% of Kampala's farmers were
commercially oriented. Maxwell's study also finds that the likelihood of farming rises with a household's
length of residence in the city: it is not an activity particularly associated with recent migrants, who in fact are
likely to lack access to convenient plots.
Ch. 43: Urbanizationin Transforming Economies 1735
add legal authority to the already considerable power of landlords in poor neigh-
borhoods, which in turn has brought additional tenure insecurity to low income
neighborhoods. In cities with rapid economic change but still incomplete prop-
erty rights, landowners may find it useful to permit high density but impermanent
housing on large tracts until they develop the land for "formal sector" purposes.
This pattern is not inherently inefficient; rather, the inefficiency arises because
land markets are not well enough established to permit lower income groups to
obtain long-term contracts. As the landlord is, in effect, the enforcing agent as
well, there is no mechanism that permits poor tenants to enter into long term
contracts with confidence. One therefore finds scattered neighborhoods of very
high housing density and very low quality, even though total urban density is
likely to be inefficiently low.
Jimenez (1985) develops a model of land tenure choice that incorporates the
cost that insecure tenure imposes on urban squatters. In his model, both squat-
ting and nonsquatting households maximize an indirect utility function which
contains income and the price of housing. Squatters have two possible budget
constraints: one if they maintain possession of their land, and the other if they
are evicted by the landowner. If they are evicted, they face the additional costs of
moving and obtaining housing in the formal sector. Squatters maximize expected
utility, which is a weighted average of the utility they will obtain in the two
states of the world (eviction and no eviction), where the weight is given by the
probability of eviction. Squatters are able to lower the probability of eviction by
engaging in (costly) community organization.
A second model by Hoy and Jimenez (1991) formalizes the disincentive to
improvements in housing stock caused by insecure tenure. The model provides
insight into squatters' decisions whether or not to improve their dwellings and
landlords' decisions whether or not to evict squatters. In this model, the stylized
sequence of events is:
(1) Squatters invade a plot of land. (2) The landowner chooses a value of de-
terrence activity (to deter investment in housing stock, since a more valuable
housing stock is more expensive to remove) and a ... rate of eviction. (3)
Squatters choose a level of housing stock conditional on the cost of eviction
and the expected eviction rate. (4) The state of the world is realized. (5)
The landowner implements his eviction rate (Hoy and Jimenez, 1991: pp.
82-83)
1736 C.M. Becker and A.R. Morrison
The principal implication of this model is that the Coase theorem is not appro-
priate in the case of squatter settlements.6 0
Not surprisingly, the market price of squatter dwellings is lower than that
of equivalent dwellings with secure land tenure. Using data from a sample of
metro Manila households, Friedman et al. (1988) find that an average squatter
dwelling would sell for 23% more if it had tenure security.6 1 This finding has an
important policy implication: governments can create wealth for the poor simply
by providing secure land tenure.
Three superb papers on Jakarta (Bertaud, 1989; Dowall and Leaf, 1990; Fer-
guson, 1992) provide a detailed analysis of the spatial consequences of land
use policies there. As in Lagos, areas of very high density coexist with un-
derutilized space in the core city (DKI Jakarta). A segmented housing market
exists, in part because the process of titling land is extremely lengthy and quite
expensive, if one considers the required bribes. But formal registration is nec-
essary to undertake housing developments which can be funded by bank loans
or whose occupants can get mortgages from the National Housing Bank. While
poor neighborhoods and villages (kampungs) have benefitted considerably from
public upgrading projects in the past decade, they generally remain outside the
formal registration system, and tenure rights are much less secure than in the
formal sector. As in Lagos, the result is to discourage housing investment-in
effect, households and landlords substitute away from risky assets (unsecured
housing) to safer investments.
Insecure property rights also mean residents do not tend to be reimbursed at
full value for their land as it is acquired ("appropriated" might be more apt) for
development. Naturally, there is a considerable value placed on secure tenure,
as Dowall and Leaf find in their hedonic regressions for Jakarta. 6 2 Jakarta in
the late 1980s was characterized by a very steep land price gradient relative to
Bangkok, Karachi, Latin American or Western cities, largely due to poor road
infrastructure. Close access to a major road increased property values by about
Rp57, 429 m2 , or slightly more than 50%. Formal land titling further increased
60 The Coase theorem implies that the "observed rate of evictions is just right. Evictions (would) occur only
if the value of land use to squatters is less than the sum of the alternative development value to the landowner
less the costs of evictions ... since squatters would outbid any competing use of the land if indeed they valued
it more" (Hoy and Jimenez, 1991: p. 79). The Coase Theorem fails because of significant imperfections in the
market for land.
61 Older squatter dwellings are valued more highly than newer ones, holding dwelling characteristics
constant. Presumably age of dwelling is negatively correlated with eviction risk.
62 Garr's (1996) study of squatting in a cemetery in central Java also finds that investments in these dwellings
appears to rise with perceived security of tenure. Willingness to pay for metered water and electricity (quite
possibly in part for the implicit property rights thereby conferred) also coincided with expansionary policies
by the local public utilities, which appeared undeterred by clear lack of formal title.
Ch. 43: Urbanizationin Transforming Economies 1737
property values by Rp22, 060 m2 for low infrastructure and Rp32554 m2 for
high infrastructure land, or about 19.4% and 19.0%, respectively. Returns to
secure title appear to be especially great at the city fringe, where kampungs are
at greatest risk of being swallowed up. Given these empirical estimates, it seems
obvious that formal land registration and improved road networks would be of
great value to all classes. But titling would also restrict potential capital gains to
developers, a far more influential group than kampung residents.
As Ferguson (1992: p. 12) discusses, the process of property development
authorization in Indonesia further exacerbates land underutilization. Roughly,
a developer submits a site plan and development schedule along with financial
documentation. Once a development permit is granted, other developers cannot
acquire and develop land in the permitted area without the consent of the orig-
inal developer. Property owners on the site do not have to sell (though there is
"national interest" pressure to do so, especially on those without formal title) to
the redeveloper immediately, and, in practice, the process of conversion is often
a lengthy one. This gradual acquisition means that much land will sit idle or in
temporary uses for substantial periods while the developer amasses the financial
means to buy out all initial holders.
To the social costs thus created must also be added the costs of delays due to
the lengthy development approval process (2.67 years on average, according to
Ferguson (1992: p. 27)), formal and "informal" fees, and costs due to restricted
land supply. Ferguson's rough estimate is that all of these costs together inflate
construction project costs by about 34% in DKI Jakarta. None of these comments
should be taken as Indonesia-specific; Jakarta's patterns are typical rather than
atypical of cities in transforming (and many industrialized) countries. Rather, the
point here is that urban land policies have not contributed to poverty reduction.6 3
The literature on land use restrictions contains considerable recognition of
their rationality from the point of view of those agents who impose them (see
Dunkerley et al., 1980). As Tewari and Kumar (1986) note, the largest benefi-
ciaries of Indian rent control laws are senior civil servants. Tewari and Kumar
also suggest (but do not formally model) the possibility that rent controls may be
efficient in a setting of limited government capacity. In their study of Bangalore,
they find that eviction is difficult, but that both landlord and tenant have mutual
threats, since the court system is paralyzed. Thus, the setting encourages landlord
and tenant to create a cooperative game. Furthermore, the ease of bribing officials
63 In general, Indonesia has been one of the most successful countries in the world in terms of combating
dire poverty; land use policy just has not been one of the policy instruments employed.
1738 C.M. Becker and A.R. Morrison
and limited applicability of laws ensures that the expected net present value of
rental payments under rent control need not be greatly restricted.
The inefficiencies of rent control arise in other areas. The system ensures con-
siderable nominal price stickiness and locks people into housing for inefficiently
long stays. Sticky nominal rents also discourage landlords from making repairs,
and hence lead to accelerated housing stock deterioration. Potential arbitrari-
ness of government officials adds as well to the riskiness of housing investment,
thereby reducing total supply.
An issue periodically raised concerns government land banking (see Rivkin
in Dunkerley et al., 1980; Farvarcque and McAuslan, 1992). Some governments
have sought to acquire urban land in advance of development, with the aim of
selling it off subsequently at low prices to users whose objectives are consistent
with those of urban planners. Its main justification would be an absence of land
and capital markets, though the presence of substantial government intervention
here would seem to stunt development of the former. Land banking is obviously a
tempting cover for favoritism and corruption; to our knowledge, no studies have
been produced which attest to its efficacy. Hannah et al. (1993) provide a nice
study of an extreme case, Korea. They looked at five "development projects"
in which the government acquired (via eminent domain, but at market prices
given existing zoning) sites, provided them with infrastructure and rezoned them
according to a master plan, and then sold parcels off to developers. The dis-
counted return to this activity for the government ranged from 70 to 550%-but
the restrictive policies that make such returns possible are also responsible for
creating an average annual real (CPI deflated) 1974-1989 land price increase of
over 12% in Seoul.
It should be noted as well that not all market distortions have clear welfare
impacts. Consider the case of lot size zoning restrictions, which are common in
Pakistan and other countries. In an aspatial model, restrictions on land consump-
tion by the wealthy will cause land prices to fall and raise welfare of the poor.
But as Pasha (1992) elaborates, the consequences are indeterminate in a standard
spatial model. If the wealthy live in or near the CBD, then lot restrictions will
cause total city size to decline, lowering both commuting costs and rents paid
by the poor. But if the wealthy live on the periphery, then restrictions will cause
them to increasingly compete with the poor for central land, raising rents and
commuting costs, and therefore lowering welfare of the poor.
Land use patterns are determined by more than conscious government pol-
icy; as Ingram and Carroll (1981) emphasize, they are also influenced by the
age of the city and by transportation technology. Briefly, improved transport
infrastructure and rising automobile ownership enable suburbanization-while
Ch. 43: Urbanization in TransfonningEconomies 1739
the periphery. Nor, for that matter, is there any free-riding by the poor in the
Tiebout sense of consuming public services but not fully bearing the cost via local
taxes, since nearly perfect zoning is enforced. Rent gradients in the apartheidcity
decline, but only within racially homogeneous areas: in general, given total land
constraints, unit land rents were higher in distant historically nonwhite areas than
in white residential areas near the CBD.
With the initially gradual and then accelerating collapse of apartheid, spatial
patterns have begun to change. White suburbanization began even prior to the end
of apartheid,as growing white urban populations and increased automobile own-
ership simultaneously put pressure on expansion and provided the means for low
cost commuting from distant suburbs. With the breakdown of apartheid, pres-
sure for white (and wealthy nonwhite) suburbanization is enhanced, but removal
of other spatial restrictions may generate substantial industrial suburbanization,
possibly taking with it substantial nonwhite workers. Reductions in the fiscal in-
dependence of exclusive suburbs will decrease incentives for income-based racial
segregation, though it may accelerate the development of very distant private
enclaves. To some extent, the changes in South Africa are an extreme case of
what has been happening elsewhere, as wealthy and politically powerful groups
seek better living standards for themselves without having to pay for services for
all, as the Tiebout model predicts. In the coming decade, South Africa's cities will
lose many of their distinctive features; at the same time, spoke and radial free-
way systems will help other large cities in transforming economies acquire some
South African attributes in terms of increasingly effective economic segregation.
In concluding this section, it should be noted that ample evidence exists that
land prices are growing very rapidly in the cities of successfully transforming
economies. But the evidence does not indicate that land prices are growing at a
systematically more rapid rate in the largest cities (see Ingram, 1980), although
land prices are of course at a much higher level in big cities than in small ones.
Nor, within large cities, does it appear that land prices are rising most rapidly in
CBDs. On the contrary, Ingram's review of Korean and Colombian cities finds
a tendency for land value gradients to decline over time in every city examined,
dramatically in the case of Bogota. 6 5 Declining land price gradients are also a
common finding in Dowall and Leaf's (1990) study of south Asian cities. In
65 A detailed study of Bogota appears in Dowall and Treffeisen (1991). They focus in particular on the
emergence of subcenters beyond the CBD. They also find a substantial elasticity of substitution between capital
and land (cr = 0.69. which is toward the upper bound for estimates from the US), indicating considerable
variation throughout the city in land use intensity. Dowall and Treffeisen find that a multicentric approach does
not improve goodness-of-fit over a monocentric model when population density gradients are estimated, but
that it does improve when land price gradients are examined.
Ch. 43: Urbanizationin Transforming Economies 1741
most cases, this is driven by an expansion of urban infrastructure, which raises
the value of land on the periphery of large cities and in smaller towns, thereby
creating conditions favorable to urban decentralization.
are the ones that live furthest from the principal employment centers". Af-
fordable transportation is thus closely linked to poverty traps in Africa, south
Asia, and parts of Latin America. In large cities in transforming countries,
5-10% of household income is typically spent on transportation, and for the
urban poor this figure can be even higher (15-16% in the Calcutta and New
Delhi figures cited in World Bank, 1986: p. 7).
· Urban transport problems are greatest in the largest cities, since mean jour-
ney distance rises with size, as jobs and housing become more distant. Low
densities in small cities mean that parking and congestion are minor prob-
lems, while people tend to live on or near a small number of main roads,
which can be easily served by public transport. Densities and hence com-
plexities rise disproportionately with city size; hence, so do expenditures on
transportation. Nor is it surprising that urban transport problems are greater
in lower income megacities, since dramatic solutions tend to be unaffordably
expensive (World Bank, 1986), and since road space tends to be a smaller
proportion of land area (Transportation Research Board, 1997). Estimates
of regional GDP costs of traffic jams (which range from 0.4% in Seoul to
2.1% in Bangkok: see World Resources Institute, 1996: p. 25) show no clear
pattern against per capita income, since wealthier cities have better transport
infrastructure-but the lost time of commuters has higher value.
* Decentralization of large "megacities" to outlying areas and to secondary
cities is inhibited largely by fears of inadequate transportation and other
infrastructure (Hamer, 1984: p. 53). Decentralization of enterprises strongly
follows expanded transportation networks virtually everywhere (Choe and
Song, 1985: p. 41 provide a detailed discussion of this in the case of Seoul's
manufacturing enterprises).
a Despite disincentives via high costs of commuting and at times poor residen-
tial services, suburbanization of higher income groups is becoming increas-
ingly prevalent. Punpuing's (1993) study of commuting by employed people
in Bangkok finds that predicted commuting time rises strongly with years of
education (or home ownership, which is highly correlated with education and
income) and age. Surprisingly, commuting length and time were not affected
by gender. In effect, very high commuting costs in Bangkok appear to be
offset by highly income elastic demand for environmental amenities that are
found only on the outskirts of the city.
· Both intraurban and intercity transportation infrastructure is expensive, and a
major drain on national budgets. Many cities devote 15-25% of their budgets
to their transport systems (World Bank, 1986). Less construction-intensive
urban transport investments, however, have extremely high rates of return (in
1744 C.M. Becker and A.R. Morrison
the 25-75% range: Barrett (1988: p. 26)), even without fully accounting for
positive externalities.
* There is significant scope for realizing cost recovery for government expen-
ditures on urban transport systems. The main constraints, especially to user
fees, but also to vehicle and fuel taxes, are political. Demand management
steps such as higher vehicle and fuel taxes, road pricing, area licensing, and
parking restraints have long been advocated for traffic control reasons as
well. In contrast, transportsubsidies tend to be self-defeating, as they imply
greater excess demand, unsustainable budget deficits, and hence ultimately
reduced public transport services.
* Throughout the transforming world, the highest return public investment ac-
tivities regarding urban transport are those that raise the efficiency of bus ser-
vices. Measures that improve private automobile efficiency have the lowest
return (and are the most regressive).
* Poor financing structures mean chronic deficits and shortages for state-owned
bus companies. This results in huge subsidies in wealthier nations, or bus
systems in which a high proportion of the fleet lies idle for want of spare
parts.
Several points of road and traffic management bear mention:
* Vehicle operating costs tend to be far greater than road construction and
maintenance costs. Nevertheless, the lack of cost recovery now typical of
most transforming countries ensures that publicly-provided infrastructure is
not supported by the (mainly private) beneficiaries. Existing arrangements
thus virtually guarantee underinvestment in road infrastructure. Moreover,
maintenance typically suffers during recessions, so that road conditions tend
to be worst in slowly growing or declining economies. 6 6
* Urban road capacity can be enhanced in the short run by several inexpen-
sive measures: improving maintenance, segregating different modes of trans-
port, creating one-way street networks, integrating traffic signal schemes,
and eliminating street trading are among the important steps here. Improved
footpaths and sidewalks are also critical, since they keep pedestrians off
roads. At least in much of Africa, inefficient road use rather than lack of
capacity is the main cause of congestion (Barrett, 1988). These management
steps are all far cheaper than new road construction but traffic and parking
management require political determination.
66 As World Bank (1988: p. 18) notes, private trucking or bus firms may not complain about poor roads,
since the high depreciation rates which result also create barriers to entry from potential competitors.
Ch. 43: Urbanizationin TransformingEconomies 1745
* It should come as no surprise that traffic management steps actually taken
often tend to come at the expense of the poor. Traffic flow in Jakarta has
improved in the past decade, despite considerable growth in automobile own-
ership, following the banning of nonmotorized vehicles (bicycle-rickshaw
becaks and pushcarts) and animals from the streets. A more egalitarian policy
would have been to have restricted automobiles and encouraged more public
buses, but this would have been at variance with Jakarta's new sophistication;
moreover, such a noncar-friendly policy might have jeopardized the new ex-
press tollway and limited the market for a national car-projects both run by
President Suharto's children.
· Enhancing bus and paratransit capacity and speeds implies improved main-
tenance, creation of high-speed express bus networks, upgraded bus transfer
depots, construction of reserved bus lanes, and paving and drainage of pri-
mary roads to improve access to low income areas (Barrett, 1988: p. 17).
Subsidized bicycle purchase schemes may also have merit.
* Ultimately, traffic and transportation demand management is integrally re-
lated to land use management and automobile use policies. The recom-
mended design structure for megacities is one of a "regional polynucleated
development pattern, with relatively high density outlying nodes connected
to the center by high capacity road and bus service" (Transportation Research
Board, 1997). Low density sprawl is difficult to serve other than by individual
vehicles and invariably adds to mean commuting distances. Unfortunately,
as we have seen, there are powerful forces working against efficient density
levels in many countries. Indeed, developers rarely have to pay the associ-
ated infrastructure costs of new housing or other construction projects; since
these costs are rarely recaptured in property taxes either, there is an effective
government subsidy implied for new construction.
In short, in much of the transforming world, there are conditions that create
transport demand above a socially efficient level. Worse, they coexist with other
conditions that repress supply to inefficiently low levels. Worst of all, while so-
cially efficient moves would help the poor, they lack political support in much of
the transforming world; hence, more expensive and less efficient options tend to
be undertaken.
The existing level of public service provision is often quite low in transforming
economies. For the countries the United Nations characterizes as "low human
development", only 72% of urban residents had reasonable access to safe drink-
1746 C.M. Becker and A.R. Morrison
ing water, while only 55% had access to sanitary means of excreta and waste
disposal as of the early 1990s (United Nations, 1994: p. 149).
In any discussion of the provision of urban public services, an important
question is, "Who is doing the providing?" Is it the central government, the mu-
nicipal government, or private firms? The level of responsibility assumed by local
governments for public service provisions varies greatly. Municipal governments
in Bogota, Calcutta, Jakarta, Nairobi and Seoul have primary responsibility for
water supply, sewerage, primary education, as well as highways and roads. In
other cities, such as Tehran, Lagos, Lima and Kinshasa, municipal governments
have few responsibilities. 6 7 Not surprisingly, given the varied responsibilities of
local governments, the allocation of local expenditures across public services
also varies enormously. Expenditures on water provision range from 5.7% of the
municipal budget in Ahmadabad to 26.1% in Bogota; expenditures on education
vary from 1.1% in Cali to 29.3% in Madras; other public services show equally
large variation (Bahl and Linn. 1992: pp. 20-27).
Although the range of services provided varies greatly across cities, most
local governments do generate the majority of revenue to finance local pub-
lic services. Of the 22 municipal budgets examined by Bahl and Linn for the
1980s, only two (Gwangju, Korea and Tunis, Tunisia) had less than 50% of
total expenditures financed with locally raised revenue; the mean share was 72%.
These revenues are generated primarily through user charges, license fees and
locally-collected taxes; the property tax is the most important revenue instru-
ment. The share of locally-generated revenue has increased significantly in many
transforming countries, including China, India and South Korea (Kojima, 1992).
Not surprisingly, income elasticities of demand for water, sewerage service,
electricity, telephones, garbage disposal, education and health services are pos-
itive (Linn, 1982: p. 632). Thus, as incomes rise, so will the demand for public
services. Even in the absence of economic growth, population growth, especially
in large urban areas, may lead to increased demand for public services. Income
and population are not the only determinants of demand for public service; the
price charged for these services also matters. In particular, pricing public services
below the marginal cost of provision is problematic for at least two reasons: it
will limit the ability of public service utilities to undertake capital expenditures
that would expand service to more users, and below-cost pricing is an inefficient
67 In Tehran, local government-while having primary responsibility for highways and roads-has sec-
ondary responsibility for water and sewerage, and no responsibility for primary education. Lagos has only
secondary responsibility for water and sewerage; Lima has no responsibility for secondary schooling and only
secondary responsibility for highways and roads: and Kinshasa has no responsibility for water, sewerage or
primary schools.
Ch. 43: Urbanizationin TransformingEconomies 1747
way to subsidize the consumption of the poor, since higher income groups also
benefit from the subsidy.
Recent studies have also found substantial willingness-to-pay for urban public
services (for a brief survey, see Crane and Daniere (1997)). While not true in
every city for every good, WTP for many services is sufficient to induce supply
from small-scale private producers (where permitted). In these cases, substan-
tial gains to market liberalization exist (enhancing competition and enabling
producers to realize scale economies); elsewhere, returns to public infrastruc-
ture investments are often quite high. Even very poor households tend to have
considerable demand for public services. Altaf and Hughes (1994), for exam-
ple, estimate a mean WTP of 4% of household expenditure for sewer access
in Ouagadougou, Burkina Faso. As anticipated, WTP rose with household in-
come, education, and whether all housing compound members were relatives.
The Ouagadougou study also suggested that an effective and cooperative govern-
ment could expand supply by creating conditions for the emergence of private
firms which would invariably use far less sophisticated (but also far cheaper)
technologies.
These findings have been mirrored in southeast Asia. Crane (1994) reports
that only 20% of Jakarta's residents get their water from municipal connections;
the rest obtain their water either from groundwater wells, from public hydrants,
or, for fully one-third of the population (Crane and Daniere, 1997), from vendors
who resell hydrant water. In Crane's survey from North Jakarta (where ground-
water is increasingly saline, making private wells of little use), only 13% of
households enjoyed municipal connections: the vast majority bought water from
vendors, spending an average of 7.5% of their income in the process. Given an
expenditure elasticity of demand for water estimated at 0.03 and a price elasticity
of around -0.5, it is unsurprising that vendors and private hydrant franchise
operators (with high costs and considerable spatial market power) price water
far-more than 14 times-above the cost through municipal connections. Mu-
nicipal system expansion and deregulation (permitting resale from connections)
would both generate efficiency and equity gains in this case, but problems in
arranging a municipal hook-up, especially by the poor, remain the core problem.
Crane et al. (1997) provide hedonic regressions of housing value as a function
of structural features, house amenities, and locational amenities. Since marginal
prices depend on quantities of characteristics consumed, they use four corrections
to remove simultaneity bias, as is now conventional. For Bangkok owners and
renters, house amenities linked to public infrastructure (piped water and sew-
erage) tend to be far more important than structural features. Legal access to
electricity alone raises a an owner-occupant's house's value by $869, or about
1748 C.M. Becker and A.R. Morrison
5%, and raises dwelling value by $100 (1-2%) for squatters. In contrast, elec-
tricity access turns out to be worth little at the margin in Jakarta, since nearly
all households have access. Access to piped water and in-house toilets are worth
more in Jakarta (in the latter case than in Bangkok) than structural characteristics:
in general, high WTP reflects scarcitiesdriven by public infrastructurepolicy.
In many and perhaps most cases, inadequate service provision reflects an
inability to collect user fees; this, in turn, is linked to limited local government
capacities (see Becker et al., 1994). Elsewhere, rents achieved by restricted ser-
vice provision are attractive both to private and public officials. But the pressures
generated by rapid growth make it difficult for even the best intentioned urban
authorities to keep up, especially when historic traditions militate against private
market solutions.
In the case of sewerage services, the problem is more complex. There is little
doubt that sewage treatment is abysmal throughout the developing world (for fig-
ures, see Kingsley et al., 1994), or that the health consequences are often severe.
But national governments tend to focus on generating economic growth, and
defer expensive solutions, especially away from capital cities. At the same time,
they assign neither responsibilities nor revenue authority to local governments,
which therefore have little incentive or capacity to treat sewage.
4.4. Environment
68 A complete review of environmental problems urban areas would include discussion of water pollution,
soil degradation and toxic contamination of soils, as well as other forms of environmental damage. It is impos-
sible to discuss the gamut of environmental problems in urban areas in this brief section; consequently, we will
use air pollution as the context in which to discuss the causes of environmental damage and some proposed
solutions,
Ch. 43: Urbanizationin Transforming Economies 1749
Air pollution can be divided into two classes: conventional and greenhouse
gases. Conventional pollutants include the immediately harmful elements of mo-
tor vehicle and industrial emissions, including carbon monoxide, nitrogen oxide,
nonmethane hydrocarbons, sulfur oxides, suspended particulate matter, and lead.
Greenhouse pollutants-which as the name implies cause damage to the earth's
ozone layer-include carbon dioxide, methane, nitrous oxide and chlorofluoro-
carbons. While the levels of air pollution in urban areas of industrialized coun-
tries have been falling for most pollutants, the reverse is true in transforming
countries. Lead emissions from automobiles, for example, are rising sharply in
transforming countries. Levels of several pollutants (particulate matter, sulfur
dioxide and carbon monoxide) in Taiwan and Korea already exceed those in
industrialized nations of similar population size (Faiz, 1992: p. 176). The race to
achieve higher recorded economic standards implies disregard for environmental
measures. On the contrary, much growth has been energy and transport-intensive,
and limited concern has been devoted to the air and water pollution implications.
Rapid growth without environmental incentives can have staggering implica-
tions, though: total energy use in southeast and east Asia is rising at an annual
rate of about 4%; CO, hydrocarbon and particulate emissions from vehicles in
already heavily polluted Bangkok are rising at 5% to 6% annually (Kingsley et
al., 1994).
Groundwater pollution and water shortages are also growing problems, es-
pecially in megacities. Braadbaart and Braadbaart (1997) focus on excessive
groundwater mining in Java, but it is a serious problem as well in Bangkok,
Hanoi, Dhaka, Manila, Delhi, Mexico City, and many Chinese cities (also see
World Resources Institute, 1996: pp. 65-66, and the brief papers by McIntosh
and Herrera in Serageldin et al., 1995). Groundwater in many places is a classic
public good and metering, while not technically difficult, requires a concerted ef-
fort on the part of authorities. Ironically, the absence of meters and illicit pumping
together ensure low revenues for municipal water authorities, who then lack the
capacity to expand water use monitoring and fee collection, thereby perpetuating
a vicious cycle.
A common misconception is that environmental degradation will cease if
externalities are fully internalized. This is not necessarily the case, since the
value to firms of polluting may exceed the value of the damage caused to other
individuals and firms. In other words, there is some economically optimal level
of pollution. In the case of urban areas of transforming economies, however, so
many environmental negative externalities are not internalized that it is clear that
degradation far exceeds its optimal level.
1750 C.M. Becker and A.R. Morrison
There are two principal causes of externalities: market failure and policy fail-
ure. Types of market failure include the lack of a market because of difficulty in
identifying or enforcing property rights (e.g., air quality), partial marketing of a
resource (e.g., timber from hillsides is sold as firewood, but watershed protection
is not sold, and the nonmarketed use is consequently ignored or undervalued),
open access to resources (tragedy of the commons), and lack of information about
environmental impacts or low cost ways to mitigate or avoid damage. Types
of policy failure include subsidies of polluting activities, nonaccountability of
public sector polluters, and ineffective management of public lands (World Bank,
1992).
What makes urban areas of transforming economies particularly susceptible
to high levels of pollution? Part of the answer clearly lies in the explanations al-
ready given. Transforming countries have less complete markets than developed
countries, including markets for property rights to environmental commodities.
The policy failure represented by nonaccountability of public sector polluters
is especially damaging when these polluters are using "dirty" technologies and
public sector producers are responsible for large shares of output. But more fac-
tors are at work. They include poverty, rapid population growth, and migration.
If urban poverty is extensive, society may be quite willing to trade off environ-
mental damage for more rapid economic growth, since the environmental impact
may not be felt for many years. 6 9 Rapid population growth and migration will
place severe strains on the environmental carrying capacity of urban areas.
In attempting to control urban pollution, governments of transforming coun-
tries face the same choice as governments in industrialized nations: whether to
use command and control (direct regulation along with monitoring and enforce-
ment mechanisms) or economic (market-based) strategies to control pollution. 7 0
The critiques of command and control systems are well known. In particular,
the approach has been criticized for being economically inefficient, since it does
not allow producers who could reduce pollution at lower cost the opportunity
to do so in place of polluters who must undertake more costly reductions in
pollution. Economists, not surprisingly, have trumpeted the advantages of the
economic approach, which can: (1) promote cost-effective means for achieving
pollution targets; (2) encourage development of pollution control technology;
(3) provide government with a source of revenue; (4) allow flexibility in pol-
71 As Yang Suzhen (in Serageldin et al., 1995) discusses, Beijing's municipal government has been among
the limited number authorities the transforming economies to level pollution fees, especially at rates which
generate substantial revenue.
72 As Bernstein (1993) notes, any agency given responsibility for implementing pollution control policies
must have clear authority, as well as adequate expertise, staff, equipment and funds to carry out its monitoring
and enforcement mission.
1752 C.M. Becker and A.R. Morrison
That air pollution should decline with income is unsurprising, and can be
generated (though not unambiguously) by a simple general equilibrium model,
as long as clean air is a normal good. That air pollution is greater in larger cities
is similarly unsurprising, but the downturn requires some explanation. In part the
pattern may reflect employment and output composition: very large cities have
higher shares in services than manufacturing than do medium-sized cities. But
other forces also may be at work: large cities may be wealthier and politically
more powerful, and hence able to demand government subsidies for abatement
equipment and measures. In any event, the topic has hardly been touched: Shukla
and Parikh's work is the first word, but hardly the last.
This section only briefly addresses the arguments underlying a longstanding and
contentious issue: Does urbanization exacerbate or alleviate poverty in develop-
ing countries? The general issue of urban poverty is ably addressed in Chapter 46
of this volume, while Kannappan (1989) also offers a superb survey of poverty
and urban labor markets.
female earnings and, over a longer period, to tighter labor markets. Since ur-
banization is linked to lower fertility and improved educational attainment, one
would therefore expect it to give rise to greater equality and more rapid growth
as well. Indeed, as Birdsall et al. note, these forces all comprise self-reinforcing
virtuous circles in the east Asian environment.7 4
One of the longest urban inequality time series is for Indonesia (Firdausy,
1993), where there appears to have been virtually no change in per capita ex-
penditure Gini coefficients from 1964 through 1990. Indonesia's rural Gini coef-
ficient appears to have declined dramatically during this period, however, from
0.34 to 0.25. For urban Indonesia, it appears that inequality increased sharply
between the mid-1960s and late 1970s, and then declined throughout the 1980s.
Income shares of the bottom 40% mirror this pattern. It is also worth noting that
most Javanese cities' Gini coefficients' trends appear to have mirrored the broad
national pattern, save for several discrepancies between 1984 and 1990. Among
the most important of these is a pattern of rising inequality in urban east Java,
which is dominated by booming Surabaya. Surabaya's Gini coefficient rose from
0.35 in 1960 to 0.41 in 1978 to 0.44 in 1990, while Gini coefficients in more
established Jakarta, Bandung and Medan were all stable or declining during this
period.
Assessing distributional patterns is a tricky matter, especially where systems
of mutual obligations exist in extended families. Even defining a "family" of
"household" is difficult, especially in polygamous societies. In much of rural
and urban Africa, sub-households headed by wives exist in a fairly indepen-
dent state, but with some mutual obligations to related units and, of course,
the husband. Surely, the definition of a "household" in these circumstances will
markedly affect distributional and poverty measures. At the same time, measures
of individual wage inequality are problematic as well. For example, Terrell and
Svejnar (1989: p. 110) in their study of Senegalese industrial workers found that
the average formal sector worker fed 10.9 people other than himself. In a sense,
it might be apt in such a setting to view intrahousehold migration as taking place
in order to equate utilities across space.
74 Other ways Birdsall et al. (1995: p. 20) hypothesize that low inequality might contribute to rapid growth
include:
* contributing to political and macroeconomic stability;
* raising x-efficiency of low-income workers;
* limiting rural/urban income gaps, which in turn limits rent-seeking migration and raises multipliers from
productivity gains;
* inducing increased savings and investments by the poor.
This latter phenomenon will result if the poor (and their businesses) are liquidity constrained: see the analytic
model of Birdsall et al. (1995b).
1754 C.M. Becker and A.R. Morrison
76 Manning (1988) provides a less sanguine view of rural employment and income growth in Indonesia.
While not terribly pessimistic, Manning does focus far more on the labor saving nature of agricultural change
than do Collier et al. (1987). Furthermore, among the poor access to machines and fertilizers is limited, per-
haps leading to rising inequality. Like Henderson and Kuncoro (1994). Manning emphasizes the considerable
centralization of Indonesian manufacturing along with the slow growth of agricultural employment.
Ch. 43: Urbanization in Transforming Economies 1757
areas tend to have owners with good marketing contacts and access to capital,
however: one should not picture these as informal sector microenterprises taking
off.
The situation in Bangladesh, a very poor and densely populated country, is
less clear. As Bhadra (1992) discusses, urban poverty as a whole appears to have
fallen since Independence and the late 1980s, but the acute poverty may not have
(depending on the survey one uses). Given Bangladesh's very high inmigration
rates, it is unsurprising that the urban poor are overwhelmingly migrants. As in
Malaysia, poverty is linked to the number working (79% have only one member
of the household working), and to occupation. A survey of squatters analyzed
by Bhadra found that 42% were engaged in transport activities (presumably,
rickshaw pulling); another study found that 29% of those classified as urban poor
were employed in transport. Construction and other unskilled activities account
for most of the rest of the urban poor's occupations. Bhadra (1992: p. 26) also
generates a series from 1976/77-1988/89 of Dhaka low income housing rents
relative to the food CPI for industrial workers: as food prices rose at an annual
rate of 11% while housing rents rose at 20% per annum during this period, he
concludes that land and hence housing rent pressures during the urban boom
were largely responsible for the persistence of high rates of acute poverty.77
Finally, Mitra (1994) regresses estimates of urban poverty incidence by Indian
state on labor productivity, industrial per capita income, and a time trend for
1960/61-1973/74. He found positive time trends for all states during this period
of economic stagnation, and inconclusive links to productivity and per capita
incomes.
Will urbanization affect demand patterns, especially in a manner that matters for
macroeconomic policy? From economy-wide CGE simulation studies such as
Becker et al.'s (1992) work on India, we know that cities are far more capital-
and import-intensive than rural areas. Government is highly urbanizing, both
because governments consume urban services (the bureaucracy and formal sector
manufactures), and because the goods and services it consumes themselves have
urban-oriented demands. Since the urban-orientation (and capital- and import-
77 Food accounted for about 60% and housing for about 15% of expenditures of all urban households during
this period (housing's share rose and food's share fell over time as well). Food shares will be higher for the
poor, naturally.
1758 C.M. Becker' and A.R. Morrison
dX i (ax, aY axi y*
dPj - pj apj ay+ap
78 Differential consumption patterns have important implications for macroeconomic policy. especially in
two-gap models in which foreign exchange or capital shortages limit total output. In this case, government re-
distributive efforts aimed at shifting demand from foreign exchange or capacity constrained sectors to demand
constrained sectors potentially could result in real GDP gains.
79 Evidence on this point is scanty. due to the fact that very few researchers have examined differences in
rural and urban consumption patterns. A possible reason is that if data are good enough to support estimation of
region-specific demand functions, then the goods market in such a country probably be sufficiently developed
and integrated so that large differences will not exist.
80 Grewe uses the Susenas dataset on income and consumption patterns in Indonesia, which has microlevel
data for over 46,000 households.
81 Overall expenditure patterns were as follows:
The term in parentheses contains the standard Slutsky terms, with nominal
income kept constant. But rising prices also generate production responses and
hence rising real incomes. As Singh et al. (1986: p. 26)) find empirically, these
profit effects tend to swamp the standard terms when agricultural prices rise.
Again, though, the implications for consumption patterns are a bit vague, as
these patterns will depend on specific policy and relative price environments.
If growing urban demand raises the relative price of agricultural goods, then one
should expect to find trends toward growing consumption shares of domestic
farm products in rural households, and declining shares among urban households.
83 Verma's (1986) analysis is state-level. But since she estimates agglomeration economies in manufactur-
ing, a large share of production presumably takes place in urban areas.
Ch. 43: Urbanizationin TransformingEconomies 1761
efficiency. For Industry 23 (Cotton Textiles)-and established consumer-goods
industry-drainage capacity, industrial water consumption asphalt roads, public
transport ... and, again, market places have maximum impact". In contrast, for
the producer-oriented Metal Products and Parts industry, power consumption,
roads and public transport, and telephone connections are most important. While
these results can be taken as only suggestive, it is difficult to avoid concluding
that the public sector provides complementary capital inputs (water supply, roads,
electricity, health and educational facilities tend to be gross complements in the
empirical work) and free but valuable substitutes as well to private industries and
public enterprises. This infrastructure, as it is likely to be strongly correlated
with city size, represents an important component of those amorphous urban
agglomeration economies.
Mitra (1997) takes a slightly different approach and estimates translog fron-
tier production functions for Indian firms in two industries, electrical machinery
(EM) and cotton and cotton textiles (CCT), and uses city population and city
work force as alternative measures of urbanization economies. OLS regressions
indicate some but limited support for "U"-shaped urbanization economies for
UM, and none for CCT. However, when efficiency indices are calculated from
the frontier functions and regressed on size of the plant's home city, the "U"-
shaped effect is more significant for EM, and sporadically significant for CCT,
though the explanatory power is low. The optimal city size for EM turns out to
have a population of 1.5-2.5 million; for CCT the optimal size is 2.5-5.0 million.
The empirical agglomeration literature is highly suggestive but not entirely
convincing, largely because of small sample sizes. The works of Henderson,
Shukla and others are fascinating, but there is a desperate need for more study
based on narrowly defined industries and larger samples. It is also difficult to
leave the literature without wondering about causality and possible aggregation
biases.
Parallel with the development of the agglomeration literature in urban and re-
gional economics has been an emerging focus over the past two decades on scale
economies in the international trade literature. Spatially oriented CGE simulation
models such as Markusen and Rutherford (1993) examine industrial production
location decisions under conditions of imperfect competition, fixed costs and dis-
crete numbers of plants. While their model examines automobile plant location in
North America, the same framework would be applicable to, say, plant location
in different types of cities in China or Brazil. Their findings suggest that output
growth may be highly sensitive to (a) whether plant sites are fixed or variable, (b)
the nature of tariff protection-but not necessarily in a continuous, linear manner,
especially if fixed costs or scale economies are substantial.
1762 C.M. Becker and A.R. Morrison
In R i = A - a In Si + (ln Si)2,
tended to be significant, and also positive in about two-thirds of cases; and (d)
that the Pareto coefficient appears to rise (a country's city system becomes less
primate) as per capita income and railroad density increase; other economic
variables did not appear to matter. Ledent (1990) examines African data, and
finds that Pareto coefficients are sensitive to the number of cities included, rising
toward unity as the number increases. Ledent finds that primacy declines with
a country's population and total area; the role of economic variables tends to
1764 C.M. Becker and A.R. Morrison
84 All the approaches analyzed in this section measure the contribution of city growth to overall economic
efficiency when that growth is caused by internal migration. To the authors' knowledge, no analysis has
combined natural population increase with efficiency analysis.
85 This is the case if there are no costs associated with migration. As Herzog and Schlottmann (1981) have
shown for the US, however, the moving, information and psychic costs of migration are nontrivial. Once costs
are included, migration will not eliminate completely interregional wage gaps.
86 Largely, one presumes, because of the questionable relevance of the concept of open unemployment in
transforming economies, especially in rural areas.
Ch. 43: Urbanizationin Transforming Economies 1767
At the macroeconomic level, three approaches have been used to gauge the
efficiency of the internal migration process, but-to the authors' knowledge-
only one of these has been applied to urban growth in transforming economies.
The first macroeconomic approach computes an efficiency ratio for migration
flows, where the ratio is defined as the absolute value of the number of net
migrants between two areas divided by the number of gross migrants between
the two areas.8 7 The usual interpretation of the efficiency ratio is that migration
is more efficient the closer is the ratio to one; a higher ratio for a given amount
of net migration indicates lower costs incurred in reallocating migrants between
regions. Galloway and Vedder (1985), however, develop a simple model in which
an optimal population distribution results in an efficiency ratio of zero, calling
into question the conventional interpretation of the migration efficiency ratio.
If diametrically opposite welfare conclusions can be obtained from the same
value of the efficiency ratio, it should be discarded as a measure of migration's
efficiency.
A second, more appropriate measure of migration's efficiency is to determine
whether migration closes regional wage and unemployment differentials. Evi-
dence that migration reduces differentials supports the contention that migration
increases allocative efficiency. For the US, Herzog et al. (1993) cite evidence
that migrants move from areas with high unemployment rates. In a simulation
exercise (also for the US), Gabriel et al. (1993) find that migration reduces but
does not eliminate interregional unemployment differentials caused by a region-
specific shock. While the evidence mustered by these authors is convincing, it is
not directly relevant to migration to urban areas of transforming economies.
A final approach to measuring the efficiency impact of migration is intro-
duced by Morrison (1993). Using data from Peru, he employs stochastic frontier
production functions to calculate the efficiency-adjusted marginal products of
labor by productive sector and Peruvian department. For migration flows to Lima
between 1976 and 1981, he then calculates foregone production in source regions
and production increases in Lima. The net result: internal migration increased
migration from 1.9 to 2.1% of 1979 GDP-a considerable efficiency gain.8 8 Note
that this efficiency gain has elements of both technical and allocative efficiency,
since the frontier production function allows for differing degrees of technical
87 If gross migration from region I to j is defined as Mij and gross migration from j to I as Mi, then the
efficiency ratio would be: Eij = I(Mij - Mji) I(Mij + Mji).
88 The percentages reported are computed at domestic relative prices, but the result remains robust if: (1)
gains and losses are calculated at world prices, and (2) negative externalities generated by the migrants in Lima
are included in the analysis.
1768 C.M. Becket and A.R. Morrison
References
Abuhadba, M. and P. Romaguera (1993). "Inter-industry wage differentials: evidence from Latin American
countries", Journal of Development Studies 30:190-205.
Acsadi, G.T.F., G. Johnson Acsadi and R.A. Bulatao, eds. (1990), Population Growth and Reproduction in
Sub-Saharan Africa (World Bank, Washington, DC).
Adams. R.H., Jr. (1992), "The effects of migration and remittances on inequality in rural Pakistan", Pakistan
Development Review 31:1189-1206.
Adelman, I. and H. Fetini (1991). "Development strategies and the environment", University of Califomia at
Berkeley Department of Agricultural and Resource Economics (CUDARE) Working Paper 587.
Adelman, I. and S. Robinson (1978a), Income Distribution Policy in Developing Countries: A Case Study of
Korea (Stanford University Press. Stanford, CA).
Adelman, I. and S. Robinson (1978b). "Migration, demographic change and income distribution in a model of
a developing country", Research in Population Economics 1:1-26 (JAI Press).
Adelman, I. and D. Sunding (1987), "Economic policy and income distribution in China",. Journal of
Comparative Economics 11:444 461.
Adelman, I. and J.E. Taylor (1990), "Is structural adjustment with a human face possible'? The case of Mexico".
Journal of Development Studies 26:387-407.
Adelman, I. and J.E. Taylor (1991), 'Multisectoral models and structural adjustment: new evidence from
Mexico", Journal of Development Studies 29:154-163.
Ades, AF. and E. Glaeser (1994), "Trade and circuses: explaining urban giants", National Bureau of Economic
Research Working Paper No. 4715, Cambridge, MA.
Ahmad, N. (1992). "Choice of location and mobility behavior of migrant households in a third world city",
Urban Studies 29:1147-1157.
Ahmad, N. (1993), "Choice of neighborhoods by mover households in Karachi", Urban Studies 30:1257-1170.
Ahmed, A. and I. Sirageldin (1993). "Socio-economic determinants of labour mobility in Pakistan", Pakistan
Development Review 32:139-157.
Aina, T.A. (1992), "Land tenure in Lagos",. Habitat International 16:3-15.
Akerlof, G. (1982), "Labor contract as partial gift exchange", Quarterly Journal of Economics 97:543-569.
Akerlof, G. (1984), "Gift exchange and efficiency-wage theory: four views", American Economic Review
74:79-83.
Alchian, A. and H. Demsetz (1972), "Production, information cost, and economic organization", American
Economic Review 62:777-795.
Alexander, C. and J. Wyeth (1994), "Cointegration and market integration: An application to the Indonesian
rice market", Journal of Development Studies 30:303-328.
Alperovich, G. (1993), "An explanatory model of city-size distribution: evidence from cross-country data",
Urban Studies 30:1591-1601.
Altaf., M.A. and J.A. Hughes (1994), "Measuring the demand for improved urban sanitation services: results
of a contingent valuation study in Ouagadougou, Burkina Faso", Urban Studies 31:1763-1776.
Amano, M. (1983), "On the Harris-Todaro model with intersectoralmigration of labour"'. Economica 50:311-
323.
Amis, P. (1984), "Squatters or tenants: the commercialization of unauthorized housing in Nairobi", World
Development 12(1):87-96.
Amsden, A.H. (1991), "Big business and urban congestion in Taiwan: the origins of small enterprise and
regionally decentralized industry (respectively)", World Development 19:1121-1135.
Andreev, E., S. Scherbov and F. Willikens (1997), The Population of Russia: Fewer and Older (University of
Groningen Faculty of Spatial Sciences, Groningen).
Ch. 43: Urbanization in Transforming Economies 1769
Angel, S. and S. Mayo (1993), "Housing: enabling markets to work", World Bank Policy Paper, Washington,
DC.
Appleton, S., P. Collier and P. Horsnell (1990), "Gender, education and employment in Cote d'Ivoire", World
Bank SDA Unit Working Paper No. 8, Washington, DC.
Aratme, N. (1992), "A study of the occupational mobility of locano migrants in metro Manila, The
Philippines", Paper given at the 1992 PAA meetings (University of Chicago, Department of Sociology,
Chicago).
Armstrong-Wright, A. (1986), "Urban transit systems", World Bank Technical Paper No. 52 (Washington,
DC).
Arnott, R.J. and M. Gersovitz (1986), "Social welfare underpinnings of urban bias and unemployment",
Economic Journal 96:413-424.
Aryee, G. (1981), "The informal manufacturing sector in Kumasi", in S. Sethuraman, ed., The Urban Informal
Sector in Developing Countries (ILO,Geneva).
Aryeetey-Attoh, S. (1992), "An analysis of household valuations and preference structures in Rio de Janeiro,
Brazil", Growth and Change 23:183-198.
Asabere, P.K. (1981), "The determinants of land values in an African city: the case of Accra, Ghana", Land
Economics 57:385-397.
Asabere, P.K. and K.O. Banahene (1983), "Population density function for Ghanaian (African) cities: an
empirial note", Journal of Urban Economics 14:370-379.
Ashenfelter, O. and R. Oaxaca (1991), "Labor market discrimination and economic development", in N. Bird-
sall and R. Sabot, eds., Unfair Advantage: Labor Market Discrimination in Developing Countries (World
Bank, Washington, DC).
Assadian, A. and J. Ondrich (1993), "Residential location, housing demand and labour supply decisions of
one- and two-earner households: the case of Bogota, Colombia", Urban Studies 30:73-86.
Bahl, R. and J. Linn (1992), Urban Public Finance in Developing Countries (Oxford University Press, New
York).
Baker, J., ed. (1990), Small Town Arica (Nordiska Afrikainstitutet, Uppsala).
Baker, J. and P.O. Pedersen, eds. (1992), The Rural-Urban Interface in Africa (Nordiska Afrikainstitutet,
Uppsala).
Banerjee, B. (1981), "Rural-urban migration and family ties: an analysis of family considerations in migration
behaviour in India", Oxford Bulletin of Economics and Statistics 43:321-355.
Banerjee, B. (1983), "The role of the informal sector in the migration process: a test of probabilistic migration
models and labour market segmentation for India", Oxford Economic Papers 35: 399-422.
Banerjee, B. (1984a), "Information flow, expectations, and job search: rural-to-urban migration process in
India", Journal of Development Economics 15:239-257.
Banerjee, B. (1984b), "The probability, size, and uses of remittances from urban to rural areas in India", Journal
of Development Economics 16:293-311.
Banerjee, B. (1986), Rural to Urban Migration and the Urban Labour Market (Himalaya for the Institute of
Economic Growth, Bombay).
Banerjee, B. (1991), "The determinants of migrating with a pre-arranged job and of the initial duration of
urban unemployment", Journal of Development Economics 36:337-351.
Banerjee, B. and G. Bucci (1994), "On-the-job search after entering urban employment: An analysis based on
Indian migrants", Oxford Bulletin of Economics and Statistics 56:33-47.
Banerjee, B. and S.M. Kanbur (1981), "On the specification and estimation of macro rural-urban migration
functions", Oxford Bulletin of Economics and Statistics 7:29.
Barat, J. (1989), "A brief review of the development of urban transportation in Brazil", International Journal of
Social Economics 16:44-57.
Barber, G.M. and W.J. Milne (1986), "Modeling internal migration in Kenya: an econometric analysis using
limited data", Unpublished manuscript (University of Toronto, Toronto).
Barber, G, W. Milne and G. Ongile (1990), "Determinants of urban labour force participation in Kenya",
Eastern Africa Economic Review 6:83-93.
Bamum, H.N. and R.H. Sabot (1977), "Education, employment probabilities and rural-urban migration in
Tanzania", Oxford Bulletin of Economics and Statistics 39:109-126.
1770 C.M. Becker and A.R. Morrisonl
Barrett, R. (1988), Urban Transport in West Africa, Technical Paper No. 81. Urban Transport Series (World
Bank, Washington DC).
Barros, R.. L. Fox and R. Mendonca (1997). "Female-headed households, poverty. and the welfare of children
in urban Brazil", Economic Development and Cultural Change 45:231-257.
Bartlett. W. (1983), 'On the dynamic instability of induced-migration unemployment in a dual economy".
Journal of Development Economics 13:85-96.
Bates, R. (1993), ' 'Urban bias': a fresh look", Journal of Development Studies 29:219-228.
Baydar, N., M.J. White, C. Simkins and 0. Babakol (1990). "Effects of agricultural development policies on
migration in peninsular Malaysia", Demography 27:97-110.
Beals, R.E., M.E. Levy and L.N. Moses (1967), "Rationality and Migration in Ghana", Review of Economics
and Statistics 49:480-486.
Becker. B.E. and Y. Gao (1989). "The Chinese urban labor system: prospects for refonnrm", Joumal of Labor
Research 10:411-428.
Becker, C.M. (1987), "Urban sector income distribution and economic development', Journal of Urban
Economics 21:127-145.
Becker, C.M. (1992), "Demographic cohort shifts and third world urbanization", Paper presented at the 39th
North American meeting of the Regional Science Association International. Chicago.
Becker, C.M. (1997), "Enterprise incentives in post-Soviet agribusiness and manufacturing", in R.E. Even-
son, ed., Proceedings of the Conference on Attaining Ukraine's Agro-industrial Potential (Yale University
Economic Growth Center, New Haven. CT).
Becker, C.M. and C. Grewe (1996), "Cohort-specific rural-urban migration in Africa", Journal of African
Economies 5:228-270.
Becker, C.M. and D. Hamley (1998), "Demographic change in the former Soviet Union during the transition
perod". World Development 26:1957-1976.
Becker, C.M. and A.R. Morrison (1988), "The determinants of urban population growth in sub-Saharan
Africa", Economic Development and Cultural Change 36:259-278.
Becker, C.M. and A.R. Morrison (1993). 'Observational equivalence in the modeling of African labor markets
and rural-urban migration". World Development 21:535-554.
Becker, C.M., J.G. Williamson and E.S. Mills (1992), Indian Urbanization and Economic Development (Johns
Hopkins University Press. Baltimore).
Becker. C.M., A.M. Hamer and A.R. Morrison (1994), Beyond Urban Bias in Africa: Urbanization in an Era
of Structural Adjustment (Heinemann, Portsmouth, NH).
Beckemnnan, W. (1992). "Economic growth and the environment: whose growth? Whose environment?", World
Development 20:481-496.
Behrman, J. and A. Deolalikar (1991), 'The poor and the social sectors during a period of macroeconomic
adjustment: empirical evidence for Jamaica", World Bank Economic Review 5:291-313.
Behrman, J.R. and B.L. Wolfe (1984a), "Micro determinants of female migration in a developing country".
Research in Population Economics 5:137-166.
Behrman, J.R. and B.L. Wolfe (1984b). "Labor force participation and earnings determinants for women in the
special conditions of developing countries". Journal of Development Economics 15:259-288.
Behrman, J.R. and B.L. Wolfe (1984c), The socioeconomic impact of schooling in a developing country".
Review of Economics and Statistics 66:296-303.
Bell, C. (1991), "Regional heterogeneity, migration, and shadow prices", Journal of Public Economics 46:1-27.
Benabou, R. (1991), "Workings of a city: location. education, and production", National Bureau of Economic
Research Technical Paper 113.
Bencivenga, V.R. and B.D. Smith (1997), "Unemployment. migration. and growth", Journal of Political
Economy 105:582-608.
Beneria, L. and S. Feldman. eds. (1992). Unequal Burden: Economic Crises, Persistent Poverty, and Women's
Work (Westview Press, Boulder and Oxford).
Bennett, J. and M. Phelps (1983), "A model of employment creation in an open developing economy", Oxford
Economic Papers 35: 373-398.
Ch. 43: Urbanizationin TransformingEconomies 1771
Bernheim, B.D. and 0. Stark (1986), "The strategic demand for children: theory and implications for fertil-
ity and migration", Migration and Development Program, Discussion Paper No. 25 (Harvard University,
Cambridge, MA).
Bernstein, J.D. (1993), "Alternative approaches to pollution control and waste management: regulatory
and economic instruments", World Bank Urban Management Programme, Urban Management and the
Environment Series, No. 3.
Bernstein, J.D. (1994), "Land use considerations in urban environmental management", Urban Management
and the Environment series, vol. 12 (World Bank, Washington, DC).
Berry, R.A. and R.H. Sabot (1978), "Labour market performance in developing countries: a survey", World
Development 6:
Bertaud, A. (1989), "The regulatory environment of urban land in Indonesia: constraints imposed on the poor",
Unpublished manuscript (World Bank, Washington, DC).
Bertaud, A. and B. Renaud (1994), Cities without land markets: lessons of the failed socialist experiment,
Discussion Paper 227 (World Bank, Washington DC).
Bertaud, A. and J. Kolko (1994), "West of Los Angeles, east of Budapest: new evidence on residential density
patterns", Urban Development Division, Unpublished manuscript (World Bank, Washington, DC).
Bhadra, D. (1992), "Migration and urban poverty: a case study of Bangladesh", Unpublished manuscript
(Cornell University, Ithaca, NY).
Bhalla, A.S. (1990), "Rural-urban disparities in India and China", World Development 18:1097-1110.
Bhatia, K.B. (1979), "Rural-urban migration and surplus labour", Oxford Economic Papers 31:403-414.
Bigsten, A. (1996), "The circular migration of smallholders in Kenya", Journal of African Economies 5:1-20.
Bigsten, A., S.K. Mugerwa (1992), "Adaption and distress in the urban economy: a study of Kampala
households", World Development 20:1423-1441.
Bilsborrow, R.E. (1993), "Issues in the measurement of female migration in developing countries", in United
Nations, ed. Migration of Women in Developing Countries (UN, New York).
Bilsborrow, R.E., T.M. McDevitt, S. Kossoudji and R. Fuller (1987), "The impact of origin community
chartacteristics on rural-urban out-migration in a developing country", Demography 24:191-210.
Birdsall, N. and R. Sabot (eds.) (1991), Unfair Advantage: Labor Market Discrimination in Developing
Countries (World Bank, Washington, DC).
Birdsall, N., D. Ross, and R. Sabot (1995a), "Inequality and growth reconsidered", Williams College, Center
for Development Economics, Research Memorandum RM-142 (Williams College, Williamstown, MS).
Birdsall, N., T. Pinckney, and R. Sabot (1995b), "Inequality, savings and growth", Williams College, Center
for Development Economics, Research Memorandum RM-148 (Williams College, Williamstown, MS).
Bissiliat, J. and J. Fenet-Rieutord (1984), Les Villes Secondaires en Afrique: Leur Role at Leurs Fonctions dans
le Developpement National et Regional, Phase 1, Senegal, Paris: Agence Francaise pour I'Amenagement et
le Developpement a l'Etranger.
Black, D., and J.V. Henderson (1997), "Urban Growth", Unpublished manuscript (Brown University Depart-
ment of Economics, Providence, RI).
Black, R. (1993), "Migration, return, and agricultural development in the Serra do Alvao, Northern Portugal",
Economic Development and Cultural Change 41:563-585.
Blau, D.M. (1986), "Self-employment, earnings, and mobility in peninsular Malaysia", World Development
14:839-852.
Blinder, A. (1973), "Wage discrimination: reduced form and structural estimates", Journal of Human Resources
8:436-455.
Bliss, C. (1992), "The design of fiscal reforms in revenue-constrained developing countries", Economic Journal
102:940-951.
Bliss, C. and N. Stem (1978), "Productivity, wages and nutrition. Part I: the theory", Journal of Development
Economics 5:331-362.
Blomqvist, A.G. (1978), "Urban job creation and unemployment in LDCs: Todaro versus Harris and Todaro",
Journal of Development Economics 5:3-18.
Boehm, T.P., H.W. Herzog Jr. and A.M. Schlottmann (1991), "Intra-urban mobility, migration, and tenure
choice", Review of Economics and Statistics 73:59-68.
1772 C.M. Becker antiA.R. Morrison
Borjas, G., S. Bronars and S. Trejo (1992), "Self-selection and internal migration in the United States', NBER
Working Paper 4002 (National Bureau of Economic Research, Cambridge, MA).
Braadbaart, O. and F. Braadbaart (1997), "Policing the urban pumping race: industrial groundwater overex-
ploitation in Indonesia". World Development 25:199-210.
Brandon, C. and R. Ramankutty (1993), Toward an Environmental Strategy for Asia (World Bank, Washington.
DC).
Braverman, A. and R. Kanbur (1987), "Urban bias and the political economy of agricultural reform", World
Development 15:1179-1187.
Brennan, E. and H. Richardson (1989), 'Asian megacity characteristics. problems and policies", International
Regional Science Review 12:117-129.
Brown, A.N. (1993). "A note on industrial adjustment and regional labor markets in Russia". Comparative
Economic Studies 35:147-157.
Brown, L.A. and A.R. Goetz (1987), "Development-related contextual effects and individual attributes in third
world migration processes: a Venezuelan example", Demography 24:497-516.
Brown. R., M. Moon and B. Zoloth (1980), "Incorporating occupational attainment in studies of male-female
earnings differentials", Journal of Human Resources 15:3-28.
Brueckner, J. (1990), "Analyzing third-world urbanization: a theoretical model with empirical evidence",
Economic Development and Cultural Change 38:587-610.
Burki, A. and Q. Abbas (1991), "Earnings functions in Pakistan's urban informal sector: a case study", Pakistan
Development Review 30:695-703.
Button, K., N. Ngoe and J. Hine (1993), "Modeling vehicle ownership and use in low income countries",
Journal of Transport Economics and Policy 27:51-267.
Byerlee, D, J.L. Tommy and H. Fatoo (1976), "Rural-urban migration in Sierra Leone", Michigan State
University, African Rural Economy Program, African Rural Economy Paper 13 (E. Lansing, MI).
Calvo, G. (1978), "Urban unemployment and wage determination in LDCs: trade unions in the Harris-Todaro
model", International Economic Review 19:65-81.
Caces, E. F. Amold, J. Fawcett and R. Gardner (1985), "Shadow households and competing auspices", Journal
of Development Economics 17:5-25.
Campbell, T. (1989), "Environmental dilemmas and the urban poor", in L.H. Jeffrey, ed., Environment and the
Poor: Development Strategies for a Common Agenda (Transaction Books, New Brunswick, NJ and Oxford).
Chan, K.W. and Y. Yunyan (1997), "Internal migration in post-Mao China: a dualistic approach". in China
Population Association, Symposium on Demography of China, Population Association and IUSSP (Beijing,
China).
Chang, K.-S. (1994), "Chinese urbanization and development before and after economic reform", World
Development 22:601-614.
Chawla, A. (1993), "Intrahousehold resource allocation: a principal-agent analysis with empirical evidence",
Ph.D. Dissertation, Department of Economics, University of Michigan.
Chen J., J. Yafen. and Z. Kaimin (1997), "Development of population aging Shanghai". in China Population
Association, Symposium on Demography of China, China Population Association and IUSSP (Beijing,
China).
Chenery, H.B. and M. Syrquin (1975), Patterns of Development, 1950-1970 (Oxford University Press.
London).
Choe, S.-C. and S. Byung-Nak (1985). "Spatial distribution of industries and important location factors in the
Seoul region", Water Supply and Urban Development Department, Discussion Paper UDD 89 (World Bank,
Washington, DC).
Cinar, E.M. (1994), "Unskilled urban migrant women and disguised employment: home-working women in
Istanbul, Turkey", World Development 22:369-380.
Clark, D. and J. Cosgrove (1991), "Amenities versus labor market opportunities: choosing the optimal distance
to move", Journal of Regional Science 31:311-328.
Clark, M.H. (1985), "Household economic strategies and support networks of the poor in Kenya", Water
Supply and Urban Development Department, Discussion Paper UDD-69 (World Bank, Washington, DC).
Cleaver, K. (1985), "The impact of price and exchange rate policies on agriculture in sub-Saharan agriculture",
Staff Working Paper No. 728 (World Bank, Washington, DC).
Ch. 43: Urbanizationin TransformingEconomies 1773
Cohen, B. and W. House (1993), "Women's urban labour market status in developing countries: how well do
they fare in Khartoum, Sudan?" Journal of Development Studies 29:461-483.
Cohen, B. and W. House (1994), "Education, experience and earnings in the labor market of a developing
economy: the case of urban Khartoum", World Development 22:1549-1565.
Cohen, M. (1993), "Megacities and the environment", Finance and Development 30:44-47.
Colburn, ED. (1993), "Exceptions to urban bias in Latin America: Cuba and Costa Rica", Journal of
Development Studies 29:60-78.
Cole, W.E. and R.D. Sanders (1983), "Interstate migration in Mexico", Journal of Development Economics
12:341-354.
Cole, W.E. and R.D. Sanders (1985), "Internal migration and urbanization in the third world", American
Economic Review 75:481-494.
Collier, P. (1979), "Migration and unemployment: a dynamic general equilibrium analysis applied to Tanzania",
Oxford Economic Papers 31:205-236.
Collier, P. and A. Bigsten (1981), "A model of educational expansion and labor market adjustment applied to
Kenya", Oxford Bulletin of Economics and Statistics 43:31-49.
Collier, P. and D. Lal (1980), "Poverty and growth in Kenya", World Bank Staff Working Paper No. 389.
Collier, P. and D. Lal (1986), Labour and Poverty in Kenya 1900-1980 (Clarendon Press, Oxford).
Collier, P., S. Radwan and S. Wangwe (1990), Labour and Poverty in Rural Tanzania (Clarendon Press,
Oxford).
Collier, W.L., S.K. Santoso, G. Wiradi and Makali (1987), Economic and Social Transformation of Rural Java
(US Agency for Agricultural Development, Jakarta).
Collier, V.C. and H. Rempel (1977), "The divergence of private from social costs in rural-urban migration: a
case study of Nairobi, Kenya", Journal of Development Studies 13:
Connolly, P. (1990), "Housing and the state in Mexico", in G. Shidlo, ed., Housing Policy in Developing
Countries (Routledge, London and New York).
Conway, P. (1994), "An atheoretic evaluation of success in structural adjustment", Economic Development and
Cultural Change 42:267-292.
Cornia, G.-A. and F Stewart (1991), "Sistema fiscal, ajuste y pobreza (The fiscal system. Adjustment and the
poor. With English summary)," Coleccion Estudios CIEPLAN 0:77-106.
Cour, J.-M. (1985), Macroeconomic Implications of Urban Growth: Interaction between Cities and Their
Hinterlands, Eastern and Southern Africa Region, Water Supply and Urban Development Region (World
Bank, Washington, DC).
Crane, R. (1992), "Voluntary income redistribution with migration", Journal of Urban Economics 31:84-98.
Crane, R. (1994), "Water markets, market reform and the urban poor: results from Jakarta, Indonesia", World
Development 22:71-83.
Crane, R. and A. Daniere, with S. Harwood (1997), "The contribution of environmental amenities to low-
income housing: a comparative study of Bangkok and Jakarta", Urban Studies 34:1495-1512.
Day, R.H., S. Dasgupta, S.K. Datta and J.B. Nugent (1987), "Instability in rural-urban migration", Economic
Journal 97:940-950.
De Cola, L. (1984), "Statistical determinants of the population for a nation's largest city", Economic
Development and Cultural Change 33:71-98.
De Cola, L. (1985), "Lognormal estimates of macroregional city-size distributions, 1950-1970", Environment
and Planning A. 17:1-16.
de la Paz L., M.H. Izazola, and J.G. de Leon (1993), "Characteristics of female migrants according to the 1990
census of Mexico", in United Nations, ed., Migration of Women in Developing Countries (UN, New York).
Deolalikar, A.B. (1988), "Nutrition and labor productivity in agriculture: estimates for South India", Review
of Economics and Statistics 70:406-413.
Dimitriou, H.T. (1990), "Transport and third world city development", in H.T. Dimitriou and G.A. Banjo, ed.,
Transport Planning for Third World Cities (Routledge, London and New York).
Doan, P.L. and B.D. Lewis (1993), "Growth linkages and urban development in Niger", Growth and Change
24:487-508.
Donato, K.M. (1991), "Current trends and patterns of female migration: evidence from Mexico", unpublished
manuscript (Louisiana State University, Department of Sociology, Baton Rouge, LA).
1774 C.M. Becker and A.R. Morrison
Dowall, D.E. (1992). Benefits of minimal land-use regulations in developing countries". Cato Journal 12:413-
423.
Dowall, D.E. and G. Clarke (1991), A framework for refonrmning urban land policies in developing countries".
Urban Management Program Policy Paper 7 (World Bank. Washington, DC).
Dowall, D.E. and M. Leaf (1990), "The price of land for housing in Jakarta: an analysis of the effects
of location, urban infrastructure, and tenure on residential plot prices", Institute of Urban and Regional
Development Working Paper 519 (University of California at Berkeley, Berkeley, CA).
Dowall, D.E. and P.A. Treffeisen (1991), "Spatial transformation in cities of the developing world: multi-
nucleation and land-capital substitution in Bogota, Colombia", Regional Science and Urban Economics
21:201-224.
Dunkerly, H., ed. (1983), Urban Land Policy: Issues and Opportunities (Oxford University Press, New York).
Dunkerley, H.B., A.A. Walters, J.M. Courtney, W.A. Doebele, D.C. Shoup and M.D. Rivkin (1978). "Urban
land policy issues and opportunities". Staff Working Paper No. 283 (World Bank. Washington, DC).
Dupont, V.(1986), Dynamique de Villes Secondaires et Afrique de l'Ouest (Ostrom, Paris).
Dureau. F. (1987), Migration et urbanisation: Le Case de la C6te d'Ivoire (Ostrom, Paris).
Eaton, J. and Z. Eckstein (1997), "Cities and growth: theory and evidence from France and Japan", Regional
Science and Urban Economics 27:443-474.
Edwards, S. and G. Tabellini (1991), "Explaining fiscal policies and inflation in developing countries", Journal
of International Money and Finance 10:S 16-S48.
Ellis, F. (1984), "Relative agricultural prices and the urban bias model: a comparative analysis of Tanzania and
Fiji", Journal of Development Studies 20:28-51.
El Shakhs, S. and H. Amirahmadi (1986), "Urbanization and spatial development in Africa: some critical
issues", African Urban Quarterly 1:3-19.
Evans, H.E. (1992), "A virtuous circle model of rural-urban development: evidence from a Kenyan small town
and its hinterland", Journal of Development Studies 28:640-667.
Evenson, R.E. and J.A. Roumasset (1986), "Markets. institutions and family size in rural Philippine
households", Journal of Philippine Development 13:141-162.
Faini, R. and J. de Melo, eds. (1993), Fiscal Issues in Adjustment in Developing Countries (St. Martin's Press,
New York, Macmillan Press, London).
Faiz, A. (1992), "Motor vehicle emissions in developing countries: relative implications for urban air quality".
in A. Kreimer and M. Munasinghe, eds., Environmental Management and Urban Vulnerability, World Bank
Discussion Paper No. 168.
Falaris, E.M. (1987), "A nested logit migration model with selectivity", International Economic Review
28:429-443.
Fallon, P. (1983), "Education and the duration of job search and unemployment in rural India", Journal of
Development Economics 12:327-3440.
Fallon, P. (1985), "The labor market in Kenya: recent evidence". Development Research Department, Labor
Market Division (World Bank. Washington, DC).
Fang, C. (1997), 'The impact of labour mobility on market development and economic growth", in China
Population Association, Symposium on Demography of China. Population Association and IUSSP (Beijing,
China).
Farvacque, C. and P. McAuslan (1992), "Reforming urban land policies and institutions in developing
countries", Urban Management Program Policy Paper 5 (World Bank, Washington, DC).
Feeney, G., F. Wang, M. Zhou and B. Xiao (1989). "Recent fertility dynamics in China: results from the 1987
one percent population survey", Population and Development Review 15:297-322.
Ferguson, B.W. (1992), "Land markets and the effect of property rights and regulations on development in
Indonesia", Unpublished manuscript (Urban Institute, Washington. DC).
Fields, G. (1975), "Rural-urban migration, urban unemployment and under-employment, and job-search
activities in LDCs". Journal of Development Economics 2:165-187.
Fields, G. (1982), "Place to place migration in Colombia", Economic Development and Cultural Change
30:539-558.
Fields, G. (1989), "On-the-job search in a labor market model: ex ante choices and ex post outcomes", Journal
of Development Economics 30:159-178.
Ch. 43. Urbanizationin Transforming Economies 1775
Fields, G. and H. Wan (1989), "Wage-setting institutions and economic growth", World Development 17:1471-
1483.
Findlay, S.E. (1987), "An interactive contextual model of migration in Ilocos None, the Philippines",
Demography 24:163--190.
Findlay, S.E. and A. Diallo (1993), "Social appearances and economic realities of female migration in rural
Mali", in United Nations, ed., Migration of Women in Developing Countries (UN, New York).
Firdausy, C.M. (1993), Critical Issues and Policy Measures to Address Urban Poverty in Indonesia, Un-
published manuscript, Seminar on Critical Issues and Policy Measures to Address Urban Poverty (Asian
Development Bank, Manila).
Firman, T. (1992), 'The spatial pattern of urban population growth in Java, 1980-1990", Bulletin of Indonesian
Economic Studies 28:95-109.
Fleck, S. (1997), "Choice or bargain? Married women's labor force participation in Hondurus", Ph.D.
Dissertation, Department of Economics, American University.
Follain, J.R. and E. Jimenez (1985), "The demand for housing characteristics in developing countries", Urban
Studies 22:421-432.
Follain, J.R., B. Renaud, and G.-C. Lim (1979), "Economic forces underlying urban decentralization trends: a
structural model for density gradients applied to Korea", Environment and Planning A 11:541-551.
Foot, D. and W. Milne (1984), "Net migration estimation in an extended, multiregional gravity model", Journal
of Regional Science 24:1.
Francis, E. and J. Hoddinott (1993), "Migration and differentiation in western Kenya: a tale of two
sub-locations", Journal of Development Studies 30:115-145.
Friedman, J., E. Jimenez and S. Mayo (1988), 'The demand for tenure security in developing countries",
Journal of Development Economics 29:185-198.
Fujita, M. (1991), "A rational expectations equilibrium model of urban growth and land markets", Journal of
Real Estate Finance and Economics 4:225-265.
Fuller, T.D., P. Lightfoot and P. Kamnuansilpa (1985). "Rural-urban mobility in Thailand: a decision-making
approach", Demography 22:565-580.
Funkhouser, E. (1997), "Demand-side and supply-side explanations for barriers to labor market mobility in
developing countries: the case of Guatemala", Economic Development and Cultural Change 45:341-366.
Gabriel, S., J. Shack-Marquez and W. Wascher (1993), "Does migration arbitrage regional labor market
differentials", Regional Science and Urban Economics 23:211-233.
Gaile, G.L. and H.R. Aspaas (1991), "Kenya's spatial dimensions of development strategy", Urban Geography
12:381-386.
Galloway, L. and R. Vedder (1985), "Migration efficiency ratios and the optimal distribution of population",
Growth and Change 16:3-7.
Garr, D.J. (1996), "Expectative land rights, house consolidation and cemetary squatting: some perspectives
from Central Java", World Development 24:1925-1933.
Gaviria, J., V. Bindlish and U. Lele (1989), 'The rural road question and Nigeria's agricultural development",
MADIA Discussion Paper 10 (World Bank, Washington, DC).
Geyer, H. and T. Kontuly (1993), "A theoretical foundation for the concept of differential urbanization",
International Regional Science Review 15:157-177.
Ghai, D. and S. Radwan, eds. (in press), Agrarian Policies and Rural Poverty in Africa (International Labour
Office, Geneva).
Gibbon, P., ed. (1993). Social Change and Economic Reform in Africa (Scandinavian Institute of African
Studies, Uppsala).
Gilbert, Alan (1990), 'The provision of public services and the debt crisis in Latin America: the case of
Bogota", Economic Geography 66:349-361.
Gilbert, Alanon (1990), "Urbanization at the periphery: reflections on the changing dynamics of housing
and employment in Latin America". in D. Drakakis-Smith, ed., Economic Growth and Urbanization in
Developing Areas (Routledge, London and New York).
Gindling, T. (1991), "Labor market segmentation and the determination of wages in the public, private-formal,
and informal sectors in San Jose, Costa Rica", Economic Development and Cultural Change 39:584-605.
1776 C.M. Becker and A.R. Morrison
Glewwe, P. (1988), "The distribution of welfare in C6te d'lvoire in 1985". Living Standards Measurement
Study, Working Paper No. 29 (World Bank, Washington, DC).
Glewwe. P and D. de Tray (1988), "The poor during adjustment: a case study of Ctte d'Ivoire", Living
Standards Measurement Study, Working Paper No. 47 (World Bank. Washington, DC).
Goldstein, A., S. Goldstein and S. Guo (1991), "Temporary migrants in Shanghai households, !984".
Demography 28:275-292.
Goldstein, S. (1990), "Urbanization in China. 1982-1987: effects of migration and reclassification", Population
and Development Review 16:673-702.
Goode, R. (1993), "Tax advice to developing countries: an historical survey", World Development 21:37-53.
Goza, EW., E. Rios-Neto and P. Vieira (1993), "The consequences of temporary out-migration for the fami-
lies left behind: the case of Jequitinhonha Valley, Brazil", in United Nations, ed., Migration of Women in
Developing Countries (UN, New York).
Greenaway, D. and O. Morrissey (1993), "Structural adjustment and liberalization in developing countries:
what lessons have we learned", Kyklos 46:241-261.
Greene, D.L. (1978), Multinucleation in Urban Spatial Structure, Unpublished Ph.D. dissertation (Johns
Hopkins University, Baltimore).
Greenwood, M. (1971), "A regression analysis of migration to urban areas of a less-developed country: the
case of India", Journal of Regional Science 11:253-262.
Greenwood, M. (1985), "Human migration: theory, models and empirical studies", Journal of Regional Science
25:521-543.
Greenwood, M. (1991), "Migration, regional equilibrium, and the estimation of compensating differentials",
American Economic Review 81:1382-1390.
Greenwood, M.J. and G.L. Hunt (1984), "Econometrically accounting for identities and restrictions in models
of interregional migration", Regional Science and Urban Economics 14:113-128.
Gregory, J. (1986), "Migration and urbanization trends", Nairobi: paper delivered at the IUSSP Seminar on
Economic Consequences of Population Trends in Africa.
Grewe, C.D. (1999), Household Demand and Composition in Indonesia, Unpublished Ph.D. dissertation
(University of Colorado, Department of Economics, Boulder, CO).
Grimes, OF. Jr. (1976). Housing for Low Income Urban Families (The Johns Hopkins University Press,
Boulder, CO).
Gronau, R. (1973), "The effect of children on the housewife's value of time", Journal of Political Economy
81:S168-S199.
Gmotaert, C. (in press), "The position of migrants in the urban informal labour markets of C6te d'Ivoire",
Journal of African Economies 1:416-445.
Gruver, G.W. and L.A. Zeager (1993), "Agricultural price policy in a less developed economy with endogenous
urban wages", Journal of Economic Development 18:27-41.
Gruver, G.W. and L.A. Zeager (1994), "Steady-state labor turnover as optimal household behavior", Journal of
Regional Science 34:75-90.
Gugler, J. (1991), "Life in a dual system revisited: urban-rural ties in Enugu, Nigeria, 1961-1987", World
Development 19:399-409.
Gupta, I. and A. Mitra (1997), "Does duration of migration matter? Poverty and occupational choice among
slum dwellers in Delhi", Unpublished manuscript (Institute of Economic Growth, Delhi).
Haessel, W.W. (1978), "Macroeconomic policy, investment and urban unemployment in less developed
countries", American Joumal of Agricultural Economics 60:29-36.
Haggblade, S. and J. Brown (1989), "Farm-nonfarm linkages in rural sub-Saharan Africa", World Development
17:1173-1201.
Hamer, A. (1988a), China Zhejiang: Challenges of Rapid Urbanization, Report No. 6612-cha (World Bank,
Washington, DC).
Hamer, A. (1988b), China Urban Housing Reform: Issues and Implement Options, Report No. 8222-cha (World
Bank, Washington, DC).
Hamer. A. M. (1984), "Decentralized urban development and industrial location behavior in Sao Paulo, Brazil".
Water Supply and Urban Development Department, Discussion Paper UDD 29 (World Bank. Washington,
DC).
Ch. 43: Urbanizationin Transforming Economies 1777
Hamer, A.M. (1985). "Bogota's unregulated subdivisions: the myths and realities of incremental housing
construction", Staff Working Paper No. 734 (World Bank, Washington, DC).
Hamid, N., I. Nabi and A. Nasim (in press), Trade, Exchange Rate, and Agricultural Pricing Policies in Pakistan
(World Bank, Washington, DC).
Hannah, L., K.-H. Kim and E.S. Mills (1993), "Land use controls and housing prices in Korea", Urban Studies
30:147-156.
Hansen, B. (1979), "Colonial economic development with an unlimited supply of land: a Ricardian case",
Economic Development and Cultural Change 27:611-628.
Hansen, E. (1983), "Why do firms locate where they do?", Water Supply and Urban Development Department,
Discussion Paper UDD 25 (World Bank, Washington, DC).
Hansen, K.T. (1982), "Lusaka's squatters: past and present", African Studies Review, 25:117-236.
Hanson, G.H. (1994), "Localization economies, vertical organization, and trade", NBER Working Paper No.
4744 (Cambridge, MA).
Hanson, P. (1933), "Local power and market reform in Russia", Communist Economies and Economic
Transformation 5:45-60.
Hardoy, J.-E. and D. Satterthwaite (1986), "Government policies and small and intermediate urban centres", in
J.E. Hardoy and D. Satterthwaite, eds., Small and Intermediate Urban Centres: Their Role in Regional and
National Development in the Third World (Westview, Boulder, CO).
Harrigan, E and P. McGregor (1993), "Equilibrium and disequilibrium perspectives on regional labor
migration", Journal of Regional Science 33:49-67.
Harris, J. and R. Sabot (1982), "Urban unemployment in LDC's: towards a more general search model", in
R.H. Sabot, ed., Migration and the Labor Market in Developing Countries (Westview, Boulder, CO).
Harris, J. and M. Todaro (1970), "Migration, unemployment and development: a two-sector analysis",
American Economic Review 60:126-142.
Hay, M.J. (1980), "A structural equations model of migration in Tunisia", Economic Development and Cultural
Change 28:345-358.
Hazari, B.R. and PM. Sgro (1991), "Urban-rural structural adjustment, urban unemployment with traded and
nontraded goods", Journal of Development Economics 35:187-196.
Hazlewood, A. (1989), Education, Work, and Pay in East Africa (Oxford University Press/Clarendon Press,
New York).
Henderson, J.V. (1979), "The economics of systems of cities", Brown University Working Paper No. 78-22.
Henderson, J. V. (1980), "A framework for international comparisons of systems of cities", World Bank, Urban
and Regional Economics Division, Urban and Regional Report No. 80-3.
Henderson, J.V. (1982a), "Urban economies of scale in Brazil", Water Supply and Urban Development
Department Discussion Paper UDD-17 (World Bank, Washington, DC).
Henderson, JV. (1982b), "The impact of government policies on urban concentration", Journal of Urban
Economics 12:280-303.
Henderson, J.V. (1986a), "Urbanization in a developing country", Journal of Development Economics 22:269-
293.
Henderson, J.V. (1986b), "The timing of regional development", Journal of Development Economics 23:275-
292.
Henderson, J.V. (1988a), Urban Development: Theory, Fact, and Illusion (Oxford University Press, New York).
Henderson, J.V (1988b), "General equilibrium modeling of systems and cities", in P Nijkamp and E. Mills,
eds., Handbooks in Regional and Urban Economics, Vol. II, (North-Holland, Amsterdam).
Henderson, J.V. (1994), "Where does an industry locate?", Journal of Urban Economics 35:83-104.
Henderson, J.V. and YM. Ioanides (1981), "Aspects of growth in a system of cities", Journal of Urban
Economics 10:117-139.
Hercowitz, Z. and D. Pines (1991),"Migration with fiscal extemalities", Journal of Public Economics 46:163-
180.
Henderson, J.V. and A. Kuncoro (1994), "Industrial centralization in Indonesia", Unpublished manuscript
(Brown University Department of Economics, Providence, RI).
Henderson, J.V., A. Kuncoro and M. Turner (1995a), "Industrial development of cities", Journal of Political
Economy 103:1067-1090.
1778 C.M. Becker and A.R. Morrison
Henderson, J.V.. A. Kuncoro and D. Nasution (1995b), The dynamics of Jabotabek development", Unpub-
lished manuscript (Brown University Department of Economics. Providence, RI).
Herzog. H. and A. Schlottmann (1981), 'Labor force migration and allocative efficiency in the United States:
the roles of information and psychic costs". Economic Inquiry 19:459-475.
Herzog, H., A. Schlottmann and T. Boehm (1993), "Migration as spatial job-search: a survey of empirical
findings",. Regional Studies 27:327-340.
Hoddinott, J. (1992), "Modeling remittance flows in Kenya", Journal of African Economies 1:206-232.
Hoddinott, J. (1994), "A model of migration and remittances applied to western Kenya", Oxford Economic
Papers 46:459-476.
Horton, S., R. Kanbur and D. Mazumdar (1994), "Labor markets in an era of adjustment: an overview", in S.
Horton, R. Kanbur and D. Mazumdar, eds., Labor Markets in an Era of Adjustment: Vol. 1, Issues Papers
(World Bank, Washington, DC).
Hotchkiss, D. and M.J. White (1993). "A simulation model of a decentralized metropolitan area with two-
worker, "traditional", and female-headed households", Journal of Urban Economics 34:159-185.
House, W. (1984), "Nairobi's informal sector: dynamic entrepreneurs or surplus labor?", Economic Develop-
ment and Cultural Change 32:277-302.
House, W. (1987), "Labor market differentiation in a developing economy: an example from urban Juba,
Southern Sudan", World Development 15:877-897.
House, W. (1992), "Priorities for urban labor market research in anglophone Africa". Joumal of Developing
Areas 27:49-67.
House, W. and H. Rempel (1980), "The determinants of interregional migration in Kenya", World Development
8:25-36.
Hoy, M. and E. Jimenez (1991), "Squatters' rights and urban development: an economic perspective",
Economica 58:79-92.
Hugo, G.J. (1993), "Migrant women in developing countries", in United Nations, ed.. Migration of Women in
Developing Countries (UN, New York).
Hyatt, D.E. and W.J. Milne (1990). "Urban and rural fertility differentials in Kenya: an econometric analysis
using micro data", Institute for Policy Analysis Working Paper (University of Toronto, Toronto. ONT).
Imai, K. (1997), "Urban structural change in Asian LD)Cs' Cities", Unpublished manuscript (University of
Colorado Department of Economics, Boulder, CO).
Infante, R. and E. Klein (1991), 'The Latin American labour market, 1950-1990", CEPAL Review 45:121-
135.
Ingram, G.K. (1980), "Land in perspective: its role in the structure of cities", Urban and Regional Economics
Division Urban and Regional Report No. 80-9 (World Bank, Washington, DC).
Ingram, G. and A. Carroll (1981), "The spatial structure of Latin American cities", Journal of Urban Economics
9:257-273.
Ishumi, A.G.M. (1984), The Urban Jobless in Eastern Africa (Scandinavian Institute of African Studies,
Uppsala).
Ito, K. (1992), "The reinstatement of local fiscal autonomy in the Republic of Korea", Developing Economies
30:404-429.
Jamal, V. (1984), "Rural urban gap and income distribution: synthesis report of seventeen African countries",
Addis Ababa, ILO/JASPA.
Jamal, V. and J. Weeks (1988), 'The vanishing rural-urban gap in sub-Saharan Africa". International Labour
Review 127:271-292.
Jimenez, E. (1982). "The value of squatter dwellings in developing countries", Economic Development and
Cultural Change 30:739-752.
Jimenez, E. (1985), "Urban squatting and community organization in developing countries", Journal of Public
Economics 27:69-92.
Jones, G. (1993), 'The role of female migration in development", in United Nations, ed., Migration of Women
in Developing Countries (UN, New York).
Joshi, H., H. Lubell and J. Mouly (1976), Abidjan: Urban Development and Employment in the Ivory Coast
(International Labour Office, Geneva).
Ch. 43: Urbanization in Transforming Economies 1779
Kaganova, 0. and M. Berezin (1993), "The market for urban real estate: development tendencies", Voprosy
Ekonomiki 1993:7 (in Russian).
Kahimbaara, J. A. (1993), "The convolution of urban planning with tradition in Lesotho, 1928-1991",
Environment and Planning A 25(7):100-120.
Kahn, J. and H. Ofek (1992), "The equilibrium distribution of population and wages in a system of cities",
Review of Regional Studies 22:201-216.
Kahnert, F (1987), "Improving urban employment and labor productivity", Discussion Paper 10 (World Bank,
Washington).
Kannappan, S. (1983), Employment Problems and the Urban Labor Market in Developing Countries, Graduate
School of Business Administration, Division of Research (University of Michigan, Ann Arbor).
Kannappan, S. (1984), "Tradition and modernity in urban employment in developing nations", Regional
Development Dialogue 5:55-62.
Kannappan, S. (1988), "Urban labor markets and development", World Bank Research Observer 3:189-206.
Kannappan, S. (1989), "Employment policy and labour markets in the developing nations: a perspective on
the state-of-the-art in the 1980s", in Bernard Salome, ed., Fighting Urban Unemployment in Developing
Countries (IECD, Paris).
Kannappan, S. (1990), "A heterodox model of employment and labor underutilization: Sri Lanka in 1971 and
1989", Unpublished manuscript, Department of Economics (Michigan State University, E. Lansing, MI).
Kano, H. (1992), "The reorientation of Indonesia's centralized budgetary system", Developing Economies
30:377-403.
Katz, E. (1997), "The Intra-household economics of voice and exit: Evaluating the feminist institutional content
of family resource allocation models", Working Paper 97-05, Barnard College, Department of Economics.
Katz, E. and O. Stark (1985a), "Desired fertility and migration in LDCs: signing the connection", Center
for Population Studies, Migration and Development Program Discussion Paper 15 (Harvard University,
Cambridge, MA).
Katz, E. and O. Stark (1985b), "On fertility, migration and remittances in LDCs", Center for Population
Studies, Migration and Development Program Discussion Paper 19 (Harvard University, Cambridge, MA).
Katz, E. and 0. Stark (1986a), "On the shadow wage of urban jobs in less developed countries", Joumal of
Urban Economics 20.
Katz, E. and O. Stark (1986b), "Labor migration and risk aversion in less developed countries", Journal of
Labor Economics 41:134-149.
Kelley, A.C. (1991), "African urbanization and city growth: perspectives, problems, and policies", Unpublished
manuscript (Duke University Department of Economics, Durham, NC).
Kelley, A.C. and R. Schmidt (1983), "Incorporating demography into general equilibrium modeling", in
A.C. Kelley, W.C. Sanderson and J.G. Willamson eds. Modeling Growing Economies in Equilibrium and
Disequilibrium.Durham, NC: Duke Univ. Press, pp. 317-337.
Kelley, A.C. and J.G. Williamson (1984), What Drives Third World City Growth? A Dynamic General
Equilibrium Approach (Princeton University Press, Princeton, NJ).
Kelley, A.C., J.G. Williamson, and R.J. Cheetham (1972), Dualistic Economic Development: Theory and
History (University of Chicago Press, Chicago).
Khasiani, S. (1991), "A model of migration motivations: migration valuations and migration dispositions
among rural youth in Kenya", Eastern Africa Economic Review 7:27-38.
Kim, S.T., C.S. Jung and K.H. Roh (1991), "Regional economic gap in Korea", Journal of Economic
Development 16:145-167.
Kingsley, G.T., B.W. Ferguson, and B.T. Bower with S.R. Dice (1994), "Managing urban environmental quality
in Asia", Technical Paper No. 220 (World Bank, Washington, DC).
Klak, T. and M. Holtzclaw (1993), "The housing, geography, and mobility of Latin American urban poor: the
prevailing model and the case of Quito, Ecuador", Growth and Change 24:247-276.
Knight, J.B, and R.H. Sabot (1981), "The returns to education: increasing with experience or decreasing with
expansion?", Oxford Bulletin of Economics and Statistics 43:51-71.
Knight, J.B. and R.H. Sabot (1982), "From migrants to proletarians: employment experience, mobility, and
wages in Tanzania", Oxford Bulletin of Economics and Statistics 44:199-226.
1780 C.M. Becker and A.R. Morrison
Knight, J.B. and R.H. Sabot (1988), "Lewis through a looking glass: public sector employment, rent-
seeking and economic growth", Economics Research Memorandum RM-108 (Williams College Center for
Development, Williamstown. MA).
Knight, J. and R.H. Sabot (1990), Education Productivity and Inequality: The East African Natural Experiment
(Oxford University Press. New York).
Knowles, J.C. and R. Anker (1981), "An analysis of income transfers in a developing country: the case of
Kenya", Journal of Development Economics 8:205-226.
Kojima. R. (1992a), "Introduction (to fiscal decentralization)", Developing Economies 30:311-314.
Kojima, R. (1992b), "The growing fiscal authority of provincial-level governments in China", Developing
Economies 30:315-346.
Konadu-Agyemang, K.O. (1991), "Reflections on the absence of squatter settlements in west African cities:
the case of Kumasi, Ghana"'. Urban Studies 28:139-151.
Kreimer, A. and M. Munasinghe, eds. (1992), Environmental Management and Urban Vulnerability, World
Bank Discussion Papers, No. 168 (World Bank. Washington, DC).
Krupnick, A.J. (1991), "Transportation and air pollution in urban areas of developed and developing countries".
Resources for the Future, Quality of the Environment Division Discussion Paper: QE91-08.
Kulaba, S. (1989), "Local government and the management of urban services in Tanzania". in R. Stren and R.
White, op. cit.
Lam, D. and R.F. Schoeni (1993), "Effects of family background on earnings and returns to schooling: evidence
from Brazil", Journal of Political Economy 101:710-740.
Lauby, J. and 0. Stark (1988), "Individual migration as a family strategy: young women in the Philippines",
Population Studies 42:473-486.
Lecaillon. J., F. Paukert, C. Morrisson and M. Germidis (1984), Income Distribution and Economic Develop-
ment (Intemational Labour Office, Geneva).
Ledent, J. (1982a), "Rural-urban migration, urbanization, and economic development", Economic Develop-
ment and Cultural Change 30:507-538.
Ledent, J. (1982b), 'The factors of urban population growth: net immigration versus natural increase",
International Regional Science Review 7: 99-126.
Ledent, J. (1990), "The distribution of city sizes in subSaharan Africa: a cross-sectional analysis", INRS-
urbanisation, Unpublished manuscript presented at the 1990 North American meetings of the Regional
Science Association (Universite du Quebec au Montreal, Montreal).
Lee, K.S. (1981), "Intra urban location of manufacturing employment in Colombia", Joumal of Urban
Economics 9:222-241.
Leigland, J. (1993), "Decentralizing the development budget process in Indonesia: progress and prospects",
Public Budgeting and Finance 13:85-101.
Leinnann, J. (1991), "Energy-environment linkages in the urban sector", World Bank Urban Management
Program Discussion Paper No. 2 (World Bank, Washington, DC).
Lent, G.E. (1978), "Experience with urban land value tax in developing countries", Bulletin for International
Fiscal Documentation 32:75-83.
Levy, M.B. and W.J. Wadycki (1974), "Education and the decision to migrate: an econometric analysis of
migration in Venezuela", Econometrica 42:377-388.
Lewis, W.A. (1954), "Economic development with unlimited supplies of labour", The Manchester School
22:139-191.
Liang, Z. and M.J. White (1996), "Internal migration in China, 1950-1988", Demography 33:375-384.
Liang, Z. and M.J. White (1997), "Market transition, government policies, and inter-provincial migration in
China: 1983-1988", Economic Development and Cultural Change 45:321-339.
Lim, G.C., J. Follain Jr. and B. Renaud (1984), "Economics of residential crowding in developing countries",
Journal of Urban Economics 16:173-186.
Lim. J.D. (1993). "Urban growth and industrial restructuring: the case of Pusan", Environment and Planning
A 25:95-109.
Lim, L.L. (1993), 'The structural determinants of female migration", in United Nations, ed., Migration of
Women in Developing Countries (UN, New York).
Ch. 43: Urbanizationin Transforming Economies 1781
Linn, J.E (1979), "Policies for efficient and equitable growth of cities in developing countries", World Bank
Staff Working Paper No. 342.
Linn, J. (1982), "The costs of urbanization in developing countries", Economic Development and Cultural
Change 30:625-648.
Lipton, M. (1976), Why Poor People Stay Poor: Urban Bias in World Development (Harvard University Press,
Cambridge, MA).
Lipton, M. (1984), "Urban bias revisited", Journal of Development Studies 20:139-166.
Lipton, M. (1991), "A reassessment of Kenya's rural and urban informal sector", World Development 19:651-
670.
Lipton, M. (1993), "Urban bias: of consequences, classes and causality", Journal of Development Studies
29:229-258.
Livas E.R. and P. Krugman (1992), "Trade policy and the third world metropolis", NBER Working Paper No.
4238 (Cambridge, MA).
Locay, L. (1990), "Economic development and the division of production between households and markets",
Journal of Political Economy 98:965-982.
Lodhi, A. and H.A. Pasha (1991), "Housing demand in developing countries: a case-study of Karachi in
Pakistan", Urban Studies 28:623-634.
Lovell, P. (1993), "The geography of economic development and racial discrimination in Brazil", Development
and Change 24:83-101.
Lowder, S. (1978), "The limitations of planned land development for low-income housing in third world",
Urban Studies 30:241-255.
Lubell, H. and C. Zarour (1990), "Resilience amidst crisis: the informal sector of Dakar", International Labour
Review 129:387-396.
Lucas, R. (1982), "Determinants of migration decisions", in Migration in Botswana: Patterns, Causes, and
Consequences. Final Report of the National Migration Study (Government Printers, Garborone, Botswana).
Lucas, R.E.B. (1997), "Internal migration in developing countries", in M. Rosenzweig and 0. Start, eds.,
Handbook of Population and Family Economics (North-Holland, Amsterdam).
Lucas, R.E.B. and O. Stark (1985), "Motivations to remit: evidence from Botswana", Journal of Political
Economy 93:901-918.
Lucas, R.E.B. and 0. Stark (1988), "Migration, remittances, and the family", Economic Development and
Cultural Change 36:465-482.
Lumbantaruan, H. (1997), "Relative efficiency of public and private enterprises in Indonesia", Boulder, CO:
Univ. of Colorado Dept. of Economics, unpubl. Ph.D. diss.
Lundberg, S. and R. Pollak (1993), "Separate spheres bargaining and the marriage market", Journal of Political
Economy 101:988-1010.
Ma, L.J.C. and M. Fan (1994), "Urbanisation from below: the growth of Jiangsu, China", Urban Studies
31:1625-1645.
Mabogunje, A. (1983), "The case for big cities", Habitat International 7:
Machado, C.C. and J.E Abreu (1991), "The elderly mobility transition in Brazil", Paper presented at University
of Colorado Program in Population Processes seminar on the Elderly Mobility Transition, Unpublished
manuscript.
Maddala, G.S., Y. Toda and N. Nozdrina (1993), "Analysis of sales prices for apartments in Moscow", Voprosy
Ekonomiki 1993:7 (in Russian).
Magnac, T. (1991), "Segmented or competitive labor markets", Econometrica 59:165-187.
Malpezzi, S. (1990), "Urban housing and financial markets: some international comparisons", Urban Studies
27:971-1022.
Malpezzi, S. (1991), "Discounted cash flow analysis: present value models of housing programmes and poli-
cies", in A.G. Tipple and K.G. Willis, eds, Housing the Poor in the Developing World: Methods of Analysis,
Case Studies and Policy (Routledge, New York).
Malpezzi, S. (1993), "Can New York and Los Angeles learn from Kumasi and Bangalore? Costs and benefits
of rent controls in developing countries", Housing Policy Debate 4:589-626.
Malpezzi, S. and G. Ball (1991), "Rent control in developing countries", Discussion Paper No. 129 (World
Bank, Washington, D.C.).
1782 C.M. Becker and A.R. Morrison
Malpezzi, S. and S. Mayo (1987), "The demand for housing in developing countries: empirical estimates from
household data". Economic Development and Cultural Change 35:687-721.
Malpezzi, S. and S.K. Mayo (1987), "User cost and housing tenure in developing countries", Journal of
Development Economics 25:197-220.
Manning, C. (1988), "Rural employment creation in Java: lessons from the green revolution and oil boom',
Population and Development Review 14:47-80.
Marcouiller, D., V. Ruiz de Castillo, and C. Woodruff (1997), "Formal measures of the informal-sector wage
gap in Mexico, El Salvador. and Peru", Economic Development and Cultural Change 45:367-392.
Markusen, J.R. and T.F. Rutherford (1993), "Discrete plant-location decisions in an applied general-
equilibrium model of trade liberalization", NBER Working Paper No. 4513 (Cambridge, MA).
Martin, M.E (1992), "Urban incomes, worker's democracy and the spring uprising", Review of Radical
Political Economics 24:136-165.
Massey, D.S. (1990),"Social structure. household strategies, and the cumulative causation of migration".
Population Index 56:33-26.
Mathieu, R. (1982), "Locational analysis of bank funded projects in Ivory Coast, Cameroon, and Mali", Water
Supply and Urban Development Department, Discussion Paper UDD 28 (World Bank, Washington, DC).
Mathur, O.P. (1982), "The role of small cities in national development re-examined". in Mathur, ed., Small
Cities and National Development (United Nations Centre for Regional Development. Nagoya, Japan).
Mathur, V. and S. Stein (1993), "The role of amenities in a general equilibrium model of regional migration
and growth", Southern Economic Journal 59:394-409.
Maxwell, D.G. (1995), "Alternative food security strategy: a household analysis of urban agriculture in
Kampala", World Development 23:1669-1681.
Mayo, S.K. and S. Angel (1993), Housing Enabling Markets to Work, A World Bank Policy Paper (World
Bank, Washington, DC).
Mayo, S. and D. Gross (1987), "Sites and services-and subsidies: the economics of housing in developing
countries", World Bank Economic Review 1:301-335.
Mayo, S., S. Malpezzi, and D. Gross (1986), "Shelter strategies for the urban poor in developing countries',
World Bank Research Observer 1:183-203.
Mazumdar, D. (1976), "The urban informal sector", World Development 4:655-679.
Mazumdar, D. (1981), The Urban Labor Market and Income Distribution: A Study of Malaysia (Oxford
University Press, New York).
Mazumdar, D. (1983), 'The rural-urban wage gap migration and the working of urban labor market: an
interpretation based on a study of the workers of Bombay city", Indian Economic Review 18:169-198.
Mazumdar, D. (1987), "Rural-urban migration in developing countries", in E.S. Mills, ed., Handbook of
Regional and Urban Economics, vol. II (Elsevier Science Publishers, New York).
Mazumdar, D. (1989), "Microeconomic issues of labor markets in developing countries: analysis and policy
implications", Economic Development Institute Seminar Paper No. 40 (World Bank, Washington, DC).
Mbuyi, K. (1989), "Kinshasa: problems of land management, infrastructure and food supply", in R. Stren and
R. White, op. cit.
McCleary, W.A. (1991), 'The design and implementation of conditionality", in V. Thomas, A. Chhibber, M.
Dailami and J. de Melo. eds., Restructuring Economics in Distress: Policy Reform and the World Bank
(Oxford University Press, New York).
McCutcheon, L. (1994), "Measuring migrant change: a look at housing in Bogota, Seoul, and Surabaya".
Journal of Developing Areas 18:357-372.
Mead, D.C. and C. Morrison (1996), "The informal sector elephant", World Development 24:1611-1619.
Megbolugbe, IF. (1989), "A hedonic index model: the housing market of Jos, Nigeria", Urban Studies 26:486-
494.
Mellor, J. (1976), The New Economics of Growth: A Strategy for India and the Developing World (Cornell
University Press, Ithaca, NY).
Mera, K. (1973), "On the urban agglomeration and economic efficiency". Economic Development and Cultural
Change 21:309-324.
Mera, K. (1984). "Measuring economic contributions of infrastructure in cities of developing countries", Water
Supply and Urban Development Department, Discussion Paper UDD 34 (World Bank, Washington, DC).
Ch. 43: Urbanizationin Transforming Economies 1783
Mills, E.S. and C.M. Becker (1986), Studies in Indian Urban Development (Oxford University Press, New
York).
Mills, E.S. and J.P. Tan (1980), "A comparison of urban population density functions in developed and
developing countries", Urban Studies 17:313-321.
Mitra, A. (1994), Urbanisation, Slums, Informal Sector Employment and Poverty (B.R. Publishing Corpora-
tion, Delhi).
Mitra, A. (1997), "Industry, agglomeration economies and urban growth", Unpublished manuscript (Institute
of Economic Growth, Delhi).
Mohan, R. (1985), "Urbanization in India's future", Population and Development Review 11:619-645.
Mohan, R. (1986), Work, Wages, and Welfare in a Developing Metropolis: Consequences of Growth in Bogota,
Colombia (Oxford University Press, New York).
Montgomery, M.R. (1987), "The impacts of urban population growth on urban labor markets and the costs of
urban service delivery: a review", in D.G. Johnson and R.D. Lee, eds., Population Growth and Economic
Development: Issues and Evidence, Social Demography Series (University of Wisconsin Press, Madison,
WI), pp. 149-188.
Montgomery, M. and E. Brown (1985), Migration and Urbanization in Sub-Saharan Africa, Population, Health
and Nutrition Department (World Bank, Washington, DC).
Moomaw, R. (1981), "Productivity and city size: a critique of the evidence", Quarterly Journal of Economics
96:675-688.
Moomaw, R. and A.M. Shatter (1996), "Urbanization and economic development: a bias toward large cities?",
Journal of Urban Economics 40:13-37.
Morley, S. (1988), "Relative wages, labor force structure, and the distribution of income in the short and long
run", Economic Development and Cultural Change 36:651-668.
Morley, S. (1994), Poverty and Inequality in Latin America: Past Evidence, Future Prospects (Overseas
Development Council, Washington, DC).
Morrison, A. (1994), "Are institutions or economic rents responsible for interindustry wage differentials?",
World Development 22:355-368.
Morrison, A. (1992), "Inter-industry wage differentials in LDCs: are institutions or economic rents responsi-
ble?", Department of Economics Working Paper (Tulane University, New Orleans, LA).
Morrison, A. (1993), "Unproductive migration reconsidered: a stochastic frontier production function frame-
work for analyzing internal migration", Oxford Economic Papers 45:501-518.
Mosley, P., J. Harrigan and J. Toye (1991), Aid and Power: The World Bank and Policy-Based Lending, vols.
1 and 2 (Routledge, London).
Munasinghe, M. (1992), Water Supply and Environmental Management: Developing World Applications
(Westview Press, Boulder and Oxford).
Muth, R. (1969), Cities and Housing (University of Chicago Press, Chicago).
Nabi, I. (1984), "Village-end considerations in rural-urban migration", Journal of Development Economics
14:129-145.
Nafziger, E.W. (1988), Inequality in Africa: Political Elites, Proletariat, Peasants and the Poor (Cambridge
University Press, Cambridge, UK).
Nakagome, M. (1989), "Urban unemployment and the spatial structure of labor markets: an examination of the
"Todaro paradox" in a spatial context", Journal of Regional Science 29:161-170.
Natskomstat National Statistical Committee of the Kyrgyz Republic (1995), Demograficheskii Ezhegodnik
Kyrgyzskoi Respubliki 1994 (Natskomstat, Bishkek).
Newbery, DM. (1988), "Road transport fuel pricing policy", Department of Applied Economics, Working
Paper: 889 (University of Cambridge).
Niedercorn, J.H. and B.V. Bechdolt, Jr. (1969), "An economic derivation of the 'gravity law' of spatial
interaction", Journal of Regional Science 9:2.
Oaxaca, R. (1973), "Male-female wage differentials in urban labor markets", International Economic Review
3:693-709.
Oberai, A.S. (1993), Population Growth, Employment and Poverty in Third-World Mega-Cities (St. Martin's
Press, New York).
1784 C. M. Becker and A.R. Morrison
Ogura, M. (1991), "Rural-urban migration in Zambia and migrant ties to home villages", Developing
Economies 29:145-165.
Oi, J.C. (1993), "Reform and urban bias in China", Journal of Development Studies 29:129-148.
Okojie, C.E.E. (1984), "Female migrants in the urban labor market: Benin City. Nigeria", Canadian Journal of
African Studies 18:547-562.
Okpala, D.G.I. (1978), "Housing standards: a constraint on urban housing production in Nigeria", Ekistics 240.
Ozo, A.O. (1986), "Residential location and intra-urban mobility in a developing country: some empirical
observations from Benin", Urban Studies 23:457-470.
Pachon Munoz, A. and S. de Hernandez (1989), "La Vivienda en Colombia, 1973-1985: La Distribucion
Espacial de la Poblacion en las Areas Metropolitanas", Boletin de Estadistica 438:245-258.
Panayotou, T. (1993), Green Markets: The Economics of Sustainable Development (ICS Press for the In-
ternational Center for Economic Growth and the Harvard Institute for International Development. San
Francisco).
Papageorgiou, Y.Y. (1990), The Isolated City State: An Economic Geography of Urban Spatial Structure
(Routledge, London).
Park, S.O. (1993),"Industrial restructuring and the spatial division of labor: the case of the Seoul metropolitan
region, the Republic of Korea", Environment and Planning A 25:81-93.
Pasha, H.A. (1992), "Maximum lot size zoning in developing countries", Urban Studies 29:1173-1181.
Paul, S. (1992), "Accountability in public services: exit, voice and control", World Development 20:1047-
1060.
Payne, G.K. and L. Payne (1991), "Urban land tenure and property rights", Urban Management Program,
Unpublished manuscript (World Bank, Washington, DC).
Pchelintsev, 0. (1993), "Regional hierarchy under threat: a spatial dimension of the socio-economic crisis in
the former Soviet Union and Russia", International Regional Science Review 15:267-280.
Peil, M. (1976), "Afican squatter settlements: a comparative study", Urban Studies 13:155-166.
Peil, M. (1979), "Housing the poor in West Africa, public and private provisions", Institute of Development
Studies Bulletin 10:28-32.
Petrakos, G.C. (1992), "Urban concentration and agglomeration economies: re-examining the relationship",
Urban Studies 29:1219-1229.
Petrakos. G. and J.C. Brada (1989), "Metropolitan concentration in developing countries", Kyklos 42:557-578.
Pfister, F. (1982), "Housing improvement and popular participation in the Upper Volta", Habitat International
6:209-214.
Phongpaichit, P. (1993), "The labour market aspects of female migration to Bangkok", in United Nations, ed.,
Migration of Women in Developing Countries (UN, New York).
Plane, D.A. (1993), "Demographic influences on migration", Regional Studies 27:375-383.
Population Reference Bureau (1997), 1997 World Population Data Sheet (PRB, Washington, DC).
Porter, R. (1993), "Toward an economic theory of the apartheid city", in L. Stetting, K.E. Svendsen and
Y. Yndgaard, eds., Global Change and Transformation: Economic Essays in Honor of Karsten Laursen
(Handelshojskolens Forlag, Copenhagen).
Portes, A. (1989), "Latin American urbanization during the years of the crisis", Latin American Research
Review 24:7-43.
Portes, A. and R. Schauffler (1993), "Competing perspectives on the Latin American informal sector",
Population and Development Review 19:33-60.
Preston, S.H. (1979), "Urban growth in developing countries: a demographic reappraisal", Population and
Development Review 5:195-215.
Pryor, F.L. (1988a), "Income distribution and economic development in Malawi", Discussion Paper No. 36
(World Bank. Washington, DC).
Pryor, FL. (1988b), "Income distribution and economic development in Madagascar", Discussion Paper No.
37 (World Bank, Washington DC).
Psacharopoulos, G. (1993), "Returns to investment in education: a global update", World Bank Policy Research
Working Paper WPS 1067 (World Bank, Washington, DC).
Puente, C.A. (1993), "The role of wage differentials in determining migration selectivity by sex: the case of
Brazil", in United Nations, ed., Migration of Women in Developing Countries (UN. New York).
Ch. 43: Urbanizationin TransformingEconomies 1785
Punpuing, S. (1993), "Correlates of commuting patterns: a case study of Bangkok, Thailand", Urban Studies
30:527-546.
Quan, Z.X. (1991), "Urbanisation in China", Urban Studies 28:41-51.
Raimondos, P. (1993), "On the Todaro paradox", Economics Letters 42:261-267.
Rakodi, C. (1988), "Urban agriculture: research questions and Zambian evidence", Journal of Modem African
Studies 26:495-515.
Rallis, T. (1988), City Transport in Developed and Developing Countries (St. Martin's Press, New York).
Ramnsdell, L. (1990), "National housing policy and the favela in Brazil", in L.S. Graham and R.H. Wilson, eds.,
The Political Economy of Brazil: Public Policies in an Era of Transition, Institute of Latin American Studies
Symposia on Latin America Series (University of Texas Press, Austin).
Rauch, J.E. (1989), "Urbanization, underemployment, and the size distribution of income: Harris-Todaro and
Kuznets's inverted U", University of California, San Diego Department of Economics Working Paper: 89-39.
Rauch, J. (1993a), "Productivity gains from geographic concentration of human capital: evidence from the
cities", Journal of Urban Economics 34:380-400.
Rauch, J.E. (1993b), "Economic development, urban underemployment, and income inequality", Canadian
Journal of Economics 26:901-918.
Ravallion, M. and D. van de Walle (1991), "Urban-rural cost-of-living differentials in a developing economy",
Journal of Urban Economics 29:113-127.
Rawski, T.G. and R.W. Mead (1998), "On the trail of China's phantom farmers", World Development 26:5.
Reccini de Lattes, Z. and S.M. Mychaszula (1993), "Female migration and labour force participation in
a medium-sized city of a highly urbanized country", in United Nations, ed., Migration of Women in
Developing Countries (UN, New York).
Redclift, M.R. (1984), "'Urban bias' and rural poverty: a Latin American perspective", Journal of Develop-
ment Studies 20:123-138.
Rempel, H. (1981), Rural-Urban Labor Migration and Urban Unemployment in Kenya, International Institute
for Applied Systems Analysis (Laxenburg, Austria).
Renaud, B. (1993), "The housing system of the former Soviet Union: does it need market reforms?", Voprosy
Ekonomiki 7:14-19 (in Russian).
Richardson, H. (1982), "Policies for strengthening small cities in developing countries", in Mathur, ed., Small
Cities and National Development, United Nations Centre for Regional Development (Nagoya, Japan).
Rickman, D. and G. Treyz (1993), "Alternative labor market closures in a regional model", Growth and Change
24:32-50.
Rietveld, P. (1992), "Methods for planning urban facilities in rural areas of developing countries: a comparison
of approaches", International Regional Science Review 15:345-356.
Rimmer, P.J. (1987), "The World Bank's urban transport policy: authorized version, revised version",
Environment and Planning A 19:1569-1577.
Riveros, L.A. (1990), "Recession, adjustment and the performance of urban labor markets in Latin America",
Canadian Journal of Development Studies 11:33-59.
Riveros, L. and L. Bouton (1994), "Common elements of efficiency wage theories: what relevance for
developing countries?", Journal of Development Studies 30:696-716.
Rodgers, G.B. (1975), "Nutritionally based wage determination in the low-income labour market", Oxford
Economic Papers 27:61-81.
Rodgers, G. (ed.) (1989), Urban Poverty and the Labour Market: Access to Jobs and Incomes in Asian and
Latin American Cities (International Labour Office, Geneva).
Rodgers, W. (1991), "The significance of access to land as a determinant of Kenya's interregional migration",
World Development 19:921-926.
Rogers, A. (1982), "Sources of urban population growth and urbanization, 1950-2000: a demographic
accounting", Economic Development and Cultural Change 30:483-506.
Rogers, A. (1990), "Requiem for the net migrant", Geographical Analysis 22:283-300.
Rogers, A. and J.G. Williamson (1982), "Migration, urbanization, and third world development: an overview",
Economic Development and Cultural Change 30:463-482.
Rogers, A. (1984), Migration, Urbanization, and Spatial Population Dynamics (Westview, Boulder).
Rondinelli, D. (1983), Secondary Cities in Developing Countries (Sage, Beverly Hills).
1786 C.M. Becker and A.R. Morrison
Root. B.D. and G.E DeJong (1991), "Family migration in a developing country", Population Studies 45:221-
233.
Rosen, K.T. and M. Resnick (1980), The size distribution of cities: an examination of the Pareto law and
primacy", Journal of Urban Economics 8. 165-186.
Rosenzweig, M.R. and O. Stark (1989), "Consumption smoothing, migration and marriage: evidence from
rural India", Journal of Political Economy 97:905-926.
Ruane. F. (1981), "On modelling the influence of sectoral policies on the spatial concentration of indus-
trial activities", Urban and Regional Economics Division, Development Economics Department Urban and
Regional Report No. 81-28 (World Bank, Washington, DC).
Rubenstein, A. (1982). "Perfect equilibrium in a bargaining model", Econometrica 50:97-110.
Sabot, R.H. (1979), Economic Development and Urban Migration: Tanzania, 1900-1971 (Clarendon Press,
Oxford).
Sabot, R.H., ed. (1982), Migration and Labor Market in Developing Countries (Westview Press, Boulder, CO).
Sabot, R.H. et al. (1981), "The Structure of wages in Kenya and Tanzania: what role has pay policy played?",
Population and Human Resources Division, Discussion Paper No. 81-55 (World Bank. Washington DC).
Sah, R.K. and J.E. Stiglitz (1992), Peasants versus City-Dwellers: Taxation and the Burden of Economic
Development (Oxford University Press/Clarendon Press, Oxford).
Safarova, G. (1996), "Reproduction of the population of St. Petersburg: medium-run perspectives", Monitoring
sotsialjno-ekonomicheskoi situatsii i sostoyaniya rynka truda S-Peterburga (Press-Analitik, St. Petersburg.
Russia) (in Russian).
Salehi Isfahani, D. (1993), "Population pressure, intensification of agriculture, and rural-urban migration",
Journal of Development Economics 40:371-384.
Sanyal, B. (1981), "Who get what, where, why and how: a critical look at the housing subsidies in Zambia",
Development and Change 12.
Sarris, A.H. and R. van den Brink (1993). Economic Policy and Household Welfare during Crisis and Adjust-
ment in Tanzania (New York University Press for the Comell University Food and Nutrition Policy Program,
New York).
Sarris, A.H. and P. Tinios (1995), "Consumption and poverty in Tanzania in 1976 and 1991". World
Development 23:1401-1419.
Sato, H. (1992), "The political economy of central budgetary transfers to states in India, 1972-1984".
Developing Economies 30:347-376.
Schmertmann, C.-P. (1992), "Estimation of historical migration rates from a single census: Interregional
migration in Brazil 1900-1980", Population-Studies 46:103-120.
Schultz, T.P. (1982), "Lifetime migration within educational strate in Venezuela: estimates of a logistic model".
Economic Development and Cultural Change 30:559-594.
Schultz, T.P. (1988), "Heterogeneous preferences and migration: self-selection. regional prices and programs,
and the behavior of migrants in Colombia", in T.P. Schultz. ed., Research in Population Economics, vol. 6
(JAI Press, Greenwich, CT).
Schwalbenberg, H.M. (1994), "Economic growth and the rise of protectionism and urban unrest in developing
economies", Open Economies Review 5:65-88.
Serageldin, I., M.A. Cohen and K.C. Sivaramakrishnan, eds. (1995), The Human Face of the Urban Envi-
ronment. Environmentally Sustainable Development Proceedings Series No. 6 (World Bank, Washington,
DC).
Sethuraman, S.U. (1981), The Urban Informal Sector in Developing Countries (International Labour Organi-
zation, Geneva).
Shabbir, T. (1991), "Earnings functions in Pakistan's urban informal sector: a case study (comment)", Pakistan
Development Review 30:704-706.
Shakhotko, L.P. (1996), The Population of the Republic of Belarus at the End of the 20th Century (Republic
of Belarus, Ministry of Statistics, Statistical Research Institute, Minsk) (in Russian).
Shapiro, C. and J. Stiglitz (1984), "Equilibrium unemployment as a worker discipline device", American
Economic Review 74:433-444.
Shaw, K. (1991), "The influence of human capital investment on migration and industry change", Journal of
Regional Science 31:397-416.
Ch. 43: Urbanization in Transforming Economies 1787
Sheppard, E. (1982), "City size distributions and spatial economic change", International Regional Science
Review 7:
Shidlo, G., ed. (1990), Housing Policy in Developing Countries (Routledge, London and New York).
Shields, N. (1980), "Women in the urban labor market in Africa: the case of Tanzania", Staff Working Paper
No. 380 (World Bank, Washington, DC).
Shixun, G. and X. Liu (1992), "Urban migration in Shanghai, 1950-1988: trends and characteristics",
Population and Development Review 18:533-548.
Shukla, V. (1996), Urbanization and Economic Growth (Oxford University Press, Delhi).
Shukla, V. and K. Parikh (1992), "The environmental consequences of urban growth", Urban Geography
13:422-449.
Shukla, V. and O. Stark (1984), "On agglomeration economies and optimal migration", Harvard Institute of
Economic Research Discussion Paper 1122 (Cambridge, MA).
Shukla, V. and O. Stark (1990), "Policy comparisons with agglomeration effects augmented dual economy
model", Journal of Urban Economics 27:1-15.
Singh, I., L. Squire and J. Strauss, eds., (1986), Agricultural Household Models (Johns Hopkins University
Press, Baltimore).
Singelmann, J. (1993), "Levels and trends of female internal migration in developing countries, 1960-1980",
in United Nations, ed., Internal Migration of Women in Developing Countries (UN, New York).
Sjaastad, L.A. (1962), "The costs and returns to human migration', Journal of Political Economy
70(Suppl.): 8 0-93.
Smith, J.P., L.A. Karoly and D. Thomas (1992), "Migration in retrospect: differences between men and wom-
en", Paper presented at the 1992 Population Association of America meetings, Denver, CO. Unpublished
manuscript.
Standing, G., ed. (1991), In search of Flexibility: The New Soviet Labour Market (Intemational Labour Office,
Geneva).
Stark, 0. (1991), The Migration of Labor (Basil Blackwell, Cambridge, MA and Oxford, UK).
Stark,O. and D. Bloom (1985), "The new economics of labor migration", American Economic Review 75:173-
178.
Stark, O. and E. Katz (1985), "A theory of remittances and migration", Harvard University Center for
Population Studies, Migration and Development Program Discussion Paper 16 (Cambridge, MA).
Stark, O. and R.E.B. Lucas (1988), "Migration, remittances, and the family", Economic Development and
Cultural Change 36:465-481.
Stark, O., M.R. Gupta and D. Levhari (1991), "Equilibrium urban unemployment in developing countries: is
migration the culprit?", Economics Letters 37:477-482.
Steel, W. and Y. Takagi (1983), "The intermediate sector, unemployment and the employment-output conflict:
a multi-sector model", Oxford Economic Papers 35:423-446.
Stigler, G. (1962), "Information in the labor market", Journal of Political Economy 70:94-105.
Stiglitz, J. (1969), "Rural-urban migration, surplus labour, and relationship between urban and rural wages",
East African Economic Review 1:1-27.
Stiglitz, J. (1974), "Alternative theories of wage determination and unemployment in LDCs: the labor turnover
model", Quarterly Journal of Economics 88:194-227.
Stiglitz, J. (1976), "The efficiency wage hypothesis, surplus labor, and the distribution of income in LDCs",
Oxford Economic Papers 28:185-207.
Stiglitz, J.E. (1982), "The structure of labor markets and shadow prices in LDCs", in R.H. Sabot, ed., Migration
and the Labor Market in Developing Countries (Westview, Boulder, CO).
Strassmann, W.P. (1991), "Housing market interventions and mobility: an international comparison", Urban
Studies 28:759-771.
Streeten, P. (1987), What Price Food? Agricultural Price Policies in Developing Countries (St. Martin's Press,
New York).
Stren, R.E. (1975), Urban Inequality and Housing Policy in Tanzania (University of California, Berkeley.
Institute of International Studies, Berkeley, CA).
Stren, R.E. (1978), Housing the Urban Poor in Africa (University of California, Berkeley. Institute of
International Studies, Berkeley, CA).
1788 C.M. Becker andA.R. Morrison
Stren, R. (1982). "Underdevelopment, urban squatting and the state bureaucracy: a case study of Tanzania".
Canadian Journal of African Studies 16:67-92.
Stren, R.E. (1989), "Administration of urban services", in R.E. Stren and R. White, eds., African Cities in
Crises: Managing Rapid Urban Growth (Westview Press, Boulder).
Stren, R.E. (1992), "African urban research since the late 1980s: response to poverty and urban growth", Urban
Studies 29:533-555.
Stren, R.E., ed. (1994a), Urban Research in the Developing World, vol. 2, Africa (University of Toronto Centre
for Urban and Community Studies. Toronto).
Stren, R.E., ed. (1994b). Urban Research in the Developing World, vol. 1, Asia (University of Toronto Centre
for Urban and Community Studies, Toronto).
Stren, R.E. and R.R. White, eds. (1989), African Cities in Crisis (Westview, Boulder).
Struyk, R.J. (1988), Assessing Housing Needs and Policy Alternatives in Developing Countries, Urban Institute
Report, 88-4 (Urban Institute Press, Washington, DC).
Struyk, R.J. and M.A. Turner (1987), "Simulating housing quality changes in developing countries: a tool for
policy analysis", World Development 15:1375-1387.
Struyk, R. and M. Turner (1991), "Econometric analysis: measuring the impact of rent controls in urban
housing markets", in A.G. Tipple and K.G. Willis. eds, Housing the Poor in the Developing World: Methods
of Analysis, Case Studies and Policy (Routledge, New York).
Summers, L. and L. Pritchett (1993), "The structural adjustment debate", American Economic Review 83:383-
389.
Svejnar, J. (1984), 'The determinants of industrial sector earnings in Senegal", Journal of Development
Economics 15:289-311.
Svejnar, J. (1989), "Models of modem-sector labor market institutions in developing countries", World
Development 17:1409-1415.
Tanzi, V. (1991), Public Finance in Developing Countries (Elgr., Aldershot, UK).
Tanzi, V. (1992), "Fiscal policy and economic reconstruction in Latin America", World Development 20:641-
657.
Telles, E. (1993), "Urban labor market segmentation and income in Brazil", Economic Development and
Cultural Change 41:231-249.
Terrell, K. and J. Svejnar (1989), The Industrial Labor Market and Economic Performance in Senegal
(Westview. Boulder, CO).
Tewari, V.K. and T.K. Kumar (1986), "Rent control in India: its economic effects and implementation in
Bangalore", Water Supply and Urban Development Department, Discussion Paper UDD 91 (World Bank,
Washington, DC).
Thobani, M. (1982). "A Nester logit model of travel mode to work and auto ownership", Urban Development
Departmnent. Discussion Paper UDD 5 (World Bank, Washington, DC).
Thomson, J.M. (1983), "Toward better urban transport planning in developing countries", Staff Working Paper
No. 600 (World Bank, Washington, DC).
Timmer, C.P. (1993), "Rural bias in the east and south-east Asian rice economy: Indonesia in comparative
perspective", Journal of Development Studies 29:149-176.
Tipple, A.G. and K.G. Willis, eds (1991), Housing the Poor in the Developing World: Methods of Analysis,
Case Studies and Policy (Routledge, London and New York).
Todaro, M.P. (1969), "A model of labor migration and urban unemployment in less developed countries".
American Economic Review 59:139-148.
Todaro, M.P. (1971), "Income expectations, rural-urban migration and employment in Africa", Intemational
Labour Review 104:387-414.
Tokman, V., ed. (1992), Beyond Regulation: The Informal Economy in Latin America (Lynne Rienner,
Boulder, CO).
Transportation Research Board (1997), "Transportation options for megacities of the developing world",
Working Paper (National Research Council, National Academy of Sciences, Washington, DC).
United Nations (1980), Patterns of Urban and Rural Population Growth (United Nations. New York).
United Nations (1982), Model Life Tables for Developing Countries (United Nations, New York).
Ch. 43: Urbanization in Transforming Economies 1789
United Nations (1988), World Population Trends and Policies: 1987 Monitoring Report (United Nations, New
York).
United Nations (1993), World Urbanization Prospects: The 1992 Revision (United Nations, New York).
United Nations (1994), Human Development Report 1994 (Oxford University Press, New York).
United Nations (1996), "Trends in urbanization and the components of urban growth", Paper presented at the
Symposium on Internal Migration and Urbanization in Developing Countries (UN Population Fund, New
York).
United Nations Development Programme (1996), Kazakstan Human Development Report 1996 (UNDP,
Almaty).
Valentine, T.R. (1993), "Drought, transfer entitlements, and income distribution: the Botswana experience",
World Development 21:109-126.
van der Gaag, J. and W. Viverberg (1988), "Wage determinants in Cte d'Ivoire", Living Standards
Measurement Study, Working Paper No. 33 (World Bank, Washington, DC).
van der Gaag, J., M. Stelener and W. Viverberg (1989), "Public-private sector wage comparisons and moon-
lighting in developing countries: evidence from Cote d'Ivoire and Peru", Living Standards Measurement
Study, Working Paper No. 52 (World Bank, Washington, DC).
van der Hoeven, R. (1984), "Zambia", in W. van Ginnekin and J.-G. Park, eds., Generating Internationally
Comparable Income Distribution Estimates (Intemational Labour Office, Geneva).
van Dijk, J., et al., eds. (1989), Migration and Labor Market Adjustment (Kluwer Academic, London).
Varshney, A. (1993), "Introduction: urban bias in perspective", Journal of Development Studies 29:3-22.
Velenchik, A.D. (1993), "Cash seeking behaviour and migration: a place-to-place migration function for C6te
d'Ivoire", Journal of African Economies 2:329-347.
Verma, S. (1986), "Urbanization and productivity in Indian states", in E. Mills and C. Becker, eds., Studies in
Indian Urban Development (Oxford University Press, New York).
Viverberg, W. and J. van der Gaag (1991), "The private wage sector in Cote d'Ivoire: homogeneity or
heterogeneity?" Southern Economic Journal 58:406-423.
Wang, Y. (1993), China: Urban Development Towards the Year 2000 (Hong Kong Institute of Asia-Pacific
Studies, Hong Kong).
Wang, Y. (1994), "China: urban development and research towards the year 2000", in R. Stren, ed., Urban
Research in the Developing World: Asia (University of Toronto Centre for Urban and Community Studies,
Toronto).
Ward, P.M. (1993), "The Latin American inner city: differences of degree or of kind?", Environment and
Planning A 25:1131-1160.
Weeks, J. (1995), "Income distribution and its implications for population trends in sub-Saharan Africa", in A.
Mafeje and S. Radwan, eds., Economic and Demographic Change in Africa (Clarendon Press, Oxford).
Weiss, A. (1980), "Job queues and layoffs in labor markets with flexible wages", Journal of Political Economy
88:526-538.
Wheaton, W.C. and H. Shishido (1981), "Urban concentration, agglomeration economies, and the level of
economic development", Economic Development and Cultural change 30:17-30.
Wheaton, W. (1981), "Housing policies and urban "markets" in developing countries: the Egyptian experi-
ence", Journal of Urban Economics 9:242-256.
Whittington, D. et al. (1993), "Household sanitation in Kumasi, Ghana: a description of current practices,
attitudes, and perceptions", World Development 21:733-748.
Widner, J.A. (1993), "The origins of agricultural policy in Ivory Coast 1960-1986", Journal of Development
Studies 29:25-59.
Williamson, J. (1988a), "Migration and urbanization", in H. Chenery and T.N. Srinivasan, eds., Handbook of
Development Economics, vol. I (North-Holland, Amsterdam).
Williamson, J. (1988b), "Migrant selectivity, urbanization, and industrial revolutions", Population and
Development Review 14:287-314.
Williamson, O. (1964), Economics of Discretionary Behavior: Managerial Objectives in a Theory of the Firm
(Prentic-Hall, Englewood Cliffs, NJ).
Willikens, E and S. Scherbov (1995), "Demographic trends in Russia", in H. van den Brekel and F Deven,
eds., Population and Family in the Low Countries 1994 (Kluwer Academic, Amsterdam).
1790 C.M. Becker and A.R. Morrison
Willis, K.G., S. Malpezzi and A.G. Tipple (1990), "An econometric and cultural analysis of rent control in
Kumasi, Ghana", Urban Studies 27:241-257.
World Bank (1984), World Development Report 1994 (World Bank, Washington, DC).
World Bank (1986), Urban Transport (World Bank, Washington, DC).
World Bank (1988a). "Report on adjustment lending", Document R88-119 (Washington, DC).
World Bank (1988b). Road Deterioration in Developing Countries (World Bank. Washington, DC).
World Bank (199()), "Report on adjustment lending II: policies for the recovery of growth". document R90-99
(Washington, DC).
World Bank (1991), Urban Policy and Economic Development: An Agenda for the 1990s, A World Bank
Policy Paper (Washington, DC).
World Bank (1992). World Development Report 1992: Development and the Environment (World Bank,
Washington, DC).
World Bank (1995), World Development Report 1995 (World Bank, Washington. DC).
World Bank Operations Evaluation Department (1993), Evaluation Results for 1991, World Bank Operations
Evaluation Study (World Bank, Washington, DC).
World Resources Institute (1996), World Resources 1996-1997 (Oxford University Press. New York).
Yamada, G. (1996), "Urban informal employment and self-employment in developing countries", Economic
Development and Cultural Change 44:289-314.
Yi, Z. and J.W. Vaupel (1989), "The impact of urbanization and delayed childbearing in China", Population
and Development Review 15:425-446.
Young, D. and D. Honghai (1997). "Urbanization, agriculture and industrialization in China: 1952-1991"'.
Unpublished manuscript (University of Alberta).
Yutopoulos, P.A. (1985), "Middle-income classes and food crises: the 'new' food-feed competition", Economic
Development and Cultural Change 33:463-483.
Zhang W., Y.Hongwen and C. Hongyan (1997), "Current changes of China's population", in China Population
Association, Symposium on Demography of China (China Population Association and IUSSP, Beijing).
Zhu, B. (1997), "The duration of stay and displacement of the floating population in the urban area: a case
study of Shanghai municipality", in China Population Association. Symposium on Demography of China
(China Population Association and IUSSP, Beijing).
Zhu, J. (1994), "Changing land policy and its impact on location growth: the experience of the Shenzhen
special economic zone, China, in the 1980s", Urban Studies 31:111611-111623.
Zhu, J. (1996), "Multilevel analysis of rural outmigration in Guangdong, China", Paper presented at the
Symposium on Internal Migration and Urbanization in Developing Countries (UN Population Fund, New
York).
Chapter44
STEPHEN MALPEZZI*
University of Wisconsin
Contents
1. Introduction 1793
1.1. Conceptual Framework 1794
2. Housing markets: The basics 1795
2.1. Property rights 1795
2.2. Housing demand 1798
2.2.1. Early literature 1798
2.2.2. Recent literature 1802
2.2.3. Cross-country demand models 1804
2.3. Housing supply 1805
2.3.1. Early cross-country research 1806
2.3.2. Market wide effects on supply 1808
2.3.3. Supply from the existing stock of housing 1809
2.3.4. Housing upgrading 1810
2.4. Tenure choice, tenure security, and mobility 1812
2.4.1. Forms of tenure 1812
2.4.2. Tenure choice 1813
2.4.3. Mobility 1814
2.5. Housing and the aggregate economy 1815
3. Related markets 1817
3.1. Housing finance 1817
3.1.1. Housing finance and financial development 1817
3.1.2. Housing finance and housing 1818
3.1.3. Institutional and systemic issues 1819
3.1.4. Mortgage design, and other technical details 1820
* The author is indebted to Edwin Mills, Bertrand Renaud and Christine Whitehead for comments on a
previous version, but bears sole responsibility for the chapter's shortcomings.
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P. Cheshire
© 1999 Elsevier Science B. V All rights reserved.
1791
1792 S. Malpezzi
Abstract
The purpose of this chapter is to survey recent research on housing markets and
policy in what used to be called the "second" and "third" worlds. We adopt
the labels "transition" economies to refer to countries as disparate as Russia
and Vietnam, and "developing" to refer to countries as disparate as Korea and
Singapore (arguably now developed) and countries like Mozambique and Laos.
It is therefore quite interesting that the bulk of the research surveyed finds that
housing market behavior is remarkably similar from place to place. Institutions
and constraints, particularly the amount of income available for housing and
other goods and services certainly do vary dramatically from place to place. And
the stakes of how well housing markets work vary from place to place. But these
differences in institutions and constraints do not obscure regularities in behavior.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1793
The first major section, on housing markets (Section 2), examines property rights,
supply, demand and tenure. Section 3 presents research on the related markets for
land, finance and infrastructure. Housing policy is covered in Section 4, includ-
ing housing subsidy systems, privatization, taxation and regulation. Section 5
concludes with a discussion of current issues and research.
JEL codes: R14, R21, R31, R38, R51, R52, 012, 018
1. Introduction
The purpose of this chapter is to survey recent research on housing markets and
housing policy in what used to be called the "second" and "third" worlds. We
will adopt the labels "transition" economies to refer to countries as disparate as
Russia and Vietnam, and "developing" to refer to countries as disparate as Korea
and Singapore (arguably now developed) and countries like Mozambique and
Laos, where the adjective "developing" is currently more a wish than a state-
ment of fact. We generally omit research devoted to "first world", "developed"
or "OECD" countries, except when such reference is fundamental or unavoid-
able.l Of course much literature relevant to housing in the "second" and "third"
worlds can be found in the technical literature on housing which has no particular
reference to place or time. 2
Some would argue-or argued a decade ago-that countries outside the so-
called developed or OECD world are so different as to obviate comparisons,
or the transferability of research results and methods, across countries. In fact,
the bulk of the research below documents that housing market behavior is re-
markably similar from place to place. Institutions and constraints, particularly the
1 See, among others, Whitehead (Chapter 40, this volume), Shephard (Chapter 41, this volume) and Olsen
(1987).
2 See, for example, Arnott (1987), Olsen (1987) and Smith et al. (1988) for reviews of the relevant technical
literature. Also omitted from this chapter is a review of the stylized facts regarding housing markets across
countries. See Malpezzi (1990) and Angel and Mayo ((1996) for such a review. Other recent reviews which
may be profitably consulted include Hoffman et al. (1991), Gilbert (1992), Rakodi (1992), World Bank (1993)
and UN Centre for Human Settlements (1996).
1794 S. Malpezzi
LandLnd_ R
~R_ Developers R_
Finance Ilers I Renters
Infrastructure dl C - Homeowners
Labor E n rs E (Income &
Materials S S Population)
amount of income available for housing and other goods and services certainly do
vary dramatically from place to place. And at a different level it could be argued
that the stakes of how well housing markets work vary from place to place. But
these differences in institutions and constraints should not obscure regularities in
behavior.
1.1. Conceptualframework
Figure 1 (from Mayo et al., 1986; Malpezzi, 1990) shows a schematic diagram of
how the housing market works. Inputs such as land, labor, finance, materials and
infrastructure are combined by supply-side agents such as landlords and develop-
ers to produce housing services. Homeowners, and to a lesser extent, renters, are
also producers, as they maintain and upgrade their houses. Relative prices inform
producers of housing services about whether to provide more or less housing, and
the input suppliers about providing more or fewer inputs.
Several other important features of housing markets are implicit in Fig. 1.
First, transactions within and across "boxes" are possible only to the extent prop-
erty rights are defined, recognized and enforced. Second, government interven-
tions can have profound effects on the operation of the housing market. Third,
fully understanding housing markets requires analysis of key input markets and
the regulatory environment, as well as revealed market behavior in the housing
market per se.
Every introductory real estate textbook emphasizes a number of other salient
features about housing. It is a large share of every country's wealth and produc-
tive capital stock. Because its stock price is large relative to incomes, it must
be financed. Some households own their own housing capital, others lease it.
Housing is fixed in location, extremely durable (slowly depreciating), and can
be viewed alternatively as a composite commodity yielding a flow of "housing
services", or as a set of individual characteristics.
Ch. 44: Economic Analysis ofHousing Markets in Developing and TransitionEconomies 1795
Each of these interesting features will be discussed in one or more sections be-
low. The section on the basics of the housing market will focus on the second and
third "boxes" in our simple conceptual framework; the section on input markets
on the first box; and the section on housing policy on the effects of government
interventions within and across boxes.
This section corresponds roughly to the first two boxes in Fig. 1. Alternatively
one could think of this section as examining the usual supply and demand dia-
gram from any principles textbook, or, more properly, multivariate supply and
demand relations. Housing market analysis is complicated by several facts. First,
the system of property rights implicit in any supply and demand diagram cannot
be taken for granted. Second, housing is a composite commodity and may be
examined in terms of its service flows or stock in some aggregate way, or in
terms of individual characteristics. Third, supply in any given period comes from
both new construction and modification of the existing stock. Fourth, transac-
tions costs of changing consumption are quite high. Fifth, housing consumption
choices are bound up with tenure choices. Sixth, housing markets are a large
part of the economy in general and the capital stock in particular, so have strong
linkages with its aggregate economy. Each of these points will be discussed at
greater or lesser length below.
Property rights are sine qua non of housing market development. Until recently,
property rights have been much neglected in the "developed" country housing
literature, but were somewhat better represented in the "traditional" developing
country housing literature. 3 Post-Perestroyka (circa 1989) the topic moved prop-
erly to the fore. Some property rights issues are common among countries, and
some issues are specific (more or less) to Africa, eastern Europe, or to some other
specific region or country.
Property rights may be defined and assigned through a formal legal system,
or by custom or tradition.4 Henceforth we use "law" to refer to both. Two ar-
3 This neglect refers to the housing literature. Economists, political scientists, and of course lawyers
have long studied property rights in other contexts. See, for example, Browder (1984), Friedman (1975) and
Moynihan (1987).
4 This section draws from Vandell's contribution to Malpezzi and Vandell (1992). See Jaffe and Louziotis
(1996) for another recent review.
1796 S. Mvfalpezzi
eas of law which particularly affect the operation of housing and real estate
markets are contract law and land use regulation. Contract law deals with the
system that defines and facilitates the transfer of property and property rights,
allocates those rights, and settles disputes. In formal systems these functions
are associated with such instruments as contracts of sale, leases, easements and
rights-of-way, operating agreements, mortgages and deeds of trust, etc. In all
countries, rich and poor, some of these functions are also affected by less for-
mal "mores and folkways of society". In many countries, including most of the
transition countries and many African countries, these systems are in flux. Land
use regulation includes the body of custom, law, regulation and case law which
governs the rights to locate certain uses in certain locations and provides stan-
dards of development and operation of those uses. Formal instruments include
zoning ordinances, building and housing codes, subdivision regulations, private
deed restrictions, environmental laws and regulations, etc.
Together, these two areas of the law render operational the notion of owner-
ship, exercise and transfer of rights in real property. A wide range of descrip-
tive studies have examined property rights in Africa (Ault and Rutman, 1979;
Kiamba, 1989); in Asia (Bromley, 1989); in Latin America (Betancur, 1987;
Gilbert, 1989); and of course formerly socialist countries (Jaffe, 1993; Pejovich,
1990). From that descriptive literature, and from analytic literature such as
Alchian and Demsetz (1973), Coase (1960), Demsetz (1967) and Williamson
(1975), a clear list of general principles has emerged. In order to maximize the
social value associated with rights in real estate, the set of laws and regulations
governing their associated property rights must possess certain characteristics.
They must be transparent and agreed on by some not-as-yet-well- defined social
consensus. They must be enforceable at a reasonable cost, with little or no uncer-
tainty. There must be some general agreement on final arbiter of disputes (most
often but not always the state).
Tenant and landlord lights must be well defined, whether with a formal or in-
formal contract. There must be clear remedies for violation by either party. These
rights and obligations will generally be freely negotiated between the parties and
represent the outcome of a competitive market process. Versacaj (1993) presents
a discussion of leases in Russia and shows how the violations of such principles
inhibits the market.
In a well-functioning system property rights will be transferrable from seller
to purchaser on payment of a consideration (Jaffe and Louziotis, 1996). The
bundle of such rights can be largely complete (fee simple, although still limited
by land use regulation), or partial, including leasehold. Specific rights include
the right to use or modify the use of the real estate, the right to derive income or
Ch. 44: Economic Analysis of Housing Markets in Developing and TransitionEconomies 1797
other benefits from its use, the right to bequeath the ownership interest, the right
not to be evicted, etc. Implicit in the contract of sale is not only an obligation
to the seller, typically to pay a specific amount for the real estate interest but
sometimes a more complex obligation such as to limit future uses or to bequeath
the property in a certain way. An explicit or implicit obligation will also exist
to maintain the property at a certain level, to maintain a use which is consistent
with existing land use regulations, to pay taxes assessed on the basis of property
ownership, to give up the property for a preemptive public purpose on payment
of compensation, and not disturb or do other damage to neighbors as a result of
use. Government constraints may be required to prevent certain adverse market
failures or imperfections (such as conflicting land uses or limitations on access
or economic productivity) or in cases in which the buyer or seller has excessive
market power.
Maximum social return to the housing stock value, generally requires liquid-
ity, or the ease of transfer of real estate interests. Landis (1986) shows that high
fees or other rights of entry to the market, restrictions on appropriate purchasers
or tenants, unreasonable constraints on use, excessively costly development stan-
dards, etc. can be counterproductive. Markets are rendered most efficient to the
extent that they are "thick", i.e., there are many transfers and prices/rents are
well established (Bikhchandri, 1986). Because the high purchase prices of real
estate requires finance for most transactions, maximum benefit is obtained when
financing is freely available at market rates. Financial innovations and reforms,
discussed below, which enhance the liquidity of the mortgage market, such as
the development of the secondary market or securitization, also enhances the
liquidity of the real estate market in general. The possibility of foreclosure in the
ownership market for the nonpayment of debt or other violation of the obligations
of ownership (such as the nonpayment of real estate taxes) is essential for effi-
cient operation (Buckley, 1990) just as is the possibility of eviction in the rental
market (Jimenez, 1984).
Property rights profoundly affect not only the efficiency of the housing mar-
ket; they also profoundly affect other social goals or questions such as the dis-
tribution of wealth or income. Concentration of ownership rights may adversely
affect on the concentration of wealth and power, as well as reduce the efficiency
of the market, especially if quasimonopoly exploitative situations are produced. 5
Solomon and Vandell (1982) discuss alternative conditions under which real
estate markets may operate which cause deviation from the competitive ideal.
5 Malpezzi (1994) argues that with a few important exceptions, mainly in taxation, housing policies which
are more efficient are also more equitable. The argument can certainly be made with respect to property rights,
e.g., the large literature linking land reforms to faster growth.
1798 S. Malpezzi
Many of the biggest property rights and regulatory issues are regarding land.
Systems can function with long-term leaseholds as well as with fee simple (Brom-
ley, 1989), but efficient markets require long and enforceable leases. In some
countries there is no history or memory of private ownership of or significant
use rights over land. Paradoxically, in such countries public landlords (central
government, large state enterprises or cities) may have true market power, and do
not respond well to external incentives. Moving to a market system has powerful
redistributive consequences which will be politically charged (Hegedus et al.,
1992; Buckley et al., 1995). More research is certainly required on the essen-
tial property rights necessary sufficient for efficient use of land. Most studies of
land reforms have focused on agricultural and/or rural land (e.g., Dorner, 1982).
Certainly we should study the effects of past major urban land reforms just as
carefully.
Table I
Sunnary of selected international housing demand studies
Follain et al. (1980) Urban Korea, 1976 Cross-section log Renter E, = 0.12 Other models reported
expenditure model. for current income. 0.42 pool tenure, stratify by
single equation OLS, for consumption: owner location.
stratifed by tenure. Ey = 0.21, 0.62. Ep is
near zero for all models.
Strassman (1980) Cartrcgena, Coloimbia. Cross-section log Renter Ey = 0.78, Other models stratify by
1978 expenditure model. owner Ev = 1.19. income; Ey generally
single equation OLS, Current income only, increases with income
stratifed by tenure. no Ep estimates. but not strictly so.
Ingram (1984) Bogota and Caii, Cross-section log Renter Ey = 0.47 One of few studies
Colombia. 1978 expenditure model. (Cali), 0.72 (Bogota); explicitly modeling
single equation OLS, Ep = -0.48 (Cali), intermetropolitan
stratifed by tenure. -0.28 (Bogota). Owner location.
Includes distance to E, = 0.76 (Cali), 0.78
workplace. (Bogota); Ep = -0.44
(Bogota).
imeneozand Keare Santa Ana and Log expenditure lodel. Renter Ey = 0.27 Permanent income
(1984) Sonsonate, El single equation OLS, (Santa Ana). 0.42 proxied by weighted
Salvador, 1980 two cross-section panel (Sonsoiuate); owner average of two years
data, stratifed by tenure. Ey = 1.05 (Santa income.
Ana).
Malpezzi and Mayo Data from Colombia. Log expenditure model. Renter and owner Price elasticities
(1985,1987) Egypt, El Salvador. single equation OLS. income elasticities, estimated for Cairo.
Ghana, hidia. Jamaica, stratified by tenure. respectively: Bogota. Beni Suef and Manila.
Korea and the 0.66, 0.75; Cali, 0.44. following Muth (19711;
Philippines, various 69: Cairo. 0.46, 0.17: most estimnsates close to
dates in the 1970s Beni Suef, 0.51. 0.42: 1, but procedure is
and the early 1980s Santa Ana, 0.48. 1.11: biased towards - 1.
Sonsonate, 0.50. 0.79:
Kumasi, 0.33, owners
NA: Bangalore, 0.58,
0.43; Kingston, 0.16,
owners NA: Seoul,
0.45, 0.44: Piusan.0.31,
0.45; Taegu, 0.44.0.47:
Kwangju, 0.62, 0.41:
other Korean cities. 0.54,
0.79: Davao, 0.88. 0.99:
Manila, 0.56, 0.57.
Ndulo 1986) Lusaka (19791 Linear expenditure Owner E, = 0.6 Also segments by
function. (presumably at means, household size, fiids
see text). larger households have
larger income
elasticities.
Mehta ad Mehta (1987) Ahmedabad (1985) Logrithmicexpeniditure Renters, Ey is 0.17 to Also segments by
function, by tenure. 0.40, Ep = -0.8: income, generally finds
owners, Er = 0.2, low income households
Ep = -0.4 have lower elasticities.
Grootacrt and Dubois Cote d'hoire (1979) Estinmae tenure choice. For Abidjan, renter Selectivity correction
11988) then rental demand Ey = 0.37, Ep = -0.25. has virtually no effect
equation with For other cities, on demand results.
selectivity correction. renter Ey = 0.28, Some problem with
Use hedonic to Ep = -0.25. interpretation of P and
decompose rent iIto P Q measures. see ext.
and Q.
Shefer 119901 Indonesia (19781 Logrithmic expenditure For all renters, Elasticities generally rise
function, by tenure. Uses E. = 0.84: for all with income class, size
expenditure to proxy owners, E = .17. of city (with exceptions).
permanent income. Also
presents results by city
size and income class.
Ch. 44: Economic Analysis of HousingMarkets in Developing and TransitionEconomies 1801
Table 1
(continued)
Lodhi and Pasha (1991) Karachi (1987-88) Linear expenditure Planned, owners: current
functions, segmented by Ey = 0.58; permanent
tenure and by formnal- Ey = 1.20.
informal development. Planned renters: current
Estimates with current Ey = 0.42, permanent
income and consumption Ey = 1.10.
(permanent income). Katchi abadi owners:
current Ey = 0.06,
permanent Ey = 0.12.
Katchi abadi renters:
current Ey = 0.09,
permanent Ey = 0.31.
Assadian and Ondrich Bogota and Cali (1978) Simultaneous model of Ey = 0.56 for head's
(1993) housing demand, work income if
location, and labor household has two
supply. earners; is 0.95 if there
is no other earner.
Lim and Lee (1993) Urban China, sixth plan Log model, province/ At provincial level, Limited comparability to
(1981-1985) municipality is the grouped Ey = 1.23 other studies because of
unit of observation. (wages), 1.32 (household aggregated data.
expenditures). At city
level, grouped Ey =
1.04 (wages), 1.30
(expendtures).
Arimah (1994) Ibadan (1987-1988) Log model, by tenure. Renter Ey = 0.88;
owner Ey = 0.56.
Chou and Shih (1995) Hong Kong (1991) Box-Cox model, by At meands, renter Tenure choice has no
tenure. Ey = 0.27 (transi- effect on renter
tory income), 0.37 demand.
(permanent income).
Notes: Most studies present a range of elasticity estimates from alternative specifications; therefore "Key
Results" entry in this table contains our judgement of representative point estimates.
estimated separately for owners and renters. Estimates of the price elasticity are
close to 1 in absolute value, ranging from -0.76 to -1.08, with the exception of
Manila owners whose price elasticity is estimated to be -0.4.
These price elasticity estimates suggest that demand is considerably more
elastic than previous estimates in the literature suggest. However, a shortcoming
of this model is that a unitary income elasticity is the null hypothesis, because a
land price coefficient of zero implies a price elasticity of 1. Note that if land prices
are measured with error, the price elasticity is biased towards -1. Therefore, the
tests of significance of land price coefficients should not be interpreted as tests of
zero price elasticity. Neither are they correct tests of unitary elasticity, because
the land's share estimate, assumed fixed for the sample, actually has a distribution
as well. Testing the micromodel price elasticities under alternative specifications
remains high on any agenda for future research.
1802 S. Malpezzi
7 Grootaert and Dubois (1988) actually estimate a reverse regression with price as dependent variable,
although they do not note that in their paper.
1804 S. Malpezzi
Daniel (1985) finds quite high income elasticities of around 1, for Hungary.
She also finds that on balance the addition of implicit housing subsidies to wage
income as an in kind transfer leads to an increase in the distribution of income.
Other studies of transition economies have examined demand indirectly,
through queues (Ciechoncinska, 1990; Andrusz, 1990; Charemza and Quandt,
1990). McMillen and Pogodzinski (1993) point out the often cited tendency for
queue to overstate effective demand because of low costs of joining the queues
may or may not be correct. They point out that there are requirements for joining
queues (e.g., based on family status), which vary with the country involved,
but which generally raise costs or present a different cost schedule to different
households.
Mayo and Stein (1995) examine these issues with data from Poland in the
1980s. Using aggregate data Mayo and Stein show that housing investment is
systematically lower in Poland and several other then-socialist economies, com-
pared to other countries with similar levels of development. They also show that
housing output prices are substantially higher than marginal costs in Poland.
Taken together these two facts confirm the existence of a shortage in Poland
during the 1980s.
Using regional (Voivodship) data, Mayo and Stein (1995) estimate several
models that show that housing queues are associated with compensating wage
differentials and labor shortages, as well as depressed migration to regions.
Pogodzinski (1993) undertakes a similar analysis and generally has similar find-
ings; higher housing queues are associated with larger regional labor shortages
in Poland. These queues are analyzed in other countries for example Renaud
(1995a, b) examines waiting lists in Russia. Charemza and Quandt (1990) also
analyze such queues.
In this section, we will examine aggregate descriptive supply (e.g., Burns and
Grebler, 1976, 1977; and related work by Renaud, 1980; Chang and Linneman,
1990; Buckley and Madhusudan, 1984). We also examine certain behavioral ev-
idence on aggregate supply elasticities, and their determinants using data from
Korea, Thailand and Malaysia. Finally, we examine supply from the existing
stock and filtering (Ferchiou, 1982).
1806 S. Malpez i
their estimates imply that as economies develop, increasing incomes fuel housing
demand; but this is in large part offset by increasing prices.
a Point estimate of price elasticity was negative, market deemed inelastic (see text).
b Point estimate of yl was negative, market deemed elastic (see text).
US estimates constructed using results from Malpezzi and Maclennan (1994), others from Table 1.
For discussion of bounds on income and price elasticities of demand see Malpezzi and Mayo (1987).
and Maclennan (1996) demonstrated that results from such models can be sensi-
tive to time period chosen. Other fruitful extensions would be to develop a stock
adjustment model and to disaggregate such analysis by region or metropolitan
area.
studies related to filtering concept (1). A number of studies have been done on the
upgrading process, which is clearly central to (3), although incomplete. Studies
of the relative price of housing at different quality levels (2) are are rare in general
and almost unkown in developing/transition economies.
Ferchiou (1982) examines data from Tunisia and Mexico, looking at chains of
moves, following Lansing et al. (1969) methodology. Generally Ferchiou finds
these chains are of limited duration; he finds that a little over two families take
part in a chain of moves before it terminates. Ferchiou notes that this is probably
a lower bound estimate, partly because of the limited size of the study area. He
also notes that construction of more expensive units initiate longer chains.
Hegedus and Tosics (1991) undertake a similar study for Hungary. Gener-
ally, they find short chains, usually less than two moves on average. They find
little systematic difference by location or by state/private construction. A conjec-
ture (by this author) is that the shortness of these chains may unsurprising give
the shifts in employment and population and repressed demand for housing in
formerly socialist Hungary.
Several studies examine filtering using some version of the so-called "stock-
user matrix" (Strassman, 1977). These basically cross-tabulate households by
income category and dwellings by value or rent. A number of these studies are
surveyed by Johnson (1985). Baer and Koo (1994) examine Korea.
later. Daniere (1992) and Malpezzi (1986) also find a nonmonatonic relationship
between tenure and income.
Other upgrading studies of note include Rakodi (1987).
This section is closely related to the discussions of property rights and hous-
ing demand from above. We will examine forms of housing tenure, the value
of tenure (Jimenez, 1982; Hoy and Jimenez, 1991), and tenure choice (Lim
et al., 1980; Tipple and Willis, 1992; Strassman, 1982; Zorn, 1988). Having
examined one method of changing consumption, upgrading, we examine the
other-moving (Strassman, 1991; Zorn, 1988).
Cairo, but not in Manila. In the latter city, rather, the probability of ownership
hardly budges with income; but the probability of renting falls significantly and
that of squatting rises. Interestingly, squatters have more in common with owners
than with renters.
Grootaert and Dubois (1988) estimate a probit model of tenure choice in Cote
d'Ivoire and find that income and lifecycle variables do the bulk of explaining
home ownership.
Several studies of Philippine data show the value of tenure. For example,
Friedman et al. (1988) estimate hedonic functions for formal and squatter hous-
ing. They find that on average a rented squatter unit in Manila would rent for
15% more were it formal; the corresponding owner premium is 25%.
A number of papers have presented evidence that in some cities, large frac-
tions of low income households own in the informal sector; as incomes rise they
rent in the formal sector; and the richest again become homeowners. Yet, such
patterns have not been scrutinized or explained carefully. Strassman (1980b, c)
suggests that availability of services such as piped water may catalyze invest-
ment by some households and make the shift to renting such units attractive
relative to current owners of informal units without such amenities. In a very
stylized version of such a world, we would observe the lowest income house-
holds owning very low quality housing, perhaps in the informal sector or with
little tenure security; past some threshold, households would begin into a higher
quality rental submarket; finally, at higher incomes and (perhaps) overcoming
financial constraints, households would be able to purchase such housing.
One variable conspicuous by its absence in most of these studies is the relative
price of each tenure form. Constructing such a variable is possible but requires
some effort (M&M, 1987b). To a household, the user cost of a rental unit is the
periodic rent paid, plus any deposits or key money payments appropriately dis-
counted, plus their own payments to others for housing services (e.g., household
maintenance expenditures). User cost for owners is even more complicated, since
it must account for financing, depreciation and inflation. User cost is another
fruitful area for future research.
2.4.3. Mobility
The range of variation in mobility is enormous. Strassman's (1991) survey finds
that in a given year less than 3% of households moved in a year in the German
Democratic Republic in 1980; in Colombo, Sri Lanka only 5% moved. At the
other extreme, in Bangkok, Thailand about 20% moved in a year, and in Seoul,
Korea an astounding 43%. Strassman (1991) argues that government interven-
tions in housing markets on balance tend to lower residential mobility. He uses
Ch. 44: Economic Analysis of HousingMarkets in Developing and TransitionEconomies 1815
data on rents and prices as a proxy for distortion in the housing market; the higher
the house price to income ratio, the lower the mobility rate. The higher the rent
to income ratio, the higher the mobility rate.
Strassman's (1991) paper sets an interesting agenda in this area. Among other
issues, even in such a simple model it's hard to imagine mobility as endogenous
while price is exogenous. And many regulations, such as rent control, that could
affect mobility can be measured directly.
In the developed country literature, in addition to income and lifecycle effects,
disequilibrium in housing consumption has been emphasized as an explanatory
variable for household mobility (e.g., Goodman, 1976; Weinberg et al., 1981).
Malpezzi (1986) constructs a measure of housing disequilibrium and uses it as
an explanatory variable in a model of planned moves and upgrades using data
from Cairo. In the event, neither housing consumption variables nor lifecycle and
income variables contribute much to explaining planned moves with that dataset.
Zorn (1988) estimates a joint model of tenure choice and mobility, focusing
on transactions costs as well as income and lifecycle. As expected, income is
positively associated with moving and owning, but negatively with moving and
renting. Transactions costs generally reduce the probability of moving, whatever
the tenure.
Housing has strong links to general development, both forward and backward
(Malpezzi, 1990; Buckley and Mayo, 1989; Renaud, 1990). In this section, we
explore the relationship between housing investment and development; hous-
ing investment and the business cycle (including employment); housing pol-
icy and structural adjustment; and alternative views of housing as a productive
investment.
In policy discussions macroeconomists will sometimes point to housing's
allegedly unfavorable incremental capital output ratio (ICOR) as evidence that re-
ducing housing investment will increase growth of the economy. While the ICOR
can be derived from a simple Harod Domar growth model, such models abstract
from the different useful lives of capital and are of little use in distinguishing
between investment in one form of capital over another. At the same time, one
can find a housing related literature that strains to find multiplier and externality
arguments for investment in housing (see Katsura, 1984, for a survey). Malpezzi
(1990) develops the argument that housing investment decisions are generally
best made on the basis of internal rate of return/present value criteria, as are
1816 S. Malpezi
investment decisions generally. Generally high and rising prices for housing can
be viewed as signals that the market requires additional investment.
A number of papers such as Strassman (1985) have examined whether hous-
ings employment multiplier is "favorable", or not. The most careful studies using
input/output analysis suggest that after indirect effects have taken into account,
housing's multiplier is more or less like other multipliers. Research on the exis-
tence of externalities such as Bums and Grebler (1977) or more recent evidence
for the US such as Green and White (1995), suggest that while health and ed-
ucation externalities may exist, in many developing and transition economies
these are probably dwarfed by the innate economic return to additional housing
investment.
This is not to imply that there are no general equilibrium or aggregate ef-
fects of housing in these economies. Buckley and Mayo (1989) present some
examples. They examine two case studies: Argentine housing policy, with special
emphasis on financial linkages, and Polish housing policy, with special emphasis
on interactions with the labor market. As of the mid-to-late 1980s, they find the
present value of welfare costs of Argentine housing subsidies through the finan-
cial system are on the order of 6% of GDP in present value terms. Buckley and
Mayo find that Poland's insufficient housing investment and ill-located housing
are equivalent to a compensating wage differential or tax of about 10% of labor
income in the late 1980s. (Both the Argentine and Polish cases are illustrative;
both countries have followed quite different policies in recent years.)
Malpezzi (1990) points out that many people working in the shelter sector
are not used to thinking of its investments as productive. This is true of many
developing country housing analysts and was certainly true under socialist cen-
tral planning, where housing was explicitly labeled "nonproductive" and was not
even counted in Net Material Product (the socialist analogue of Gross National
Product). But of course shelter and infrastructure investments are, in fact, by
definition productive: they are investment in an asset that yields a flow of services
over time. To label such investment as "consumption", is quite common, but
incorrect. The same criteria which governs choice of other investments governs
housing. Arguments about externalities, indirect contributions to labor produc-
tivity, and employment multipliers obscure this central point. Malpezzi (1990)
discusses these issues in some detail, including the role housing market reform
can play in structural adjustment.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1817
3. Related markets
This section is primarily about the first box in Fig. 1, above. Why examine the
input markets for housing, when this is rarely done explicitly for many other
goals? Because of its high cost in relation to incomes, housing must be financed.
Because of its locational fixity, housing markets are profoundly affected by the
operation of land markets, and by infrastructure.
3.1. Housingfinance
Bertrand Renaud (1984) put it best: "Cities are built the way they are financed".
In this section, we will examine housing finance, not only as a key input to
housing development (Struyk and Turner, 1986), but also housing finance as a
key element in financial development more generally. We will investigate certain
systemic and institutional issues (Buckley, 1994; Renaud, 1990), as well as more
technical details, such as instrument design (Buckley et al., 1993; Chiquier and
Renaud, 1992; Sandilands, 1980). We will examine the effects of trying to run
housing finance as a subsidy system rather than as a system for true intermedia-
tion (Renaud, 1990, 1993; Buckley, 1991). We will also discuss the key role the
evolution of housing finance plays in the reform of formerly socialist economies
(Buckley et al., 1993; Renaud, 1991).
Housing is the largest asset owned by most households. Housing is always
financed, in the sense that virtually all owners of housing capital must pay for
their units over several periods. Even households which own their units "free and
clear" finance the unit in the sense that holding such a large asset has a financial
opportunity cost.
But in most countries only a small share of this potential finance, roughly
equal to the value of the underlying assets, is in the form of mortgages or other
formal sector finance. Goldsmith (1985) shows that in both developing and de-
veloped countries formal sector finance is only a small part of the total.
GNP per capita under $1000. For middle income countries ($1000-10,000 per
capita), the average is about 16%. For countries above $10,000, the average is
about 25%. Of course there is great variation within groups, but the overall trend
is quite clear.
Buckley (1991) illustrates the spillovers that an inappropriate housing finance
system can have with the case of Argentina during the 1980s.10 Almost all hous-
ing finance came from two sources: a housing fund based on wage taxes (FON-
AVI) and the National Mortgage Bank (BHN). Roughly, FONAVI targeted lower
income households, and BHN middle income households. Buckley showed that
in the system in place in the 1980s, given that payments are only partially in-
dexed, long grace periods, and poor foreclosure and other recovery practices,
only 2-5% of every dollar FONAVI invested was ever recovered. During the
period Buckley studied BHN was also largely dependent on central government
transfers and forced deposits from local authorities at negative rates. Together
FONAVI and BHN transferred about 2% of GNP into a narrow segment of the
housing market.
Buckley (1991) used a simple consumer surplus model to demonstrate the
ineffiency of such transfers. The present value of the net cost of these distortions
approached 6% of GNP, and the large expenditures may have accounted for up
to one-third of Argentina's inflation during the period.
Many similar examples are documented in the literature. Sandilands (1980)
and Silveira (1989) examine Brazil, another country with experience of a large
housing fund.
Early studies of housing finance in developing countries emphasized the role
of the deposit taking institutions in housing finance. Implicitly the model for
many was the US savings and loan system and to some extent Britain's building
society system. However, in developing countries and transition economies as
well as in the developed world, the trend has been to move away from deposit
taking institutions, and to break the direct link between small scale saving and
mortgage lending. Secondary institutions and bond finance are becoming more
important as proximate sources of funds (see Renaud, 1995).
Of course the starting point for understanding land markets is to refer back to our
previous discussion of property rights, and tenure issues. See also Alan Evans'
contribution in Chapter 42 of this volume.
Here we examine several other land related issues, especially land pricing,
effects of infrastructure, and location. We also discuss development vs. redevel-
opment, and land development and environmental issues (see Joseph Gyourko's
contribution to this volume). We also examine regulatory issues (Bertaud, 1992;
Dowall, 1989b; M&M, 1997a).
12 More detailed reviews of such issues can be found in Renaud (1989) and Buckley (1996).
1822 S. Malpezzi
Hamer et al. (1993) examines urban land markets in China. Many of the
issues in China are conceptually similar, although they are played out against
the backdrop of one of the most rapidly growing urban populations in the world.
(In contrast to Russia's stable urban population.) The authors point out the short-
comings of various land market "reforms" during the 1980s, which focused on
control and limiting the size of cities, and to a lesser extent on government
revenue and the land needs of state enterprises. Overall allocative efficiency
was not highlighted in these reforms. Reform has been pushed further by the
growing obviousness of problems arising from poor allocation of land and fiscal
problems of both local governments and enterprises. Hamer et al. point out that
the post-1979 reintroduction of foreign investment brought in a group of actors
accustomed to well defined property rights and market or market-like mecha-
nisms. After 1984, a number of Chinese cities embarked on a range of land use
management experiments, including the well known Special Economic Zones
in several southern and coastal cities. Less well known are changes such as the
creation of profit oriented real estate development companies (mainly publicly
owned) which develop commercial real estate and sell at prices including land
location premia: and introduction of the range of development fees, taxes and
lease premia for specific tenure rights. China is also introducing land and building
cadastres which define the location, use and rights of specific parcels. Dowall
(1993) provides a convenient summary of some of the recent policy changes and
requirements for the new two land markets.
Perhaps the most pathological case of land market regulation in any large
country was South Africa's Apartheid system. Turok (1994a, b) presents a con-
cise description. Brueckner (1996) analyzes the welfare gains from dismantling
Apartheid in the context of the standard urban model.
15 With a number of associates, Dowall (1991c) performed similar analyses in some 30 markets. Dowall
(1997) provides a review, see also Dowall and Clarke (1991).
1826 S. Malpezzi
Table 3
Land price gradients
Asabere (
1981) Kumasi, Ghana Individuals, infrastruc- -0.037 Also found nonuinal PL
1970-1979 ture whether sold to rose 1.1% per month,
Ashante/non-Ashante, 1970- 1979 (of CPI).
type of neighborhood.
size of plot.
Sol and Kim (1998) Korea six Land price gradient -0.027 (Seoull to Finds green belt
cities greenbelt dummies. -0.290 (Kwang-ju) a strongly binding
constraint.
J. Eckert data reported Moscow. Russia Simple gradient Q 1992:-0017 Moscow is first ever case
in Bertaud and Renaud 1992 estimatedd by this Q 1992: -0.034 of steepening gradient
(1994) author from graphs). as market emerges.
(Need to find figures)
J. Eckert data reported Krakow, Poland Simple gradient 1992: -0.10 (Need to find figures)
in Bertaud and Renaud 1992 (estimated by this
(1994) author from graphs).
DowalI (1989) Karachi, Pakistan Dummy for large plots -0.058 Large plots sell at large
1985 and developed plots. premium and developed
plots worth more than
twice undeveloped.
16 The numerical gradient estimates in Table 2 were constructed by this author from data presented front
Fig. 3 of Bertaud and Renaud's (1994, 1997). from Eckert.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1827
3.3. Infrastructure
See also the contributions by Eberts and McMillen (Chapter 38), Small and
G6mez-Ibfiez (Chapter 46) in this volume, and World Bank (1994). Here, we
will focus on transport, water and sanitation, power and housing.
The provision of infrastructure and related services-transport, water, sani-
tation and so forth-is a traditional public sector activity, and one of particular
importance to low income households. Directly, households benefit from several
types of infrastructure through saving time and money (for example, publicly
supplied water rates versus user charges) and through improved living conditions.
Often, infrastructure investments encourage new construction and upgrading of
existing housing, including the provision of more houses to rent. Households also
benefit indirectly from infrastructure investments, if these are seen as legitimizing
previously illegal or informal settlements (discussed in the previous section).
Like land and finance, infrastructure for housing generally needs to be con-
sidered in conjunction with infrastructure for other uses. Roads, electricity, water
and sanitation are at some level all shared by households and firms, or are if
economies of scale are taken advantage of. In a series of studies of infrastructure
in Nigeria, Thailand and Indonesia, Lee and his associates have examined the ef-
ficiency losses from inappropriate infrastructure policies, with a particular focus
on manufacturing, although many of the arguments can be generalized to other
sectors (need more here). See, for example, Anas et al. (1996), Kessides (1993)
and Lee et al. (1996).
Government policies on the supply and pricing of urban infrastructure are
characterized by various conflicting tendencies. For example, governments have
taken the view that (a) water and sanitation (and sometimes other types of in-
frastructure) are merit goods; (b) infrastructure has significant externalities; (c)
low income households may, out of ignorance, seriously underestimate the bene-
fits of improved water and sanitation; and (d) some of these services involve large
1828 S. Malpezzi
4. Housing policy
This section is primarily about how the elements in Fig. 1, relate to each other.
In particular, how government policies profoundly affect these relationships.
The previous discussions of housing markets and related input markets, per-
force incorporated some discussion of public policy issues. In this section we
examine public policy more systematically. We begin with a stylized history of
thinking in the area, in both theory and practice. We examine council/public
housing, an idea brought to its ultimate conclusion in socialist countries. We
then introduce some alternative policy paradigms, sites and services: (Turner,
1972; Mayo and Gross, 1987), and upgrading: Indonesia's Kampung Improve-
ment Program (KIP), for example (Payne, 1984). We compare and contrast these
approaches with what some term the market wide approach and others term
the enabling approach (UN Centre for Human Settlements, 1990; World Bank,
1830 S. Malpezzi
large part of the communist elite's compensation, selling public units to sitting
tenants at nominal prices, in effect, capitalizes the value of the previous subsidy
and exacerbates perceived inequities in the distribution of income and wealth.
Among many other interesting points, the authors discuss the interaction be-
tween rent policy and selling price policy. The current situation, with low con-
trolled rents and low selling prices, bleeds local treasuries and (as just argued)
perpetuates perceived inequities. If, on the other hand, rents are kept low and sell-
ing prices are raised to market levels, privatization will be hindered as households
have strong incentives to keep renting. Conversely raising rents while keeping
selling prices low encourages privatization. Hegediis et al. (1992) favor raising
both selling prices and rents to market levels, to stem local government losses,
and to avoid distorting tenure decisions in either direction. This would also wipe
out current subsidies, which could be replaced with more equitable and tar-
geted housing allowances. However, they are not optimistic about the political
feasibility of raising rents and selling prices simultaneously.
A different point of view is expressed in Buckley et al. (1995). They argued
that moving to market rents is so unlikely for political reasons that the inequities
in the current distribution of units should be ignored, especially since the most
desirable units had already been privatized, i.e., the worst inequities could no
longer be addressed. Another argument they raised in favor of low rent-low
price policy was that governments would be hard pressed to come up with the
money for housing allowances. They therefore recommended a strategy not too
far removed from current practice: keep rents and selling prices low; effectively
giving units away would, they argue, allow a faster move to a market. The most
detailed and comprehensive roadmap to privatization on a wide scale remains the
study by Renaud et al. (1993) for Russia; see also Kaganova (1994) and Kosareva
and Struyk (1992).
4.2. Subsidies
Subsidies and public actions reduce the cost of something to particular recipients.
A common type of subsidy is a payment to someone for a particular purpose-
such as an allowance used for rent. But subsidies are also created when govern-
ment makes rules which change the price that someone has to pay for a good or
service-such as rent control.
In most countries, market as well as socialist, the pattern of subsidies that
has grown up over time has little to do with explicitly articulated policy objec-
tives. For example, most countries pay some sort of lip service to the notion that
housing subsidies should be at least partly targeted to low income households.
Ch. 44: Economic Analysis of Housing Markets in Developing and TransitionEconomies 1833
17 See, for example, Agrawal (1988), Piggott (1984), Trollegard (1989) and Nicholson and Willis (1990).
1834 S. Malpezzi
by lower income tenants, and only a fraction of eligible low income tenants
obtained public housing, leading to violations of horizontal and vertical equity.
Ondiege (1986) carried out a similar study for Kenyan public housing. He found
an average deadweight loss of only a little over 2%. But Ondiege also noted the
wide variance in outcomes across households, implying poor horizontal equity
and perverse vertical equity (the largest benefits go to the richest households).
Ravallion (1989) analyzes the welfare costs of stylized housing programs in
Indonesia, but is only able to present ex ante costs and benefits under different
assumptions of how binding standards would be.
Such welfare analyses are difficult to carry out in transition economies be-
cause of the lack of market comparators (e.g., a private market reasonably close
to equilibrium that can be used to estimate market prices for units and market
demand for households). Papers which present rough magnitudes based on cost
data include Wang (1991) and Pudney and Wang (1995).
A related issue that has not received the attention it deserves in developing
and transition economies is the extent to which public housing expenditure-for
owner occupied or rental units-simply crowds out private. For example, Murray
(1983) found that for every 100 public rental housing units built in the US during
the 1970s, private construction was reduced by about 85 units.'8 From the studies
done so far, in the developed and developing countries, it seems that publicly built
housing or subsidized new construction has not given good value for money. But
often the largest and most problematic expenditures never appear on the budget.
The dominance of off-budget housing expenditures through the financial sys-
tem and tax code in developed countries is well known. For example, in the
US the value of the mortgage interest deduction tax subsidy alone is at least
three times the size of on-budget federal housing expenditures. Off-budget ex-
penditures are often as much, or more, a central feature of housing policies
in developing and transition countries. In developed countries tax expenditures
receive much of the attention. In developing countries there are often large im-
plicit subsidies in the provision of land for shelter projects, although these are
somewhat self-limiting, as large implicit subsidies limit their scale. Such housing
finance subsidies are often "off the books"; Buckley and Mayo (1989) discuss
the example of Argentina. In transition economies, the subsidy implied by non-
market housing systems, already discussed, are large and only partly carried on
government budgets.
18 But more recent work by Murray 1993) highlights the fact that the degree of substitution varies markedly
by type of housing program.
Ch 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1835
Another important class of off-budget expenditures related to housing are en-
ergy related. These are particularly problematic in the former Soviet and eastern
European countries, where residential energy use was unpriced and spectacularly
wasteful. Renaud et al. (1993) calculate that Russian residential energy prices are
less than 10% of world prices, for example.
Housing allowances require a certain level of administrative capability, as well
as reasonably reliable income data. Many countries would have great difficulty
with the latter if not the former, given the nature of urban labor markets. Little
research has been undertaken on this important practical issue.
A number of authors suggest changes in regulation are often among the most
pressing areas for reform (Hannah et al., 1989; Dowall, 1992). Regulatory reform
can play a key role in the three areas just discussed, i.e., increasing the sup-
ply of finance, infrastructure and developable land. Zoning, taxes, rent controls
and building standards are other obvious regulatory areas to study for possible
change. Governments must carefully weigh the costs and benefits, and the distrib-
utional consequences, of regulation. Regulation should strive for a "level playing
field" insofar as is practical. Land regulation has already been briefly discussed;
we now discuss the issue more broadly, based on a simple framework laid out in
Hannah et al. (1989) and M&M (1997a).
While there is nothing technically difficult about doing so, hardly ever are the
effects of all the numerous taxes, regulations and subsidies added up. In the US
the "user cost" literature takes this approach, usually focusing on the interaction
between taxes, inflation and finance.' 9 In a number of developing countries, we
are beginning to adopt a variant of the same approach.
In this framework, there are three entities from whose point of view housing
policies and programs are evaluated: the economy, housing suppliers (or devel-
opers) and households. The exact incidence of the various costs and benefits
of government interventions can be a subtle issue. For example, although the
incidence of the property tax appears straightforward-property owners pay the
property tax-some portion of the tax could be shifted to tenants (for rental
property) or to the owners of capital generally (if capital markets were well
integrated).2 0 Incidence can depend on the competitiveness of the market, the
state of transactions costs and knowledge in the market, the efficiency of financial
markets in a country, and the time frame-in other words, it is rarely settled and
unambiguous. Hannah et al. (1989) adopt a simple approach, where the entire
cost or benefit is assigned to one participant. They point out that if our knowl-
edge of actual incidence improves, it would not be difficult to build in more
sophisticated treatment of incidence.
Hannah et al. (1989) and Malpezzi and Mayo (1997a) use the case of Malaysia
to illustrate the simple "incentives model". Malaysia's Special Low Cost Housing
Program was designed to induce private developers to build low-cost housing.
But analysis using this model demonstrated that despite strong demand, on bal-
ance government regulations still cost the developer money, raised costs and re-
duced supply. Many of these regulations yielded little or no benefit to consumers
or anyone else.
Similar models have been applied to Turkey (Baharagolu et al., 1997), Korea
(Kim K-H., 1991) and Ethiopia (Erbach et al., 1996). In each case, the application
has highlighted several interventions, especially on the regulatory front, that have
had unintended consequences, driving up housing costs, often disproportionately
at the low end of the market.
the aggregate supply of housing. Despite many studies which imply controls,
qualitatively reduce returns to rental investors, given the myriad ways real world
regimes work and ways around controls (legal and illegal) the size of the net
aggregate effect on supply remains unknown. Malpezzi and Ball (1993) found
that countries with stricter rent control regimes invested less in housing, in the
aggregate; but while they controlled for demand (income and demographics),
they were unable to control for other constraints on housing markets (e.g., land
use constraints, financial constraints). Since these may well be correlated with the
strength of controls, precise quantitative measures of the effects of rent controls
per se await future research.
The evidence todate casts doubt on controls' effectiveness as income transfer
mechanisms. In Cairo and Bangalore, for example, no relationship was found
between the benefits gained from reduced rent and household income, because
rent control is not well targeted to low income groups. In Kumasi and Rio, bene-
fits were found to be somewhat progressive (Malpezzi and Ball, 1991; Malpezzi,
1986; Malpezzi and Tewari, 1991; Malpezzi et al., 1990; Silveira and Malpezzi,
1991).
Another questionable assumption behind redistribution as a rationale for con-
trols is the notion that landlords are rich and tenants are poor. In Cairo, Kumasi
and Bangalore, the income of tenants and landlords was compared; and, while the
landlords' median income was higher in all three, there was significant overlap.
In Cairo, for example, about 25% of tenants had incomes that were higher than
the landlord median, and about 25% of landlords had incomes lower than the
tenant median. There is no guarantee the transfers will only occur from high
income landlords to low income tenants.
Rent control issues in formerly socialist countries are somewhat different.
While construction is now more or less decontrolled, rents on public and private
housing are still controlled. For example, in an unpublished presentation, Jan
Brzeski explained that in Poland current proposals for rent reform are not based
on letting rents seek market levels but on rents based on costs. Each year rents
may be increased by a fraction of the difference between current rents and 6%
of "replacement costs". Apart from the poor incentives inherent in cost-based
systems, in practice these replacement costs are poorly estimated. Tolley (1991)
provides a detailed analysis of cost-based rents in the Chinese context.
Generally, tenants in formally socialist countries have very strong security of
tenure, although if a tenant dies, and there are no parties with rights of succession
to the tenure, private property owners may now repossess the property, resetting
rents to market or selling. State rental property was devolved to municipalities.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1839
This has had serious repercussions for municipal finance, as rents do not cover
maintenance costs.
Alternatives for decontrol are analyzed in Malpezzi and Ball (1991), follow-
ing Arnott (1981). In addition to analysis of changes in controls per se, Malpezzi
and Ball emphasize the need for collateral reform in land, finance and housing
development regulations, for decontrol to work. Decontrol in an inelastic market
(due for example to other distortions) will largely raise rents with little supply
response, and lead to political pressure for reimposition of controls.
Reversing the normal order of such a discussion, we go from the specific to the
general. First, several "live" housing issues are briefly outlined by geographical
region. Because of the size of the countries and the scope of the issues they
face, we discuss India and China separately. Next we briefly examine research by
"modes of analysis", with particular reference to recent research on comparative
indicators. A more general research agenda concludes the chapter.
Current research on China focuses on how to reform prices and move from a
command and control system to a market oriented system in land and housing
markets. Other research points to the severe problems China faces managing a
huge money-losing public stock as the private housing market takes off. Chi-
na's rapid urbanization (see Becker and Morrison, Chapter 43 of this volume)
adds further urgency to the research agenda. Even more than in other countries,
housing issues in China are tied up with a host of other micro- and macroreform
issues, because of housing's link to employment (through enterprise housing) and
the fact that a large share of Chinese wages have traditionally been paid in-kind
Ch. 44. Economic Analysis of Housing Markets in Developing and TransitionEconomies 1841
rather than in cash (Tolley, 1991; Renaud and Bertaud, 1989; Hamer, 1996; Fong,
1989; Lim and Lee, 1993; Zhang, 1986).
Much recent research on India focuses on financial innovations such as replac-
ing state directed credit with private and quasiprivate institutions like the Housing
Finance and Development Corporation (HDFC). India also urgently needs ad-
ditional research on the regulatory environment for land and housing. Some
empirical research has been undertaken on rent control, but little has been done
on land use regulation or the regulation framework for housing finance, apart
from mainly descriptive analyses of the Urban Land Ceiling Act (Malpezzi and
Tewari, 1991; Mehta and Mehta, 1989; Mohan, 1992; Munjee, (date); Buckley,
1990; Acharya, 1987).
Remarkably little research has been published for other south Asian markets,
with the exception of Pakistan. Several papers have examined demand (Pasha and
Ghaus, 1988) and land issues, namely, rebulatory issues and the relative roles of
public versus land development in Pakistan (Dowall, 1989b, 1991a; Pasha, 1992).
The research agenda in east Asia is somewhat different. Here the focus is
on price bubbles and their effects, on households and the aggregate economy
(Chang, 1990). Other studies examine the order of liberalization of housing and
financial markets, and compares the "tigers" and "semitigers" with slower growth
economies like the Philippines and not-yet-liberalizing economies such as Burma.
(Bertaud and Malpezzi, 1994; Mayo, (date); Green et al., 1994; Kim K.-H.,
1990a, b, 1991a, 1993). Events at the time of this 1997 writing suggest fur-
ther research on the relationship between property lending, underwriting and
development incentives, and the aggregate economy would prove particularly
fruitful.
In central and eastern Europe, and the states of the former Soviet Union,
research focuses on developing property rights, the distributional consequences
of existing subsidies, and different methods of privatization. Issues of political
economy loom large here, as does research relevant to the development of true
financial intermediaries to replace conduit institutions, and the effects of broader
reform dynamics on housing, and vice versa (Sillince, 1990; Hegedus and Tosics,
1991; Daniel, 1989; Renaud, 1995a, b; Alexeev, 1988a; Kosareva and Struyk,
1993; Jaffe, 1989).
Current issues in northern Africa and the mideast include the effects of small
but heavily subsidized public housing programs and financial institutions that
offer even less transparent subsidies to a selected few households (Tipple, 1993;
Dehesh, 1994). Daniere (1992) examines tenure choice, and Struyk (1988b) va-
cancy rates. Research has been carried out on finance (Landeau, 1987) and the
informal sector (Mayo et al., 1980; Payne, 1980; Yonder, 1987). Particularly sen-
1842 S. Malpezzi
sitive and important housing issues arise relative to the development of politically
volatile parts of this region, e.g., Gaza and the West Bank.
Despite recent growth in countries such as Uganda, sub-Saharan Africa is
still in many respects a region in economic disarray, and the housing market is no
exception (Malpezzi and Sa-Aadu, 1996). Analysis of property rights, currently
in flux in many countries, is particularly useful (Ault and Rutman, 1979; Asabere,
1981; Mabogunje, 1992; Besley, 1993). Other literature addresses the conse-
quences of extremely disrupted and delclining economies for the housing market
(Amis and Lloyd, 1990). Other research focuses on "governance" issues, deep
subsidies to small minorities (Awotona, 1987; Megbolugbe, 1983). Many papers
make particular reference to South Africa, especially to apartheid's unwinding
and the housing market ties to locational issues (Turok, 1994a, b; Brueckner,
1996; Mayo, 1993a, b; Hoek Smit, 1992).
Latin America and the Caribbean have spawned significant research on fi-
nance under extreme inflation. Chile's experience with housing under stabiliza-
tion and structural adjustment is much studied (Morande, 1992; Renaud, 1988;
Rojas and Greene, 1995). Squatting and tenure issues also come to the fore
(Gilbert, 1983, 1989, 1993). One of the best integrated analyses of housing,
land and labor markets in any developing or developed market remains the "City
Study" of Bogota and Calli, Colombia (Ingram, 1984; Mohan, 1994).
Twenty years ago, many doubted that market-based models of housing markets
had much applicability to developing countries. Furthermore, it has been increas-
ingly clear that in many respects the "distance" between so-called developing
countries at middle and low levels of development are often greater than the
"distance" between particular pairs of developing countries. To give a simple
concrete example, the World Development Report reports a US life expectancy
of 77 years for the US, which is average for "high income countries". Jamaica,
Costa Rica, Sri Lanka and Jordan are just a few of the developing countries
with life expectancies over 70 years. Hong Kong, a country until recently often
classified as developing, has a life expectancy of 78 years. Contrast these coun-
tries with Mozambique, Sierra Leone, Uganda, Senegal and a half dozen other
countries with life expectancy under 50 years. One hardly has to develop the idea
that the countries we label "transition" are similarly diverse. Among the countries
of the former Soviet Union alone, per capita income ranges from around $3000
in Russia to under $500 in Azerbaijan and Tajikistan.
Ch. 44: Economic Analysis of Housing Markets in Developing and TransitionEconomies 1843
Thus, it is perhaps remarkable that this chapter documents that, in general,
market models have been fruitfully applied to so many and diverse places. But
the chapter also documents that widely divergent market conditions, institutions
and constraints requires a tailoring of models to the particular case at hand. A
number of examples now exist of solid case studies of markets, including analysis
of household survey data. See especially the studies of Cairo by Mayo et al.
(1982), of Indonesia by Struyk et al. (1990), of Bangkok by Angel et al. (1986),
and by Dowall (1989a), and of Cameroon by the Government of Cameroon in
1990. References on data collection and study design include Jones and Ward
(1993), Tipple and Willis (1991), Malpezzi et al. (1982) and Malpezzi (1984,
1988). Guides to regulatory research include the so-called Bertaud model, and
the "Malaysia model" of incentives (Hannah et al., 1989; M&M, 1995), and rent
control (Malpezzi and Rydell, 1986; Malpezzi et al., 1988). Examples of institu-
tional and property rights research include Cifuentes et al. (1984), Harsman and
Quigley (1991) and Page and Struyk (1990).
Many elements of the general research agenda are discussed in the preceding
paragraphs, and to some extent within the body of the chapter. More research is
needed on behavioral parameters and their determinants; particularly on the sup-
ply side. Additional research is needed on institutions and the role of governance.
There is a large agenda related to property rights. Most of the research surveyed
above is static; we need to learn more about dynamics, especially of reform.
1844 S. alpezzi
examined as part of an overall real estate market which also includes commercial
uses. This perspective has been largely neglected in the developing and transition
country literature. Paradoxically, in developing countries the distinction between
housing and commercial property is less distinct than in richer economies, since
housing unit and place of work often overlap (ILO, 1995).
The industrial organization of the housing market remains a fertile area for
research. Analysis of distributional as well as efficiency outcomes from privati-
zation of public housing in socialist and other countries would remain high on any
list. Despite the large literature on the production and consumption efficiencies of
public housing vouchers and other programs in the US and a few other countries,
remarkably little empirical research has been done in developing countries or the
former socialist economies. Even more remains to be done on the analysis of
off-budget expenditures, particularly subsidies in the tax and financial systems.
Many of these countries have not yet built up large tax expenditures as has the
US and other developed countries.
References
Aaron, H. (1975), Who Pays the Property Tax? (The Brookings Institute, DC).
Aaron, H.J. and G.M. von Furstenberg (1971), "The inefficiency of transfers in kind: the case of housing
assistance", Western Economic Journal 9:184-191.
Acharya, B.P. (1987), "The urban land ceiling act: a critique of the 1976 legislation", Habitat International
11:39-51.
Agrawal, N. (1988), "The economic effects of public housing in Australia", The Economic Record 254-267.
Agarwala, R. (1983), Price Distortions and Growth in Developing Countries, World Bank Staff Working Paper
No. 575.
Ahmad, N. (1994), "A joint model of tenure choice and demand for housing in the city of Karachi", Urban
Studies 31:1691-1706.
Akbar, M. (1990), "Rate of return to investment in housing attributes", Pakistan Journal of Applied Econ.
9:71-92.
Alchian, A.A. and H. Demsetz (1973), "The property rights paradigm", Journal ofEconomic History 33:16-27.
Alexeev, M. (1988a), "Market vs. rationing: the case of Soviet housing", Review of Economics and Statistics.
Alexeev, M. (1988b), "The effect of housing allocation on social inequality: a Soviet perspective", Journal of
Comparative Economics 12:228-234.
Alexeev, M. (1990), "Distribution of ahousing subsidies in the USSR, with some Soviet-Hungarian compar-
isons", Comparative Economic Studies 32:138-157.
Amis, P. (1984), "Squatters and tenants: the commercialization of unauthorized housing in Nairobi", World
Development 12:87-96.
Amis, P. and P. Lloyd (1990), Housing Africa's Poor (Manchester University Press, UK).
Anas, A., K.S. Lee and M. Murray (1996), Infrastructure Bottlenecks, Private Provision, and Industrial
Productivity: A Study of Indonesian and Thai Cities, World Bank, Policy Research Working Paper No.
1603.
Andrusz, G.D. (1990), "Housing policy in the Soviet Union", in: J.A.A. Sillince, ed., Housing Policies in
Eastern Europe and the Soviet Union (Routledge, New York).
Angel, S. and S. Chuated (1990), "The down-market trend in housing production in Bangkok, 1980-87", Third
World Planning Review 12:1-20.
1846 S. Malpezzi
Angel, S. and S.K. Mayo (1996), Enabling Policies and Their Effects on Housing Sector Perfotnnance: A
Global Comparison. Paper presented to the Habitat II Conference, Istanbul, Turkey.
Angel, S. et al. (1986), The Land and Housing Markets of Bangkok: Strategies for Public Sector Participation,
PADCO. Thailand National Housing Authority, and ADB.
Annez, P. and W.C. Wheaton (1984), "Economic development and the housing sector: a cross-national model",
Economic Development and Cultural Change 32:749-746.
Antwi, A. (1995), An Economic Perspective of Land Policy Implementation in a Developing Economy: The
Case of Accra (Royal Institute of Chartered Surveyors, Ghana).
Archer, R.W. (1989), An Outline of Urban Land Policy for the Developing Countries of Asia (Asian Institute
of Technology, Bangkok).
Arimah, B.C. (1992a). "An empirical analysis of the demand for housing attributes in a Third World City",
Land Economics 68:366-379.
Arimah, B.C. (1992b), "Hedonic prices and the demand for housing attributes in a Third World city: the case
of Ibadan, Nigeria", Urban Studies 29:639-651.
Arimah, B.C. (1994), 'The income elasticity of demand in a sub-Saharan African housing market: evidence
from Ibadan, Nigeria". Environment and Planning A 26:107-119.
Amott, R.J. (1987), "Economic theory and housing", in: E.S. Mills, ed., Handbook of Regional and Urban
Economics, North Holland, V.2.
Arnott, R. and N. Johnston (1981), Rent Control and Options for Decontrol in Ontario (Ontario Economic
Council, Toronto).
Asabere, P. (1981a), "The determinants of land values in an African city: the case of Accra, Ghana", Land
Economics.
Asabere. P. (1981b). "The price of land in a chiefdom: empirical evidence on a traditional African city,
Kumasi", Journal of Regional Science: 529-539.
Asabere, P.K. and K. Owusu-Banahene (1983), "Population density function for Ghanaian (African) cities: an
empirical note", Journal of Urban Economics 14:370-379.
Asabere, K., C.F. Sirmans and P.E. Colwell (1982), "The intensity of residential land use in Accra, Ghana",
Journal of Urban Economics 11:190-198.
Ashenfelter, 0. and D. Genesove (1992), "'Testing for price anomalies in real estate auctions". The American
Economic Review 82:501-505.
Assadian, A. and J. Ondrich (1993), "Residential location, housing demand and labor supply decisions of one-
and two-earner households: the case of Bogota, Columbia", Urban Studies 30:73-86.
Ault, D.E. and G.L. Rutman (1979). "The development of individual rights to property in tribal Africa", Journal
of Law and Economics 22:163-182.
Awotona, A. (1987), "Housing policy in Nigeria-government policies for housing Nigeria's urban poor and
the working class: laudable promises, great expectations, colossal failure", Habitat International 11:89-103.
Baer, W.C. (1991), "Filtering and third-world housing policy", Third World Planning Review 13:69-82.
Baer, W. and C. Koo (1994), "Housing turnover in developing nations: the case of Seoul". Journal of Planning
Education and Research 13:104-118.
Baharoglu, D., L.M. Hannah and S. Malpezzi (1997), Getting Housing Incentives Right in Turkey, University
of Wisconsin, Center for Urban Land Economics Research Working Paper.
Baken. R.-J. and J. Van der Linden (1993), "Getting the incentives right: banking on the formal private sector-
a critique of World Bank thinking on low-income housing delivery in third world cities", Third World
Planning Review 15:1-22.
Bariandiaran, E. and E. Mills (1983), Macroeconomic Analysis of the Chilean Housing Program, Mimeo.
Barlow, J. and A. King (1992), "The state, the market, and competitive strategy: the house building industry in
the United Kingdom, France, and Sweden", Environment and Planning A (UK) 24:3811-400.
Barrett, R. (1983), "A brief review of World Bank experience in traffic management", World Bank.
Barro, R.J. and X. Sala-i-Martin (1995), Economic Growth (McGraw Hill, New York).
Barry, C., G. Casteneda and J. Lipscomb (1994), "The structure of mortgage markets in Mexico and prospects
for their securitization". Journal of Housing Research 5:173-204.
Bartone, C. (1990), "Water quality and urbanization in Latin America", Water International 15:3-14.
Bartone, C. (1991), "Environmental challenge in third world cities", APA Journal: 411-415.
Ch. 44: Economic Analysis of Housing Markets in Developing and TransitionEconomies 1847
Bater, J.H. (1994), "Housing developments in Moscow in the 1960s", Post-Soviet Geography 35:309-328.
Behbehani, R., V. Shetty Pendakur and A. Armstrong-Wright (1984), Singapore Area Licensing Scheme: A
Review of the Impact (Water Supply and Urban Development Department, World Bank).
Belkina, T. (1994), "Housing statistics and the condition of the housing sector", Problems of Economics
Transition 37:56-71.
Benninger, C. (1986), Formal and Semi-formal Strategies of Land Acquisition. Poona, India (Centre for
Development Studies).
Bertaud, A. (1992), "The impact of land-use regulations on land supply, consumption and price", Regional
Development Dialogue 13:35-40.
Bertaud, A. (1997), The Spatial Distribution of Population in Cracow: A City's Structure Under the Conflict-
ing Influences of Land Markets, Zoning Regulations and its Socialist Past. Paper presented to the Lincoln
Institute of Land Policy Conference on Land Prices, Land Information Systems, and the Market for Land
Information (Cambridge, Mass.).
Bertaud, A. and S. Malpezzi (1994), Measuring the Costs and Benefits of Urban Land Use Regulation: A
Simple Model with an Application to Malaysia (Center for Urban Land Economics Research, University of
Wisconsin).
Bertaud, A. and B. Renaud (1994), Cities Without Land Markets: Lessons of the Failed Socialist Experiment,
World Bank Discussion Paper No. 227.
Bertaud, A. and B. Renaud. (1997), "Socialist cities without land markets", Journal of Urban Economics
41:137-151.
Bertaud, A., M.-A. Bertaud and J. Wright (1988), Efficiency in Land Use and Infrastructure Design: An
Application of the Bertaud Model, World Bank, INU Discussion Paper.
Bertaud, M.-A. and D. Lucius (1989), Land Use, Building Codes, and Infrastructure Standards as Barriers to
Affordable Housing in Developing Countries (World Bank Urban Development Department).
Besley, T. (1993), Property Rights and Investment Incentives: Theory and Microevidence from Ghana
(Research Program in Development Studies, Princeton University).
Besley, T. (1995), "Finance, credit and insurance", in: J. Behrman and T.N. Srinivasan, eds., Handbook of
Development Economics, vol. 3(A) (Elsevier).
Betancur, J.J. (1987), "Spontaneous settlement housing in Latin America: a critical examination", Environment
and Behavior 19:286-310.
Bikhchandri, S. (1986), Market Games With Few Traders, Ph.D. Thesis, Stanford University School of
Business.
Binswanger, H.P. and K. Deininger (1993), "South African land policy: the legacy of history and current
options", World Development 21:1451-1475.
Blaesser, B.W. (1981), Clandestine Development in Colombia (U.S. Agency for International Development).
Blinder, A. (1987), Hard Heads and Soft Hearts (Excellent chapter on the environment, Addison Wesley).
Boehm, T.P. (1981), "Tenure choice and expected mobility: a synthesis", Journal of Urban Economics 10.
Boleat, M. (1985), National Housing Finance Systems: A Comparative Study. London: Croom Helm.
Briscoe, J. (1984), "Water supply and health in developing countries: selective primary care revisited",
American Journal of Public Health: 1009-1013.
Bromley, D.W. (1989), "Property relations and economic development: the other land reform", World
Development 17:867-877.
Browder, J.O. and J.R. Bohland (1995), "Patterns of development on the metropolitan fringe: urban fringe
expansion in Bangkok, Jakarta and Santiago", Journal of the American Planning Association 61:310-327.
Browder, O.L. (1984), Basic Property Law, Fourth Ed. (West Publishing, St. Paul).
Brueckner, J.K. (1996), "Welfare gains from removing land-use distortions: an analysis of urban change in
post-apartheid South Africa", Journal of Regional Science 36:91-109.
Buckley, R.M. (1985), Pricing Federal Credit Programs: An Application of the Option Pricing Perspective to
the AID Housing Guaranty Program, Urban Institute Working Paper.
Buckley, R.M. (1990), "Housing finance in India", Mimeo (World Bank).
Buckley, R.M. (1991), "The measurement and targeting of housing finance subsidies: the case of Argentina",
Public Finance 46:355-372.
1848 S. Malpezzi
Buckley, R.M. (1994), "Housing finance in developing countries: the role of credible contracts", Economic
Development and Cultural Change 42:317-332.
Buckley, R.M. (1996), Housing Finance in Developing Countries (Oxford).
Buckley, R.M. and A. Dokeniya (1989), Inflation, Monetary Balances and the Aggregate Production Function,
World Bank, INU Discussion Paper No. 55.
Buckley, R.M. and E.N. Gurenko (1995), Housing Finance in Reforming Economies: Measuring Macro
Constraints on Moving to Market (World Bank).
Buckley, R.M. and E.N. Gurenko (1997), "Housing and income distribution in Russia: Zhivago's legacy"'',
World Bank Research Observer 12:19-32.
Buckley. R.M. and R. Madhasudhan (1984). The Macroeconomics of Housing's Role in the Economy: An
International Analysis. Paper presented to the American Real Estate and Urban Economics Association.
Buckley, R.M. and S.K. Mayo (1989), "Housing policy in developing countries: evaluating the macroeconomic
impacts", Review of Regional and Development Studies 1.
Buckley, R.M. and B. Renaud (1988), Urban Finance in Post-Depression Latin America: The Solvency Status
of Urban Borrowers, INU Discussion Paper No. 22.
Buckley, R.M. and R.J. Struyk (1985), An Economic Analysis of AID's Housing Guaranty Program. Urban
Institute Working Paper.
Buckley, R.M., B. Lipman and T. Persaud (1993a), "Mortgage design under inflation and real wage uncertainty:
the use of dual indexed instruments", World Development 21:455-464.
Buckley, R.M., P.H. Hendershott and K.E. Villani (1993b), Ponzies to Zombies: Housing Finance in Previously
Centrally Planned Economies (World Bank).
Buckley, R.M., Z. Daniel and M. Thalwitz (1993c), The Welfare Cost and Transparency of Housing Policy in
Socialist Economies: The Hungarian Experience (Urban Development Division, World Bank).
Buckley, R.M., P.H. Hendershott and K.E. Villani (1995), "Rapid housing privatization in reforming
economies: pay the special dividend now", Journal of Real Estate Finance and Economics 10:63-80.
Bums, L.S. and L. Grebler (1976), "Resource Allocation to Housing Investment: A Comparative Intemational
Study", Economic Development and Cultural Change 25:95-121.
Bums, L.S. and L. Grebler (1977), The Housing of Nations: Advice and Policy in a Comparative Framework
(Macmillan, London).
Caimcross, S. and J. Kinnear (1988), Measurement of the Elasticity of Domestic Water Demand: A Study of
Water Vendors and their Clients. in Urban Sudan (London School of Hygiene and Tropical Medicine).
Cameroon, Republic of (1990), Evaluation of Real Housing Needs in Cameroon (Ministry of Town Planning
and Housing, Yaounde, Cameroon).
Chang. C.-O. and C.W.R. Ward (1993), Forward Pricing and the Housing Market: The Pre-Sales Housing Sys-
tem in Taiwan. Paper presented to the American Real Estate and Urban Economics Association International
Conference, Mystic, CT.
Chang, C.-O. and P.PY. Lai (1993), Indicators of Real Estate Cycles in Taiwan. Working Paper, National
Chengchi University.
Chang, C.-O. and P. Linneman (1990), 'Forecasting housing investment in developing countries", Growth and
Change 21:59-72.
Chang, C.-O., C.-C. Lin and S.C.L. Farr (1994), A Study of Real Estate Transaction Price in Taipei
Metropolitan Area (National Chengchi University).
Charemza, W. and R. Quandt (1990), "Dual housing markets in a centrally planned economy", Kyklos 43:411-
436.
Cheung, Y.-L., S.-K. Tsang and S.-C. Mak (1995), "The causal relationships between residential property
prices and rentals in Hong Kong: 1982-1991", Journal of Real Estate Finance and Economics 10:23-36.
Chiquier, L. and B. Renaud (1992), Alternative Mortgage Instruments in Distorted Housing Systems: How
Useful is the Dual Rate Adjustable Mortgage? (EMTIN Technical Department. World Bank).
Choko, M.H. and C. Guangting, eds., (1994), China: The Challenge of Urban Housing (Meridien, Laval,
Quebec).
Chomchan, S. and S. Silapacharanan (1990), Land Use Development in Bangkok and Major Cities: Directions
and Problems. Thailand Development Research Institute Foundation. Background Report No. 5-1.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1849
Chou, W.L. and Y.C. Shih (1995), "Hong Kong housing markets: overview, tenure choice, and housing
demand", Journal of Real Estate Finance and Economics 10:7-22.
Chow, Y.-F. and N. Wong (1997), Property Value, User Cost, and Rent: An Investigation of the Residential
Property Market in Hong Kong (Chinese University of Hong Kong).
Ciechoncinska, M. (1990), "The social image of urban housing-dreams and realities: the case of Poland",
Journal of Urban Affairs 12:157-172.
Cifuentes V., J.R. Stevenson and R.L. Paredes (1984), Housing and Urban Development in Bogota: The
Institutional Backdrop (World Bank, Washington).
Clapham, D. (1995), "Privatization and the east European housing model", Urban Studies 32:679-694.
Coase, R.H. (1960), "The problem of social cost", Journal of Law and Economics 3:1-44.
Cook, D.B. (1984), "Building codes and regulations in low income settlements", in: P.J. Richards and A.M.
Thomson, eds., The Urban Poor: The Provision of Communal Services (Croom Helm).
Daniel, Z. (1983), "Public housing, personal income and central redistribution in Hungary", Acta Oeconomica
31:87-104.
Daniel, Z. (1985), "The effects of housing allocation on social inequality in Hungary", Journal of Comparative
Economics.
Daniel, Z. (1989), "Housing demand in a shortage economy: results of a Hungarian survey", Acta Oeconomica
41:157-179.
Daniel, Z. and G. Partos (1989), Housing Reforms in Hungary: History and Lessons of Experience. Paper
presented to the Joint Seminar on the Urban Housing Reforms in China, Beijing.
Daniel, Z. and A. Semjen (1987), "Housing shortage and rents: the Hungarian experience", Economics of
Planning.
Daniell, J. and R. Struyk (1994), "Housing privatisation in Moscow-who privatize and why", International
Journal of Urban and Regional Research 18:510-526.
Daniere, A.G. (1992), "Determinants of tenure choice in the Third World: an empirical study of Cairo and
Manila", Journal of Housing Economics 2.
Daniere, A.G. (1994), "Estimating willingness to pay for housing attributes-an application to Cairo and
Manila", Regional Science and Urban Economics 26:577-600.
Deaton, A.S. (1984), The Demand for Personal Travel in Developing Countries: Pricing and Policy Analysis.
Paper presented to the Transportation Research Board, Washington, D.C.
De Leeuw, F. (1971), "The demand for housing: a review of the cross-section evidence", Review of Economics
and Statistics 53:1-10.
De Leeuw, E and L. Ozanne (1981), "Housing", in: H. Aaron and J. Pechman, eds., How Taxes Affect
Economic Behavior (The Brookings Institute, DC).
De Soto, H. (1989), The Other Path: The Invisible Revolution in the Third World (Harper and Row, New York).
Dehesh, A. (1994), "Developmental instability in Iran: its impact on housing since 1962", Cities 11:409-424.
Demsetz, H. (1967), "Towards a theory of property rights", American Economic Review 57:347-373.
Dhareshwar, A. (1987), Transportation Sector Bibliography. World Bank Transportation Department Report
TRP8.
Diamond, D.B. Jr. (1978), "A note on inflation and relative tenure prices", Journal of the American Real Estate
and Urban Economics Association 6:438-450.
Diamond, D.B. Jr. (1993), "Restructuring housing finance in Hungary", Housing Finance International,
December, pp. 10-42.
Diamond, D.B. Jr. and M.J. Lea (1992), "The decline of special circuits in developed country housing finance",
Housing Policy Debate 3:747-778.
Diamond, D.B. Jr. and M.J. Lea (1993), "Housing finance in developed countries: an international comparison
of efficiency", in B. Turner and C. Whitehead, eds., Housing Finance in the 1990s (Gavle, Sweden: The
National Swedish Institute for Building Research).
DiPasquale, D. and W.C. Wheaton (1992), "The markets for real estate assets and space: a conceptual
framework", Journal of the American real Estate and Urban Economics Association 20:181-198.
Dorner, P. (1992), Latin American Land Reforms in Theory and Practice (University of Wisconsin Press).
Dowall, D. (1988), "The land and housing market assessment: an important tool", Habitat International 12:135-
149.
1850 S. Malpezzi
Dowall, D. (1989a), "Bangkok: a profile of an efficiently performing housing market", Urban Studies: 327-339.
Dowall, D. (1989b), Karachi Land and Housing Market Study (UNCHS and PADCO, Karachi Development
Authority).
Dowall, D. (1991a), "Comparing Karachi's informal and formal housing delivery systems", Cities 8:217-227.
Dowall, D. (1991b), The Land Market Assessment: A New Tool for Urban Management. Jointly published
by the United Nations Development Program, The World Bank, and the United Nations Centre for Human
Settlements (Habitat).
Dowall, D.E. (1991c), "A second look at the Bangkok land and housing market", Urban Studies 29:25-37.
Dowall, D.E. (1991d), "The Karachi development authority: failing to get the prices right", Land Economics
67:463-471.
Dowall, D. (1992), "Benefits of minimal land-use regulations in developing countries", Cato Journal 12:413-
423.
Dowall, D. (1993), "Establishing urban land markets in the People's Republic of China", Journal of The
American Planning Association 59:1.82-192.
Dowall, D.E. (1997), A Retrospective Look at the Land Market Assessment. Paper presented to the Lincoln
Institute of Land Policy, Cambridge, November.
Dowall, D.E. and M. Leaf. (1991), "The price of land for housing in Jakarta", Urban Studies 28:707-722.
Eberts, Chapter 38 this volume.
Edwards, M. (1982), "Cities of tenants: renting among the urban poor in Latin America", in: A. Gilbert et al.,
eds., Urbanization in Contemporary Latin America (Wiley, New York).
Edwards, M. (1990), "Rental housing and the urban poor: Africa and Latin America compared", in: A. Phillip
and P. Lloyd, eds., Housing Africa's Poor (Manchester University Press, Manchester).
Erbach, G.E., K. Moges and J.K. Bachmann (1996), Cost Benefit Analysis of Representative Housing
Investents (Ministry of Works and Urban Development, Government of Ethiopia).
European Conference of Ministers of Transport (1995), Transport Infrastructure in Central and Eastern
European Countries (OECD, Paris).
Evans, A.W. (Chapter 42, this volume).
Evers, H. (1976), "Urban expansion and land ownership in underdeveloped societies", in: J. Walton and L.
Masotti, eds., The City in Comparative Perspective (Sage).
Fannie Mae (1991), Housing Finance Arrangements: A Comparative Analysis. Conference Proceedings.
Fannie Mae (1992), Housing Finance for Sustainable Development. Conference Proceedings.
Farvaque, C. and P. McAuslan (1992), Urban Land Policies and Institutions in Developing Countries. World
Bank/UNDP Urban Management Program Paper UMP-5.
Ferchiou, R. (1982), "The indirect effects of new housing construction in developing countries", Urban Studies
19:167-176.
Ferguson, B., J. Rubenstein and V. Domingez Vial (1996), "The design of direct demand subsidies for housing
in Latin America", Review of Regional and Urban Studies 8:202-219.
Fitzwilliam Memorandum, The (1991), Land Value Changes and the Impact of Urban Policy upon Land
Valorisation Processes in Developing Countries. Conclusions of an International Research Workshop
at Fitzwilliam College, Cambridge; reprinted in International Journal of Urban and Regional Research
15:623-628.
Follain, JR. and E. Jimenez (1985a), "Estimating the demand for housing characteristics: a survey and
critique", Regional Science and Urban Economics 15.
Follain, J.R. and E. Jimenez (1985b), "The demand for housing characteristics in developing countries", Urban
Studies 22.
Follain, J.R. and S. Malpezzi (1980), Dissecting Housing Value and Rent (Urban Institute).
Follain, J.R., G.-C. Lim and B. Renaud (1980), "The demand for housing in developing countries: the case of
Korea", Journal of Urban Economics 7.
Follain, J.R., G.-C. Lim and B. Renaud (1982), "Housing crowding in developing countries and willingness to
pay for additional space: the case of Korea", Journal of Development Economics 11:249-272.
Follain, J.R., E. Jimenez, S. Malpezzi and S.K. Mayo (1983), "Housing in Korea: recent trends and recurrent
problems", in Gill-Chin Lim, ed., Urban Planning and Spatial Strategies in Rapidly Changing Societies
(Consortium on Urban and Regional Policies in Developing Countries, Princeton).
Ch. 44: Economic Analysis of Housing Markets in Developing and TransitionEconomies 1851
Hall, D., ed. (1993), Transport and Economic Deveiopment in the New Central and Eastern Europe (Belhaven
Press, London and Halsted Press, New York).
Hamer, A.M. (1996), The World Bank's Involvement in Urban Housing Reforms in China. Paper presented to
the American Real Estate and Urban Economics Association International Conference, Orlando.
Hamer, A.M. et al. (1993), China: Urban Land Management in an Emerging Market Economy (World Bank.
Washington, DC).
Hannah, L., A. Bertaud, S. Malpezzi and S. Mayo (1989), Malaysia: The Housing Sector; Getting the Incentives
Right. World Bank Sector Report No. 7292-MA.
Hannah, L., K.-H. Kim and E.S. Mills (1993), "Land use controls and housing prices in Korea", Urban Studies
30:147-156.
Hanushek. E. and J. Quigley (1978), "An explicit model of intra-metropolitan mobility", Land Economics
411-429.
Hardoy, J.E. and D. Satterthwaite (1989). Squatter Citizen: Life in the Urban Third World (Earthscan, London).
Hardoy, J., S. Caimcross and D. Satterthwaite, eds. (1990), The Poor Die Young: Housing and Health in Third
World Cities (Earthscan Publications, London).
Harsman, B. and J.M. Quigley (1991), Housing Markets and Housing Institutions: An International Compari-
son (Kluwer Academic Press, Dordrecht).
Hegedus, J. and I. Tosics (1991), "Filtering in socialist housing systems-results of vacancy chain surveys in
Hungary", Urban Geography 1:1-18.
Hegedus, J. and I. Tosics (1994), 'The poor, the rich and the transformation of urban space", Urban Studies
31:989-993.
Hegedus, J., K. Mark. R. Struyk and I. Tosics (1992), The Privatization Dilemma in Budapest's Public
Rental Housing Sector. Paper presented to the American Real Estate and Urban Economics Association
international meeting.
Hegedus, J., K. Mark, R. Struyk and I. Tosics (1994), "Tenant satisfaction with public housing management:
Budapest in transition", Housing Studies 9:315-328.
Hinds, M. (1990), Issues in the Introduction of Market Forces in Eastem European Socialist Economies. World
Bank: Europe, Middle East and North Africa Region Internal Discussion Paper No. IDP-0057.
Hoek-Smit, M. (1992), Housing Supply Systems and Housing Outcomes in Southern Africa. Paper prepared
for the American Real Estate and Urban Economics Association/USC International Real Estate Conference,
Los Angeles.
Hoffman, M., B. Haupt and R.J. Struyk (1991), International Housing Markets: What We Know; What We
Need to Know. Fannie Mae Working Paper.
Howe, H. and P. Musgrove (1977), "An analysis of ECIEL budget data for Bogota, Caracas, Guyaquil and
Lima", in: C. Lluch et al., eds., Patterns in Household Demand and Savings (Oxford University Press,
Oxford).
Howenstine, E.J. (1957), "Appraising the role of housing in economic development", International Labor
Review 75.
Hoy, M. and E. Jimenez (1991), "Squatter rights and urban development: an economic perspective",
Econonomica 58:79-92.
Ihlanfeldt, K.R. (1981), "An empirical investigation of alternative approaches to estimating equilibrium
demand for housing". Journal of Urban Economics 9:97-105.
International Labour Office (1995), Shelter Provision and Employment Generation. Geneva.
Ingram, G.K. (1984), Housing Demand in the Developing Metropolis: Estimates from Bogota and Cali,
Colombia. Staff Working Paper No. 633.
Ingram, G.K. and A. Carroll (1981), "The spatial structure of Latin American cities", Journal of Urban
Economics 9:257-273.
International Journal of Urban and Regional Research (1987), Special Volume on Eastern Europe 11.
Ivanicka, K., M. Zubkova, E. Sindlerova and D. Spirkova (1996), Development of Housing Finance in Vise-
grad Countries. Paper presented to the American Real Estate and Urban Economics Association International
Conference, Orlando.
Jaffe, A.J. (1989), "Concepts of property, theories of housing, and the choice of housing policy", Netherlands
Journal of Housing and Environmental Research 4.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1853
Jaffe, A.J. (1989), Property Rights and Market Behavior in Eastern European Housing Reforms. Paper pre-
sented to the January (1993) meeting of American Real Estate and Urban Economics Association,
Anaheim,
CA.
Jaffe, A.J. and D. Louziotis, Jr. (1996), "Property rights and economic efficiency:
a survey of institutional
factors", Journal of Real Estate Literature 4:137-159.
Jimenez, E. (1982a), "The economics of self help housing: theory and some evidence
from a developing
country", Journal of Urban Economics 11:205-228.
Jimenez, E. (1982b), "Thevalue of squatter dwellings in developing countries", Economic
Development and
Cultural Change 30.
Jimenez, E. (1984), "Tenure security and urban squatting", Review of Economics and Statistics
66:556-567.
Jimenez, E. (1985), "Urban squatting and community organization in developing countries",
Journal of Public
Economics 27:69-92.
Jimenez, E. (1987a), Pricing Policy in the Social Sectors: Cost Recovery for Education
and Health in
Developing Countries (Johns Hopkins University Press).
Jimenez, E. (1987b), "The magnitude and determinants of home improvement in self-help
housing: Manila's
Tondo project", Land Economics 70-83.
Jimenez, E and D. Keare (1984), "Housing consumption and permanent income
in developing countries:
estimates from panel data in El Salvador", Journal of Urban Economics 15:172-194.
Johnson, T.E. Jr. (1985), A Study of Upwards Filtering of Housing Stock as a Consequence
of Informal
Settlement Upgrading in Developing Countries (Graduate School of Design, Harvard
University).
Johnson, T.E. Jr. (1987), "Upward filtering of the housing stock", Habitat International
1 1:173-190.
Jones, G. and P. Ward. eds., (1994), Methodology for Land Market and Housing Analysis
(University College
Press, London).
Kaganova, O.Z. (1994), On the Program of the Land Policy and Reform of Taxation
in the Cities of Russia
Proposed by Adherents of Henry George's Ideas. Discussion Paper.
Kalbermatten, J. et al. (1980), Appropriate Technology for Water Supply and Sanitation:
A Summary of
Technical and Economic Options (World Bank).
Kalbermatten, J., DeAnne Julius, D. Mara and C. Gunnerson (1982), Appropriate Sanitation
Alternatives: A
Technical and Economic Appraisal (Johns Hopkins University Press).
Karadimov, N. (1992), The Reorientation of Urban Development and Housing Policy in
Central and East-
European Countries in their Transition to a Market Economy: General and Specific
Tendencies and
Phenomena of the Experience of the Republic of Bulgaria (Ministry of Regional
Development, Housing
Policy and Construction, Sofia).
Kasongo, B.A. and A.G. Tipple (1990), "An analysis of policy towards squatters
in Kitwe, Zambia", Third
World Planning Review 12:147-165.
Katsura, H.M. (1984), Economic Effects of Housing Investment. Urban Institute Contract
Report.
Katsura, H.M. and R.J. Struyk (1991), "Selling eastern Europe's social housing stock:
proceed with caution",
Housing Policy Debate 2:1251-1271.
Katzman, M. (1977), "Income and price elasticities of demand for water in developing
countries", Water
Resources Bulletin 13:47-55.
Kaufman, D. and J. Quigley (1987), "The consumption benefits of investment in infrastructure:
the evaluation
of sites and services programs in underdeveloped countries", Journal of Development
Economics 25:263-
284.
Kessides, C. (1993), The Contributions of Infrastructure to Economic Development. World
Bank Discussion
Paper No. 213.
Kiamba, M. (1989), "The introduction and evolution of private landed property in Kenya", Development and
Change 20:121-148.
Kidokoro, T. (1992), "Development control systems for housing development in southeast
Asian cities",
Regional Development Dialogue 12:64-87.
Kim, J. (1995), "Demand for rental housing in Korea with some estimation issues", Journal of Urban
Planning
and Development-Asce 121:57-74.
Kim, J.-H. (1990), "Korean housing policies: review and future directions", in: Hyundai Research
Institute,
Proceedings of (the) International Housing Conference on Korean Housing Policies, Seoul.
1854 S. Maipezi
Kim. J.-H. (1991). Housing Program Evaluation Using the Present Value Model: A Korean Experience. Paper
presented to the World Bank seminar on Korean Housing Markets and Policy.
Kim, K.-H. (1990a). "An analysis of inefficiency due to inadequate mortgage financing: the case of Seoul,
Korea", Journal of Urban Economics 28:371-390.
Kim, K.-H. (1990b), "Housing finance and housing related taxes in Korea", in Hyundai Research Institute,
Proceedings of (the) International Housing Conference on Korean Housing Policies, Seoul.
Kim, K.-H. (1993), "Housing prices, affordability and government policy in Korea", Journal of Real Estate
Finance and Economics 6:55-72.
Kim, K.-H. (1994), Financing Shelter and Urban Development: Towards Habitat II (UN Centre for Human
Settlements, Nairobi).
Kim, K.-H and K.-Y. Kim (1992), Policy for Public Housing Finance in Korea. Paper presented at the
International Symposium on Housing, Korea National Housing Corporation, Seoul.
Kim, K.-H. and E.S. Mills (1988), "Korean development and urbanization: prospects and problems", World
Development 16:157-167.
Kim, K.-H. and S.-H. Suh (1989), An Analysis of Optimality of Housing Investment in Korea. Paper presented
to the Korea International Economic Association.
Kim, K.-H. and S.-H. Suh (1993), "Speculation and house price bubbles in the Korean and Japanese real estate
markets", Journal of real estate Finance and Economics 6:73-88.
Kim, S.-J. (1992). "A model of rental housing choices in the Korean market", Urban Studies 29:1247-1263.
Kingsley, G.T. and R.J. Strnyk (1990), Institutional Refonn and the Housing Sector: Thinking Through the
Transition. Urban Institute Working Paper, U.I. Project No. 3941.
Kingsley, G.T. and R.J. Struyk (1992), Progress in Privatization: Transforming Eastern Europe's Social
Housing (The Urban Institute for USAID).
Klak, T. and M. Holtzclaw (1993), "The housing, geography, and mobility of Latin American urban poor: the
prevailing model and the case of Quito, Ecuador", Growth and Change 24:247-276.
Kornai. J. (1990), The Road to a Free Economy (Norton).
Kornai, J. (1992), The Socialist System: The Political Economy of Communism (Princeton University Press).
Kosareva, N. and R. Struyk (1993), "Housing privatization in the Russian Federation"'. Housing Policy Debate
4:81-100.
Kosareva, N., O. Pchelintsev and G. Ronkin (1991), "On the path to housing reform-analysis and forecast",
Problems of Economics 34:60-82.
Kravis, I.B., A. Heston and R. Summers (1982), World Product and Income: International Comparisons of
Real Gross Product (Johns Hopkins University Press).
Kuznets, S. (1961), "Quantitative aspects of the economic growth of nations. VI: long term trends in capital
formation proportions", Economic Development and Cultural Change 9:3-124.
Lakshmanan, T.R., L. Chatterjee and P. Kroll (1978). "Housing consumption and level of development: a cross
national comparison", Economic Geography 54:222-233.
Landeau, J.-F. (1987), "Tunisia: a case study in analyzing the affordability of mortgage loans", African Urban
Quarterly.
Landis, J.D. (1983), How Competitive Are Urban Housing Markets? Ph.D. Dissertation (University of
California, Berkeley).
Landis, J.D. (1986), "Land regulation and the price of new housing: lessons from three California cities".
Journal of the American Institute of Planners: 9-21.
Larbi, W.O. (1995), The Urban Development Land Process and Urban Land Policies in Ghana (Royal Institute
of Chartered Surveyors).
Lea, M.J. and S.A. Bernstein (1995), International Housing Finance Sourcebook 1995 (Intenmational Union of
Housing Finance Institutions, Chicago).
Lee, K.S. (1988). Infrastructure Constraints on Industrial Growth in Thailand. INURD Working Paper No.
88-2.
Lee, K.S. (1992), "Spatial policy and infrastructure constraints on industrial growth in Thailand", Review of
Urban and Regional Development Studies 4.
Lee, KS. and A. Anas (1992), "Costs of deficient infrastructure: the case of Nigerian manufacturing", Urban
Studies 29:1071-1092.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1855
Lee, K.S., A. Anas and G.-T. Oh (1996), Costs of Infrastructure Deficiencies in Manufacturing in Indonesia,
Nigeria and Thailand. World Bank, Policy Research Working Paper No. 1604.
Lee, L.-F. and R. Trost (1978), "Estimation of some limited dependent variable models
with application to
housing demand", Journal of Econometrics: 357-382.
Leibenstein, H. (1966), "Allocative efficiency versus X-efficiency", American Economic
Review.
Li, M. (1977), "A logit model of homeownership", Econometrica 45:1081-1097.
Lim, G.-C. (1987), "Land markets and public policy: a Korean case study", Habitat International
11:73-81.
Lim, G.-C. and M.-H. Lee (1993), "Housing consumption in urban China", Journal of Real
Estate Finance and
Economics 6:89-102.
Lim, G.-C., J.R. Follain and B. Renaud (1980), "The determinants of homeownership in
a developing
economy", Urban Studies 17.
Lim, G.-C., J.R. Follain, Jr and B. Renaud (1984), "Economics of residential crowding
in developing
countries", Journal of Urban Economics: 16:173-186.
Lin, C-C.S. (1993), "The relationship between rents and prices of owner-occupied housing
in Taiwan", Journal
of Real Estate Finance and Economics 6:25-54.
Linneman, P. (1981),'"The demand for residence site characteristics", Journal of Urban Economics:
129-148.
Linneman, P.D. and I.E. Megbolugbe (1994), "Privatisation and housing policy", Urban Studies
31:635-651.
Lira, R. (1989), Viviendas Subsidiadas en Paises en Desarollo: Eficiencia y Movilacion de Recursos
(Institudo
de Estudios Urbanos, Universidad Pontifica Catolica de Chile).
Lluch, C., A. Powell and R. Williams (1977), Patterns in Household Demand and Savings (Oxford
University
Press, Oxford).
Lodhi, A. and H.A. Pasha (1991), "Housing demand in developing countries: a case-study
of Karachi in
Pakistan", Urban Studies 28:623-634.
Mabogunje, A.L. (1992), Perspective on Urban Land and Urban Management Policies in sub-Saharan
Africa.
World Bank Technical Paper No. 196.
Mackay, C.J. (1996), "The development of housing policy in South Africa in the post
apartheid period",
Housing Studies 11:133-146.
Maclennan, D. (1989), Privatization, Deregulation and Competition in the British Housing System,
1979 to
1989. Mimeo.
Maddison, A. (1995), Explaining the Economic Performance of Nations (Edward
Elgar).
Malpezzi, S. (1984), Analyzing an Urban Housing Survey. World Bank: Infrastructure and
Urban Development
Department Discussion Paper No. UDD-52.
Malpezzi, S. (1986), Rent Control and Housing Market Equilibrium: Theory and Evidence
from Cairo, Egypt.
Unpublished Ph.D. Dissertation (The George Washington University).
Malpezzi, S. (1988), Analyzing Incentives in Housing Programs: Evaluating Costs and Benefits
with a Present
Value Model. World Bank: INU Discussion Paper No. 23.
Malpezzi, S. (1990), "Urban housing and financial markets: some international comparisons",
Urban Studies
27:971-1022.
Malpezzi, S. (1991), "Discounted cash flow analysis: present value analysis of housing programmes
and
policies", in: A.G. Tipple and K.G. Willis, eds., Housing the Poor in the Developing
World: Methods of
Analysis, Case Studies and Policies, pp. 208-233 (Routledge Publishing).
Malpezzi, S. (1992), Rental Housing in Developing Countries: Issues and Constraints. United
Nations Centre
for Human Settlements, Rental Housing: Proceedings of an Expert Group Meeting (Nairobi).
Malpezzi, S. (1993), "Can New York and Los Angeles learn from Kumasi and Bangalore?
A comparison of
costs and benefits of rent controls", Housing Policy Debate 4:589-626.
Malpezzi, S. (1994), "'Getting the incentives right': a reply to Robert-Jan Baken and Jan Van
Der Linden",
Third World Planning Review 16:451-466.
Malpezzi, S. (in press), "Welfare analysis of rent control with side payments: a natural experiment
in Cairo,
Egypt", Regional Science and Urban Economics.
Malpezzi, S. and C.P. Rydell (1986), Rent Controls in Developing Countries: A Framework for
Analysis. World
Bank, Water Supply and Urban Development Department Discussion Paper No. 102.
Malpezzi, S. and D. Maclennan (1996), The Long Run Price Elasticity of Supply of New
Construction in
the United States and the United Kingdom. Paper presented to the European Network for Housing
Re-
1856 S. Malpezzi
search/American Real Estate and Urban Econonmics Association Conference, Glasgow, September (1994).
Revised, University of Wisconsin, Center for Urban Land Economics Research Working Paper.
Malpezzi, S. and G. Ball (1991), Rent Control in Developing Countries. World Bank Discussion Paper No.
129.
Malpezzi. S. and G. Ball (1993), "Measuring the urban policy environment: an exploratory analysis using rent
controls", Habitat Intemational 17:39-52.
Malpezzi, S. and J. Sa-Aadu (1996), "What have African housing policies wrought'?', Real Estate Economics.
Malpezzi, S. and K. Vandell (1993), Housing Markets and Policy in Market Economies: Lessons for Formerly
Socialist Economies. Paper prepared for the American Real Estate and Urban Economics Association,
Mystic, CN, October.
Malpezzi, S. and L. Ozanne (1980), Evaluation of FMRs in the Section 8 New Construction Program: An
Application of Hedonic Price Indexes (Urban Institute).
Malpezzi, S. and S.K. Mayo (1987a), "The demand for housing in developing countries", Economic
Development and Cultural Change 35:687-721.
Malpezzi, S. and S.K. Mayo (1987b), "User cost and housing tenure in developing countries". Journal of
Development Economics 25:197-220.
Malpezzi, S. and S.K. Mayo (1994), A Model Design for a Developing Country Housing Market Study.
University of Wisconsin-Madison, Center for Urban Land Economics Research Working Paper.
Malpezzi, S. and S.K. Mayo (1997a), "Getting housing incentives right: a case study of the effects of regulation,
taxes and subsidies on housing supply in Malaysia". Land Economics 73:372-391.
Malpezzi, S. and S.K. Mayo (1997b), "Housing and urban development indicators: a good idea whose time has
returned", Real Estate Economics 25:1-11.
Malpezzi, S. and V.Tewari (1991), Costs and Benefits of Rent Regulation in Bangalore, India. World Bank,
Infrastructure and Urban Development Department Discussion Paper No. 82 (160 pp).
Malpezzi, S., M. Bamberger and S.K. Mayo (1982), Planning an Urban Housing Survey. World Bank:
Infrastructure and Urban Development Department Discussion Paper No. UDD-42.
Malpezzi, S. and S.K. Mayo and D.J. Gross (1985), Housing Demand in Developing Countries. World Bank
Staff Working Paper No. 733.
Malpezzi, S., S.K. Mayo, R. Silveira and C. Quintos (1988), Measuring the Costs and Benefits of Rent Control:
Case Study Design. INU Discussion Paper No. 24.
Malpezzi, S., G. Tipple and K. Willis (1990), Costs and Benefits of Rent Control: A Case Study in Kumasi,
Ghana. World Bank Discussion Paper No. 74.
Mathey, K. (1990). "An appraisal of Sandinista housing policies", Latin American Perspectives 17:76-99.
Mayo, S.K. (1981), "Theory and estimation in the economics of housing demand", Journal of Urban Economics
10:95-116.
Mayo, S.K. (1986), "Sources of inefficiency in subsidized housing programs: a comparison of U.S. and German
experience", Journal of Urban Economics 20:229-249.
Mayo, S.K. (1990), Housing Finance Development: Experiences in Malaysia and Thailand, and Implications
for Indonesia (World Bank).
Mayo, S.K. (1993), Housing Policy Reform in South Africa: Intemational Perspectives and Domestic
Imperatives (World Bank).
Mayo, S.K. (1993), South African Housing Sector Performance in International Perspective. Paper presented
to the 21st IAHS Housing Congress, Capetown.
Mayo, S.K. and D.J. Gross (1987), "Sites and services-and subsidies: the economics of low cost housing in
developing countries", World Bank Economic Review 1:301-335.
Mayo, S.K. and J. Stein (1988), Housing and Labor Market Distortions in Poland: Linkages and Policy
Implications. INU Discussion Paper No. 25.
Mayo, S.K. and J. Stein (1995), "Housing and labor market distortions in Poland: linkages and policy
implications", Joumal of Housing Economics 4:153-182.
Mayo, S.K. et al. (1982), Informal Housing in Egypt (Abt Associates).
Mayo, S.K., S. Malpezzi and D.J. Gross (1986), "Shelter strategies for the urban poor in developing countries".
World Bank Research Observer 1:183-203.
Ch. 44: Economic Ainalysis of Housing Markers in Developing and Transition Economies 1857
Mayo, S.K., S. Angel, M. Heller and W. Stephens (1992), The Housing Indicators Program: Extensive Survey
Preliminary Results (Urban Development Division, World Bank).
McLure. C. (1977), "The "new view" of the property tax: a caveat", National
Tax
McMillen, D.P. and J.M. Pogodzinski (1993), Land Use in Planned Economies: Journal.
A Survey of Eastern Europe.
Paper prepared for the American Real Estate and Urban Economics Association
International Conference,
Mystic, CT.
McPhail, A.A. (1993), "The "five percent rule' for improved water service: can
households afford more?",
World Development 21:963-973.
Megbolugbe, IF. (1983), 'The hopes and failures of public housing in Nigeria:
a case study of Kulende and
Adewole housing estates, Ilorin", Third World Planning Review 5:349-369.
Mehta, M. and D. Mehta (1989a), A Report on Housing Finance Systems
in Metropolitan Areas of India
(USAID, New Delhi).
Mehta, M. and D. Mehta (1989b), Metropolitan Housing Market: A Study of
Ahmedabad (Sage Publications,
New Delhi).
Merrett, S. (1984), "The Assessment of housing consumption requirements in
developing countries", Third
World Planning Review 6.
Mills, E.S. (1995), Housing: The Step-Child of the Privatization and Deregulation
Movement. Paper prepared
for the UNU/WIDER Conference on Human Settlements in the Changing Global
Political and Economic
Process.
Mills, G., D. Sherwell and M. van Rooyen (1995), "Integrated housing policy
in South Africa: cost constraints
on built fonn", Environment and Planning B: Planning and Design
22:5-20.
Mitra, B.C. (1987), Land Supply for Low Income Housing in Delhi (Institute
for Housing Studies, Working
Paper, Rotterdam).
Mohan, R. (1989), Housing and Urban Development in India: The Need for
Innovative Strategies. Mimeo
(New Delhi).
Mohan, R. (1992), "Housing and urban development policy issues for the 1990s.
1", Economic and Political
Weekly 27:1913-1920.
Mohan, R. (1994), Understanding the Developing Country Metropolis: Lessons
from the City Study of Bogota
and Cali, Colombia (Oxford University Press, Oxford).
Mok, H.M.K., P.P.K. Chan and Y.-S. Cho (1995), "A Hedonic price model
for private properties in Hong
Kong', Journal of Real Estate Finance and Economics 10:37-48.
Morande, F.G. (1992), "The dynamics of real asset prices, the real exchange rate
trade reforms and foreign
capital inflows: Chile, 1976-1989", Journal of Development Economics 39:111-139.
Moser, C.O.N. and L. Peake, eds., (1987), Women, Human Settlements and Housing
(Tavistock, London).
Moynihan. C.J. (1987), Introduction to the Law of Real Property. Second
Edition (West Publishing Co., St.
Paul).
Mozolin, M. (1994), "The geography of housing values in the transformation to
a market economy--a case
study of Moscow", Urban Geography 15:107-127.
Munjee, N.M. (1992), "Recent developments in Indian housing finance", Housing
Finance International.
Murray, M.P. (1983). "Subsidized and unsubsidized housing starts: 1961-1977", Review
of Economics and
Statistics 65:590-597.
Murray, M.P. (1993), Subsidized and Unsubsidized Housing Stocks 1935-1987: Crowding
Out and Cointegra-
tion (Department of Economics, Bates College).
Muth, R.F. (1971), "The derived demand for urban residential land", Urban
Studies 8:243-254.
Nayani, P. (1987), Process of Illegal Subdivision of Land and Housing Consolidation.
New Delhi School of
Archictecture, Working Paper.
Ndulo, M. (1986), "Patterns of housing demand in a low income economy: the
case of urban Zambia", Eastern
Africa Economic Review 2.
Nelson, K.P. (1992), "Housing assistance needs and the housing stock". Journal of the American Planning
Association 58:85-102.
Nicholson, M. and K.G. Willis (1990),"The scale and distribution of financial subsidies to owner-occupied
households: a comparison", Journal of Economic and Social Measurement
16:71-85.
1858 S. Malpezzi
Okpala, D.C.I. (1985), Urban Low Income Shelter in Africa: A Review of Provision and Delivery Policies,
Programmes and Strategies (United Nations Center for Human Settlements, Addis Ababa).
Okpala, D.C.I. (1986), Investments in Human Settlements Development in the National Development Plans of
African Economies. Addis Ababa: UNCHS (Habitat) Discussion Paper.
Okpala, D.C.I. (1987), "Received concepts and theories in African urbanization", Urban Studies 24:137-150.
Okpala, D.C.I. (1994), "Financing housing in developing countries-A review of the pitfalls and potentials in
the development of formal housing finance systems", Urban Studies 31:1571-1786.
Okun, D.A. (1982), "Financing water supply systems", in: E.J. Shiller and R.L. Droste, eds., Water Supply and
Sanitation in Developing Countries (Ann Arbor Science).
Olsen, E.O. (1969), "A competitive theory of the housing market", American Economic Review: 612-622.
Olsen, E.O. (1987), "The demand and supply of housing services: a critical review of the empirical literature",
in: E.S. Mills, ed., Handbook of Regional and Urban Economics, vol. 2 (Elsevier, Amsterdam).
Olsen, E.O. (1988), "What do economists know about rent control?", The Journal of Real Estate Finance and
Economics 1:295-308.
Olsen, E.O. and D.M. Barton (1983), "The benefits and costs of public housing in New York City", Journal of
Public Economics 20:299-332.
Ondiege, P. (1986), Implicit Benefits in the Rental Public Housing Sector in Nairobi. Paper presented to the
International Conference of Urban Shelter in Developing Countries, London, September 1-4.
Page, D. and R.J. Struyk (1990), Measuring the Housing Sector: Results from the International Housing Market
Survey (The Urban Institute).
Parry, David L., Maya T. Koleva and Eugeni M. Popov (1990), "Bulgarian housing markets and mortgage
financing", Housing Finance International, May 1990, pp. 27-30.
Pasha, H.A. (1992), "Maximum lot size zoning in developing countries", Urban Studies 29:1173-1181.
Pasha, H.A. and M.S. Butts (1996), "Demand for housing attributes in developing countries: a case study of
Pakistan", Urban Studies 33:1141-1154.
Pasha, H.A. and A. Ghaus (1988), "The demand for housing in Pakistan", Pakistan Journal of Applied
Economics 7:83-99.
Payne, G. (1980), "The Gecekondus of Ankara", Process of Architecture 15:78-115.
Payne, G., ed. (1984), Low Income Housing in the Developing World (Wiley).
Payne, G. (1989), Informal Housing and Land Subdivisions in Third World Cities: A Review of the Literature
(Centre for Development and Environmental Planning, Oxford Polytechnic).
Peattie, L. (1987), "Affordability", Habitat International 11:69-76.
Pejovich, S. (1990), The Economics of Property Rights: Towards a Theory of Comparative Systems (Kluwer
Academic Press, Dordrecht).
Peng, R. and W.C. Wheaton (1994), "Effects of restrictive land supply on housing in Hong Kong: an ecometric
analysis", Journal of Housing Research 5:263-291.
Piggott, J. (1984), "The value of tenant benefits from UK council housing subsidies", Economic Journal
94:384-389.
Pinto-Lima, L. (1990), "The financing of housing in Brazil", Housing Finance International 5:25-26.
Pogodzinski, J.M. (1991), "Reform of the Polish housing market: issues and methods", Mimeo.
Pogodzinski, J.M. (1993), The Effect of Housing Market Disequilibrium on the Supply of Labor: Evidence
From Poland, 1989-1990.
Pogodzinski, J.M. (1994), "The value of 'full' ownership rights: a Hedonic study of Warsaw", Mimeo.
Polinsky, A.M. (1977), "The demand for housing: a study in specification and grouping", Econometrica
42:1533-1551.
Pollak, R. and T.Wales (1981), "Demographic variables in demand analysis", Econometrica 49:1533-1551.
Pornchokchai, S. (1992), Bangkok Slums: Review and Recommendations (Agency for Real Estate Affairs,
Bangkok).
Pudney, S. and L. Wang (1995), "Housing reform in urban China-efficiency, distribution and the implications
for social security", Economica 62:141-160.
Quigley, J.M. (1982), "Nonlinear budget constraints and consumer demand: an application to public programs
for residential housing", Journal of Urban Economics 12:177-201.
Rakodi, C. (1987), "Upgrading in Chawama, Lusaka: displacement or differentiation?", Urban Studies 25.
Ch. 44: Economic Analysis of Housing Markets in Developing and Transition Economies 1859
Rakodi, C. (1992), "Housing Markets in third world cities: research and policy in the 1990s", World
Development 20:39-55.
Rakodi, C. (1995a), "From a settler history to an African present-housing markets in Harare, Zimbabwe",
Enviroment and Planning 13:91-116.
Rakodi, C. (1995b), "Housing finance for lower income urban households in Zimbabwe", Housing Studies
10:199-227.
Rakodi, C. and P. Withers (1995), Home ownership and commodification of housing in Zimbabwe",
International Journal of Urban and Regional Research 19:250-271.
Ravillion, M. (1989), "The welfare cost of housing standards: theory with application to Jakarta", Journal of
Urban Economics 26:197-211.
Renaud, B. (1980), "Resource allocation to housing investment: comments and further results", Economic
Development and Cultural Change 28:189-199.
Renaud, B. (1984), Housing and Financial Institutions in Developing Countries. World Bank Staff Working
Paper No. 658.
Renaud, B. (1988a), Compounding Financial Repression with Rigid Urban Regulations: Lessons of the Korean
Housing Market. World Bank, Infrastructure and Urban Development Department Discussion Paper No. 21.
Renaud, B. (1988b), Housing Under Economic Structural Adjustment in Chile. INURD Working Paper No.
88-3.
Renaud, B. (1988c), "The future of the housing finance system in Korea", Mimeograph.
Renaud, B. (1989a), Affordable Housing, Housing Sector Performance and Behavior of the Price-to-Income
Ratio: International Evidence and Theoretical Analysis. Paper Presented to the Center of Urban Studies and
Urban Planning, University of Hong Kong.
Renaud, B. (1989b), Understanding the Collateral Qualities of Housing for Financial Development: The
Korean "Chonsei" as Effective Response to Financial Sector Shortcomings. INU Discussion Paper No. 49.
Renaud, B. (1990), "Housing affordability and housing finance: an international perspective", in: Hyundai
Research Institute, Proceedings of (the) International Housing Conference on Korean Housing Policies,
Seoul.
Renaud, B. (1991), Housing Reform in Socialist Economies. World Bank Discussion Paper No. 125.
Renaud, B. (1992), The Housing System of the Former Soviet Union: Why Do the Soviets Need Housing
Markets? Paper presented to the Fannie Mae University Colloquium Series, UCLA.
Renaud, B. (1993), "Confronting a distorted housing market: can Korean policies break with the past?", in: L.B.
Krause and F-K. Park, eds., Social Issues in Korea: Korean and American Perspectives (Korea Development
Institute, Seoul).
Renaud, B. (1995a), "The real estate economy and the design of Russian housing reforms, Part I", Urban
Studies 32:1247-1264.
Renaud, B. (1995b), "The real estate economy and the design of Russian housing reforms, Part II", Urban
Studies 32:1437-1451.
Renaud, B. (1997), Financial Markets and the Financing of Social Housing: The View from Developing Coun-
tries. Paper presented to the Conference on Social Housing Finance in the European Union, Nunspeet, The
Netherlands.
Renaud, B. and A. Bertaud (1989), The Urban Housing Reforms in China: Review of City Experiments and
Analytical Issues. Paper presented at the Joint Seminar on Urban Housing Reforms, Beijing.
Renaud, B. and R.M. Buckley (1987), Housing Finance in Developing Countries: A Framework for Bank
Operations (World Bank).
Renaud, B. and R.M. Buckley (1988), Urban Finance in Post-Depression Latin America. INU Discussion Paper
No. 22.
Renaud, B. and M. Gronicki (1990), Bulgaria: Aide Memoire, World Bank Housing Reconnaissance Mission.
Renaud, B. et al. (1993), Russia-Housing Reform and Privatization: Strategy and Transition Issues. World
Bank Sector Report 11868-RU.
Renaud, B., F. Pretorius and B. Pasadilla (1997), Markets at Work: Dynamics of the Residential Real Estate
Market in Hong Kong (Hong Kong University Press, Hong Kong).
Rodwin, L., ed. (1987), Shelter, Settlement and Development (Allen and Unwin, London).
1860 S. Malpezzi
Rojas, E. and M. Greene (1995), "Reaching the poor: lessons from the Chilean housing experience",
Environment and Urbanization 7:31-50.
Rosen, H.S. (1979), "Housing decisions and the U.S. income tax", Journal of Public Economics 1-23.
Rosen, S. (1974), "Hedonic prices and implicit markets", Journal of Political Economy 177-201.
Salazar-Carrilo, J. and I. Tirado De Alonso (1989), "Housing services in national accounts: the case of Latin
America", International Journal for Housing Science and its Applications 13:297-300.
Sandilands, R.J. (1980), Monetary Correction and Housing Finance in Colombia, Brazil and Chile (Gover
Publishing, Famborough, UK).
Sanyal, B. (1981), "Who gets what, where, why and how: a critical look at housing subsidies in Zambia",
Development and Change 12:409-440.
Sanyal, B. (1987), "Problems of cost-recovery in development projects: experience of the Lusaka squatter
upgrading and site/service project", Urban Studies 24:285-295.
Schlyter, A. (1985), "Housing strategies: the case of Zimbabwe", Trialog 6:20-29.
Shefer, D. (1990), "The demand for housing and permanent income in Indonesia", Urban Studies 27:259-272.
Shephard, S. (Chapter 41, this volume).
Shidlo, G., ed. (1990), Housing Policy in Developing Countries (Routledge).
Shidlo, G. (1994), Housing Regulatory Reform in Mexico. World Bank Infrastructure Note HS-8.
Sillince, J.A., ed., Housing Policies in Eastern Europe and the Soviet Union (Routledge, London).
Silveira, R. (1989), The Evolution of Rent Control in Brazil, 1921-1989. INURD Working Paper No. 89-8.
Silveira, R. and S. Malpezzi (1991), Costs and Benefits of Rent Control in Rio De Jainero. World Bank,
Infrastructure and Urban Development Department Discussion Paper No. 83, (97 pp).
Simkins, C., M. Oelofse and D. Gardner (1992), "Designing a market-driven state-aided housing programme
for South Africa", Urban Forum 3:13-38.
Small, K.A. and J.A. Gomez-Ibanez (Chapter 46, this volume).
Smith, L.B., K. Rosen and G. Fallis (1988), "Recent developments in economic models of housing markets".
Journal of Economic Literature 26:29-64.
Solomon, A. and K. Vandell (1982), "Alternative perspectives on neighborhood decline", Journal of the
American Planning Association 45:81-91.
Son, J.Y. and K.H. Kim (1998), "Analysis of urban land shortages: the case of Korean cities", Journal of Urban
Economics 43:362-384.
Stedman, S.J. (1992), Botswana: The Political Economy of Democratic Development (L. Rienner Publishers,
Boulder).
Strassman, W.P. (1977), "Housing priorities in developing countries: a planning model", Land Economics 53.
Strassman, W.P. (1980a), Guidelines for Estimating Employment Generation through Shelter Sector Assistance
(Michigan State University).
Strassman, W.P. (1980b), "Housing improvement in an opportune setting: Cartagena, Colombia", Land
Economics.
Strassman, W.P. (1980c), "Shelter improvement in Lima, Peru", Mimeo.
Strassman, W.P. (1982), The Transformation of Urban Housing: The Experience of Upgrading in Cartegena
(Johns Hopkins University Press).
Strassman, W.P. (1984), "The timing of urban infrastructure and housing improvements by owner occupants",
World Development 12:743-754.
Strassman, W.P. (1985), "Employment in construction: multicountry estimates of costs and substitution
elasticities for small dwellings", Economic Development and Cultural Change 33:395-414.
Strassman, W.P. (1991), "Housing market interventions and mobility: an international comparison", Urban
Studies 28:759-771.
Strassman, W.P., A. Blunt and R. Thomas (1992), "Land prices and housing in Manila", Urban Studies 31:267-
285.
Straszheim, M. (1975), An Econometric Analysis of the Urban Housing Market (NBER).
Struyk, R.J. (1982), "Upgrading existing dwellings: an element in the housing strategies of developing
countries", Journal of Developing Areas 17:69-76.
Struyk, R.J. (1987), Assessing Housing Needs and Policy Alternatives in Developing Countries (Urban
Institute Press, Washington, DC).
Ch. 44: Economic Analysis of Housing Markets in Developing and TransitionEconomies 1861
Struyk, R.J. (1988a), The Distribution of tenant benefits from rent control in urban Jordan", Land Economics
64.
Struyk, R.J. (1988b), "Understanding high vacancy rates in a developing country: Jordan", Journal of
Developing Areas 22:373-380.
Struyk, R.J. (1993), Implementing Housing Allowances in Russia: Rationalizing the Rental Sector (Urban
Institute Press, Washington, DC).
Struyk, R.J., ed. (1995), Economic Restructuring in the Former Soviet Bloc: Evidence from the Housing Sector
(Urban Institute Press, Washington, DC).
Struyk, R.J., ed. (1996), Economic Restructuring in the Former Soviet Bloc: Evidence from the Housing Sector
(Urban Institute Press, Washington, DC).
Struyk, R.J. and N.B. Kosareva (1993a), "Developing the Russian housing finance system", Housing Finance
International 7:2-10.
Struyk, R.J. and N.B. Kosareva (1993b), The Russian Housing Market in Transition. Prepared by the Urban
Institute for U.S. Agency for International Development.
Struyk, R.J. and N.B. Kosareva (1994), Transition in the Russian Housing Sector: 1991-1994 (The Urban
Institute Press, Washington, DC).
Struyk, R.J. and R. Lynn (1983), "Determinants of housing investment in slum areas: Tondo and other locations
in metropolitan Manila", Land Economics 59:444-454.
Struyk, R.J. and M. Turner (1986), Housing Finance and Quality in Two Developing Countries (Urban Institute
Press, Washington, DC).
Struyk, R.J., M. Hoffman and H. Katsura (1990), The Market for Shelter in Indonesian Cities (Urban Institute
Press, Washington, DC).
Struyk, R.J., K. Mark and J.B. Telgarsky (1991), "Private management for eastern Europe's state rental
housing", Journal of Housing Economics 1.
Struyk, R.J., J. Daniell, A. Puzanov and N.B. Kosareva (1993a), Income Adjustments When Raising Rents
in Post-Soviet Economies: Housing Allowances or Wage Increases? Analysis for Moscow (Urban Institute
Press, Washington, DC).
Struyk, R.J., N.D. Kosareva, Daniell, Hanson and Mikelsons (1993b), Implementing Housing Allowances in
Russia. Rationalizing the Rental Sector (University Press of America (UPA), Lanham).
Struyk, R.J., et al. (1994), Implementing Housing Allowances in Russia: Rationalising the Rental Sector
(Urban Institute Press, Washington DC).
Telgarsky, J.P. and R.J. Struyk (1991), Toward a Market-Oriented Housing Sector in eastern Europe: Devel-
opments in Bulgaria, Czechoslovakia, Hungary, Poland, Romania, and Yugoslavia (Urban Institute Press,
Washington, DC).
Tewari, V.K. and T. Krishna Kumar (1987), "Rent control in India: its economic effects and implementation in
Bangalore", Nagarlok 19:55-85.
Thalwitz, M. (1993), "Poland: building a new system of housing finance", Housing Finance International,
December, pp. 13-42.
Thomson, M.J. (1983), Toward Better Urban Transport Planning in Developing Countries. World Bank Staff
Working Paper No. 600.
Tideman, N. et al. (1990), Open letter to Mikhail Gorbachev on land markets in Russia.
Tipple, A.G. (1983), "Housing policy and culture in Kumasi", African Urban Studies 15:17-30.
Tipple, A.G. (1993), "Housing in Algeria-recent research at CAROD-a review article", Third World
Planning Review 15:55-62.
Tipple, A.G. (1994), "The need for new urban housing in sub-Saharan Africa-problem or opportunity",
African Affairs 93:587-610.
Tipple, A.G. and K.G. Willis (1991a), "Tenure choice in Kumasi, Ghana", Third World Planning Review
13:27-45.
Tipple, A.G. and K.G. Willis, eds. (1991b), Housing the Poor in the Developing World: Methods of Analysis,
Case Studies and Policy (Routledge, London).
Tipple, A.G. and K.G. Willis (1992), "Who owns, who rents: tenure choice in a west African city", in: L.
Kilmartin and H. Singh, eds., Housing in the Third World: Analyses and Solutions (Concept Publishing,
New Delhi).
1862 S. Malpezzi
Tolley, G.S. (1991), Urban Housing Refonn in China: An Economic Analysis. World Bank Discussion Paper
No. 123.
Tolley, G.S., ed. (In press), Urban Land and Housing Reform in Socialist and Formerly Socialist Countries.
Tomann, H. (1992), Towards a Housing Market in Eastern Germany. Paper presented to the American Real
Estate and Urban Economics Association, Los Angeles.
Tomlinson, R. (1992), "Competing urban agendas in South Africa", Urban Forum 3:97-110.
Trollegaard, S. (1989), Distributional Aspects of Housing and Taxation Policies. United Nations, Economic
Commission for Europe, Committee on Housing, Building and Planning, monograph.
Turner, B.. J. Hegedus and I. Tosics, eds. (1992), The Reform of Housing in Eastern Europe and the Soviet
Union (Routledge, London, New York).
Turner, J.F.C. (1972), Freedom to Build (Macmillan, New York).
Turok, I. (1994a), "Urban planning in the transition from apartheid-Part I: the legacy of social control", Third
World Planning Review 65:243-259.
Turok, I. (1994b), "Urban planning in the transition from apartheid-Part II: towards reconstruction", Third
World Planning Review 65:355-374.
United Nations (1993), Housing Policy Guidelines: The Experience of ECE with Special Reference to
Countries in Transition (E.93.II.E.15).
U.N. Centre for Human Settlements (Habitat) (1989), Improving Income and Housing: Employment Genera-
tion in Low-Income Settlements (Nairobi).
UNCHS (Habitat) (1990), The Global Strategy for Shelter to the Year 2000 (Nairobi, Kenya).
UNCHS (Habitat) (1996), An Urbanizing World: Global Report on Human Settlements, 1996. New York:
Oxford University Press.
UN Centre for Human Settlements, and the World Bank (1993), The Housing Indicators Program-Volume I:
Report of the Executive Director (Nairobi).
UN Centre for Human Settlements, and the World Bank (1993), The Housing Indicators Program-Volume I:
Indicator Tables (Nairobi).
UN Centre for Human Settlements, and the World Bank (1993), The Housing Indicators Program--Volume
III: Preliminary Findings (Nairobi).
UN Centre for Human Settlements, and the World Bank (1993), The Housing Indicators Program-Volume
IV: The Extensive Survey Instrument (Nairobi).
Urban Edge, The (1983), Third World Faces Great Subway Debate. World Bank 7.
U.S. Agency for International Development (1981), Housing and Health: An Analysis for Use in the Planning,
Design, and Evaluation of Low-Income Housing Programs (U.S. A.I.D).
Valladares, L. do P. (1978), "Working the system: squatter response to resettlement in Rio de Janeiro",
International Journal of Urban and Regional Research 2:12-25.
van der Linden, J. (1986), The Sites and Services Approach Reviewed (Gower, Aldershot).
van Vliet, W. (1990), International Handbook of Housing Policies and Practices (Greenwood Press, Westport,
CT).
van Vliet, W. and J. van Weesep, eds. (1990), Government and Housing: Developments in Seven Countries
(Sage Publications, Urban Affairs Annual Reviews, Vol. 36).
Vandell, K.D. and T.J. Riddiough (1991), On the Use of Auctions as a Disposition Strategy for RTC Real
Estate Assets: A Policy Perspective. University of Wisconsin, Department of Real Estate and Urban Land
Economics Working Paper.
Vanderporten, B. (1992), "Strategic behavior in pooled condominium auctions", Journal of Urban Economics
31:123-137.
Vernez, G. (1973), Bogota's Pirate Settlements. University of California at Berkeley, Working Paper.
Versacaj, J.P. (1993), "Leasing real estate in the Russian Federation", Real Estate Review: 70-75.
Wadhva, K. (1983), "An evaluation of the urban land ceiling legislation: a case study of Ahmedabad", Nagarlok
15:76-86.
Walters, A.A. (1979), The benefits of minibuses: the case of Kuala Lumpur", Journal of Transport Economics
and Policy.
Wang, Y. (1991), "The size of housing subsidies in China", Review of Urban and Regional Development
Studies 3.
Ch. 44: Economic Analysis of Housing Markets in Developing and TransitionEconomies 1863
Ward, P. (1989), "Land values and valorisation processes in Latin American cities: a research agenda", Bulletin
of Latin American Research 8:47-66.
Ward, P., E. Jimenez and G. Jones (1993), "Residential land price changes in Mexican cities and the
affordability of land for low-income groups", Urban Studies 30:1521-1542.
Warford, J. and R. Turvey (1974), Lahore Water Supply-Tariff Study. World Bank, Public Utilities Report
No. PUN 12.
Weinberg, D.H., J. Friedman and S.K. Mayo (1981), "Intraurban residential mobility: the role of transactions
costs, market imperfections and household disequilibrium", Journal of Urban Economics 9:332-348.
Wells, J. (1986), The Construction Industry in Developing Countries: Alternative Strategies for Development
(Croom Helm).
Wheaton, W.C. (1979), "Public policy and the 'shortage' of housing in Egypt", Mimeograph.
Wheaton, W.C. (1981), "Housing policies and urban 'markets' in developing countries: the Egyptian
experience", Journal of Urban Economics 9:242-256.
Whitehead, C. (Chapter 40, this volume).
Whittington, D., J. Briscoe, X. Mu and W. Barron (1990), "Estimating the willingness to pay for water services
in developing countries: a case study of the contingent valuation method in Haiti", Economic Development
and Cultural Change 38:293-311.
Whittington, D., D.T. Lauria, A.M. Wright, K. Choe, J.A. Hughes and V. Swarna (1992), Household De-
mand for Improved Sanitation Services: A Case Study of Kumasi, Ghana (UNDP-World Bank Water and
Sanitation Program).
Whittington, D., D.T. Lauria, K. Choe, J.A. Hughes, V. Swarna and A.M. Wright (1993), "Household sanita-
tion in Kumasi, Ghana: a description of current practices, attitudes and perceptions", World Development
21:733-748.
Williams, M. and M. Ondera (1982), "Public expenditures on housing in less developed countries", Journal of
Economic Development 7:127-133.
Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications (Free Press).
Willis, K.G. and A.G. Tipple (1991), "Economics of multihabitation: housing conditions, household occupancy
and household structure under rent control, inflation and nonmarketability of ownership rights", World
Development 19:1705-1720.
Willis, K., S. Malpezzi and G. Tipple (1990), "An econometric and cultural analysis of rent control in Kumasi,
Ghana", Urban Studies 27:241-257.
Wong, Y.-C. and P.-W. Liu (1988), "The distribution of benefits among public housing tenants in Hong Kong
and related policy issues", Journal of Urban Economics 23:1-20.
Woodfield, A. (1989), Housing and Economic Adjustment (Taylor and Francis for the United Nations, New
York).
World Bank (1981), The Road Maintenance Problem and International Assistance (Washington).
World Bank (1986), Urban Transport: A World Bank Policy Study.
World Bank (1990), Ghana Housing Sector Review. Report No. 8099-GH.
World Bank (1992), China: Implementation Options for Urban Housing Reform (World Bank, Washington,
DC).
World Bank (1993a), China: Urban Land Management in an Emerging Market Economy (World Bank,
Washington, DC).
World Bank (1993b), Housing: Enabling Markets to Work. World Bank Policy Paper.
World Bank (1993c), Russia Housing Reform and Privatization: Strategy and Transition Issues. (By B Renaud
et al.) Report 11868-RU, 2 Vols.
World Bank (1994), World Development Report 1994: Infrastructure for Development (Oxford University
Press, Oxford).
Yap, K.S., ed. (1992), Low Income Housing in Bangkok: A Review of Some Housing Submarkets. Bangkok,
Asian Institute of Technology, HDS Monograph No. 25.
YapiDiahou, A. (1995), "The informal housing sector in the metropolis of Abidjan, Ivory Coast", Environment
and Urbanization 7:11-30.
Yonder, A. (1987), "Informal land and housing market: the case of Istanbul", Journal of the American Planning
Association 53:213-219.
1864 s. Malpezzi
Yoon, I.-S. (1994), Housing in a Newly Industrialized Economy: The Case of South Korea (Brookfield, Vt.,
Aldershot, UK and Ashgate, Avebury, Sydney).
Yu, E-L. and S.-M. Li. (1985), "The welfare cost of Hong Kong's public housing program", Urban Studies
22:33-140.
Zaroff, B. and D.A. Okun (1984), "Water vending in developing countries", Aqua 5:289-295.
Zearley, T.L. (1993), "Creating an enabling environment for housing: recent reforms in Mexico", Housing
Policy Debate 4:239-249.
Zhang, J. (1986), Urban Construction Reconsidered: China's Cities. Issues and Solutions. World Bank, UDD
Discussion Paper No. 101.
Zorn, P. (1988), "An analysis of household mobility and tenure choice: an empirical study of Korea", Journal
of Urban Economics 24:113-128.
Chapter45
Contents
1. Introduction 1866
2. Conceptual and measurement issues 1869
2.1. What is poverty? 1869
2.2. Poverty line 1870
2.3. Unit of analysis 1873
2.4. Equivalence scales 1874
2.5. Measurement of poverty 1876
2.5.1. Cardinal measurement 1876
2.5.2. Ordinal measurement 1879
2.5.3. Caveats on international comparisons 1880
3. Nature, characteristics and correlates of poverty 1881
3.1. Rural poverty 1882
3.2. Urban poverty 1886
4. Gender and poverty 1890
4.1. Access to land and employment 1894
4.2. Access to capital and other related inputs 1896
5. Poverty and population growth 1898
6. Poverty and economic growth 1902
7. Strategies and policies 1910
7.1. Macroeconomic policies 1912
7.2. Trade and industrialization policies 1914
7.3. Land reform 1916
* We would like to gratefully acknowledge the helpful comments of Arsenio Balisacan, Jere Behrman,
Douglas Brooks, Malcolm Dowling, Frank Harrigan, S. Iftekhar Hossain, Emmanuel Jimenez, Haider A.
Khan, Michael Lipton, Edwin Mills, and S. Rashid. The authors acknowledge the valuable research assistance
given by Isabelita Alba and Emma Murray, editorial assistance by Cherry Zafaralla, and secretarial support by
Helen K. Buencamino, Zenaida M. Acacio and Anna Liza P. Silverio. The views expressed in the paper are
the authors' own and do not reflect the official position of the Asian Development Bank.
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P. Cheshire
© 1999 Elsevier Science B. V All rights reserved.
1865
1866 E.M. Pernia and M.G. Quibria
Abstract
This chapter begins with the dimensions of the poverty problem in developing
countries, and then deals with conceptual and measurement issues. It next dis-
cusses the nature, characteristics and correlates of rural and urban poverty. It
looks into the issue of gender and poverty, then turns to the relationship between
poverty and population growth. It also examines the crucial nexus between eco-
nomic growth and poverty. Before concluding, the chapter discusses strategies
and policies for poverty reduction.
1. Introduction
1 It is now widely recognized that there are serious shortcomings in perceiving poverty only in terms of
income or consumption of commodities. Poverty, as a reflection of human ill-being, has many dimensions.
Sen (1985) argues that human well-being extends beyond possession of commodities or generation of utilities;
it should be measured in terms of "functioning" and "capabilities". However, even if one confines oneself to
basic capabilities and human functioning, the task of operationalizing capabilities for measuring poverty is not
an easy one. For more on this issue, see Section 2.
1868 E.M. Perniaand M.G. Quibria
He added:
All the other dishonors are chivalry itself by comparison. Poverty blights
whole cities; spreads horrible pestilence; strikes dead the very souls of all
who come within sight, sound, or smell of it. What you call crime is nothing:
a murder here and a theft there, a blow now and a curse then. What do
they matter? They are only the accidents and the illnesses of life. There
are millions of poor people, abject people, dirty people, ill-fed, ill-clothed
people. They poison us totally and physically; they kill the happiness of
society; they force us to do away with our own liberties and to organize
unnatural cruelties for fear they should rise against us and drag us down
into their abyss. Only fools fear crime; we all fear poverty. (Shaw, 1913, as
reproduced in Smith, 1970: p. 65)
Poverty has been described as "a matter of deprivation" (Sen, 1981). This de-
privation can be absolute, infringing on basic sustenance of life; or it could be
relative, in relation to the standard of living enjoyed by a reference group in
the society with higher incomes. Two concepts of poverty, absolute and relative,
reflecting two types of deprivation, have been distinguished in the literature.
Absolute poverty is defined as the inability to attain a minimal standard of
living. 4 This notion of minimal standard is, however, not fixed but varies across
countries. What is considered an absolute necessity in one country may be con-
sidered a luxury in another. As a result, items of consumption which are deemed
as luxury in Bangladesh may be considered a necessity in the US. Relative
poverty, on the other hand, is defined as the inability to attain a given contem-
porary standard of living, which is often defined in terms of some indicator of
well-being such as the median income. In the US, it was proposed that families
with one half of the median income be classified as poor (Fuchs, 1965). Similarly,
the poor may be defined, as has been suggested for developed countries, as the
lowest four deciles in the national income distribution. There is a disconcert-
ing aspect about this notion of relative poverty, as it seems to predetermine the
extent of relative poverty. It has been pointed out, no matter how much income
increases, "in the absence of a marked shift toward greater equality in the income
distribution the poor, in a relative sense, will always be with us" (Sawhill, 1988:
p. 1076). In other words, it may be possible to eradicate absolute poverty, but it is,
3 This section draws on Quibria (1991).
4 Atkinson (1989) distinguishes between two different concepts of poverty. One is the standard of living
concept, where one is perceived as poor if one's total expenditure or the consumption of specific goods falls
below a given threshold. The other is the right to a minimum level of resources, where one is perceived as poor
if one's income falls below a specified minimum level. In the second approach, the right to a minimum income
is critical, and not its disposal. In this chapter, emphasis is on the first approach, which seems to have gained
greater acceptability.
1870 E.M. Pernli and M.G. Quibria
however, never possible to eradicate such a type of relative poverty. Despite this
nagging aspect about relative poverty, it should be emphasized that it is perfectly
legitimate 5 -and, as has been suggested, is consistent with the Rawlsian criterion
of justice-for government and policy makers to be concerned with the lower
segments of the population in the income scale (Rawls, 1971).
In developing countries today, the term poverty is used essentially to indicate
absolute poverty. In developed countries, with increasing material wealth and vir-
tual elimination of hunger, this concept has lost much of its significance. Poverty
in most advanced industrial nations is now defined in relative terms as a matter
of deviation from social and economic norms.
Finally, it appears that there is some measure of ambiguity about whether
poverty should be viewed from a welfarist, or a nonwelfarist perspective. In
other words, is it concerned with the level of resources of the poor or with
their economic well-being? Sen (1977) argues that the subject should adopt a
nonwelfarist approach and focus on the level of resources, not the family's sense
of well-being.6 This perspective, however, has been far from universally adopted.
The poverty line defines the minimum acceptable standard of living for the so-
ciety. When an individual is poor, he/she has a standard of living that is below
the minimum acceptable level. Thus, the definition of the poverty line is based
on two concepts: (i) the concept of a standard of living; and (ii) the concept of a
minimum acceptable level. The standard of living is a multidimensional concept,
which encompasses all the commodities an individual consumes and all the activ-
ities in which he or she participates. To define a poverty line that incorporates this
multidimensional aspect of standard of living, one ought to specify a minimum
acceptable level for all the different dimensions. 7 These dimensions relate to
various basic needs such as food, health, education, clothing, fuel and transport.
5 J. Rawls (1971), an American philosopher, presents a systematic moral theory in his book. A Theo) of
Justice, which is in opposition to utilitarianism. The principal hypothesis of the book is that a system is just
if and only if it accords with the principles which would be agreed on by rational people making a kind of
social contract. The book further argues for principles that emphasize basic liberties and improvement for the
worst-off individuals in the society.
6 In terms of notation, if y indicates income and u (y) the utility from income, then the subject is concerned
with and not with u(y). See Atkinson (1989, Chapter 1).
7 Although the basic needs approach to poverty is couched in specific commodity terms and adopts a
nonwelfarist approach, it can be easily translated into welfarist terms. Or in other words, there is behind the
apparent nonwelfarist approach an embedded welfarist perspective. To put it technically, assume x° specifies
the basket of "basic needs" commodities. Then the utility level associated with x0 is u(x°). By solving the
optimizing problem of the consumer, one can define the money metric utility function: e(p, x° ) = [min pr:
u(x) > u(x°)], where p is the level of prices.
Ch. 45: Poverty in Developing Countries 1871
If a household fails to achieve an adequate level for any one of these basic needs,
it is counted as poor. However, there are a number of practical difficulties in
implementing this multiple- index approach. To agree on an acceptable level for
one item of basic needs is not easy, but to agree on many different items is next
to impossible. Furthermore, the failure to satisfy an item of basic needs may
arise from individual preference than from stringency of income. A well-to-do
household may be considered poor for its idiosyncrasies of preference than for
its deprivation due to lack of income. All these difficulties have led to a greater
reliance on the single indicator of poverty.
There is a disadvantage in a single index of living standard, often summarized
in total household income or expenditure. Income, which includes both savings
and dissavings, is in some sense considered a better conceptual measure of living
standard. It is argued that it is possible for an individual to choose a level of living
below his or her income. But this is decided on volition rather than dictated by
a lack of income. In this sense, income should be the indicator of poverty, since
it is a measure of the opportunities open to a household and is not influenced
by the consumption decisions made. 8 However, there are a number of problems
with using income as a proxy of family welfare. First, income data are often
unreliable, since there are problems in identifying and assessing production for
home consumption and for sale in developing countries. Second, income may
understate the level of living. A family may dissave and its current standard
of living may exceed current income. Expenditure may then be the more ap-
propriate indicator of welfare. Third, the level of living may exceed income if
there is sharing. An elderly parent living with his children may benefit from their
expenditures. Fourth, income may, in some instances, overstate the standard of
living. This happens when money alone is not sufficient to buy the necessities, for
example, when there is rationing or unavailability of goods. For all these reasons,
poverty incidence is now calculated not on the basis of household income but
household expenditure-in many cases by consumption expenditure of privately
supplied goods.
There are two general approaches to arrive at a poverty line at the operational
level. One is to determine a bundle of basic goods, consisting of food and non-
food items, and then calculate the income required to purchase this bundle at
current prices. Denoting the food items by vector xf, purchasable at pf prices
and the nonfood items by the vector x nf , purchasable at prices pnf, the poverty
line is defined as: (pfx f + pnfxnf) where X(>1) is a provision for inefficient
expenditure or waste.
8 This notion is advocated by those who subscribe to the minimum rights approach, as noted earlier.
1872 E.M. Pernia and M.G. Quibria
There are three problems with this approach. First, the notion of a basic good
is a matter of value judgment. What will be considered a basic good by one
investigator will be deemed as nonbasic by another. Second, some basic goods are
provided from the public budget. The Indian norm on the poverty line excludes
these public goods, the implicit assumption being that these public goods are
provided adequately to all sections of the population. It is, however, questionable
whether this is correct, especially in the case of health, education, water supply
and sanitation. Third, prices of many commodities vary according to location,
year and the socioeconomic status of the consumer. The poor often pay more for
an inferior variety; the rich often pay less for a superior variety. These issues are
important if comparisons are made over time and space.
The other approach determines the poverty line more directly on the basis of
food requirements. It was used by Orshansky (1965) to develop the poverty line
for the US and is now widely used for calculating poverty in developing coun-
tries. First, the minimum expenditure necessary to satisfy the minimum nutri-
tional intake should be determined. However, the determination of the minimum
nutritional intake is a formidable task. Many economists in developing countries
working in this area apply the Food and Agriculture Organization/World Health
Organization guidelines on minimum "energy requirements". The minimum food
expenditure thus calculated is then multiplied by an appropriate proportion to
allow for nonfood requirements. In more precise terms, denoting the food items
by the vector xf , purchasable at pf prices, the minimum food expenditure is given
by pfx f . Then the poverty norm becomes: pfx f where 0(>1) is a multiplier
to incorporate nonfood items and to allow for inefficient expenditure. But the
important and still unresolved issue is, what is the appropriate multiplier? That
is, how does one determine 0? Orshansky (1965) used (the reciprocal of) the
average ratio of food expenditure to the total expenditure in the population.
There are at least two major problems with the nutrition-based food approach.
First, in the case of food requirements, while physiological needs may be a good
starting point, it is difficult to determine x f with any precision. Nutritional needs
depend on the location, climate as well as the occupation and age of the indi-
vidual. They vary from person to person, and even from one day to another for
the same person. Therefore, any nutritional statement can only be probabilistic;
i.e., at a certain consumption level, there is a certain probability that a particular
person is inadequately fed. Second, there is a disparity between expert recom-
mendation and actual consumption. Consider a hypothetical example where an
individual chooses to spend her income on things other than meeting her energy
requirements, even though she has the means to do so. It can be argued that such
individuals are not poor since they could have chosen not to be poor.
Ch. 45: Poverty in Developing Countries 1873
In poverty analysis, the unit of analysis should ideally be the individual as it is,
in the ultimate analysis, an enquiry into the issue of the welfare of an individual.
However, in practice, most empirical studies take the household as the basic unit
of analysis.12 This choice is dictated more by the availability of data than by any
other considerations. Income-expenditure data of an individual within a family
are very difficult to obtain. There may be members within the family with zero
recorded cash income but they may not be without resources. Indeed, records
9 The distinction between the standard of living approach and the minimum rights approach is critical here.
10 A welfarist approach is implicitly taken here.
11 That is, if one adopts the minimum rights approach, individual preference is irrelevant. However, if the
standard of living approach is adopted, individual preference can make the difference between rich and poor.
12 There may be many different definitions of a household. The definition which has become increasingly
popular is the so-called "kitchen" definition, i.e., people living at the same address with common housekeeping
and having a common meal.
1874 E.M. Pernia and M.G. Quibria
of income transfers among family members are difficult to obtain. Besides data,
another important consideration for taking the household as the unit of analysis
is that it can take into account items of expenditure which have "public good"
characteristics for family members. If some assets are indivisible or if there are
economies of scale in consumption, then the division of family resources by the
size of the household will underestimate the individual level of economic well-
being. However, it may be possible that, within the family, there are unequal
standards of living between members. There is some empirical evidence which
shows disparity in consumption between sexes within the household.13 In light
of this fact, there is some doubt on the use of household as the proper unit of
analysis for poverty measurement.1 4
The index numbers which relate the poverty lines for different household sizes
and types are referred to as equivalence scales.15 These are used to make welfare
comparisons between households with different size and demographic charac-
teristics. What cost of living indices are to welfare comparisons for a given
household facing different prices, equivalence scales are to the poverty line for
comparing poverty across households. While the cost of living indices are based
on the assumption of unchanging tastes, equivalence scales are based on the
premise of differing tastes between households due to variations in observable
characteristics.
Three main approaches have been adopted for the determination of the equiv-
alence scales for different types of households. One is the survey of individual
assessments of needs. As the method uses survey questionnaires to elicit house-
hold preferences, it introduces a strong subjective element in the construction
of equivalence scales. The other is empirical investigation of the expenditure
behavior of households. Here again different assumptions about the effect of
demographic variables on the cost function leads to different equivalence scales.
And still another is the nutritional and physiological studies. The estimated scales
are likely to vary considerably across time and space as nutritional needs depend
13 For a discussion of this evidence, see Section 4. If such gender disparities exist, then the conventional
measures based on the assumption of an equal distribution of resources between household members may lead
to distortions in the true nature and extent of poverty. See also Haddad and Kanbur (1990).
14 Atkinson (1991) states that there is a strong case for measuring poverty on an individual basis from the
minimum rights perspective. It is by posing the poverty problem on an individual basis that important gender
issues such as the feminization of poverty can be meaningfully addressed.
15 Operationally, in most cases an equivalence scale provides a set of conversion factors that are used to
convert a household of given size and demographic composition to its adult male equivalent.
Ch. 45: Poverty in Developing Countries 1875
Table I
The Amsterdam scale
16 As noted earlier, most equivalence scales assign adult females and children a value less than one in temls
of adult male equivalent, the reason being that children and females tend to consume less than adult males.
It is often assumed that this difference in valuation mirrors a difference in needs as women and children can
achieve the same level of well-being by consuming less than adult men.
1876 E.M. Perniaand M.G. Quibria
2.5.1. Cardinalmeasurement
All single-index poverty measures count an individual or household to be poor
if its income or expenditure falls below a given poverty line. To illustrate these
measures, let us assume that there are n households and let the income of unit i
be yi. If the incomes are arranged in ascending order and the poverty line is z,
then:
<
Y1 Y2 Y3' Yq < Z < Yq+l ... Yn
Denote by p the set of the poor and q the cardinality of the set p. That is, there are
q units below the poverty line. The most commonly used measure of poverty-
the headcount ratio- is the number of income-receiving units as a fraction of the
total population. If q is the number of the poor in a population of size n, then the
headcount ratio is given by H q/n. While it is useful as a summary measure
of poverty, it reveals nothing about the depth or severity of poverty. That is, how
poor are the poor? 18
A group of measures that can address this issue is the poverty gap measures.
The poverty gap of an individual i is given by g = (z - yi) and then the aggregate
poverty gap is given by
g = Y(Z - Yi).
iep
The corresponding poverty gap ratios are denoted by gi/z and g/z. The aggregate
poverty gap is often normalized by the number of the poor to obtain the average
poverty gap (g/q). Sen (1977) focuses on a slightly different normalization to
obtain the income-gap ratio:
G = (1/qz) (z - yi).
ip
17 For an empirical demonstration on how the different definitions of an "average" household leads to
different levels of potential error in poverty estimates, see Bhalla and Vashistha (1988).
18 The headcount measure has been subject to other criticisms as well. For example, it has been argued that
poverty is not really a discrete, either or condition; and as such one does not shed the affliction associated with
poverty by crossing an income line (see Watts, 1991). Whether one accepts this criticism depends on one's
notion about poverty.
Ch. 45: Poverty in Developing Countries 1877
Expressed as a percentage of the poverty line, the income gap ratio measures
the proportion by which the average income (or consumption) of the poor would
have to increase in order to eliminate poverty completely.
H and G are, in a sense, complementary. H captures the number of people in
poverty but not its depth, while G measures the depth of poverty but is insensitive
to the number involved. Neither measure is sensitive to redistribution of income
within poor units. That is, if a dollar of income is taken away from the poorest
unit and is given to a richer unit that is still below the poverty line, then G will
remain unchanged. In other words, G does not reflect the severity of poverty.
Sen (1976) has noted the desirable properties of a poverty measure, which
include the sensitivity of the measure to the number of poor, to the depth of
poverty, as well as to the distribution of income among the poor. He has formal-
ized the requirements in the following two axioms of monotonicity and transfer.
Monotonicity axiom implies that other things remaining the same, a reduction
in the income of any poor household will increase the poverty measure. Transfer
axiom implies that other things remaining the same, a transfer of income between
two poor households-from a poorer household to a richer one-will increase
the poverty measure. Simply put, any increase in inequality among the poor, due
to one or a series of regressive transfers, must be reflected in a higher level of
poverty.
It can be readily seen that H violates both these axioms. Since H measures
only the number of households under poverty, a fall in the income of any of
these households leaves the measure unchanged. It thus runs contrary to the
monotonicity axiom. Similarly, H is unaffected by a transfer of income from any
poor household to any nonpoor household, and this violates the transfer axiom.
It can be seen by inspection that the income gap measure satisfies the mono-
tonicity axiom but does not satisfy the transfer axiom. A fall in the income of any
household under the poverty line will increase the poverty measure. The mea-
sure is, however, unaltered by any mean-preserving transfer of income among
households below the poverty line. For example, a direct transfer from a poor
household to another less poor household will increase the poverty gap of the
first household by exactly as much as it will reduce that of the second household,
leaving the measure unchanged. In addition, G is insensitive to any change in
the number of poor households: if the population below the poverty line were
duplicated with the same characteristics, this would leave G unchanged.
Sen (1976) has proposed a distributionally sensitive index which combines
the properties of H and G in an ingenious way. This index is given by:
S = H[G + (1 - G)d],
1878 E.1M. Pernia and M.G. Quibria
where Cp stands for the coefficient of variation of income among the poor.20
The FGT class of measures is consistent with the monotonicity axiom for
a > 0 and with the transfer axiom for any a > 1. For a > 2, the FGT class
of measures also satisfies the transfer sensitivity axiom (Kakwani, 1980), which
requires that a (regressive) transfer from a poor to a less poor household will
result in an increase in the poverty measure, but the rate of increase will be
smaller, the higher the initial incomes of the two households. In addition to the
above desirable properties, the FGT class of measures has been found to be of
great advantage in empirical application.
19 The Gini-coefficient is an aggregate measure of inequality, which can vary from zero to one, the former
implying perfect equality and the latter perfect inequality.
20 Note that the squared coefficient of variation is defined as the variance (where each observation has equal
probability) divided by the square of the mean.
Ch. 45: Poverty in Developing Countries 1879
As mentioned earlier, the F(a) index has an attractive "decomposability"
property, which implies: If a population of n units is divided into m subgroups
with the number of units in each subgroup as ni, then the aggregate index F(a)
for the entire population is a weighted average of the aggregate indices of the
subgroups, the weights being their population shares:
F(a) = (njl/n)F(a).
2.5.2. Ordinalmeasurement21
There is a degree of arbitrariness in any cardinal measure of poverty. This arbi-
trariness relates to the choice of the poverty line and to the specific functional
form of the poverty measure. However, for many purposes, there is no need
for precision of a cardinal measure; ordinal ranking of income distributions in
terms of poverty will suffice. This ordinal ranking, which does not depend on
one particular line or one particular poverty measure, utilizes the technique of
stochastic dominance analysis. The following provides an intuitive explanation
of the idea behind stochastic dominance analysis.
Consider a range of poverty lines, the maximum being defined by the highest
permissible income level. Then poverty has unambiguously declined between
two dates (say, before and after the policy change) if the cumulative distribution
of income for the latter date lies nowhere above that for the former date over
the entire interval up to the maximum allowable poverty line. This is called
the first-order dominance condition. 2 2 If the result is unambiguous, then poverty
has declined irrespective of the poverty line and the poverty measures adopted.
21 For technical details, see Atkinson (1989: Chap. 2), Foster and Shorrocks (1988) and Ravallion (1993).
22 The dominance conditions have been discussed, among others, by Atkinson (1987, 1989) and Foster and
Shorrocks (1988). For a nontechnical exposition, see Ravallion (1993).
1880 E.M. Pernia and M.G. Quibria
However, if the cumulative distribution curves intersect within the range of per-
missible poverty lines, then the results are ambiguous; in other words, it implies
that different poverty lines and measures may provide different answers.
If the first-order dominance condition fails, the stronger second- order domi-
nance condition may prove useful. However, to apply this, one needs to impose
more structure on the poverty measure. If we restrict our attention to additive
measures that reflect the depth of poverty, such as FGT measures for which
a > 1, the second-order dominance condition states that poverty will have fallen
if the area under the old cumulative distribution function is greater than that
under the new distribution over the entire range of admissible poverty lines.
As is obvious, the second-order dominance condition requires more restrictions
than the first- order dominance condition to achieve an unambiguous ranking of
distributions.
mechanism that, on balance, enhances the general well-being of both rural and
urban households in the long run. 23 Still, if externalities are taken into account,
the improvement in general well-being may be arguable. For instance, the exter-
nal diseconomies on earlier urban residents (e.g., congestion costs) and on the
rural population (e.g., loss of dynamic agents of change) may be greater than the
income gains to migrants at the margin.
Poverty in developing countries has for a long time been largely a rural phe-
nomenon, principally because the majority of the population in these countries is
rural-based, earning significantly lower incomes and with relatively less access to
public services than their urban counterparts. For example, the proportion of poor
people living in rural areas is reported to be about 96% in Kenya, 86% in C6te
d'Ivoire, 91% in Indonesia, 80% in India and Thailand, 70% in the Philippines,
66% in Guatemala and 59% in Panama (World Bank, 1990). Despite significant
growth in agricultural production and employment over the past three decades,
rural poverty has persisted in many developing countries, particularly in sub-
Saharan Africa and south Asia. In terms of incidence, anywhere from 40 to 60%
of the rural population is considered poor in these countries.
Rural poverty can be characterized in many ways. Among the principal char-
acteristics or correlates are: (i) limited access to land and irrigation facilities; (ii)
slow adoption of modern technology; (iii) large dependency burden; (iv) limited
human capital; (v) concentration in backward regions; and (vi) concentration
among minority and ethnic groups (Asian Development Bank, 1992; Quibria,
1993).
The rural poor usually have very little or no land at all. They also rarely
benefit from available irrigation facilities. For instance, in Bangladesh in 1987
the extremely poor owned less than half as much land on average as did the
nonpoor at 0.4 versus 0.9 hectares (Hossain et al., 1994). Additionally, the poor
had merely one-quarter of their land irrigated while the nonpoor had one-third of
their land irrigated. In India, households with no land had a poverty incidence of
more than 40% in contrast to those cultivating more than 4 hectares which had
an incidence of less than 15% (Dev et al., 1994). In Nepal, land ownership in
the hills averaged 0.3 hectares for the poor versus 0.5 hectares for the nonpoor;
23 This is the basic rationale why attempts to stem migration directly tend to be ill-conceived and coun-
terproductive. Admittedly, migration can be "excessive" in terms of social consequences at destination and
origin. A better approach to this problem is through indirect measures, such as reforming policies that have
undesirable spatial biases and providing alternative destination choices, if not improving basic conditions at
origin (Pemia and Paderanga. 1981; Renaud, 1981; Tolley and Thomas, 1987).
Ch. 45: Poverty in Developing Countries 1883
in the Terai plains the corresponding figures were 1.1 versus 3.1 hectares (Asian
Development Bank, 1992; Quibria, 1994).
Also, in the Philippines, a great majority of poor farmers own small-size
rainfed plots or work as tenants (Balisacan, 1995). The situation in Thailand is
roughly comparable (Krongkaew et al., 1995). In Sri Lanka, most poor farmers
owned less than 0.4 hectares of land, although with adequate rainfall and op-
portunities for growing and selling high-value crops (e.g., vegetables and fruits)
a holding of only 0.1 hectare could provide a household sufficient earnings to
escape poverty (Gunatilleke et al., 1994). This suggests that low productivity and
unfavorable pattern of land use combine with limited landholding to contribute
to rural poverty. In Latin America, the average size of farms diminished from 2.4
hectares in 1950 to 2.1 hectares in 1980 (de Janvry and Sadoulet, 1989). It is in
small farms where the bulk of rural poverty in Latin America is concentrated.
Although limited access to land is an important feature of rural poverty, tenur-
ial status is not necessarily an indicator of poverty (Asian Development Bank,
1992). In the Philippines, for instance, tenancy is not significantly related to
poverty, but in Bangladesh the opposite appears to be the case. In India, the
relatively high income farmers lease land from both large and small landowners.
Moreover, it is not at all clear that tenants are slower than owner-operators in the
adoption of modern techniques.
Poor farmers are comparatively slow in adopting modem agricultural tech-
niques (Asian Development Bank, 1992; Quibria, 1994a). For example, Bangla-
deshi poor farmers planted 31% of their land with modern rice varieties compared
with 45% in the case of the nonpoor. Similarly, poor Filipino farmers employ
agricultural techniques which are antiquated relative to those used by the more
affluent farmers. This slowness in the adoption of modem technology stems from
several factors, including institutional barriers, a scarcity of information partly
due to low literacy among the poor, and a lack of access to credit. These con-
straining factors affect different types of farmers differently within and among
regions of a given country. Hence, there is wide variation across countries in the
rate of adoption of modem agricultural techniques.
Poor households are typically large in size and have few income earners
relative to nonearners or dependents (Asian Development Bank, 1992; Quibria,
1994a). In Bangladesh, the extremely poor households are slightly smaller in
size than the nonpoor households, but they have a larger proportion of depen-
dent children, fewer working males, and a higher child-woman ratio. In Nepal,
the poor have larger families and more children but roughly the same share of
female-headed households as the nonpoor. In Pakistan, households with many
dependents and headed by old persons are likely to be poor. Large family size
1884 E.M. Pernia and M.G. Quibria
Table 2
Indicators of education in developing countries
Per cent illiterate Gross enrolment ratio (Per cent of school age)
(Aged 15+) Primary Secondary
1959-1963 1970 1992 1960 1970 1992 1960 1970 1992
Sub-Saharan Africa NA 73 50 e 58 NA 66 67 NA 18
Benin 95 84 77 27 36 66 2 5 12
Burundi 86 80 67 18 30 69 1 2 6
Burkina Faso NA 92 83 NA 13 31 NA 1 8
Chad 94 89 55 17 35 65 NA 2 7
Ghana 73a 69 39 38 64 74 5 14 38
Guinea 93 86 67 30 33 42 2 13 10
Kenya 81 68 26 47 58 95 2 9 29
Madagascar NA 46 20 52 90 92 4 12 NA
Mali 98 92 73 10 22 25 1 5 7
Mozambique NA 78 63 48 47 60 2 5 8
Niger 99 96 88 5 14 29 NA 1 6
Nigeria 85 75 47 36 37 76 4 4 20
Rwanda 84 68 43 49 68 71 2 2 8
Somalia 99 97 76 9 59 NA 1 6 10
Sudan 87 83 57 25 47 50 3 14 22
Uganda 75 59 41 49 38 71 3 4 13
Zambia 72 48 25 42 90 97 2 13 31
Zimbabwe 61 45 17 96 74 119 6 7 48
Per cent illiterate Gross enrolment ratio (Per cent of school age)
(Aged 15+) Primary Secondary
1959-1963 1970 1992 1960 1970 1992 1960 1970 1992
with many young children also characterizes poor households in the Philippines.
In Thailand, rural poor households are not only large in size and with many young
dependents but are also commonly headed by younger persons compared with
the nonpoor. Family size, dependency and poverty will be discussed further in
Section 5 on poverty and population growth.
The poor are lacking in assets, and even the only asset that they possess-
human capital or labor power-is limited. The poor typically have low literacy,
lower even in rural than in urban areas. Table 2 shows indicators of literacy and
schooling during different periods for various countries in sub-Saharan Africa,
Asia and Latin America. It is clear that, at the aggregate level, the poorer coun-
tries have had slower educational advancement over time than the better-off
countries.
1886 E.1M. Perniaand M.G. Quibria
Very few literates among the poor have secondary or higher education (Asian
Development Bank, 1992; Quibria, 1994a). In India, more than 70% of heads
of poor households are illiterate and less than 3% of them have education above
the secondary level. In Indonesia, net school enrollment rates are lower among
the poor than among the nonpoor in all age groups, with the disparity increasing
with the age of the child. In Bangladesh, Nepal, Pakistan and similar low income
countries, rural poor households are also headed by individuals with low literacy
or no formal education at all. In the Philippines, Thailand and other middle in-
come developing countries, the poor also have lower educational attainment than
the nonpoor, but the difference is smaller than in the lower income countries.
Rural poverty is typically concentrated in regions or provinces with low qual-
ity of land, inadequate water control, and limited integration into the market
(Asian Development Bank, 1992; Quibria, 1994a). In China, following economic
liberalization in 1978, agricultural growth picked up markedly with a consequent
reduction in rural poverty. But in areas where land quality was poor, water control
limited, and integration into labor and product markets circumscribed, poverty
persisted. In India's Bihar state, poverty incidence is 50% higher than the national
average and more than twelvefold that in the most prosperous state, Himachal
Pradesh. In Pakistan, poverty incidence is higher in Punjab and Baluchistan than
in the northwest Frontier Province and Sindh. In the Philippines, the poorest
region, Bicol, has 70% more people below the poverty line than the national
average, and more than threefold the national capital region's percentage of poor
(Balisacan, 1995). In Indonesia, west Nusatenggara's poverty incidence is more
than four times that of Jakarta (Tjondronegoro et al., 1994). In Thailand, poverty
has traditionally been concentrated in the dry northeast region (Krongkaew et al.,
1994). These pronounced interregional variations clearly suggest that a national
development strategy must incorporate appropriate regional thrusts.
To the extent that the minority and ethnic groups have limited or no access to
common resources and public services, they comprise a large proportion of the
poor (Asian Development Bank, 1992; Quibria, 1994a). In India, the so-called
scheduled castes and tribes represented more than one-third of the rural poor in
1983 although they accounted for less than one-eighth of the rural population.
In many other countries minority ethnic groups usually have a higher poverty
incidence than do other social groups or the population as a whole.
24 This refers to the annual rise in the proportion of the population that is urban, which is to be distinguished
from the rate of increase in urban population; the latter is about two to three times the former.
1888 E.M. Peinia and M.G. Quibria
N
rJ
NJ 1 -
N NNl_
<0
Z
C
_
N 00 *1t
N 00
r,
*N 0
_ 00: 00 00 N 00 00 Nr Z 00 N.
_ N
- 00 Cl N. N 00 R B0 00 ,o 00 0
I
< <2
z
3I
I
u
D xo
P
o
_
.
z z
U Y
.s
O 0r C 0000 N
°000 _0
N -0 N
0 zz z z z
0.
0 _ x0
M uY
F:
00
.0
-
- X
(_ l_
r SI^ X X'r v
0 i
z mrr
-tz zeer rr
ZZ 0 c0 0 N - N 0
.' E8R 000Nd~agn00
9 =I
_5 2 No 00 N rN N * C
NN N 00
Z
_ r- - N
N Z N N 0 N C 0 00
_
_3 O0, 00 N N 00 < < N N 800
1
_ vNN_ .0N
C- - -Ip
) 0
N 4r 00>
00 00
o
Y Bg &t .$ < r
xN00
r
C XN
J:D00N0 'C 00
N 0XNNN
00 N00''
0 00rN
O0 N v, - N Na 00 000 N 00 0
0 0 0
00000
r0 °0 0 00wr0
CŽr-
r.'-
0 _ 0rJ d
0 0
_ N N l .
zerrx O N ZZ
J m vl
0 '0 'o C
LS
Ch. 45: Poverty in Developing Countries
1889
010000 V00
0 X0-000
""~CO
o.0 n m oo )C
0
00
$ Et x
R-R-00000000-0-0$
X °°
0 >- 0100000-0-R "> 9
0000400 >
00000-00
'no
Ch,
00 00 00 0- 00 0 -CO
F
00 00000 000 00. 000 00000000 d0
d '
.= e~~r
0.0
.t
o oV
0- 000- 0000 0000 0000 00 0
0
0 - 00 00 0 0
00 -0 '00000-0 0000000
0 00 00 0- 0 0
- - 0
7 o00-000000000
0~~~\
0 00 00 00
00 00 00 0. 00X%~ga
000 000
7 700 0 0 0
_oo
0 0 00
0s 00 00 0
0
00 o jd
0- 0 00 00 0- 0
-0000
0 0 0 0 00 mf sNen ~ ONm 40 _ 0
a o
d0 ° °
7
0 00~ 9 ~ ~~ ~
10 0 '
0 0 0 0
0 a
0 ~~~ "
0~
00 .0 0. ~m
-. o0 )0S
, moO
0 , 0, 0
o~ ~m
00 000w uo t
"o au O Z r9:
1890 E.M. PerniaandM.G. Quibria
has been proposed as follows: (i) self-employed; (ii) casual laborers hired by a
sequence of employers on a daily contract; (iii) wage workers on longer contracts
with SSEs; (iv) high-wage formal sector whose boundary is defined partly by
enterprise size and partly by ownership; and (v) public sector where wages are
administratively determined (Fields, 1990). Urban poor laborers fall under the
first three categories.
Finally, as in the case of the rural poor, the urban poor are disproportion-
ately represented among the disadvantaged subgroups, such as castes and ethnic
groups (Asian Development Bank, 1994). Several of them are migrants from the
most depressed rural areas. They reside in slums and squatter settlements lacking
in basic services such as water and sanitation (Pernia and Alabastro, 1997), as can
be gleaned from national data on various developing countries in Table 3. Since
these settlements are typically illegal, they are not provided with health services
or schools in the same manner as legal neighborhoods (Wegelin, 1994). In most
cases, because of government plans to relocate squatters to permanent and le-
gal sites, their settlements are not furnished with even the most basic services
or amenities (Mills and Pemia, 1994). This situation may persist indefinitely if
governments lack the wherewithal or the will to provide the permanent sites and
services, or if the squatters resist being relocated. In many cases, shortly after
squatters are relocated to designated legal settlements, they move back to the
illegal sites after selling or subleasing the subsidized dwelling provided by the
government.
In sum, poverty essentially connotes low earnings which are caused by the
lack of ownership of, or access to, physical and human assets. In turn, this is
attributable to unequal distribution of physical capital, such as land and basic
infrastructure, and of opportunities for investment in human capital. Before dis-
cussing the link between economic growth and poverty, we look first at the issues
of gender inequality and population growth in relation to poverty.
26 Studies which indicate such disparities include Agarwal (1986), Banerjee (1983), Behrman (1988a,b),
Chen et al. (1981), Sen and Sengupta (1983), Sen (1988) and Taylor and Faruque (1983). Studies which find
no such intrahousehold consumption disparity include Basu (1989, 1993), Behrman and Deolalikar (1990),
Das Gupta (1987) and Harris (1990). It may be noted that the studies that claim to find gender disparities in
intrahousehold consumption do not generally consider possible gender disparities in energy expenditures that
might reverse the perceived direction of gender discrimination, as in Pitt et al. (1990).
27 For a sample of these studies in Asian developing countries, see Senauer et al. (1988) for the Philippines,
and Senauer et al. (1986) for Sri Lanka.
1892 E.M. Pernia and M.G. Quibria
tality by arguing that "over 100 million women are missing" in Asia: if gender
differences in mortality in Asia and north Africa were identical to those in sub-
Saharan Africa, then there would be 100 million more women in the world than
are reported today. 28 Sen argues that the bias against women in intrahousehold
allocation accounts for a substantial part of this difference in mortality. This in-
trahousehold allocation encompasses a wide gamut of resources, including food
consumption, medical care and other health inputs.
As in food allocation, similar inequities are said to exist between sexes in the
availability of medical care.2 9 However, Deolalikar (1991) notes that there is little
gender bias-if any, it is toward girls-in health care in Indonesia. An indirect
evidence of the inequity in health care between sexes is provided by the respec-
tive morbidity rates, but available evidence is sparse. Two notable studies in this
respect-by Chen et al. ( 1981 ) and Deolalikar (1991)-seem to confirm a similar
geographical pattern of gender disparity: females have higher morbidity in south
Asia while there is no discernible difference in southeast Asia. In a recent study
of health status and physical functioning, Strauss et al. (1993) found that women
report more problems and at an earlier age than men, despite greater longevity of
women; and that females suffer from a greater incidence of functional disabilities
than males. While this is true for all ages in Jamaica, it is true for older people in
Malaysia and Bangladesh.
Regarding educational attainment, Behrman and Schneider (1994) note wide-
spread inequalities in Asian developing countries. The inequalities in expected
years of schooling (given by a weighted average of enrollment rates for primary,
secondary, and tertiary levels multiplied by the number of years required for
completing the respective levels) are once again more acute in south Asia (ex-
28 To arrive at serious quantitative estimates of "missing women", as Sen (1992: p. 124) has himself noted.
requires "proper demographic models of births and deaths", with "clear speculation of the possible coun-
terfactual scenarios". It has been argued that the simple Sen-type calculation, based on comparisons of the
crude ratios of women to men in two populations, tends to exaggerate the number of missing women. Klasen
(1994) has noted that Sen's estimate is to some extent dependent on the demographic peculiarity of African
populations that have a considerably low ratio of males to females at birth. As Coale (1991) has argued, the
expected sex-ratio in a nondiscriminating society is not a constant but varies with population growth and the
sex-specific mortality pattern in such a population. He notes that in a rapidly growing population, the cohorts
of young people far exceed older cohorts. and since males outnumber females by about five to 6% at birth.
the male excess will cause the overall sex ratio to tilt in favor of males. Using the Model Tables "west". Coale
arrives at a different estimate of 60 million missing women. As Klasen has shown, an alternative counterfactual
assumption regarding the expected sex-ratio leads to a different estimate of missing women. He also shows
that the application of the Model Life Tables "west" would imply that there is a problem of "missing men" in
South Africa, the Indian state of Kerala, and to a lesser extent in sub-Saharan Africa. While the application of
sophisticated demographic models may have improved the estimates. the wide divergences among estimates
only prove that estimating "missing women" is still far from an exact science.
29 See. for example, Chen et al. (1981): Kynch and Sen (1983); Taylor and Faruque (1983); Das Gupta
(1987); Miller (1981); and Alderman and Gertler (1989).
Ch. 45: Poverty in Developing Countries 1893
cept in Sri Lanka) compared with southeast and east Asia. Similar disparities
exist in terms of literacy, where men seem to be ahead of women in most de-
veloping countries. These disparities are more pronounced in poorer developing
countries.3 0
An important policy issue in this regard is: how important are individual pref-
erences, vis-d-vis public policies, in determining the gender gap in schooling
attainment and cognitive achievement? If gender gaps reflect essentially indi-
vidual preferences, then there is little that public policies can do to eliminate
or reduce the gap. Recent studies by Alderman et al. (1992, 1993), however,
indicate that the gender gap in cognitive achievement in rural Pakistan for the
age group 20-25 was about 25% that of males for the same age group. For this
age group, nearly 50% of the gender gap could have been reduced by eliminating
the inequality in access to single-sex schools. In other words, a large part of the
gender disparity reflects a bias in policy rather than individual preferences.
However, if one takes a broader perspective on individual well- being as be-
ing determined by economic and noneconomic factors, including one's level of
participation in society, then disparity in terms of traditional poverty indicators,
as discussed earlier, provides only a partial basis for judgment. In many devel-
oping countries, in addition to the disparities in economic indicators between
men and women, there are social, cultural and psychological factors which limit
women's full participation in societal activities. 3 1 For example, women often
have limited access to employment opportunities, and there is often a consid-
erable differentiation in the nature of jobs held by men and women. Women are
generally engaged in inferior jobs-which are unskilled, have low security, and
provide less opportunity for advancement. 3 2 Accordingly, the welfare level being
enjoyed by women is likely to be lower than that enjoyed by men. Moreover, as is
often suggested, women have to work longer hours as they engage in both wage
and nonwage employment, the latter usually consisting of household chores. Ac-
are not particularly suited to become tenants. Therefore, inheritance laws and the
general perception of landlords tend to conspire toward excluding women from
having access to agricultural land. However, when women have access to land,
there is no evidence that they are discriminated against in the contract. The terms
of contract are usually uniform across a particular village or locality, with little
or no variation between tenants.
Many women, including poor landless women, are engaged in homestead
agriculture. This includes vegetable and fish cultivation, poultry and livestock
raising, tree plantation, and crop processing. Most poor women have a small
homestead, although in the traditional social environments of poor developing
countries, they are often more constrained, as compared with poor men, by lack
of technical information, extension services, credit, and good seeds (Safilos-
Rothschild, 1990; Bardhan, 1995).
Despite increasing participation, it is widely believed that females are gener-
ally more disadvantaged than males in labor markets. In many countries, females
used to be assigned to specialized farm tasks; but women are increasingly making
inroads into operations previously considered to be the exclusive domain of men.
This segregation into different sectors of employment may have much to do with
physiological and biological differentiation, differential skills and training, and
the sociocultural environment of the country. Nevertheless, with economic devel-
opment, the social environment is undergoing rapid changes in many developing
countries.
Labor market discrimination takes many forms, including: wage discrimina-
tion-paying differential wages to equally productive workers; employment
discrimination-hiring members of a favored group; occupational segregation-
the assignment of members of different groups to different jobs; crowding-the
confinement of less-favored groups to undesirable occupations; and statistical
discrimination-selection of individuals for employment and training on the ba-
sis of attributes (such as gender) considered to be related to productivity.
Gender disparity in returns to labor has been the subject of a number of stud-
ies. A summary of studies that provides estimates of returns to schooling is given
by Psacharopoulos (1988). As Behrman and Zhang (1994) note, these estimates
do not necessarily indicate a one-way differential-in some cases, the estimates
indicate no substantial disparity in the labor market returns to schooling, but
in others they indicate substantial differences favoring either women or men.
Behrman and Zhang conclude that, women, generally though not always, have
received labor market returns in Asia at least of the same order of magnitude as
have men.
1896 E.M. Pernia and M.G. Quibria
A few studies have attempted to decompose the wage differential into differ-
ences due to observed human capital and wage discrimination. A study on Taiwan
by Gannicott (1986) indicates a gender wage differential of 0.44, of which one
third can be attributed to observed human capital and two-thirds to discrimination
(i.e., the unexplained residual). While such studies are potentially interesting, as
Behrman and Zhang (1994: p. 26) note, many of these suffer from a number of
analytical problems that include
inferring dynamic patterns from cross-sectional data, often mixing the ef-
fects on the value of time (the wage rate) with those on hours worked by
using earnings, ignoring the impact of grade repetition and school dropout,
and ignoring the possibility that individuals' background (e.g., abilities,
motivations, habits and family connections) directly increases both their
schooling and their labor market earnings, so that schooling in such studies
partially represents the effects of unobserved background in addition to the
effects of schooling itself.
In developing countries many parents are so poor that their fecundity is im-
paired owing to inadequate nutrition. The more general case, however, is that
low income families are larger in size than high income ones, even allowing for
higher mortality among the former. The natural consequence is a distribution
of consumption that is worse than the distribution of family incomes (Boulier,
1977). There may be attenuating forces such as economies of scale in household
consumption, additional work effort on the part of parents, and productive con-
tribution of children net of their consumption needs.3 6 But these countervailing
effects of family size do not seem to be large and significant enough to matter
(Cassen, 1976). Moreover, if small families are in a relatively better position to
take advantage of such income-enhancing mechanisms as education and health
investments, migration and labor-market mobility, large families would be in a
relatively worse position.
The extended family system that is quite common among poor households
especially in rural areas may work for or against better income distribution. If
the arrangement enables a family to earn in excess of its requirements and thus
accumulate physical or human capital, it can be better off. Otherwise, the family
would be worse off, and this will be reflected in a further deterioration of income
distribution.
The simple intergenerational effect of family size is via parents' income which
is a key determinant of an individual's income. Since poor parents have more
children than the rich, unequal distribution of income is transmitted intergener-
ationally by differential fertility (see also Potter, 1979). Cassen (1976: p. 811)
adds:
There are more complex intergenerational effects, such as those that operate
through property inheritance if several children can inherit property, then
those with more children will divide their property into smaller fractions;
35 This section draws on Pernia (1982) and Deolalikar and Pernia (1993).
36 It is also argued that the cost to parents of additional children are fairly low if adult equivalence scales
are adopted in the computation of cost (Rodgers, 1984).
Ch. 45: Poverty in Developing Countries 1899
and if they have less property to start with, or worse, if their property is
diminished by sale in their lifetimes while that of the rich increases by
purchase, there will be a tendency for the distribution of property to become
more unequal over time and therefore probably, the distribution of income
also.
100
90
80
o
C\
70
0 60
50
40
>
0
30
o
0. 20
10
0
1 1.5 2 2.5 3 3.5 4
Lagged Population Growth (%, 1980-90)
Fig. 1. Poverty incidence and population in developing countries.
in particular, is usually pronounced. Chenery et al. (1974), for example, find that
a 1% increment to population growth rate is associated with a 1.6% fall in the
income share of the poorest 40% of households. On the other hand, Rodgers
(1983) concludes from a cross- sectional exercise that the effects of population
growth on inequality are not so important. Nevertheless, a finding similar to that
of Chenery et al. (1974) is reported by Ahluwalia (1976), showing, in addition, a
positive relation between population growth and the income share of the richest
quintile.
Using longitudinal district-level data on population, agricultural production,
and infrastructure from north India, Evenson (1984) finds evidence of a wors-
ening functional distribution of income with increasing population density. In
districts where population density increased fastest, the poorest groups gener-
Ch. 45: Poverty in Developing Countries 1901
ally comprising the landless laborers experienced the largest decline in real in-
comes, while the richest groups, typically large land owners, experienced a sharp
increase in the rents paid to them.
There is ample evidence showing that bigger households have larger house-
hold incomes but lower incomes per capita than smaller households (e.g.,
Kuznets, 1976; Lipton and de Haan, 1997). Poverty incidence tends to rise with
family size and the evidence is that this relationship is monotonic. In the Philip-
pines, for example, 1975 data showed that only 9% of one-person families were
below the poverty line, but the proportion of poor was 34% for four-person fam-
ilies, 52% for six-person families, and 65% for 10-or-more person households
(Pernia, 1982). On the whole, poverty incidence was below the average for up to
five-member families and above the average for larger-size families. The effects
of high fertility on poverty are also documented by village studies in Java (Penny
and Singarimbun, 1973) and by an urban study for Kerala (Scott and Matthew,
1983). In rural areas these effects often operate through landlessness, 37 unem-
ployment, and low output or wages, as shown by studies in India, Bangladesh,
and the Philippines (Das Gupta, 1978; Arthur and McNicoll, 1978; Kikuchi and
Hayami, 1980).
Recent studies in the Philippines and Thailand also show that households
with many children are at great risk of being poor and that a high rate of child
bearing restricts women's ability to participate in the labor market and contribute
to household income (East-West Population Institute, 1993). Using longitudinal
household survey data from India, Gaiha and Deolalikar (1993) find that not
only are larger families likely to be poor at any given point in time, but they are
also likely to experience chronic poverty (defined over nine years of longitudi-
nal data). In other words, an important distinguishing characteristic of chronic
poverty viz., households whose incomes fall below the poverty line continuously
every year is large numbers of children. To the extent that some proportion of
these children represent unwanted births, family planning programs that reduce
the number of unwanted births can have a significant effect on the alleviation of
chronic poverty.
The fact that large families are generally poor has obvious implications for
family welfare. Household budget studies in developing countries typically show
that the consumption basket tends to be heavily weighted by items that satisfy
such basic human needs as food, clothing and shelter. For the average Filipino
family, for example, about 78% of total expenditures are accounted for by these
37 For example, in Bangladesh, landlessness increased from an estimated 7.3% of the farm labor force in
1951 to 26.1% in 1977. Real agricultural wages in the 1970s were allegedly below what they had been in the
1830s (Khan, 1984).
1902 E.M. Perniaand M.G. Quibria
basic necessities, with food alone claiming half of the total (Cabafiero, 1978).
The food share is higher in rural than in urban areas, and is also larger the poorer
and bigger is the household (Valenzona, 1976). Moreover, these studies suggest
that a smaller number of children would raise the absolute consumption of a low
income household. Thus, increases in household consumption and savings, in-
cluding investment in human capital, are important opportunity costs of children
(Mueller, 1972).
Notwithstanding arguments that high population growth or fertility can have
favorable effects on poverty and income distribution (as when families try to cope
with poverty by having many children), the empirical evidence on the adverse
effects seems overwhelming (Deolalikar and Pernia, 1993). It is often argued that
the poor have large families because they perceive the marginal benefit to exceed
the marginal cost of an additional child. But such an advantage may be illusory
and short-term at best. The fact more often is that, in the long run, families with
many children save less and invest less in the human capital (i.e., education,
health and nutrition) of their children. Thus, poverty tends to be transmitted in-
tergenerationally. High fertility not only impoverishes a household in the present
but significantly increases the likelihood of that household's children being in
poverty as adults owing to their scant endowments of human capital, which is
frequently the poor's only asset.
Because the poor typically have higher fertility and infant mortality rates than
the nonpoor, government efforts at slowing population growth via reductions
in both infant mortality and fertility disproportionately benefit the poor. Con-
versely, a weak population policy, or lack of a population-control strategy, hurts
the poor more than the nonpoor. The experience of the more economically and
demographically successful developing countries clearly suggests that a family
planning program can be highly effective if it is supported or complemented by
programs to lower infant/child mortality and promote female literacy and school-
ing, at least up to the lower secondary level (Deolalikar and Pernia, 1993; Cassen,
1994).
80
oC
Gz 70
' 60
-f 50
_ 40
00
0.0 2.0 4.0 6.0 8.0 10.0
Lagged GDP Growth (%O,1980-90)
Fig. 2. Poverty incidence and GDP growth in developing countries.
Table 4
Economic growth and poverty in developing regions
(a) Growth of real per capita income in industrial and developing countries, 1960-1991
(Average annual percentage change)
Percentage of
population below the Number of poor
Region poverty line (Millions)
1985 1990 1985 1990
Note: The poverty line used here-$370 annual income per capita in 1985 purchasing power parity
dollars-is based on estimates of poverty lines from a number of countries with low average incomes.
In 1990 prices, the poverty line would be approximately $420 annual income per capita. The estimates
for 1985 have been updated from those in World Bank (1990) to incorporate new data and to ensure
comparability across years.
a Estimates.
b Does not include the former USSR.
Source: World Bank (1992), Tables 1.1 and 1.2 (Poverty data based on Ravallion et al. (1992)).
Ch. 45: Poverty in Developing Countries 1905
70
60
50
I.2
40
Q
30
O
C.)
t 20
o
0
10
0
0.0 2.0 4.0 6.0 8.0 10.0
GDP Growth (%, 1980-90)
Fig. 3. Poverty incidence and GDP growth in Asia.
. n
1Z
10
8
i)
a 6
0
72 40
°. 0
-2 ... . ::::::::::::;:
1980s, poverty incidence remained unchanged at about 48% during the 1980s
and the number of poor increased markedly from 184 to 216 million. Similarly,
in Latin America and the Caribbean, per capita incomes also contracted in the
1980s and both poverty incidence and the number of poor people increased
significantly, from 22 to 26% and from 87 to 108 million, respectively. 4 0
The important association between poverty and economic growth is exempli-
fied more clearly by the experiences of developing countries in Asia, a region
which has been the scene of dramatic economic and social transformations. As
noted earlier, poverty incidence at a point in time is largely a reflection of pre-
vious economic growth performance. This is again illustrated quite clearly with
data for several Asian developing countries in Fig. 3. Putting it differently, the
rate of reduction in poverty is contemporaneously associated with GDP growth
40 Psacharopoulos et al. (1995) report a similar increase in poverty in Latin America during the 1980s
although their numbers are somewhat different from those reported here. The differences, however, are largely
due to the differences in the poverty line. Despite the differences, what emerges clearly from their analysis
is that there is a strong correlation between growth and poverty reduction. Psacharopoulos et al. note that
economies that grew in the 1980s, for example, Colombia and Costa Rica, performed better with respect to
poverty reduction and income inequality than those that stagnated. In particular, countries that failed to achieve
macroeconomic stability such as Brazil and Peru experienced substantial increases in poverty.
Ch. 45: Poverty in Developing Countries 1907
over time, as shown in Fig. 4. This cross-country evidence is indirect, however,
inasmuch as individual economies start the growth process at different points and
follow different growth paths. For more direct evidence, one needs to examine
individual countries. This indicates that economic growth reduces poverty most
of the time, although not all the time (Fields, 1995).
There is compelling evidence from the experiences of such newly industri-
alizing economies (NIEs) as Korea and Taiwan and such upcoming NIEs as
Malaysia, Indonesia and Thailand that economic growth leads to poverty decline.
In these countries poverty incidence has been cut to half or even less in 10-
15 years (Table 5). In the Republic of Korea and Taiwan poverty incidence is
reported to be below 5%. In Malaysia, Indonesia and Thailand where rapid eco-
nomic growth has been more recent than in the NIEs, poverty incidence ranges
from 15 to 20%. In countries with relatively poor-to-modest economic perfor-
mance, such as south Asian countries and the Philippines, poverty reduction has
been correspondingly slow and somewhat unclear. Poverty incidence in these
countries remains at 30-45%. In some cases, the number of poor people, if not
the incidence itself, appears to be on the rise.
Economic growth in the NIEs and near-NIEs has not only been rapid and
sustained but, more importantly, it has been broad-based and pervasive, that is,
reaching the poor as workers and as consumers (World Bank, 1993). Because
growth in these economies has been labor-intensive, many job opportunities have
been created leading to full or near-full employment and, consequently, rising
real earnings. Moreover, a part of the increase in national incomes has been
utilized to improve the access of the poor to social services. Since consumption
by the poor of these services (education, health, and family planning) is also
simultaneously an investment in their human capital, the quality and productivity
of their labor have been enhanced.
The development strategy of these successful Asian developing countries ba-
sically pursued agricultural development and industrialization policies based on
market incentives and outward orientation. The association between outward
orientation and sustained economic growth alongside poverty reduction is not
accidental (Quibria and Srinivasan, 1993). Sustained economic growth is feasible
for a developing country with a limited domestic market only if opportunities
in international markets are effectively exploited. A developing country with
surplus labor and scarce capital will naturally have a comparative advantage in
activities that are labor intensive. Accordingly, a strategy of promoting export-
oriented, labor-intensive industries not only ensures sustained growth but also al-
leviates poverty in both rural and urban areas. Furthermore, an outward oriented,
L908 E.M. Perniaand M. G. Quibria
cia a a a ci a <a a a <a,
o4acia-$_ cicg
'^Rz
5 z a_ a a aA a
_ A
F
a a a ci a c aO
i a a a a c aa a a
a
B
p
axxx
a a a a aG a a a a
g
a a a a a8a a a aaaa°a a cia - a
Ol
C4 Z
O ^i
N ^ ^ ^ ^ ^ ^
a a a a a° a a a°, R
_______________
c^ a ao a Z a^
.9
a a c a a
aa
R sEiR g z
a
t4
0 _ ._ _
a
a
a a ci ci a ci a a
ff a^g a 7a~a6~a7
.q a
<cig ,A<
-q
.5
a _
_______ t .X ^. ;
a
: -
D
H
-C
_3
a ORZ
a 7 zc X t <:
a
*0
-~~~~
a
4
o
8z Z
S.A,
t m
~'h~ s
_
Xaciiaciaa^
___ ____
-at-e
.- __~2
a a^ a a a a a ci ci a a a ci ci
o 6. ^_z
7 V, r a 0
c C^ c a^ a^ a^ ciciaacic Zc
z tN c >>
a
aO gi a vi c a Z
-E
=
-' e^-o
a a a cX a
o a^ tS a a,,,,a gE
a^ a a a
aa a a^ a^ a a a a a^
o7
Rs~s
oARAR
N o
a A6b
_ _ _
a ci a a a a a a: ci < <c ae a
c
a0 ao a^ a a^ a co a aa^ c^
A888
[ - r ^ I c
-
a U4 a h]
.e <
Ch. 45: Poverty in Developing Countries 1909
2$
O ON O ON ON ON O O ON O O ON ON
R E 8 O ON E RR R
ON
NC_N N N- NNN
e
0
NO
cNO N o
N ON ON < N ON N N NO N N N N N,
a
CO
8888888888gg
- N N Z N N N N N N - N - N a
_N ON ON ON ON
_
CO
ON ON O ON O O
N O ON
ON -o ON N N rN ON N N oN N N ON -
.
_Ks
ON
- NO N N N ° ON N N ON - N -
N ON N N N N N ON No N NN O
ON
0
%aaF
o o>u
o
N
ON
2$
m
g~8z
$ I oo>_ N O NON ON ON ON
N O N N O NN O N ON
S
-
ooR o No ON N N gg ON ONON ON ON ON ON ON ON -d
a
0
_ _ _ S
O
s
< >s
.0 zN N
2N O N ON ND . ON Ni -
o
-
sO
N N_ N N N N N N
0
£ t mN _ ON ON ON ON ON ON ONON ON ON ON ON ON ON o
0
ON ON ON ON ON ON ON ON ON ON ON ON ON O 0 C
m
In% ~~~I - 0
°oo -
g
CO
OON.~ ZO.O O :i
O 0I 00--iIOOI
ONN N N N ON N N, N N NO ON
O ON
o o ^ t
oxD~~
N 0C _ 0 C ON~
6 S oW
0 _ 0.. 000 ~3 0 CO
V1t
m
2$ OO2$02$2$~ .
r £t
0COZV
.
a,
°ff -
v) t 00 m
f ON N ON ON N O N N N N ON ON O ON
v) N
C CO
c U NO E o 0* E °
1910 E.M. Perniaand M.G. Quibria
liberal economic policy environment also encourages the inflow of foreign direct
investment in industries that are typically labor intensive and export oriented.
The economic growth policy of the dynamic Asian developing countries has
been complemented by a sound social investment policy, particularly in the areas
of education, health and family planning, with favorable consequences for the
quantity and quality of human capital and overall economic productivity. It is
also important to note that such social investment has enhanced the status and
productivity of women at home and in the labor market. Moreover, better ed-
ucated, healthier and more productive women have tended to have fewer and
higher quality children, further improving prospects for succeeding generations.
Thus, economic growth and human resource development leading to poverty
reduction have been promoted in a mutually reinforcing manner. a process aptly
referred to as a "virtuous circle" (World Bank, 1990).
Direct poverty alleviation or social safety net programs have been necessary
in many countries, especially in certain sectors where chronic poverty exists or in
periods of adjustment and transition with the implementation of economic policy
reforms. As will be discussed below, these programs have typically involved
significant budgetary spending which would have been difficult without rapid
economic growth, and probably feasible only with foreign aid or borrowings.
In sum, economic growth that is outward oriented and labor intensive is likely
to be also sustained, rapid and broad-based and, hence, poverty reducing. In
addition, such growth provides the wherewithal for improved social services and
investment in human capital, not to mention direct poverty alleviation programs.
This has been the path taken by the Asian NIEs and near-NIEs. Conversely, eco-
nomic growth that is inward oriented and capital intensive tends to be erratic and
weak, with the benefits being captured by the upper income segments of society
and limited trickle-down to low income groups. This has been the experience of
most of the other Asian developing economies, at least up until the early 1990s.
By and large, this conclusion is applicable to the growth and poverty experience
of many African and Latin American countries.
Families may split up, leaving one parent with inadequate income. A sizable
number of the poor show upward mobility to the ranks of the nonpoor (this does
not mean that this poverty is not a matter of concern but that these people are not
part of a permanent "under class").
Such mobility requires careful interpretation. It may arise in connection with
the life cycle. 41 If lifetime income is above the poverty level, there is no lifetime
poverty for this individual. If the capital market is perfect and if the individual
can borrow against future income at a constant interest rate, then he or she can
maintain a constant consumption. In this case, by using a consumption-based
poverty line, it can be concluded that there is no lifetime poverty. However, if
an income-based poverty line is used, there would be an indication of poverty.
If capital markets are imperfect (credits are rationed at a fixed interest rate), the
individual will not be able to rise above the poverty line. Similarly, the lack of a
perfect insurance market may lead to temporary poverty.
Without perfect capital and insurance markets, some people who may be
nonpoor on a lifetime basis may face transient poverty. Those who suffer from
transient poverty may be helped by transfer and other programs or perhaps by
public interventions to remove the relevant market imperfections. But those who
are under chronic poverty, either because of lack of assets or lack of employable
skills, will not be helped by the possible remedies of market imperfections. They
cannot be helped by mere income transfers: they will need more assets, possibly
made available through a program of asset redistribution. In an agrarian setting,
this may involve a more egalitarian distribution of land, but implementation of
land reform programs is fraught with many practical difficulties, as discussed
below.
One of the most lively debates about the causes and cures of poverty has
focused on the relationship between economic growth and poverty alleviation,
which has been discussed earlier. The so-called "trickle-down" view argues that
rapid economic growth is both a necessary and sufficient condition for poverty
reduction. An opposing view, often referred to as the "basic needs approach",
argues that while growth may be necessary, it is by itself insufficient and there
is a need for some intervention in favor of the poor. 42 The distinction between
chronic and transient poverty can shed some further light on this matter. As far
41 Rowntree (1901) finds that the life of a laborer is marked by five alternating periods of want and
comparative plenty, the periods of want being childhood, when he himself had children, and old age.
42 There is yet another view which argues that the relationship between provision of basic needs and growth
need not necessarily be competitive but can be complementary. According to this view, considerations of
optimal growth entail some measure of balance between growth and basic needs. See, for example, Quibria
(1982).
1912 E.M. Perniaand M.G. Quibria
as transient poverty is concerned, fast growth and economic dynamism can help,
but do not have any direct bearing on the problem. As far as chronic poverty is
concerned, rapid growth and economic dynamism can help in the long run but
require direct intervention in the form of asset transfers as well as investment in
human capital formation through education and training.
It is widely agreed now that economywide and sector strategies and policies
have significant impacts on levels and trends of poverty, often much greater than
the effects of special programs directed at poverty alleviation per se (Quibria
and Srinivasan, 1993). These policies include macroeconomic policies (fiscal and
monetary), foreign trade and exchange rate policies, sector policies and various
income and asset transfer policies.
There are two types of macroeconomic policies for developing countries: (i)
stabilization policies that are meant to return the economy to an equilibrium
growth path from which it has deviated due to shocks; and (ii) adjustment policies
that attempt to change the current path of the economy to a new equilibrium
path. The two types of policies often use the same instruments such as cur-
rency devaluation and fiscal contraction. Sometimes the two types of objectives
are intermingled; sometimes adjustment programs are preceded by stabilization
programs (Behrman, 1993).
Macroeconomic policies affect the real incomes of the poor, primarily through
their impact on the returns to their assets-principally labor, on the prices they
face with respect to the goods and services they produce and sell as well as those
they buy, and on the transfer incomes they receive. An adjustment program which
often relies on devaluation of the domestic currency is likely to increase the price
of tradables and decrease the price of nontradables. If the tradable sector is more
labor intensive than the nontradable sector, the real wage will rise and the poor
whose asset is labor is likely to benefit from such an increase (Behrman, 1993).
However, a priori theorizing about the signs and the magnitudes of these effects
may be mere speculation. In the final analysis, the issue of the relation between
macroeconomic policy and poverty is essentially an empirical one.
However, it is extremely difficult empirically to isolate the impact of macro-
economic policies on poverty. It is expected that during adjustment, economic
growth slows down, which in turn exacerbates the poverty problem in the short
run; however, in the longer term as economic growth gathers momentum, the
poor are likely to be more than compensated for their short-run losses. Neverthe-
Ch. 45: Poverty in Developing Countries 1913
less, while one can expect a dynamic tradeoff between adjustment and poverty,
this is far from empirically established.
Distributional changes can also influence the final outcome in this regard. A
favorable distributional shift in Indonesia and an unfavorable distributional shift
in the Philippines reduced poverty in the former and increased it in the latter
(Balisacan, 1995; Ravallion and Huppi, 1991). However, these distributional im-
pacts of adjustment are largely dependent on labor market flexibility, trade and
investment regimes, and other economic and social conditions.
Similarly, poverty outcomes also depend on the degree and composition of
public expenditure reduction. If public expenditure reduction translates into heavy
reduction of social sector spending, it is likely to affect the poor adversely in both
the short and long run. However, if the expenditure reduction is accompanied by
increasing social sector spending, as has happened in some countries, the adverse
effect of slower short-run growth can be more than offset by the prospects of
longer-term gains for the poor.
Behrman (1993) reviews the evidence from a cross-country statistical analysis
of various indicators of living conditions (in particular, labor and social indi-
cators) in developing countries and how changes in them have been associated
with participation in international organization-supported adjustment programs.
The cross-country evidence suggests that reduction of economywide distortions,
such as overvalued exchange rates and industrial protection, improved the prices
of agricultural exportables while only marginally affecting the prices of importa-
bles. The short-run supply response was fairly small. However, the adjustment
programs had little impact on labor markets and social sectors. There was some
evidence that there were some positive effects on rural income growth and neg-
ative effects on longer-run human capital investments. These results are some-
what less deleterious than those often claimed by many skeptics of adjustment
programs.
Behrman (1993) also provides an in-depth analysis of three country stud-
ies. The Indonesian study indicates that the country experienced a substantial
decrease in poverty during 1984-1987 when it undertook a major adjustment pro-
gram. The poverty reduction that took place reflected in part increased prices for
exportable cash crops, and in part increased government support in terms of poli-
cies that augmented the consumption of the poor during the adjustment. Finally,
Indonesia had rapid growth for two decades before adjustment and investments
made during those two decades also paid off during the adjustment.
The Jamaican experience of the mid-1980s suggests very little evidence of
the short-run negative effects of adjustment programs on the income and human
resources of the poor. The analysis of the C6te d'Ivoire adjustment program in the
1914 E.M. Perniaand M.G. Quibria
early 1980s contrasts sharply with that of Indonesia in that poverty is estimated
to have risen faster than the rate of fall of mean real per capita income. This sug-
gests a combination of negative growth and more unequal income distribution.
Although C6te d'Ivoire had also experienced rapid growth in the 1960s and much
of the 1970s, substantial declines in the world prices of its major exports in the
late 1970s precipitated the adjustment. The poverty situation was aggravated by
the withdrawal of the support prices for exportables grown by the poor farmers.
Finally, it may be noted that structural adjustment policies have received un-
favorable reviews from many poverty analysts. A number of studies that claim
to show the negative effects of adjustment on poverty (e.g., Cornia et al., 1987)
are based on erroneous and simplistic comparisons of poverty indicators "be-
fore" and "after" adjustment. If the adjustment is intended to correct a poten-
tially unsustainable situation, then the relevant comparison is between the actual
after-adjustment paths of the poverty indicators and the counterfactual paths had
there been no adjustment (Srinivasan, 1993). To undertake such an analysis, one
requires some manipulation of applied general equilibrium models (see, e.g.,
Bourguignon et al., 1991, 1992). However, the construction of a counterfactual
scenario is dependent on the structure of the model and on the various parameters
assumed for the model. While an applied general equilibrium model may be
conceptually the ideal tool to explore these issues, it falls short of the ideal in
actual applications. 4 3
43 Some of the problems associated with applied general equilibrium models in this regard were succinctly
expressed by Srinivasan (1992: p. 24): "Unfortunately, the models involve many other parameters, and rela-
tively few of the parameters used in the model are derived from econometric analysis of the data from the
country being modeled. Besides a pure Walrasian AGE model does not incorporate the fundamental monetary
and financial variables of the macroeconomics of stabilization and structural adjustment policies. Eclectic
models that combine macroeconomics with Walrasian microeconomics ... appear promising for distributional
analysis. However, it is essential, though not simple, to incorporate the "political-economy" of structural
adjustment and stabilization in a credible way in such models".
Ch. 45: Poverty in Developing Countries 1915
A critical element of trade and industrial policies of the more dynamic de-
veloping economies has been that policies did not penalize the traded goods
sector. Market forces had to a large extent been allowed to determine the real
exchange rate and lead to an efficient allocation of resources. Endowed with
abundant labor resources, the Asian NIEs, and later their southeast Asian neigh-
bors, emphasized production and exports of low value- added, labor-intensive
manufacturing goods. Their success in exports of labor-intensive manufacturing
led to a dramatic reduction in poverty in these countries. Subsequently, however,
as labor costs rose, these economies responded by moving up the export product
hierarchy.
There are a number of possible routes through which a liberal trading en-
vironment contributes to an efficient allocation of resources and rapid growth.
In removing barriers to market entry, trade liberalization is a powerful antidote
to rent seeking and monopoly. By encouraging greater competition and a more
efficient allocation of resources, trade liberalization increases output, investment
and employment in sectors where a high degree of competition already exists.
Further, trade liberalization creates the opportunity for further specialization and
the exploitation of economies of scale where such economies exist. All these
contribute to economic growth and poverty reduction.
A strategy of import substitution, based on high tariffs, quantitative restric-
tions and over-valued exchange rates, often creates an inefficient industrial struc-
ture, characterized by a high degree of monopoly and a narrow ownership struc-
ture. Without sufficient internal or external competition, this type of industrial
structure is conducive neither to growth nor to domestic capital accumulation. It
is also true that, on balance, long-term capital, especially in the form of foreign
direct investment, is unlikely to be attracted to economies in which there is an
array of trading and associated bureaucratic restrictions and regulations. It is no
coincidence that economies in southeast Asia that have comparatively liberal
trading regimes have demonstrated significant successes in attracting foreign
direct investment that contributed in significant ways to the growth and poverty-
reduction performance of these countries. Moreover, to the extent that technolog-
ical progress is embodied in new capital goods produced abroad, growth will be
impeded by protectionist regimes that tax or otherwise raise the cost of foreign
direct investment or technology licensing.
In sum, trade and industrialization policies of the successful developing
economies, in particular those in Asia, were such that they introduced a regime
of internal and external competition for these economies, a neutral incentive
structure that did not penalize exports, and a conducive economic environment
where savings and investment (including foreign) flourished. All these helped to
1916 E.M. Perniaand M.G. Quibria
generate not only a high rate of income growth but also a high rate of employment
growth that helped to foster a dramatic reduction in poverty.
Land reform is often advocated as an effective antidote to poverty. The term land
reform has been used in many different senses, but in the following it is defined
to encompass both land distribution and tenancy reform. The former implies
distribution of land from households with holdings higher than a stipulated upper
limit to households owning little or no land; the latter entails changing the terms
and conditions defining the tenancy contract.
Tenancy legislation appears to have no sound theoretical or empirical basis.
Share tenants, who comprise the great majority of agricultural tenants, work hard,
maintain the land, and adopt new practices quickly. When laws make tenancy ille-
gal, as the Asian experience indicates, the landlord is found to use wage laborers
and this leads to a less efficient arrangement. It is therefore argued that tenancy
regulations should be eliminated to allow a wider range of contractual choices
and to induce larger farms to lease out more land to the landless (Otsuka, 1993).
To make the tenancy option attractive to the larger landowners and to ensure
economic success for the tenants, the latter should be given public support in
the form of easy access to capital and technical know-how on farm operation
and management. It is argued that rather than suppressing the tenancy market,
as many reforms have attempted to do, efforts should be made to make it func-
tion efficiently, as a well-functioning tenancy market can contribute to both the
equity and efficiency of the rural economy. In particular, tenancy deregulation is
expected to open up for the landless tenants the agricultural ladder out of poverty.
As regards redistributive land reform, there is a sound rationale underlying
such a policy. First, there is no evidence of significant economies of scale in agri-
culture; therefore, there can be no general presumption against small farms. For
the rice economies of developing countries, the skill-intensive and scale-neutral
nature of rice cultivation makes the case for land redistribution even stronger.
Second while small farms are at a disadvantage in perceiving and adopting new
technology, this difference is not particularly significant empirically. Indeed, the
receptiveness of small farmers to extension services dilutes most objections to
small farms. Finally the fragmented and oligopolistic nature of the land market
in most developing countries further strengthens the case for land reform. In
this connection, it has been noted that land reform can affect the local political
structure, as it gives "voice" to the poor who are induced to participate in the local
44 This subsection draws on Rashid and Quibria (1995).
Ch. 45: Poverty in Developing Countries 1917
government institutions and in the management of local public goods (Bardhan,
1996).
Land reform, which improves distributional equity, appears to have many
strong arguments in its favor on grounds of efficiency. The traditional argu-
ments-such as the efficiency wage argument, (in terms of better health and a
more effective labor force) and the labor-absorption argument (in terms of the
greater labor intensity of small farms)-appear to be well founded. And while
the impact of land reform on gender equity or on the environment are uncertain,
there is no presumption against land reform on these grounds. On balance, land
reform can have a salutary impact on poverty alleviation by promoting social
equity.
Despite its putative benefits, land reform has been far from universally adopted.
The Asian experience provides a simple answer to this apparent paradox. In
Taiwan, land reform was conducted efficiently by an alien military power-
a precedent that carries no force for democratic regimes. In the Republic of
Korea, land reform was instigated by a military government and succeeded in
bankrupting the landlord class. No agricultural surplus was generated to help to-
ward industrialization, and the Republic of Korea's industrial success came over
a decade later owing to seemingly unrelated circumstances. In Thailand, land
reform has affected marginal farmers, yet the economy has grown rapidly. In In-
donesia, no land reform at all has taken place and yet the economy has performed
well. The above indicates that land reform is neither necessary nor sufficient for
success in economic development or poverty alleviation. In addition, land reform
is difficult to implement in democratic societies where landlords tend to exert
more political power than tenants and landless farmers.
Finally, in most agrarian societies especially in Asia, while there may exist a
marked inequality in terms of land ownerships, there is much less inequality in
terms of operational holdings (Otsuka, 1993). This fact tends to weaken much of
the equity argument in favor of land reform. Moreover, in many poor countries
(such as Bangladesh), the large size of the population makes further redistribution
economically infeasible. In these countries, the main plank of antipoverty policy
in rural areas needs to be based on enhancing land productivity through new
seed-fertilizer and irrigation innovations, and not through further redistribution.
7.4. Infrastructure
the productivity of the factors of production, namely, land, labor, capital and
technology (Jimenez, 1992; World Bank, 1994). In addition, both infrastructure
services provide consumption utility that directly enhances well-being and raises
living standards.
ficient. In developing countries, the poor often use fuelwood for cooking and
kerosene or candles for lighting, thus leaving them out of subsidized electricity.
They also have to buy water from private vendors who charge prices that are
often a multiple of those charged by public utility agencies (World Bank, 1988).
Within a system of user fees the poor can be protected through block pricing
for utilities, whereby higher unit prices are charged at higher consumption levels
such that the nonpoor pay full cost (Jimenez, 1992).
Third, while the trend is toward less government intervention and more private
sector participation in the economy, poverty reduction efforts will continue to
require public sector involvement in planning, budgetary provision, coordination,
and monitoring (Lanjouw, 1995). The government's direct role will be called for
in the provision of such basic facilities (with limited scope for cost recovery)
as rural roads and irrigation systems. The Malaysian experience in privatization
is highly instructive. The privatization program enabled the government to shift
increased attention and resources to poverty-linked infrastructure. For example,
the share of rural roads in the aggregate road network nearly doubled to 32%
between 1965 and 1990, and rural poverty incidence declined from 53% in 1973
to 19% in 1989 (Lanjouw, 1995).
7.5. Credit
Because poor households usually do not possess any physical capital, credit pro-
vision is considered as an important means of alleviating poverty. Siamwalla
(1993) suggests three ways in which the poor may benefit from credit: (i) it may
enable them to cope with the consequences of poverty; (ii) it may act as a way
out of poverty; and (iii) even if credit is not made available directly to them but to
nonpoor households in the community, it may create activities with trickle-down
benefits to them.
Based on his analysis of rural credit market of developing Asia, Siamwalla
(1993) arrives at the following conclusions. First, the provision of credit ap-
pears to be ill-suited to deal with the consequences of poverty. More effective
mechanisms are those that enhance human capital (basic education, health and
nutrition), as mentioned above, and those that reduce the risks faced by the poor,
such as public works projects during periods of crop failure or food subsidies
in times of natural disasters. In these cases, public intervention is more in the
nature of a grant rather than a loan. Second, credit provision seems better suited
to assist the poor engage in gainful economic activities that will enable them
to move out of poverty. For instance, credit may be used to purchase new tools
or adopt new technology to make them more efficient in producing goods for
Ch. 45: Poverty in Developing Countries 1921
the market. Higher earnings then allow them to make profits and amortize their
loans.
Third, while like many other government policies and programs, general pro-
vision of credit benefits the richer households, the impact often does not stop
there but generates subsequent rounds of effects. These additional or "trickle-
down" impacts can benefit the poor by increasing the demand for their labor ser-
vices or goods they produce. Finally, credit should be regarded as playing merely
a supportive role. It can help in poverty alleviation only if there are profitable
investment opportunities and viable projects available to the poor. However, if
they do not exist, provision of credit is unlikely to alleviate poverty, especially if
it is chronic.
It is often suggested that the poor have to rely on informal sector credit at
exorbitant terms as they are shut off from formal sector credit owing to lack
of collateral. This traditional view about informal credit has been increasingly
challenged by economists (Bottomley, 1963; Adams et al., 1984; Ghate et al.,
1992) who suggest that the informal credit market is often highly competitive,
and therefore informal sector interest rates do not contain monopoly profits but
high administrative costs of small loans. Moreover, in many countries, the poorer
tenants receive loans from landlords in the context of an interlinked rural trans-
action where the tenants receive loans at below-market rates (Taslim, 1988) for
being involved in other land and/or labor market transactions with the landlord.
The exclusion from the formal sector credit that the poor often suffer because
of lack of collateral can be overcome by group-based lending, as exemplified
by the Grameen Bank in Bangladesh. This type of credit program has achieved
almost full recovery, although their impact on poverty alleviation may not be
commensurate with their success in loan recovery.
In sum, it is worth recapitulating the distinction between chronic and tran-
sient poverty. While transient poverty is often the outcome of market failures,
chronic poverty reflects low factor productivity (due to low level of technology)
and low levels of physical and human capital. It is therefore widely accepted that
while credit can redress transient poverty by addressing market failures, it cannot
address chronic poverty unless accompanied by new technological innovations
and/or investments in physical and human capital.4 5
45 This view has recently been contested by Hoff (1996) who argues that with imperfect risk markets and
constraints on borrowing for consumption needs, poor households can increase their tradeoffs in expected
incomes for reduced risks. Therefore, the policies which are designed for short-term poverty alleviation-such
as food-for-work programs-can have long- term impacts. These impacts include increased long-term average
earnings of the poor and reduction in inequality in asset holdings. While it is clear that credit can help the
transient poor from sliding into chronic poverty, it is not clear how it can help reduce chronic poverty.
1922 E.M. Perniaand M. G. Quibria
7.7. Targeting
Targeting is beset with other practical difficulties. 4 6 First, the efficiency rank-
ings of poverty programs are not unique but change with alterations in budgetary
outlays. Second, maximum impact on poverty is subject to different interpreta-
tions and will depend to a large extent on the poverty measure adopted. Third,
cost-effectiveness is often used as the principal criterion for "better" targeting-
excluding the nonpoor and including the poor as far as possible. However, "bet-
ter" targeting does not necessarily imply greater impact on poverty if extensive
participation entails high administrative costs and, hence, reduced cost effective-
ness. Finally, targeting can be accompanied by an erosion of political support for
antipoverty programs. This can happen if the focus of the program in question
changes from one of a universal coverage to that of a targeted one (Besley and
Kanbur, 1991). However, this political economy perspective is somewhat sim-
plified as the supporters for antipoverty programs cover a broader constituency
beyond the direct beneficiaries.
8. Concluding remarks
46 An overview of various approaches to targeting can be found in Besley and Kanbur (1991).
Ch. 45: Poverty in Developing Countries 1925
strong argument for broad-based, labor-intensive economic growth as opposed
to narrow, sectoral approaches to poverty alleviation.
It is often argued that women are poorer than men in developing countries
and, therefore, there is a need for gender-focused policies on grounds of equity
and efficiency. On grounds of equity, redressing gender discrimination is viewed
as tantamount to poverty alleviation. On grounds of efficiency, the issue of sub-
optimality in investment in human capital due to market failure is seen as more
serious for females than it is for males. However, gender equity requires that
females participate not only in economic but also in noneconomic dimensions
of life. A good deal of gender disparity stems from inequity in intrahousehold
allocation of resources. To the extent that intrahousehold allocation is governed
by societal norms and values, it cannot be directly influenced by the government.
Public policy can influence economic factors, through economic incentives and
fiscal measures, but not necessarily sociocultural norms and values. This explains
why gender inequity has tended to perpetuate itself.
A direct consequence of rapid population growth in developing countries
is that low income families are typically larger than high income ones, even
allowing for higher mortality among the former. However, because of the in-
teractive links between fertility and poverty, one cannot unambiguously say that
high fertility directly results in poverty. Rather, the evidence suggests that rapid
population growth may influence poverty by impinging on its correlates or char-
acteristics, such as inadequate human capital, lack of physical assets such as land
or tools, and low earnings, or, at the macrolevel, weak economic growth and
income inequality. Families with many children tend to save less and invest less
in physical assets and, more importantly, in the human capital of their children.
This means that poverty tends to be transmitted intergenerationally. Thus, poli-
cies to slow population growth-via raising women's education, reducing infant
mortality, and promoting suitable family planning programs-disproportionately
benefit the poor. Conversely, a nonexistent or weak population policy hurts the
poor more than the nonpoor.
The generally accepted view is that, in a long-run context, economic growth
performance is the most reliable determinant of poverty trends. For economic
growth to make a meaningful dent on poverty it has to be rapid and sustained. In
turn, rapid and sustained growth is feasible with economic policies that are mar-
ket based and outward oriented. For developing countries, such economic growth
is almost always labor intensive and broad-based, implying maximum utilization
of labor, the poor's principal, if not sole, asset. Not only is economic growth
critical for employment generation, it also raises the government's revenue and
1926 E.M. Perniaand M. G. Quibria
References
Adams. D.W., D. Graham and J. von Pische, eds. (1984). Undermining Rural Development with Cheap Credit
(Westview, Boulder, CO).
Agarwal, B. (1986). "Womnen, poverty and agricultural growth in India". Journal of Peasant Studies 13:165-
220.
Ahlburg, D.A. (1994), "Population growth and poverty", in: R. Cassen and contributors, Population and
Development: Old Debates, New Conclusions (Transaction Publishers, New Brunswick and Oxford).
Ahluwalia, M.S. (1976), "Inequality, poverty and development", Journal of Development Economics 3:307-
342.
Ahluwalia, M.S., N.G. Carter and H.B. Chenery (1979), "Growth and poverty in developing countries", Journal
of Development Economics 6:299-341.
Alderman H. and P. Gertler (1989), "The substitutability of public and private medical care for the treatment
of children's illnesses in Pakistan", Living Standards Measurement Study (LSMS) Working Paper 56 (The
World Bank, Washington, DC).
Alderman, H., J. Behrman, D. Ross and R. Sabot (1992), "The gender gap in cognitive achievement in a poor
rural economy", rmimeo (Williams College).
Alderman, H., J. Behrman, S. Khan, D. Ross, and R. Sabot (1993), "Public schooling expenditures in rural
Pakistan: Efficiently targeting girls and a lagging region", mimeo (Bryn Mawr College).
Altimir, O. (1982), 'The extent of poverty in Latin America". Staff Working Paper No. 522 (World Bank,
Washington. DC).
Arthur, W.B. and G. McNicoll (1978), "Samuelson, population and intergenerational transfers". International
Economic Review 19:241-246.
Asian Development Bank (1992), Asian Development Outlook 1992 (Oxford University Press, Hong Kong).
Asian Development Bank (1994), Asian Development Outlook 1994 (Oxford University Press, Hong Kong).
Atkinson, A.B. (1987). "On the measurement of poverty", Econometrica 55:749-764.
Atkinson, A.B. (1989), Poverty and Social Security (Harvester Wheatsheaf, London).
Atkinson, A.B. (1991), "Comparing poverty rates internationally: lessons from recent studies in developed
countries", The World Bank Economic Review 5:3-21.
Balisacan, A.M. (1994). "Chapter on the Philippines", in: M.G. Quibria, Rural Poverty in Developing Asia,
Vol. 2 (Asian Development Bank, Manila).
Balisacan, A.M. (1995). "Anatomy of poverty during adjustment: The case of the Philippines", Economic
Development and Cultural Change 44:33-62.
Banerjee, N. (1983), "Women and poverty: report on workshop", Economic and Political Weekly 18:1693-
1698.
Bardhan, K. (1993), "Women and rural poverty in some Asian cases", in: M.G. Quibria, Rural Poverty in Asia:
Priority Issues and Policy Options (Oxford University Press, Hong Kong).
Bardhan, P. (1995), "Research on poverty and development twenty years after, 'Redistribution with Growth' ",
in: M. Bruno and B. Pleskovic, eds., Annual World Bank Conference on Development Economics (World
Bank. Washington. DC).
Bardhan, P. (1996), "Efficiency, equity and poverty alleviation: policy issues in less developed countries",
Economic Journal 106:1344-1356.
Basu, A.M. (1989), "Is discrimination in food really necessary for explaining sex differentials in childhood
mortality?", Population Studies 43:193-210.
Basu, A.M. (1993), "How pervasive are sex differentials in childhood nutritional levels in South Asia?", Social
Biology 4:25-37.
Baxter, M. (1986), New Developments in Agricultural Extension (The World Bank, Washington, D.C.).
Behrman, J.R. (1988a), "Intrahousehold allocation of nutrients in rural India: are boys favored? Do parents
exhibit inequality aversion?", Oxford Economic Papers 40:32-54.
Behrman, J.R. (1988b), "Nutrition. health, birth order and seasonality: Intrahousehold allocation among
children in rural India", Journal of Development Economics 28:43-62.
Behrman, J.R. (1990), 'The action of human resources and poverty on one another", Living Standards
Measurement Study (LSMS) Working Paper 74 (The World Bank, Washington, DC).
Ch. 45: Poverty in Developing Countries 1929
Behrman, J.R. (1993), "Macroeconomic policies and rural poverty: issues and research strategies", in: M.G.
Quibria, Rural Poverty in Asia: Priority Issues and Policy Options (Oxford University Press, Hong Kong).
Behrman, J.R. and A.B. Deolalikar (1990), "The intrahousehold demand for nutrients in rural South India:
individual estimates, fixed effects and permanent income", Journal of Human Resources 24:665-697.
Behrman, J.R. and M.R. Rosenzweig (1994), "Caveat emptor: cross-country data on education and labor
force", Journal of Development Economics 44:147-171.
Behrman, J.R. and R. Schneider (1994), "An international perspective on schooling investments in the last
quarter century in some fast growing east and Southeast Asian countries", Asian Development Review 12:1-
50.
Behrman, J.R. and Z. Zhang (1994), "Gender issues and employment in Asia", Paper presented in the Asian
Development Bank Round Table on Employment Creation for Broad-Based Growth (Manila, Philippines).
Behrman, J.R., A.B. Deolalikar and B.L. Wolfe (1988), "Nutrients: impacts and determinants", The World
Bank Economic Review 2:299-320.
Besley, T. and R. Kanbur (1991), "Principles of targetting", in M. Lipton and Jacques Van Der Gaag, eds.,
Including the Poor (The World Bank, Washington, DC).
Bhagwati, J. (1988), "Poverty and public policy", World Development 16:539-555.
Bhalla, S. and P.S. Vashistha (1988), "Income distribution in India: a reexamination", in: T.N. Srinivasan and
P.K. Bardhan, eds., Rural Poverty in South Asia (Columbia University Press, New York).
Boserup, E. (1990), "Population, the status of women and rural development", in: G. McNicoll and M. Cain,
eds., Rural Development and Population (Oxford University Press, Oxford).
Bottomley, A. (1963), "The costs of administering private loans in underdeveloped rural areas", Oxford
Economic Papers 15:154-163.
Boulier, B.L. (1977), "Population policy and income distribution", in: C.R. Frank and R.C. Webb, eds., Income
Distribution and Growth in less Developed Countries (The Brookings Institution, Washington, DC).
Bourguignon, F., J. de Melo and A. Suwa (1991), "Distributional effects of adjustment policies: simulations
for archetype economies in Africa and Latin America", The World Bank Economic Review 5:339-366.
Bourguignon, E, W. Branson and J. de Melo (1992), "Adjustment and income distribution: a micro-macro
model for counterfactual analysis", Journal of Development Economics 38:17-39.
Cabaiero, T.A. (1978), "The shadow price of children in Laguna households", Philippine Economic Journal
17:62-83.
Cassen, R. (1976), "Population and development: a survey", World Development 4:785-830.
Cassen, R. (1994), "Population policy: a new consensus", Policy Essay No. 12 (Overseas Development
Council, Washington, DC).
Chen, L.C., E. Huq and S. DeSouza (1981), "Sex bias in the family allocation of food and health care in
Bangladesh", Population and Development Review 17:55-70.
Chen, S., G. Datt and M. Ravallion (1993), "Is poverty increasing in the developing world?", mimeo (The
World Bank, Washington, DC).
Chenery, H.L. et al. (1974), Redistribution with Growth (Oxford University Press for the World Bank and the
Institute of Development Studies, University of Sussex, London).
Coale, A.J. (1991), "Excess female mortality and the balance of sexes in the population: an estimate of the
number of 'missing females' ", Population and Development Review 17:517-523.
Cochrane, S. (1979), Fertility and education: What do we really know? (Johns Hopkins University Press,
Baltimore).
Cochrane, S., D. O'Hara and J. Leslie (1980), "The effects of education on health", Staff Working Papers, No.
405 (The World Bank, Washington, DC).
Cornia, G.A., R. Jolly and E Stewart (1987), Adjustment with a Human Face (Clarendon Press, Oxford).
Das Gupta, M. (1978), "Production relations and population: Rampur", Journal of Development Studies
14:177-185.
Das Gupta, M. (1987), "Selective discrimination against female children in rural Punjab, India", Population
and Development Review 13:77-100.
de Janvry, A. and E. Sadoulet (1989), "Investment strategies to combat rural poverty: a proposal for Latin
America", World Development 17:1203-1222.
1930 E.M. Perniaand M.G. Quibria
Deaton, A. (1989). Looking for boy-girl discrimination in household expenditure data", World Bank
Economic Review 3:1-15.
Deaton, A. (1995), "Inequality within and between households in growing and aging economies", in: M.G.
Quibria. Critical Issues in Asian Development: Theories. Experiences and Policies (Oxford University Press,
Hong Kong).
Deolalikar, A.B. (1991), "Intrahousehold allocation of health inputs and distribution of health outcomes among
Indonesian children", in: Robert McNamara Fellowship Program: 10th Anniversary Publication (The World
Bank, Washington DC: Economic Development Institute).
Deolalikar, A.B. (1994), "Special employment programs and poverty alleviation". Paper presented at the Asian
Development Bank's Development Round Table on Employment Creation for Broad-Based Growth, 12-14
October 1994, Manila, Philippines.
Deolalikar, A.B. and F.M. Pemia (1993). "Population growth in developing Asia: trends and implications for
economic growth, ecology and poverty" (Asian Development Bank, Manila).
Dev. S.M., K.S. Parikh and M.H. Suryanarayana (1994) "Chapter on India", in: M.G. Quibria. Rural Poverty
in Developing Asia, Vol. 1 (Asian Development Bank, Manila).
Directorate-General of Budget, Accounting and Statistics, Executive Yuam (1992). Social Indicators of the
Republic of China (Republic of China).
East-West Population Institute (1993), "Research program on priority health and population issues", Final
report submitted to the Asian Development Bank.
Evenson, R.E. (1984), "Population growth, infrastructural development, technology and welfare in rural north
India", Paper prepared for the IUSSP Seminar on Population. Food and Rural Development (New Delhi,
India).
Evenson, R.E. (1993), "Institutions and rural poverty in Asia", in: M.G. Quibria, Rural Poverty in Asia: Priority
Issues and Policy Options (Oxford University Press, Hong Kong).
Fields, G.S. (1990), "Labor market modeling and the urban informal sector: theory and evidence", in: D.
Tumham, B. Salome and Antoine Schwartz, eds., The infonnal sector revisited (OECD. Paris).
Fields, G.S. (1995a), "Income distribution in developing economies: conceptual, data, and policy issues in
broad-based growth". in: M.G. Quibria. Critical Issues in Asian Development: Theories Experiences and
Policies (Oxford University Press, Manila).
Foster, A. and M.R. Rosenzweig (1993a), "Information flows and distribution in the labor markets in low-
income countries", World Bank Economic Review and World Bank Research Observer. Proceedings of the
World Bank Annual Conference on Development Economics 1992. pp. 173-203.
Foster, A. and M.R. Rosenzweig (1993b), "Information, learning. and wage rates in low-income rural areas",
Joumal of Human Resources 28:759-790.
Foster, A. and M.R. Rosenzweig (1994a). "A test for moral hazard in the labor market: contractual
arrangements, effort, and health", Review of Economics and Statistics 76:213-217.
Foster, A. and M.R. Rosenzweig (1994b), "Comparative advantage, information and the allocation of workers
to tasks: evidence from an agricultural labor market", mimeo (University of Pennsylvania).
Foster. J. and A.F. Shorrocks (1988), "Poverty orderings", Econometrica 56:173-77.
Foster, J.E., J. Greer and E. Thorbecke (1984). "A class of decomposable poverty measures", Econometrica
52:761-766.
Fuchs, V.(1965), 'Toward a theory of poverty", Task force economic growth and opportunity: The concept of
poverty (U.S. Chamber of Commerce, Washington, DC).
Gaiha. R. and A.B. Deolalikar (1993), "Persistent. expected and innate poverty: estimates for semi-arid rural
south India, 1975-1984", Cambridge Journal of Economics 18:409-422.
Gannicott, K. (1986), "Woman, wages and discrimination: some evidence from Taiwan", Economic Develop-
ment and Cultural Change 39:721-730.
Gertler, P. andO. Rahman (1994), "Social infrastructure and urban poverty", in: E.M. Pemia, Urban Poverty
in Asia: A Survey of Critical Issues (Oxford University Press, Hong Kong).
Ghate, P., A. Das Gupta, M. Lamberte. N. Poapongsakom, D. Prabowo and A. Rahman (1992), Informal
finance: Some findings from Asia (Asian Development Bank: and Oxford University Press, Manila).
Glover, D. and K. Kusterer (1990), Small farmers, big business: Contract farming and rural development (St.
Martins, New York).
CIt 45: Poverty in Developing Countries 1931
Greenwood, M.J. (1975), "Research on internal migration in the U.S.: a survey", Journal of Economic
Literature 13:397-433.
Gunatilleke. G., M. Perera, R.A.M.C. Wanigaratne, R.E. Fernando, W.D. Lakshman, J.K.M.D. Chandrasiri and
R.D. Wanigaratne (1994), "Chapter on Sri Lanka", in: M.G. Quibria, Rural Poverty in developing Asia, Vol.
1 (Asian Development Bank. Manila).
Haddad, L. and R. Kanbur (1990), "How serious is the neglect of intrahousehold inequality". Economic Journal
100:866-881.
Harris, B. (1990), "The intrafamily distribution of hunger in south Asia", in: J. Dreze and A. Sen, eds., The
Political Economy of Hunger, vol. 1 (Clarendon Press, Oxford).
Hoff, K. (1996) "Comment on 'Political Economy of Alleviating Poverty Theory and Institutions. by Timothy
Besley' ", in: M. Bruno and B. Pleskovic, eds., Annual World Bank Conference on Development Economics
(World Bank, Washington DC).
Hossain, M., R. Mannan, H.Z. Rahman and B. Sen (1994), "Chapter on Bangladesh", in: M.G. Quibria, Rural
Poverty in Developing Asia, Vol. 1 (Asian Development Bank, Manila).
Jamison, D. and L. Lau (1982), Farmer education and farm efficiency (Johns Hopkins University Press,
Baltimore).
Jimenez, E. (1987) Pricing Policy in the Social Sectors: Cost Recovery for Education and Health in Developing
Countries (The Johns Hopkins University Press for the World Bank, Baltimore).
Jimenez, E. (1992), "Human and physical infrastructure: public investment and pricing policies in develop-
ing countries", Paper prepared for the Asian Development Bank's Conference on Development Economics
(Manila, Philippines).
Kakwani, N.C. (1980), "On a class of poverty measures", Econometrica 48:437-446.
Khan, A.R. (1984), "Population growth and access to land: an Asian perspective", Paper prepared for the
IUSSP Seminar on Population, Food and Rural Development (New Delhi, India).
Khandker, S. (1996), " Grameen Bank: impact, costs, and program sustainability", Asian Development Review,
14:97-130.
Kikuchi, M. and Y. Hayami (1980), "Inducements to institutional innovations in an agrarian community",
Economic Development and Cultural Change 29:21-36.
Klasen, S. (1994), "'Missing women' reconsidered", World Development 22:1061-1071.
Krongkaew, M., P. Tinakom and S. Suphachalasai (1994), "Chapter on Thailand", in: M.G. Quibria, Rural
Poverty in Developing Asia, Vol. 2 (Asian Development Bank, Manila).
Kuznets, S. (1976), "Demographic aspects of the size distribution of income: an exploratory essay", Economic
Development and Cultural Change 25:1-94.
Kynch, J. and A.K. Sen (1983), "Indian women: well-being and survival", Cambridge Journal of Economics
7:363-380.
Lipton, M. and A. de Haan (1997), "Poverty in emerging Asia", Background paper for Emerging Asia: Changes
and Challenges (Asian Development Bank).
Lipton, M. and M. Ravallion (1995), "Poverty and Policy", in: J. Behrman and T.N. Srinivasan, eds., Handbook
of Development Economics (Elsevier Science BV).
Lanjouw, P. (1995), "Infrastructure: a ladder for the poor", Finance and Development 32:33-35.
Mathur, O.P. (1994), "The state of India's urban poverty", Asian Development Review 12:32-67.
Mazumdar, D. (1987), "Rural-urban migration in developing countries", in: E.S. Mills, ed., Handbook of
Regional and Urban Economics (Elsevier, Amsterdam; also in World Bank Reprint Series No. 422).
McNicoll, G. (1984), "Consequences of rapid population growth: overview and assessment", Population and
Development Review 10:177-240.
Meyer, C.A. (1991), "A hierarchy model of associative farming", Journal of Development Economics 34:371-
383.
Miller, B.D. (1981), The Endangered Sex (Cornell University Press, New York).
Mills, E. and C. Becker (1986), Studies in Indian Urban Development (Oxford University Press for the World
Bank, Oxford).
Mills, E. and E.M. Pernia (1994), "Introduction and overview", in: E.M. Pernia, Urban Poverty in Asia: A
Survey of Critical Issues (Oxford University Press, Hong Kong).
1932 E.M. Pernia and M.G. Quibria
Mueller, E. (1972), "Economic cost and value of children: conceptualization and measurement", in: J.T.
Fawcett, ed., The Satisfactions and Costs of Children: Theories, Concepts and Methods (East-West Center,
Honolulu).
Orshansky, M. (1965), "Who's who among the poor: a demographic view of poverty", Social Security Bulletin
28:3-32.
Otsuka. K. (1993), "Land tenure and rural poverty", in: M.G. Quibria, Rural Poverty in Asia: Priority Issues
and Policy Options (Oxford University Press, Hong Kong).
Penny, D.H. and M. Singarimbun (1973), "Population and poverty in rural Java: some economic arithmetic
from Sriharjo", Mimeograph 41 (Comell International Agricultural Development, Ithaca, New York).
Pernia, E.M. (1982), "Micro-level implications of population growth", in: A.N. Herrin, V.P. Paqueo and E.M.
Pernia, eds., Essays on the Economics of Fertility, Population Growth and Public Intervention in a De-
veloping Country: The Philippines, University of the Philippines School of Economics Discussion Paper
8212.
Pernia, E. M., ed. (1994), Urban Poverty in Asia: A Survey of Critical Issues (Oxford University Press for the
Asian Development Bank, Hong Kong).
Pernia, E. M. and S. Alabastro (1997), "Aspects of urban water and sanitation in the context of rapid
urbanizaton in developing Asia", Economic Staff Paper No. 56 (Asian Development Bank, Manila).
Pernia, E. M. and C. Paderanga (1981), "Philippines: the impact of national policy on urban systems", in: M.
Honjo, ed., Urbanization and Regional Development, pp. 149-177 (Maruzen Asia for UNCRD, Nagoya,
Japan, Singapore).
Pitt, M. and S. Khandker (1994), "Household and intrahousehold impacts of the Grameen and similar targeted
credit programs in Bangladesh", mimeo (The World Bank, Washington, DC).
Pitt, M., M.R. Rosenzweig and Md. N. Hassan (1990), "Productivity, health and inequality in the intrahouse-
hold distribution of food in low-income countries", American Economic Review 80:1139-1156.
Potter, J.E. (1979), "Demographic factors and income distribution in Latin America", Proceedings of the Con-
ference on Economic and Demographic Change: Issues for the 1980s, Helsinki, 1978, Liege: International
Union for the Scientific Study of Population I.
Psacharopoulos, G. (1985), "Returns to education: a further international update and implications", Journal of
Human Resources 20:583-604.
Psacharopoulos, G. (1988), "Education and development: a review", The World Bank Research Observer 3:99-
116.
Psacharopoulos, G., S. Morley, A. Fiszbein, H. Lee and W. Wood (1995), "Poverty and income inequality in
Latin America", Review of Income and Wealth 4:245-264.
Quibria, M.G. (1982), "An analytical defense of basic needs: the optimal savings perspective", World
Development 10:285-291.
Quibria, M.G. (1991), "Understanding poverty: an introduction to conceptual and measurement issues", Asian
Development Review 9:90-112.
Quibria, M.G., ed. (1993), Rural Poverty in Asia: Priority Issues and Policy Options (Oxford University Press
for the Asian Development Bank, Hong Kong).
Quibria, M.G., ed. (1994a), Rural Poverty in Developing Asia, vol. 1 (Asian Development Bank, Manila).
Quibria, M.G., ed. (1994b), Rural Poverty in Developing Asia, vol. 2 (Asian Development Bank, Manila).
Quibria, M.G. (1995a), Critical issues in Asian Development: Theories, Experiences and Policies (Oxford
University Press for the Asian Development Bank, Hong Kong).
Quibria, M.G. (1995b), "Gender and poverty: issues and policies (with special reference to Asian developing
countries)", Journal of Economic Surveys (forthcoming).
Quibria, M.G., and T.N. Srinivasan (1993), "Introduction", in: M.G. Quibria, Rural Poverty in Asia: Priority
Issues and Policy Options (Oxford University Press, Hong Kong).
Rashid, S. and M.G. Quibria (1995), "Is land reform passe? With special reference to Asian agriculture", in:
M.G. Quibria, Critical Issues in Asian Development: Theories, Experiences and Policies (Oxford University
Press for the Asian Development Bank, Hong Kong).
Ravallion, M. (1993), Poverty Comparisons. Fundamentals of Pure and Applied Economics, vol. 56 (Harvard
Academic Press, Chur Switzerland).
Ch. 45: Poverty in Developing Countries 1933
Ravallion, M. and Huppi, M. (1991), "Measuring changes in poverty: a methodological case study of Indonesia
during an adjustment period", World Bank Economic Review 5:57-84.
Ravallion, M., G. Datt and D. Van De Walle (1991), "Quantifying absolute poverty in the developing world ",
Review of Income and Wealth 37:345-361.
Ravallion, M., G. Datt and S. Chen (1992), New Estimates of Aggregate Poverty Measures for the Developing
World, 1985-89 (The World Bank, Population and Human Resources Department, Washington, DC).
Rawls, J. (1971), A Theory of Justice (Harvard University Press, Cambridge, MA).
Renaud, B. (1981), National Urbanization Policy in Developing Countries (Oxford University Press, Oxford).
Robinson, M. and D. Snodgrass (1987), "The role of institutional credit in Indonesia's rice intensification
program", mimeo, Development Discussion Paper No. 248 (Harvard Institute for International Development,
Cambridge, MA).
Rodgers, G. (1983), "Population growth, inequality and poverty", Paper prepared for the UN Expert Group on
Population, Resources, Environment and Development, Geneva.
Rodgers, G. (1984), Poverty and Population: Approaches and Evidence (International Labour Organisation,
Geneva).
Roumasset, J.R. and J. Smith (1981), "Population, technological change and landless workers", Population and
Development Review 7:401-420.
Rowntree, B.S. (1901), Poverty: A Study of Town Life (Macmillan, London).
Safilos-Rothschild, C. (1990), "Gender and rural poverty in Asia", Paper prepared for the International Seminar
on Poverty Alleviation through Agricultural Projects, Organized by the Economic Development Institute of
the World Bank, the Asian Development Bank and the Centre on Integrated Rural Development for Asia and
the Pacific.
Saito, K. and D. Spurling (1991), Agricultural Extension and Research for Women Farmers (The World Bank,
Washington, DC).
Sawhill, I.V. (1988), "Poverty in the U.S.: why is it so persistent?", Journal of Economic Literature 26:1076.
Schultz, T.P. (1989), "Women and development: objectives, frameworks and policy interventions", mimeo
(Yale University).
Scott, W. and N.T. Matthew (1983), Levels of Living and Poverty in Kerala (United Nations Research Institute
for Social Development, Geneva).
Sen, A.K. (1976), "Poverty: an ordinal approach to measurement", Econometrica 44:219-231.
Sen, A.K. (1977), "On weights and measures: informational constraints in social welfare analysis", Economet-
rica 45:1539-1572.
Sen, A.K. (1981), Poverty and Famines (Clarendon Press, Oxford).
Sen, A.K. (1985), Capabilities and Commodities (North Holland, Amsterdam).
Sen , A.K. (1988), "Family and food: sex bias in poverty", in: P. Bardhan and T.N. Srinivasan, eds., Rural
Poverty in South Asia (Columbia University, New York).
Sen, A.K. (1990), "More than 100 million women are missing", New York Review of Books 37:61-66.
Sen, A.K. (1992), Inequality Reexamined (Clarendon Press, Oxford).
Sen, A.K. and S. Sengupta (1983), "Malnutrition of rural children and the sex bias", Economic and Political
Weekly 18.
Senauer, B., D. Sahn and H. Alderman (1986), "Effects of the value of time on food consumption patterns in
developing countries: evidence from Sri Lanka", American Journal of Agricultural Economics 86:920-927.
Senauer, B., M. Garcia and E. Jacinto (1988), "Determinants of the intrahousehold allocation of food in the
rural Philippines", American Journal of Agricultural Economics 70:170-180.
Siamwalla, A. (1993), "Rural credit and rural poverty", in: M.G. Quibria, Rural Poverty in Asia: Priority Isues
and Policy Options (Oxford University Press for the Asian Development Bank, Hong Kong).
Sirageldin, I.A. (1975), "The demographic aspects of income distribution", in: W. Robinson, ed., Population
and Development Planning (The Population Council, New York).
Sjaastad, L. (1962), "The costs and returns of human migration", Journal of Physical Economies 70:80-93.
Smith, W.S. ed. (1970), Bernard Shaw's Plays (W.W. Norton &Co., Inc., New York).
Srinivasan, T.N. (1992), "Income distribution and the macroeconomy: some conceptual and measurement
issues", Journal of Philippine Development 19:1-25.
1934 E.M. Perniaand M.G. Quibria
Srinivasan, T.N. (1993), "Rural poverty: conceptual, measurement and policy issues", in: M.G. Quibria. Rural
Poverty in Asia: Priority Isues and Policy Options (Oxford University Press for the Asian Development
Bank, Hong Kong).
Stone, R. (1954), The Measurement of Consumers' Expenditure and Behavior in the United Kingdom. 1920-
38, vol. 1 (Cambridge University Press, Cambridge).
Strauss, J., P. Gertler. O. Rahman, and K. Fox (1993), "Gender and life cycle differentials in the patterns and
determinants of adult health". Joumal of Human Resources 28:791-837.
Strauss, J. and D. Thomas (1995). "Human resources: empirical modeling of household and family decisions".
in: J. Behrman and T.N. Srinivasan, eds., Handbook of Development Economics, vol. 3A (North Holland
Publishing Company, Amsterdam).
Streeten, P. (1990). "Poverty: concepts and measurement", Bangladesh Development Studies 18:5-6.
Tabatabai, H. (1993), The Incidence of Poverty in Developing Countries: An ILO Compendium of Data
(Intemational Labour Organisation, Geneva).
Taslim, M.A. (1988). "Tenancy and interlocked markets: issues and some evidence", World Development
16:655-666.
Taylor, G. and R. Faruque (1983), Child and Maternal Health Services in Rural India (Johns Hopkins
University. Baltimore).
Tjondronegoro, S.M.P., I. Soejono and J. Hardjono (1994), "Chapter on Indonesia", in: M.G. Quibria, Rural
Poverty in Developing Asia, Vol. 2 (Asian Development Bank, Manila).
Todaro, M.P. (1976), Internal Migration in Developing Countries: A Review of Theory, Evidence, Methodol-
ogy and Research Priorities (Intemational Labour Organisation, Geneva).
Tolley, G. S. and V. Thomas, eds. (1987), The Economics of Urbanization and Urban Policies in Developing
Countries (World Bank. Washington. DC).
United Nations (1992), World Population Monitoring 1991. Population Studies No. 126. Sales No. E.92.XIII.2.
United Nations (1994), Population Distribution and Migration, Paper prepared for the UN International
Conference on Population and Development (Cairo, Egypt).
UNDP (1995), Human Development Report 1995 (Oxford University Press, Oxford).
Valenzona, R. (1976). "Poverty measurement and nutrition", University of the Philippines School of Eco-
nomics Discussion Paper 76-27.
Watts, H.W. (1991), "An economic definition of poverty", in: D.P. Moynihan, ed., On Understanding Poverty
(Basic Books, New York).
Wegelin, E. (1994), "Urban shelter, municipal services, and the poor", in: E.M. Pernia.
World Bank (1994), Social Indicators of Development 1994.
World Bank (1995), Social Indicators of Development 1995.
World Bank (1980), World Development Report 1980 (Oxford University Press, Oxford).
World Bank (1984), World Development Report 1984 (Oxford University Press, Oxford).
World Bank (1988), World Development Report 1988 (Oxford University Press, Oxford).
World Bank (1990). World Development Report 1990 (Oxford University Press, Oxford).
World Bank (1992), World Development Report 1992 (Oxford University Press, Oxford).
World Bank (1993), World Development Report 1993 (Oxford University Press, Oxford).
World Bank (1994), World Development Report 1994 (Oxford University Press, Oxford).
World Bank (1995), World Development Report 1995 (Oxford University Press, Oxford).
World Bank (1996), World Development Report 1996 (Oxford University Press, Oxford).
Chapter46
URBAN TRANSPORTATION
KENNETH A. SMALL
JOSI A. G6MEZ-IBANEZ*
Harvard University
Contents
1. Introduction 1939
2. Highway infrastructure costs and congestion 1940
2.1. Congestion severity and growth 1940
2.2. Financing new infrastructure 1944
2.3. Other supply-side measures 1946
2.4. Demand-side measures for congestion relief 1948
2.4.1. Land use controls 1949
2.4.2. Public transit 1950
2.4.3. Transportation demand management (TDM) 1951
2.4.4. Pricing measures 1952
2.5. Private toll roads 1955
2.6. Conclusion 1957
3. Motor vehicle air pollution 1958
3.1. Damage estimates 1958
3.2. The relationship between congestion and pollution 1960
3.3. Technological controls 1960
3.4. Transportation control measures 1962
3.5. Pricing policies 1963
3.6. Conclusion 1964
4. Motor vehicle accidents 1965
4.1. Magnitude of accident costs 1965
* We are grateful to Pia Koskenoja for research assistance, and to David Anderson, Graham Crampton,
Amihai Glazer, Phil Goodwin, Yoshitsugu Kanemoto, Herbert Mohring and Clifford Winston for comments
on an earlier draft.
Handbook of Regional and Urban Economics. Edited by E.S. Mills and P Cheshire
© 1999 Elsevier Science B. V All rights reserved.
1937
1938 K.A. Small and J.A. Gdmez-bdgiez
Abstract
The capacity of urban road and street systems is determined in large part by the
desire to accommodate peak period flows. Investment decisions thus involve a
tradeoff between infrastructure and congestion costs, and so it makes sense to
discuss infrastructure expansion and congestion as a single policy issue.
I Total delay, extra fuel consumption. and the ratio of registered vehicles to population are taken from
Schrank et al. (1993: Tables 11, 16, 17). We have applied a per person value of time of $6.00 per hour, based
on the consensus of studies as reviewed by Small (1992: pp. 43-44) that commuting time is valued at roughly
half the wage rate, and multiplied by an assumed vehicle occupancy of 1.1,
Ch. 46: Urban Transportation 1941
studied by Schrank et al. (1993: Table 6), all but three showed an increase be-
tween 1982 and 1990 in a "congestion index" that reflects ratios of daily volume
to capacity on selected road segments; between 1986 and 1990, the average
area suffered an estimated 18% increase in vehicle hours of delay (their Table
14). Anecdotal evidence abounds for increases in congestion on specific roads. 2
London, famous for its all-day traffic congestion, experienced steady declines
in average speeds from 1968 through 1990, although this appears to have been
reversed by the recession of the early 1990s ("Traffic speeds in central and outer
London", 1995).
Speeds faced by the average person, however, have not necessarily declined.
US data suggest that average trip times and speeds have changed little since
the mid-1970s (Gordon et al., 1991; Gordon and Richardson, 1994). Journey-to-
work statistics from the US Census show, for example, that the mean commuting
time in the largest 15 metropolitan areas increased only slightly, from 26.0 to
26.6 minutes, between 1980 and 1990. 3 This finding need not conflict with the
findings of worsening traffic on specific roads:
... not only is there no contradiction but the two phenomena are causally
related. Rational commuters will, sooner or later, seek to escape congestion
by changing the location of their homes and/or their jobs .... The process is
facilitated by the decentralizing location decisions of firms seeking to move
closer to suburban labor pools. (Gordon et al., 1991: p. 419)
2 See for example Meyer (1994), European Conference of Ministers of Transport (1995: pp. 72-76), Levy
(1994) and Daniere (1995).
3 Rosetti and Eversole (1993: Tables 2-1, 4-13). These figures exclude Boston's mean travel time, that
increased from 23.4 to 24.2 minutes. The average masks substantial differences: in the largest area (New York:
The mean commute time decreased by 7.7%, whereas in the second largest (Los Angeles) it increased by
11.9%; it also increased in each of the other 13 areas, by an average of 5.4%. Figures are for the Consolidated
Metropolitan Statistical Area if there is one, otherwise for the Metropolitan Statistical Area. For recent com-
parisons, these census data may be more reliable than those from the National Personal Transportation Survey
(NPTS), since the survey methodology for the latter changed between the latest two surveys (1983 and 1990),
possibly causing the 1990 data to be overweighted with newer cars (Lave, 1996).
1942 K.A. Small and JA. Gdmez-lbicez
4 US Census Bureau (1998: Table 1020) and the American Automobile Manufacturers Association (1997:
p. 65).
5 Similarly, Pucher (1995a: p. 101) reports that during the decade of the 1980s private car ownership per
capita increased faster in almost every European country than in the US, and fastest in those starting with the
lowest levels. Pucher also reports results of a survey of 93 European cities by Sharman and Dasgupta (1993),
indicating that private car traffic increased by 30-35% per decade in the 1970s and 1980s.
6 See Winston and Bosworth (1992: Fig. 8-2) and Gramlich (1994: Fig. 3).
Ch. 46: Urban Transportation 1943
-a
Cl' IOO aq I
a 00-Cl Clnla> ao rd a
a
ol l a I CA a d
tr- I a
m
a -.
> oa
t t
r-
C
N C
-
a a a> N C
:t: ~F
C= m m h0 = =o m o C gl 0
a
a
o
ot I vi ce 0 o0 -
a
vn
n r o-
a o. o o . .
.o 0 - ma>l c Cl,
C C lj r C-
ln o mm m
N
a> o 'a
,c In
IC
o
a> Cl C, o
- - a>~ r- 00 'D O
Cl C a C l a Cl ON 00
r-
Cil n00C a> N
\O
a
. '6
a
C4
a a
e-
3a aC
>
,t Vn m CD C c l, tCl r-'a oCo
Cl > a> on m o
-a It a> on Cl 'a
aIa o
.a Cl o o a > ' cn a>
bmi zt C C3) a >m al a C
-I- 10 -i0 `oC
a a>
Ch
a a
z a : Cl ~ ~ C
a>=
a on
an a> - 'a C ,-I t n m C
on
oo oo0 Cl t 00 Cl 00
r Ol CO - m
0m aOD on - -t C Cl
Cl rl
-,- - m on -a
LI) DO
a
'I
0 0 a O
o
<ao aa 0
a
° 5a 'aa -aa N
a~cs
C am o n o n a
a a 4 a
a
aO.
1944 K.A. Small and J.A. G6mez-lbafez
Table 2
Per capita passenger car ownership, passenger travel, and road network: the US and 17 European nations
a Source for population: US Census Bureau (1994: Table 2). Source for other figures: US Federal Highway
Administration (various years: Tables MV-1, MV-201A, VM-1, HM-10).
b Member nations of the European Conference of Ministers of Transport (ECMT): Belgium, Denmark, West
Germany, Greece, France, Ireland, Italy, Luxembourg, the Netherlands, the UK, Spain, Portugal, Norway,
Sweden, Switzerland, Austria, and Finland. For passenger travel, the figures exclude Greece, Ireland and
Luxembourg; and the UK includes only Great Britain (populations are adjusted accordingly). Source for
population: UN, Demographic Yearbook (1975, 1985, 1992); plus Europa Publications, The Europa World
Year Book (1992), for Great Britain. Source for registered passenger cars: ECMT (1993: Table 3-3-1), for
1970-1980; Banister and Berechman (1993: p. 16) for 1990. (Banister and Berechman's data are from the
same source but contain more recent years.) Source for km of road: ECMT (1993: Table 3-2-1); 1989 data are
used for 1990. Source for passenger travel: ECMT (1995: p. 30).
by 1995 (US Census Bureau, 1997: Tables 14, 1009). However, congestion also
depends on local population growth and on the amount that each vehicle is driven.
Trends in usage per vehicle are unclear: in the US, the reported annual distance
traveled per vehicle rose from the years 1950 to 1970, declined in 1980, then
rose again by 23% to a peak value of 18,700 km in 1993 (American Automobile
Manufacturers Association, 1997: p. 64). In other nations, there is more room for
continued growth in car ownership since current levels are well below the natural
saturation level of one per potential driver.
Road building programs have not kept pace with vehicle traffic in many devel-
oped countries due to a combination of political opposition and financial con-
Ch. 46: Urban Transportation 1945
straints. Much of the traffic growth has occurred in densely populated urban-
ized areas, where road construction is not only very expensive but increasingly
opposed by environmentalists and by residents of neighborhoods along the pro-
posed right of way. Furthermore, past expansion sometimes relied on taxes whose
bases are not growing commensurately with traffic.
In Europe, for example, fuel tax revenues have more than covered highway
expenditures in most nations (Pucher, 1995a), but physical constraints and politi-
cal opposition prevent the expansion of highways in many congested cities. In the
US, a major problem has been the historic tie between highway investment funds
and fuel tax revenues. Table 3 shows the average US fuel tax rates in nominal and
real terms. Since 1960, the real tax rate per vehicle mile has been cut in half as
state legislatures have raised nominal tax rates by less than inflation, and as the
fleet average fuel efficiency has grown.
One response has been to finance new roads at the time of land development,
in the form of fees or exactions charged to the land developers. The rationale
Table 3
Average motor fuel tax rates: US
a Source: US FHWA (1995: Table FE-101A). For 1985 and later, rate shown is a weighted average of different
rates applied to gasoline and diesel fuel. The weights are aggregate fuel consumption, from the US FHWA
(1990, 1995: Table MF-2); and the US FHWA, Highway Statistics, Summary to 1985 (1987: p. 9, Table MF-
221) (US Government Printing Office, Washington, DC). For 1990, during which the tax rate went up in
December, the rate shown is the average by month, i.e., it is 11/12 times the initial rate plus 1/12 times the final
rate.
b State gasoline and diesel tax revenues (American Automobile Manufacturers Association, 1993: p. 78; 1996:
p. 82) divided by motor vehicle fuel consumed (American Automobile Manufacturers Association, 1997: p.
64).
c Previous column divided by the consumer price index relative to that for 1995. Price index is from US Census
Bureau (1997: Table 752).
d Previous column divided by fleet average fuel efficiency in miles per gallon for cars, from the US Census
Bureau (1998: Table 1053; 1981: Table 1092) (the latter used for 1960 only).
1946 K.A. Small anaJ.A. G6me-lb&iez
but generally not encouraging. Giuliano et al. (1990) estimate, for example, that
only about 4% of peak-period workers using a new HOV lane in Orange County,
California, joined carpools because of the new lane.
TSM measures may include other direct controls on vehicle movements. Ex-
amples include controlled left turn lanes, oneway streets, limitations on curbside
parking, and partial or complete bans on vehicles in designated parts of cities.
Such vehicle bans are used successfully in many European cities, especially in
historic areas with narrow streets (Hass-Klau, 1990). The evidence from Ger-
many and the UK shows that they substantially increase pedestrian volumes and
that total retail trade, while sometimes dropping for a time, remains healthy and
often increases over a period of a few years (Hass-Klau, 1993).
Recently, a great deal of attention and money have been lavished on intelligent
transportation systems (ITS), which attempt to make road travel more efficient
using new information technology. Some of these measures are making modest
improvements in people's trips by helping them find their way, or by routing
them around areas with especially high congestion. However, simulation studies
frequently show that once the market penetration of information systems exceeds
a critical fraction, further dissemination has little effect on aggregate costs and
may even raise them by facilitating routing decisions that minimize private but
not social cost (Ben-Akiva et al., 1991; Mahmassani and Jayakrishnan, 1991;
Arnott et al., 1991; Emmerink et al., 1996). More ambitious efforts aimed at
increasing highway capacity by automating vehicle spacing are still preliminary
and appear to be very expensive.
Managing the existing infrastructure more efficiently to increase its person-
carrying capacity, then, seems a promising direction for traffic policy, but one
with only limited potential beyond current practice.
Increasing the supply of highway infrastructure is only one way of dealing with
supply-demand imbalance. Another way is to manipulate demand. Terminology
is not uniform, but such demand-side measures are often divided into four cat-
egories: land use controls, public transit enhancement, transportation demand
management and pricing measures.
Ch. 46: Urban Transportation 1949
2.4.1. Land use controls
It is natural to look to land use policies as a starting point for influencing travel
demand since the demand for transportation is derived from the spatial separation
of land uses.
Many urban planners have linked congestion to local imbalances between em-
ployment and housing, arguing that high-density employment centers often lack
nearby housing suitable for their workers (Cervero, 1989a, b; Nowlan and Stew-
art, 1991). For this reason, some cities are trying to achieve a closer balance by
restricting suburban sprawl, by channeling growth into outlying satellite towns,
or by dispersing job growth to locations where residential sprawl has already
occurred.
In the US, however, the ratio of jobs to housing is reasonably balanced in
most cities, and the imbalances that do occur tend to be eliminated over time
by market forces (Giuliano, 1991). Indeed, Small and Song (1992) show that in
the Los Angeles region, employment and housing are sufficiently interspersed
so that the average commute in 1980 could be reduced by two-thirds without
changing land use patterns, just by reallocating workers to different houses. This
suggests that factors other than availability of housing are causing people to live
at some distance from their jobs. Such factors probably include public services,
amenities, racial segregation, frequent job searches (Rouwendal and Rietveld,
1994; Crampton, 1997) and the complexities of job and housing choices by two-
worker households (Kim, 1995).
Comparable calculations for Tokyo reveal a pattern of residential choices
much more constrained by the available locations of housing (Merriman et al.,
1995). Tokyo is a residentially dispersed metropolitan area with strong job con-
centrations in the central areas. This land use pattern, combined with high public
transit usage, produces a much greater average commuting time than in any US
city. Furthermore, two-worker families are less common in Japan than in most
developed western nations. In such circumstances, the journey to work is more
likely to be the dominant consideration in the choice of housing location.
Is there a directly observable link from land use patterns to the amount of road
traffic? Newman and Kenworthy (1989, 1991) use data from 32 large metropol-
itan areas in developed countries to argue that the major determinant of auto-
mobile usage is in fact the density of development. They show that metropolitan
areas with residential densities above 30-40 persons per hectare, such as those
in western Europe and Japan, have much lower levels of automobile usage and
rely more heavily on public transportation than their lower density counterparts
in Canada, Australia and the US. However, they do not control for the effects of
1950 K.A. Small and J.A. G6dnez-lbdiiez
other variables, such as income levels, that might influence the mix of automobile
and transit usage (Gordon and Richardson, 1989; G6mez-Ibfiez, 1989).
But even if Newman and Kenworthy (1989, 1991) are right, does their finding
lead to useful policy conclusions? High urban densities in Europe and east Asia
are probably due to a scarcity of habitable land relative to total population, and
to the fact that many of their cities experienced their greatest population growth
before the automobile era was fully upon them. More recent development on the
peripheries of European cities has densities and patterns of transport use quite
similar to those in the US. Thus, it is unclear that international differences in
urban densities are primarily policy-driven.
Furthermore, once low density development becomes common it is extremely
difficult to reverse. Downs (1992) systematically examines the prospects for us-
ing land use regulation to reduce transportation demand in US metropolitan areas.
Downs defines a prototype city, complete with suburban subcenters, and applies
empirical travel propensities to model various land use changes. He finds only
a tiny effect on travel patterns from any regional land use policies that are even
remotely politically feasible. In part, this is due to the fragmented governmental
structure in US metropolitan areas and the ferocity with which local governments
protect their powers of land use planning.
Finally, most advocates of using land use policies to influence transporta-
tion choices ignore the information that land markets are providing about eco-
nomic efficiency. Urban job concentrations have deep roots in agglomeration
economies, which in turn underlie the economic functioning of large cities. Al-
though there may well be market failures in land development that prevent opti-
mal land use, heavy-handed regulation that subverts land markets could exact a
high cost. There seems to be little empirical research on how business efficiency
is affected by policies aimed at limiting the spatial concentration of industry.
Given the interest in such policies, their widespread use outside of North Amer-
ica, and their potentially far-reaching effects, such research could have a high
payoff.
2.4.4. Pricingmeasures
Most of the policies discussed above elicit only small changes in behavior be-
cause they change the relative advantages to travelers of different modes, times,
or routes only slightly. Even HOV lanes, potentially a big time saver, do not
perform as well as hoped because of the additional time and loss in flexibility
involved in sharing rides, especially where trips are dispersed and work schedules
are flexible. The policies also do nothing to reduce the attractiveness of peak-
hour vehicle use for categories of trips not targeted directly, and thus tend to be
undermined by latent demand.
Pricing measures, by contrast, can significantly deter trips of all types, while
permitting flexible travel patterns. They encourage people to use ride-sharing or
mass transit when feasible, but allow them to switch to the car at other times.
A variety of forms of pricing have been proposed and tested, including raising
parking charges or charging tolls for use of congested roads.
Parking charges have been studied the most and have considerable potential
(Giuliano and Small, 1995). Shoup (1994) reports that parking is free for 99% of
all automobile trips in the US. His research on seven case studies from Los An-
geles, Washington (DC) and Ottawa, demonstrates that charging realistic market
Ch. 46: Urban Transportation 1953
rates for parking in large cities can reduce the proportion of commuters who
drive alone by an average of 25%. Shoup has advocated a relatively unintrusive
method of charging for employee parking known as "cashing out free parking".
Under Shoup's proposal, employers who offer free parking are required to offer
an equivalent cash bonus as an alternative. A limited form of cashing out was
legislated in California in 1992, but the concept is severely inhibited by US tax
laws that exempt free parking-but not cash subsidies-from employees' taxable
income.
One disadvantage of using parking charges to reduce congestion is that they
do not deter through trips, which may in fact increase to fill the street capacity
vacated by former free parkers. Of course, there is a market argument for pricing
parking independent of its effect on congestion (as a means of optimizing the use
of resources to provide parking spaces). Furthermore, the presence of congestion
and tax distortions makes the practice of including free parking as a standard
fringe benefit of employment less socially efficient or desirable.
A more direct approach is to price roads at or near marginal congestion cost,
a policy known as "congestion pricing". Vickrey's (1955) eloquent advocacy
stimulated academic, but little practical, interest in congestion pricing for many
decades. Practical interest has arisen recently, however, due to a combination of
government needs for additional revenue sources, failures of other measures to
reduce congestion, and new technology that makes it feasible to price roads with-
out toll booths (Small et al., 1989: Chap. 5). A thorough, policy-oriented review
is provided by National Research Council (1994). Key theoretical treatments
include Walters (1961), Vickrey (1969) and Arnott et al. (1993).
Empirical evidence confirms that congestion pricing, even when far from
optimally designed, can substantially reduce congestion if the charges are high
and well targeted. The most dramatic example is Singapore, where from 1975
through 1994, a peak charge for entering the central area was assessed, first
during the morning peak only and then during both morning and afternoon peaks.
(Beginning in 1995 the charge was extended to all day on weekdays and a charge
system was imposed on two expressways that approach the central area.) The
entry charge varied roughly from US$1.50 to 2.50 per day over those years. Peak
traffic into the central area was reduced immediately by 47% (G6mez-Ibdfiez
and Small, 1994: Chap. 3), although there is debate about how much of this was
due to other measures such as taxes and quotas on new cars and increases in
parking fees. Singapore converted from a system of paper licenses displayed in
windshields to electronic collection of charges in 1998 (Phang and Asher, 1997;
Phang and Toh, 1997). The electronic system will eventually permit Singapore
to vary charges more finely by time-of-day, location and level of congestion.
1954 K.A. Small and J.A. Gdmez-lbdnez
Low charges or charges that vary little by time-of-day have much less effect.
Examples include three Norwegian cities which recently established toll rings
around their central areas. The charges (roughly US$0.72 to 1.60 per round-trip)
are small relative to incomes and vary little by time-of-day. Effects on traffic have
been correspondingly small, probably less than 5%. The most interesting effects
are in one of the cities (Trondheim), where the charge is lifted at 5 pm, causing
some travellers to delay their trips and some retailers to extend their shop hours
(G6mez-Ibfiez and Small, 1994: Chap. 4).
Simulation models also suggest that reasonably fine-tuned congestion pricing
policies could significantly reduce congestion (Keeler and Small, 1977; Harvey,
1994; Anderson and Mohring, 1996). Prices would of course vary widely; Harvey
(1994: Tables 3-4), for example, estimates average peak period fees of $0.10
and $0.15 per mile for the San Francisco and Los Angeles metropolitan areas,
respectively. Harvey's fees are not necessarily optimal but were chosen to reduce
congestion to a specified level; they imply roughly $2 and $3 per daily round-
trip for a typical worker traveling during the peak in both directions. Harvey also
estimates that a $3 per day employee parking fee would have a somewhat smaller
effect on overall vehicle travel.
Unfortunately, congestion pricing has not proven to be politically popular.
The basic problem is that the efficiency gains for most motorists seem small
and uncertain while the charges they will have to pay are large and very visible
(Altshuler et al., 1979; G6mez-Ibiiez, 1992; Giuliano, 1992; Rom, 1994). The
charges are, of course, a transfer from motorists to the government, and in theory
could be rebated in some way so that the motorists would be better off than be-
fore. But motorists seldom understand this, or trust that governments will rebate
in this way.7
It is thus understandable that the more grandiose plans for congestion pricing
once advanced in Hong Kong, London, the Netherlands, Stockholm, and Cam-
bridge, England, have been abandoned or delayed. It seems likely that experience
in the near future will be limited to small demonstrations or special cases where
time-varying prices can reinforce policies adopted for other purposes. Two ex-
amples are discussed in the next subsection: one where time-varying tolls are
needed to finance new infrastructure, and another where a revenue-neutral change
to a time-varying toll structure can solve a well-defined peaking problem on an
existing toll road. A third possibility is a concept known as high-occupancy/toll
(HOT) lanes, in which unused capacity in an HOV lane is released for single-
7 In other cases the benefits may simply be small because long distance automobile commuters are self-
selected from those with relatively low values of time (Calfee and Winston, 1998).
Ch. 46: Urban Transportation 1955
occupant cars for a fee, giving them in effect a time-varying toll since only
during the peak period would they need to take the tolled lane at all. The first
such facility opened in the Los Angeles area in 1995 (Small and G6mez-Ibdfiez,
1998) and the following year some existing HOV lanes near San Diego were
converted to HOT lanes (Duve, 1994; Kawada et. al., 1997).
One of the keys to political viability of any road pricing scheme will be
to establish a clear link between toll revenues and expenditures on things that
citizens want. The Norwegian toll rings were designed to finance an explicit set
of infrastructure projects. Equally telling, a planned toll ring for Stockholm, part
of a package involving new highways and rail improvements, was at least tem-
porarily abandoned due to public opposition to one of the highway projects being
financed by it. Jones (1991) finds that public opinion in London is much more
favorable toward road pricing if it is presented as part of a package including
improvements to public transportation. Many commentators in the US believe
that some or all of the revenues from road pricing would need to be offset by
tax decreases for the concept to be viable. Recent survey evidence from the Los
Angeles region shows support rising from 40 to 49% when revenue is used to
reduce other taxes (Harrington and Krupnik, 1996). Small (1992b) employs the
principle of tax offsets as part of an attempt to describe a congestion pricing
package that would be Pareto-improving to broad classes of people, in the sense
that the average person in each identifiable income class or interest group would
be made better off.
roads, but there is increasing interest in private urban roads. Urban toll roads
offer greater traffic potential but they also typically cost much more to build and
may face severe competition from untolled alternatives in the off peak (Johansen,
1989; G6mez-Ibaihez and Meyer, 1993).
The advantages of private provision can be considerable. By tapping private
capital markets and issuing private debt, private roads can provide a way around
borrowing contraints that affect some governments, particularly in developing
countries. Lower construction and operating costs are another possibility, al-
though the evidence for them is limited (G6mez-Ibfiez and Meyer, 1993: pp.
201-203). Pricing structures that may be politically or legally infeasible for a
public roadway are sometimes possible for a private enterprise. Other quite un-
expected forms of innovation, such as novel construction methods and routes,
may occur as happened in several California proposals. Environmental or other
objections to specific sites may in some cases be overcome by private entrepre-
neurs with the flexibility and motivation to do so; and if a project appears too
vulnerable to delay to be economical, the private company may be quicker to cut
its losses and abandon the plan before it becomes a costly embarassment.
On the other hand, private ownership can create problems as well. One of
these stems from the fact that private roads are usually financed by tolls while
public roads are often financed by broad-based taxes. The use of tolls avoids the
collection and deadweight losses of taxation, but it also causes an inefficient al-
location of traffic between the private toll road and any competing free, or nearly
free, alternative routes. The private highway will carry too little traffic which,
given that it is of higher quality, will produce greater congestion and higher
accident rates overall than if the entire system were optimally managed. Traffic
misallocation is especially a problem where tolls are very high, as in Spain and
Mexico (G6mez-Ibfiez and Meyer, 1993: pp. 141-142, 156-157; Ruster, 1997).
Furthermore, in the presence of free roads the investment signals provided by
profitability of a proposed new road link do not accurately reflect its overall effect
on welfare (Mills, 1995).
Further problems arise where the owner of a private highway has so much
monopoly power that authorities turn to regulation. Will the potential efficien-
cies of unregulated private enterprise be realized if competition is lacking? Can
regulation be imposed without destroying the advantages of flexibility and in-
novation? Given the resistance to time-of-day pricing of public highways, for
example, how likely is it that a public regulatory body would permit such a
pricing structure on a regulated highway?
On the last question, there were only two cases as of 1996 of toll roads where
congestion pricing was used. One is on the French Autoroute Al southbound
Ch. 46: Urban Transportation 1957
from Lille to Paris, that is operated by a publicly owned but independent road
corporation; here a revenue-neutral time-of-day differential, on Sundays only,
was introduced successfully in 1992 (G6mez-Ibfifiez and Small, 1994: Chap. 5).
The other is the new HOT lanes, mentioned earlier, that were built in the median
of the existing Riverside Freeway (State Route 91) in the Los Angeles region. A
private consortium built these lanes under a franchise agreement with the State of
California that allows the consortium flexibility to adjust its time-of-day pricing
schedule subject only to a cap on its overall rate of return (G6mez-Ibiiez and
Meyer, 1993: pp. 173-176; Fielding, 1994).
The form of franchising is crucial to reaping the advantages of pricing flexibil-
ity. European and developing nations have typically regulated toll rates, leading
to the usual problems of political interference with economic decisions. For ex-
ample, France developed an elaborate system of cross-subsidies both within and
across its toll road operating companies, public as well as private; the result is that
investment decisions do not face any real market discipline, and there is political
pressure to charge toll rates that are relatively uniform across the entire country.
In contrast, California adopted rate of return regulation for its recent experiment
with private highways, which in theory should preserve more incentives for ef-
ficiency. A number of researchers are beginning to explore other franchising or
regulatory regimes that would protect the public from potential monopoly abuse
by the private toll road operator while still preserving the operator's flexibility to
implement time-of-day or other congestion pricing regimes (Engel et al., 1997;
Fielding and Klein, 1993). Clearly, the best franchising or regulatory framework
will depend in complex ways on the specific local context. This is a topic ripe
for research using tools from the fields of industrial organization and of law and
economics.
2.6. Conclusion
tion in the foreseeable future. Nevertheless, two avenues toward the dissemina-
tion of road pricing seem to be gradually opening. The first is targeted, limited-
purpose projects such as those in France and California which can command
support by solving problems peculiar to a local situation in a manner that is
viewed as experimental. The second is private highways, to which governments
are occasionally turning for financial reasons and which may be allowed pric-
ing flexibility if policy-makers understand its importance. Both avenues lead
to small-scale demonstrations of principles and technology. As confidence is
gained, technology and financial imperatives will encourage more road operators
to experiment with pricing flexibility.
Meanwhile, it may be comforting that congestion is a self-limiting phenom-
enon. When it gets too bad, people find ways to avoid or mitigate it-perhaps
a better car stereo, mobile phone, or relocation outside the congested city core.
Some of these measures are harmless enough but others arguably create further
negative externalities, a possibility that adds weight to economists' appeals to
adopt efficient measures that tackle congestion directly.
The environmental costs of motor vehicles are another great concern. The air pol-
lution problem is especially well documented and motor vehicles are important
contributors to it. In Europe and the US, motor vehicles typically account for 32-
98% of national emissions of carbon monoxide (CO), volatile organic chemicals
(VOCs, primarily hydrocarbons), and nitrogen oxides (NOx) (Small and Kazimi,
1995: Table 1). Transportation also accounts for an important fraction of emis-
sions of "greenhouse gases", especially carbon dioxide (CO 2), that are believed
to be having significant long-term effects on the earth's climate (Cline, 1991).
Resulting policy issues are discussed by many authors including Button (1993),
Hensher (1993), Whitelegg (1993) and Harrington et al. (1995).
The environmental costs of motor vehicles are hard to measure and, as one would
expect, vary according to local conditions. Most of the available estimates are for
the health and materials-damage effects of local (or tropospheric) air pollution
rather than for the damage from depletion of the ozone layer by greenhouse
gases. Reviewing studies from Europe, Australia and the US, for example, Quinet
(1994: p. 58) reports that estimated aggregate social costs average around 0.3 and
Ch. 46: Urban Transportation 1959
0.4% of GNP for noise and local air pollution, respectively. Small and Kazimi
(1995) estimate health costs from local air pollution caused by the average on-
road automobile in the Los Angeles region in 1992 to be $0.03 per vehicle mile
using middle-range assumptions. Costs of truck emissions are higher by a factor
of 16. The Small-Kazimi estimates are within the rather wide range estimated by
McCubbin and Delucchi (1996), who similarly find trucks to be more damaging
by a factor of at least 10. In both studies, most of the health costs are from
increased mortality due to inhalable particulates, both those directly emitted and
those indirectly formed in the atmosphere from VOCs, NOx and sulfur oxides
(SOx). The rest of the costs are due mainly to minor illnesses from ozone that is
formed in the atmosphere from VOCs and NO,.
Estimates of damage from greenhouse-gas emissions are speculative given the
extreme scientific uncertainty and the very long-term nature of the effects. Some
policy analysts have used the cost of control as a proxy for damage costs, al-
though this is very controversial. The basic idea is that the costs of the emissions
control schemes considered in recent international negotiations should, if policy
makers are rational, reflect their assessment of what the damage costs are likely
to be. One widely discussed emission control scenario involves stabilizing CO 2
emissions at 1990 levels by the year 2000 and at 20% below those levels by the
year 2010. Manne and Richels (1992) estimate that the marginal CO 2 reduction
costs implied by such a scenario rise over time and then stabilize at around $208
per ton of carbon in the US (1990 prices). This is more than twice the size of
a carbon tax proposed but never implemented in the European Union, and is
also within the range of other estimates of carbon taxes required to meet various
reductions (Quinet, 1994: p. 49). Updated to 1992 prices, this figure is equivalent
to 3.1 cents per vehicle mile for an automobile with the 1992 US average fuel
economy (Small and Kazimi, 1995: p. 28). But of course this emissions control
scheme is both stringent and arbitrary: the damage costs for greenhouse gases
may actually be much lower than these control costs imply (Prud'homme, 1997;
G6mez-Ibfiez, 1997).
Policy-makers have been partially successful in improving air quality. In the
US, the ambient levels of most pollutants have been reduced steadily since the
1960s, although targets for achieving health-based standards have slipped by
decades and there has been little improvement in ozone, for which motor vehi-
cles bear substantial responsibility (Calvert et al., 1993; Harrington et al., 1995).
Europe has lagged behind the US in emission controls on motor vehicles (Small
and Kazimi, 1995: pp. 9-11). In Japan, high population densities and levels of
automobile usage make air pollution a continuing problem. Tangible policies
bringing about significant reductions in CO 2 emissions, other than the high fuel
1960 K.A. Small and J.A. G6mez-lbez
taxes already in place for other reasons, have yet to be realized in any part of the
world.
8 These episodes are known as "off-cycle" or "open-loop" events. For careful documentation, see Calvert
et al. (1993), and references in Small and Kazimi (1995).
1962 K.A. Small and J.A. Gdmez-lbdAie
Clean air legislation promotes several of the TDM policies also used for conges-
tion relief. In the clean air context they are usually called "transportation control
measures", or TCMs, the term used in the US Clean Air Act. Hall (1995) finds
that TCMs have only limited ability to reduce motor vehicle travel compared
with pricing policies, in agreement with our earlier discussion of TDM policies
toward congestion. Employer-based policies are especially ineffective in control-
Ch. 46: Urban Transportation 1963
ling pollution because work trips are an even smaller proportion of total travel
(less than one-fourth in the US) than they are of peak-period travel.
3.5. Pricingpolicies
In theory, one might charge motorists for air pollution on the basis of actual
vehicle emissions in use. Assuming this will not be feasible for some time, an
approximation would be a fee proportional to distance traveled, with the rate
determined by the estimated emissions rates (as measured, for example, by an
annual test of the vehicle or by the manufacturer's new-vehicle test results). This
proposal would not eliminate the need for more reliable inspections, but it would
strongly encourage drivers to minimize their use of high-polluting vehicles. For
CO 2 emissions, the most direct pricing policy would be a fuel tax whose rate is
based on the carbon content of the fuel.
A number of studies have examined the quantitative impacts of pricing poli-
cies. Harvey (1994) tests distance- and emissions-based pollution fees in his
San Francisco and Los Angeles simulations described earlier. The pollution fees
he tests are rather small, about $110 per year per vehicle in Los Angeles, so
have correspondingly small effects. Employee parking fees of $3 per day reduce
emissions by only around 1.4-2.1%, presumably because the fees affect only the
minority of trips that are to or from work and because they fail to discourage new
trips from taking advantage of any improvement in travel conditions.
Harvey also tests a gasoline tax increase of $2 per gallon in his San Francisco
simulations. Such a tax is equivalent to $670 per ton carbon, 9 more than three
times the amount implied by international agreements for CO 2 reduction dis-
cussed earlier. The gasoline tax increase costs the average vehicle owner around
$1000 per year l° and is predicted to have modest impacts on both vehicle miles
of travel (reduced by 8.1%) and on emissions (VOC, CO and NOx all reduced
by around 7.7%). The gasoline tax is a rather blunt instrument for dealing with
local pollution, however, because it is not emissions-based. The tax is best tar-
geted for CO 2 control, and elicits an estimated 36% reduction in CO 2 emissions.
This reduction is more than four times the estimated reduction in total vehicle
travel because motorists shift toward vehicles and driving patterns that are more
fuel-efficient.
9 This is based on 1 ton carbon content for every 335 gallons of refined petroleum product, calculated from
Manne and Richels (1992: p. 59).
10 Based on average annual fuel consumption of 513 gallons per year for US passenger cars in 1993
(American Automobile Manufacturers Association, 1995: p. 66).
1964 K.A. Small and J.A. Gomez-bdcez
3.6. Conclusion
Estimates suggest that the marginal social costs of motor vehicle use due to
air pollution are a small fraction of the marginal cost of urban driving. Thus,
if motorists had to pay for the air pollution damages they cause, total vehicle
travel would not be affected much. This finding implies that measures to reduce
emissions per vehicle mile are likely to be the most efficient means of motor
vehicle air pollution control.
The principal measure for reducing emissions per vehicle mile to date has
been to set maximum emissions standards for new vehicles. This has worked
reasonably well so far, but further progress will require stronger incentives on
individual vehicle owners to maintain their vehicles so as to keep emissions
low. Such incentives affect individual drivers in a more direct way than previous
Ch. 46: Urban Transportation 1965
policies, and as a result most governments have been reluctant to move in this
direction.
The problem of global CO 2 pollution is rather different and far more un-
certain. Nevertheless, current evidence does not provide a compelling case for
drastic across the board reductions in vehicle travel. Rather, incentives for fuel
conservation seem called for, and the evidence suggests that the primary response
would be through improved fuel efficiency of vehicles. Once again, a technical
solution underlies the most promising approach.
Economic theory has not provided the same ready consensus with regard to the
problem of motor vehicle accidents that it has for congestion and pollution. Yet
the aggregate costs of accidents appear to be far higher than those of pollution,
and comparable to those of congestion even in urban areas (Small, 1992a: Table
3.2). The difference in treatment may be due to the elegance with which conges-
tion and pollution can be formulated as externalities. For motor vehicle accidents,
the externality formulation is certainly possible (Jansson, 1994), but the nature
of the externality in practice is less clear. Meanwhile other potential market
failures-arising from insurance, incentives in the legal system, and the like-
are prominent. Such issues in the evaluation of motor vehicle accidents make the
subject fascinating and potentially very productive for economic analysis.
In a widely cited and comprehensive study, Miller (1993) estimates that motor
vehicle accidents cost the US $333 billion in 1988, or an average of $0.164 per
vehicle mile.lT In another careful study, Newbery (1988) estimates the total 1984
accident costs in the UK at £26 billion (in 1986 prices) or US$0.22 per vehicle
mile. 12 These US and UK estimates correspond to 7 and 5% of GNP, about three
to four times higher than the typical estimates of 1.5-2% found by Quinet (1994:
pp. 37-39, 58) in a review of other studies-which include rural areas, where
accident costs are quite high because higher rural traffic speeds increase accident
severity. These cost estimates are relative to a world with no accidents at all, a
11 Total motor vehicle in the US was 2,025.6 billion vehicle miles, according to the US FHWA (1988: Table
M-1).
12 This applies the 1986 exchange rate of £1 = US$1.47, from US Council of Economic Advisors, Annual
Report, 1995, in Economic Report of the President,Feb. 1995 (Washington: US Government Printing Office:
Table B- 112.) Vehicle kilometers traveled are inferred from Newbery's Tables 4 and 5 to be 273.8 billion.
1966 K.A. Snull and J.A. G6mez-lbdiez
Empirical evidence is needed to decide the extent to which family and relatives'
altruism is health-specific. The second argument is that the individual may al-
ready take her family's and friends' concern for her into account in her own
utility function.
The willingness to pay to prevent nonfatal injuries is analyzed carefully by
Jones-Lee et al. (1995) based on survey data from the UK. Their results suggest
that people are unable to provide reliable answers when asked directly, through
what are called "contingent valuation" questions. But credible self-consistent
answers are given to more indirect questions which ask people to choose among
hypothetical treatments for speficied road injuries, some of which have better
probable outcomes than others but ancillary fatal risks. From these questions,
the authors make a strong case that the average valuation of a nonfatal injury
classified as "serious" in the UK-a weighted average of injuries ranging from
some requiring 3-4 months for full recovery to some causing severe permanent
disability-is about 9.5% of the valuation of a fatal injury.
Policy toward motor vehicle safety has often been discussed under the assump-
tion that people do not behave as rational consumers when evaluating safety.
However, empirical research has clearly established that people are willing to pay
for safety improvements, and is mostly consistent with the conclusion that they
evaluate costs and benefits rationally. Examples of this research include studies
of the demand for specific safety features on vehicles (Arnould and Grabowski,
1981; Mannering and Winston, 1995), for vehicle models that rank high on an
1968 K.A. Small and J.A. G6mez-lbdfez
index of crash worthiness (McCarthy, 1990), and for models with good records
of actual safety outcomes (Winston and Mannering, 1984).
A related aspect of consumer behavior is the extent to which people may com-
pensate for safer vehicles by engaging in more dangerous behavior-or equiva-
lently, the extent to which they offset safety hazards by being careful. Peltzman
(1975) posited that the potentially beneficial effects of government-mandated
safety features in new vehicles would be partially or fully offset because drivers
would respond by driving more dangerously. (Safety features could even raise
accident costs overall if such drivers then injure pedestrians and cyclists more
frequently.) Many researchers have tested the "Peltzman effect" empirically; a
good example is Chirinko and Harper (1993), who also provide a useful re-
view. A common finding is that compensating behavior occurs but does not fully
offset the original safety improvement. Such studies are frequently plagued by
problems of changing vehicle mix and by potential endogeneity of the variables
measuring the use of safety equipment.
4.4. Externalities
How do accident costs vary with traffic volumes? The extent to which accident
costs are externalities depends critically on the extent to which accident rates
and their severity are affected by traffic volume. The problem is similar to that
of congestion; if adding additional vehicles to the traffic stream increases the
accident rate per vehicle, then a form of congestion externality exists even if
each traveler fully bears the average accident risk (Vickrey, 1968).
Like the relationship between traffic volumes and speeds, the relationship be-
tween volumes and accident rates could be a simple static one, or it might depend
in complex ways on the dynamics of traffic movements. It may be important to
know how different vehicle types or driver behaviors affect costs. At the simplest
static level, the question may be reduced to asking what is the elasticity of total
accident cost on a road with respect to its traffic volume; if the elasticity exceeds
1, there is a negative externality because each user imposes costs on other users,
causing social marginal cost to exceed private marginal cost.
Ch. 46: Urban Transportation 1969
Until recently, most researchers had to make assumptions about the relation-
ship between accident costs and traffic volume since there was little empirical
evidence available. Vickrey (1968) argues from impressionistic evidence for an
elasticity of 1.5; Jones-Lee (1990) follows official British practice by assuming
1.0 (i.e., linearity); Newbery (1988) splits the difference at 1.25.
Some empirical work is emerging which suggests a more complex story in
which the elasticity could actually be less than 1, particularly for accidents seri-
ous enough to cause a fatality or injury rather than only property damage. Zho
and Sisopiku (1997) find that accident rates on a 26 km segment of an urban
expressway in Detroit have a "U"-shaped relation to the volume/capacity ratio;
that is, accidents per vehicle mile are lowest at intermediate volumes. But when
injury and fatality accidents are broken out separately, accident rates decline
fairly dramatically with traffic volume. Similarly, in a Norwejian study of mostly
rural roads, Fridstr0m and Ingebrigsten (1991) find an elasticity of only 0.47,
implying accident rates decrease with volume.
The problem with measuring the relationship between accident costs and traf-
fic volume is that traffic densities are highly correlated with time-of-day and
degree of urbanization, both of which affect accident rates and severities strongly.
For example, in Zho and Sisopiku's (1997) study, many of the hours with low
volumes are in the evening or on weekends, when visibility and alcohol probably
play a role in elevating accident rates. Unfortunately, researchers have not yet
done a careful multivariate study of the determinates of accident rates, even
though this is obviously a critical issue in measuring safety externalities.
How much of accident costs are borne by nonmotorists? Pedestrians and cyclists
appear to account for considerably more than half of the motor vehicle deaths
in the UK, according to figures reported by Jones-Lee (1990: p. 51). In the US
they account for 16% (US Census Bureau, 1995: Table 1033). The government
pays for some medical expenses through various programs of health or indigent
care. Both of these represent potential externalities if there is no compensation
mechanism by which motorists are faced with these costs.
Does insurance affect users' perceptions of the accident costs resultingfrom their
driving? It is typical for government publications to include insurance premiums
as a fixed cost of driving, as for example in the US Census Bureau (1995: Table
1038). The belief that people do not perceive these costs as being variable under-
lies "pay at the pump" proposals to convert automobile insurance to a surcharge
on the gasoline tax.
1970 K.A. Small and J.A. G6mez-lbdfiez
How do tort and criminal law affect incentives? The legal system is another way
that parties may be charged for accidents resulting from their driving decisions.
The literature in law and economics provides a theoretical framework for ana-
lyzing such effects; see, for example, Calabresi (1970), Brown (1973), Shavell
(1987) and Boyer and Dionne (1987). Empirical work that applies these ideas
Ch. 46: Urban Transportation 1971
to motor vehicle accidents is limited, but one interesting study is Calfee and
Winston's (1993) analysis of liability damages for pain and suffering. In the case
of automobile safety, they find that allowing such damages to be recovered may
in fact reduce welfare by forcing drivers as a group to buy insurance that they do
not really want.
Do people seek out a certain level of risk taking? It is possible that people who
drive in an aggressive or dangerous manner are using the road system to express
emotions that would otherwise find another outlet. The predominance of young
males in accident statistics is certainly suggestive. If so, an evaluation of safety
improvements would have to take into account how costly those alternative out-
lets of destructive behavior would be to society. If they are equally costly, the
safety improvement has no net social benefit. To address such a question requires
going well outside the usual framework of economic analysis to incorporate
psychological motivations for behavior.
4.5. Conclusion
Accidents constitute one of the largest costs of motor vehicle use, and one that
appears highly susceptible to public policy. Accidents almost certainly entail
externalities, although their exact nature and magnitude are largely unknown.
Valuing accident costs involves sophisticated concepts and difficult empirical
challenges, but researchers seem to be succeeding. The behavior and technol-
ogy that produce accidents interact strongly with insurance and legal institutions,
adding to the challenge of performing accurate analysis of incentives and policy
effects. Some types of consumer behavior-such as drunk driving-are proba-
bly not entirely rational, and determined regulatory approaches toward them can
be quite effective. Others-such as purchase and use of safety equipment-are
rational and thus readily susceptible to economic or legal policies designed to
fine-tune the incentives acting on individuals.
Three topics seem both important and feasible areas for futher economic
analysis. The first is exactly how accident rates and severity vary with traffic
volumes. The second is how people perceive insurance costs, and whether, or
not, any institutional changes could make those perceptions more consonant with
efficiency. The third is the incentives produced by the legal system.
The historical evidence suggests that there is considerable room for address-
ing accidents with technology and targeted behavioral incentives. We suspect that
efficient policies would lead to various adjustments in vehicle mix, technology
and driving behavior. At this point there seems little reason to believe that the
1972 K.A. Small and J.A. G6mez-lbdriez
costs of driving would dramatically rise or that people would adjust by markedly
reducing their use of motor vehicles.
5. Public transportation
13 See American Public Transit Association (1987: p. 32; 1993: p. 64; 1995: p. 88). The main problem in
reporting transit ridership is the distinction between linked and unlinked trips. An unlinked trip consists of a
ride on one transit vehicle, whereas a linked trip may include two or more vehicles connected by a transfer.
Measuring patronage by unlinked trips creates misleading comparisons whenever there is a change in the
proportion of travelers who transfer, such as occurs on the introduction of a rail transit system supplemented
with feeder buses.
1974 K.A. Smnll and J.A. G6mez-Ibdfiez
is probably the major reason why public transit ridership is still increasing in
some of them. To give just two examples, annual real subsidies grew by 45%
in Canada between 1980 and 1992, with ridership growing just 7% (Perl and
Pucher, 1995: Table 2). In West Germany between 1980 and 1993, annual sub-
sidies increased by 67% in real terms while ridership held essentially constant
(Pucher and Kurth, 1995: pp. 118, 126).
Even where public transit ridership has stabilized or increased, moreover, car
ownership and use are increasing faster as we demonstrated in Section 2.1. As a
result, public transportation's share of total trips is declining in Europe as in the
US.
Analogous trends are evident in the newly emerging economies of eastern Eu-
rope and in the developing countries of Asia, Latin America and Africa. Pucher
(1995b: p. 220) examines four eastern European countries, finding that between
1985 and 1993 public transportation ridership fell dramatically in three (East
Germany, Hungary and Poland) and rose in only one (Czechoslovakia). In the
developing world, trends are more complex. Transit ridership in the poorest coun-
tries is growing fairly rapidly as the population shifts from human or animal-
powered modes to motorized modes, primarily bus. In the large cities of such
countries, public transportation often accounts for 80% or more of all motorized
trips, although a much smaller share of total trips. Among the richer developing
countries, however, the transition to motorized forms of travel is nearly complete
and the shift from public transportation to the automobile has begun in earnest.
Public transportation ridership is often still stable or growing slowly, usually
without the aid of significant subsidies. But automobile ownership and use is
growing far more rapidly so that public transportation's share of motorized trips,
although still higher than in Europe, is falling steadily.
argued that the higher transit usage in European cities is largely due to higher
densities and other factors under the control of planners. As noted in Section 2.4,
however, other factors besides land use are involved and it may not be desirable or
politically feasible to force households and firms into denser locations to achieve
only small reductions in automobile use.
The most obvious method of arresting the decline in transit usage, and the one
most widely adopted, is to subsidize transit. Three arguments are often advanced
in support of such subsidies. The first is that motorists do not pay the full marginal
social costs of automobile use, and thus transit subsidies are necessary to insure
that travel choices between public and private modes are not distorted. The better
solution would be to price automobile use directly, of course. If it is administra-
tively difficult or politically impossible to correct the mispricing of automobile
use, however, then subsidizing automobile's competitor, public transportation,
may be a useful second-best corrective.
The second argument is that public transportation is characterized by
economies of scale so that fares set at marginal cost will be insufficient to cover
total cost. Scale economies are more likely in rail transit than bus transit if only
the costs of the public transportation firm are considered. Cost functions of bus
firms show only modest economies of scale and then only for very small firms
or firms that have excess capacity (Berechman, 1993: pp. 120-127). In con-
trast, cost functions for rail transit and for multimodal enterprises generally show
economies of scale, traffic density or scope (Berechman, 1993: pp. 127-128;
Viton, 1992).
Economies of scale can be found in bus as well as rail, however, if one
includes the value of time spent by the travelers. The reason is that greater
passenger densities permit some combination of more frequent service, greater
vehicle utilization, and more direct routes (Mohring, 1972; Turvey and Mohring,
1975; Nash, 1988; Jansson, 1993). This effect is so pronounced that some sim-
ulation models have predicted surprisingly large modal shifts to public transit
resulting from simultaneous marginal-cost pricing of peak automobile and transit
trips (Viton, 1983); the initial effect of much higher peak automobile prices is
further accentuated by resulting improvements in transit service quality and/or
fare reductions.
The final argument for subsidies is that poor and disadvantaged households
tend to be more dependent on public transportation than the rest of the population.
Ch. 46: Urban Transportation 1977
Subsidies therefore may be an important means of aiding these groups or insuring
that they have a minimum level of mobility needed to participate in society.
Transit subsidies bring their own problems, however, as reviewed insightfully
by Wachs (1989). First, public transportation demand is often fairly insensitive to
price or service improvements so that substantial subsidies are needed to have a
significant effect on ridership. Estimates of the direct price elasticity of demand
generally range from -0.1 to -0.5 in developed countries with many studies
reporting results around -0.3; the elasticity of demand with respect to travel
time, service frequency or service miles, is perhaps a bit higher in absolute value
but less than 1.15 Estimates of the cross-elasticity of demand for automobile trips
with respect to public transportation fares or service are even smaller because
only a fraction of transit riders are diverted from automobiles and because the
starting base of automobile use is so large (Kemp, 1973). For example, estimates
of cross-elasticities for eight Australian cities are all less than 0.02 in absolute
value (Dodgson, 1986: Table 4). The modest direct elasticities suggest that the
efficiency gains from pricing at marginal cost instead of average cost in the face
of increasing returns to scale may be modest. The even smaller cross-elasticities
suggest that transit subsidies are a costly way to correct for the mispricing of
automobile use.
Another difficulty is that many public transportation riders are not poor or
disadvantaged, so that transit subsidies are a relatively inefficient means of re-
distributing income (Meyer and G6mez-Ibfiez, 1981). In the smaller and less
congested US metropolitan areas, the majority of transport users are generally
too poor to own a car or too young, elderly or infirm to drive (Pucher et al.,
1983). But in larger and more congested metropolitan areas, many public trans-
portation riders are commuters to the central city, who are wealthier than the
average metropolitan resident. In both small and large metropolitan areas, more-
over, many poor households own cars or use taxis because public transportation,
despite subsidies, does not serve well the trips they need to make.
Perhaps the most serious problem is that a large portion of the subsidies is
absorbed by reduced productivity or higher wage rates instead of being used to
improve service or reduce fares. This is supported by both cross-sectional and
time-series studies of transit firms receiving different levels of subsidy. For ex-
ample, Bly et al. (1980) examine data on transit subsidies and performance in 59
cities in 17 developed countries between 1966 and 1976. Their analysis suggests
that a 1% increase in subsidies was associated with a 0.4-0.6% increase in the
15 See Goodwin (1992: p. 160), Oum et al. (1992: p. 148), Chan and Ou (1978) and Berechman (1993: pp.
38-39).
1978 K.A. Snuill and J.A. G6mez-lbciiez
real costs of providing a vehicle mile of service. They also test specifications in
which either the subsidy or productivity variables are lagged; the model performs
better when the subsidy variable is lagged, suggesting that the causality was
from subsidies to costs rather than vice versa. Anderson (1983), Pucher et al.
(1983) and Perry and Babitsky (1986) find similar results in analyses of cross-
sections of US public transportation companies, as does Cervero (1984) for a
pooled time-series cross-section of California transit systems.
Similarly, Lave (1991) examines costs using annual data from 62 large US
bus firms. He finds that unit costs increased by an average of 1.4% per year in
the era before federal subsidies were available (1950-1964), compared to 2.1%
per year during the decade of federal capital grants (1965-1974) and 3.1% per
year when the federal government subsidized both capital and operating expenses
(1975-1985). Larger firms experienced more cost inflation, suggesting to Lave
that workers at such firms may have used their cities' vulnerability to transit
strikes to extract more of the subsidies in the form of higher wages.
How much of subsidy funds have been absorbed by higher costs? For the US,
Pickrell (1985a) uses simple accounting identities to calculate that 61% of the
increase in real (net of inflation) subsidies between 1970 and 1980 was used to
cover an increase in the real cost of providing a vehicle mile of service, while
only 14% was used to reduce real fares and 9% to increase the number of vehicle
miles of service offered.
There is some evidence that the rate of increase in unit costs has slowed since
the mid-1980s, perhaps because taxpayers have become reluctant to finance fur-
ther subsidy increases. For example, the average operating expense per vehicle
hour in the US declined 7% in real terms between 1986 and 1992 (American
Public Transit Association, 1993). Similarly, in a detailed study of Boston's pub-
lic transit system G6mez-Ibfiez (1996) finds that real operating expenses per
vehicle hour increased rapidly in the 1970s, but then declined in the 1980s after
the state legislature strengthened management's position on key labor contract
issues.
Pickrell also finds that costs were typically far higher than forecast. Capital
cost overruns, ranging from 17 to 150%, occurred in all seven systems where
comparisons were possible. About half was due to construction delays, the rest to
underestimates of the real unit prices of inputs or the quantities of inputs required.
Operating costs were also higher than forecast in five of the six cases where such
comparisons were possible, in large measure because train speeds were slower
than forecast.
Other researchers have examined the experience of individual metropolitan
areas that built new rail systems or extensions, and argued that bus improvements
would have been a less costly method of attracting or retaining ridership. Kain
(1994: p. 16) estimates the demand for transit ridership using time-series data
for the Atlanta metropolitan area; he then predicts that between 1980 and 1993
Atlanta could have attracted 9% more ridership at only 37% of the cost if it had
reduced fares and improved service on its bus system instead of building its new
rail transit system. Using a similar approach, G6mez-Ibfiiez (1996) finds that
Boston's policy of extending its rail lines further into the suburbs adds $10.68
in new deficit spending per new ride attracted, while a policy of reducing real
fares on the existing system would add only $2.60 to the deficit for each new ride
attracted.
Nor is there evidence that these costly rail systems have significant impacts
on automobile traffic. When the Bay Area Rapid Transit (BART) line opened
between San Francisco and Oakland in 1973, 8750 daily automobile trips on
the parallel bridge were diverted; but soon after, 7000 new automobile trips
appeared on the bridge (Sherret, 1975). Zurich's ambitious regional light rail
system, which opened in 1990, has succeeded in substantially increasing transit
usage, yet with no observable reduction in traffic flow at the city boundaries
(OECD, 1994: p. 97). The reason is probably the "fundamental law" of traffic
congestion, discussed earlier: new peak-period motorists soon take the place of
those who switch to transit.
Less developed countries are facing a similar debate about the appropriate role
of rail and bus transit, although in a context in which the vast majority of travelers
are still using public rather than private transportation. Many major metropolitan
areas in developing countries have built or planned new rail systems since the
late 1960s, when Mexico City opened the first modem heavy rail transit system
in the developing world.
A recent survey prepared for Britain's Transport and Road Research Labo-
ratory finds that developing nations have experienced many of the same rider-
ship shortfalls and cost overruns as the US (Halcrow Fox and Associates, 1989;
Fouracre et al., 1990). The study examined 21 large cities in developing countries
Ch. 46: Urban Transportation 1981
with new rail systems: 13 in full or partial operation and eight under construction
or in advanced stages of planning. Of the 13 operational systems, only three
(Hong Kong, Porto Allegre and Singapore) were built within 10% of their fore-
cast capital costs while six had overruns of more than 50%. The most common
reason for overruns was failure to accurately predict the difficulty of construc-
tion in a densely developed urban environment. Similarly, forecast ridership was
approximately achieved in only two of the cities (Manila and Tunis), while in
the remaining seven cases for which comparisons were possible actual ridership
fell short of forecasts by 20-90%. Although the new rail lines are often heavily
patronized by international standards-Mexico City has twice as many rail pas-
sengers as London-they still capture only 10-20% of all urban passenger trips,
and the limited available data suggest no noticeable diversion from automobiles
(Halcrow Fox and Associates, 1989: p. 7.11).
On the positive side, three of the 10 cities with complete financial data in-
dicated that fares covered operating costs and made some contribution toward
depreciation. Furthermore, 10 of the 13 completed systems are estimated to have
produced a social rate of return above 10%, in most cases due to large time sav-
ings by bus users who switched to rail. The largest social rates of return (above
15%) were in cities that had high rail patronage and relatively high values of
time (Hong Kong and Singapore), or that had moderate patronage and very low
construction costs (Cairo). The researchers did not compare the returns on rail
investment with those that might have been achieved from improving the bus
systems.
Table 4
Percentage changes in local bus service characteristics: the UK, 1985/86 to 1993/94
Deregulated areas
All areas Metropolitan areas London
Even more discouraging was the substantial reduction in local bus ridership.
Ridership fell by 27% in all deregulated areas and 36% in the metropolitan coun-
ties. Some decline in ridership would have been expected even in the absence
of policy reforms, given that real incomes were rising in the UK. But the rate of
ridership loss was greater than in the previous decade, reinforcing the supposition
that riders have been made worse off.
Many researchers have attempted to assess who won and lost from privatiza-
tion and deregulation in the UK and whether, on the whole, social welfare was
improved. The key difficulty is to isolate the effects of reforms from the effects
of subsidy cuts and other adverse trends.
One approach is to use metropolitan London as a control. White (1990, 1995),
Mackie et al. (1995) and others have argued that the policy of gradually contract-
ing out adopted in London proved to be superior to the more sudden and complete
privatization and deregulation adopted in the rest of the country. As shown in
Table 4, metropolitan London enjoyed as large a service increase and almost as
large a unit cost saving as did the deregulated areas, but it suffered much smaller
fare increases and losses of patronage.
Such comparisons are suggestive but far from conclusive given the difficulty
in controlling for other differences between London and the rest of the country. In
particular, government subsidies were cut far less in London, as shown in Table 5.
Seven years after deregulation, London transit was still subsidized at 39% of fare
receipts, whereas in the other metropolitan areas subsidies had been cut dras-
tically to just 18% of receipts. Employment growth was stronger in downtown
London than in the rest of the UK, moreover, and London was probably losing
fewer bus riders to automobiles because driving and parking is so much more
difficult and costly there.
Ch. 46: Urban Transportation 1985
Table 5
Government revenue support for local bus service: the UK
Deregulated areas
All areas Metropolitan areas London
1985/86
Revenue support 510 319 207
Passenger receipts 1944 640 426
Support as percentage of receipts 26.2 49.8 48.5
1992/93
Revenue support 237 114 167
Passenger receipts 1725 626 432
Support as percentage of receipts 13.7 18.2 38.6
tive to riders and profitable for operators. The fact that these services appeared
only after deregulation or illegally also suggests that regulatory barriers may
be very important impediments. In the American suburbs, moreover. Cervero
(1997) discovered a variety of paratransit-like services have been growing, in-
cluding shared-ride taxi services, airport limousines and subscription vans to
ferry children to after-school activities. These services still account for only a
tiny percentage of total trips, but they suggest interesting possibilities.
The second hopeful development is improvements in computing and satellite-
based automatic vehicle locating (AVL) systems. Low cost scheduling software
linked to geographic information systems is readily available (Stone et al., 1993).
Teal (1993) contends that current technology makes us much more likely to re-
alize the promise of dial-a-ride than we were in the 1970s. Nelesson and Howe
(1995) argue that instead of providing on-demand door-to-door service, as in the
old dial-a-ride, it might be much less expensive and almost as convenient to pro-
vide on-demand service to a series of conveniently located telephone-equipped
stops or nodes. The potential is still largely unproven but the time seems ripe for
a new series of demonstrations.
5.7. Conclusion
6. Conclusions
References
Altshuler,; A.A. and J.A. G6mez-lbAiez, A.M. Howitt (1993), Regulation for Revenue: the Political Econ-
omy of Land Use Exactions (Brookings Institution and Lincoln Institute of Land Policy, Washington,
DC/Cambridge, MA).
Altshuler. A.A., J.P. Womak and J.R. Pucher (1979). The Urban Transportation System: Politics and Policy
Innovation (MIT Press, Cambridge, MA).
American Automobile Manufacturers Association (1993, 1995-1998), Motor Vehicle Facts and Figures
(Detroit).
American Public Transit Association (1987, 1993, 1995), Transit Fact Book (Washington. DC).
Anderson, D. and H. Mohring (1996), Congestion Costs and Congestion Pricing for the Twin Cities, Report
No. MN/RC-96/32 to Minnesota Department of Transportation, (Department of Economics, University of
Minnesota, St. Paul. Minneapolis).
Ch. 46: Urban Transportation 1991
Anderson, S. (1983), "The effect of government ownership and subsidy on performance: evidence from the
bus transit industry", Transportation Research A 17A: 191-200.
Apogee Research (1994), Costs and Effectiveness of Transportation Control Measures (TCMs): A Review and
Analysis of the Literature (National Association of Regional Councils, Washington, DC).
Arnmott, R., A. de Palma and R. Lindsey (1991), "Does providing information to drivers reduce traffic
congestion?", Transportation Research 25A:309-318.
Amott, R., A. de Palma and R. Lindsey (1993), "A structural model of peak-period congestion: a traffic
bottleneck with elastic demand", American economic review 83:161-179.
Amould, R. and H. Grabowski (1981), "Auto safety regulation: an analysis of market failure", Bell Journal of
Economics 12:27-48.
Banister, D. (1993), "Investing in transportation infrastructure", in: D. Banister and J. Berechman, eds.,
Transport in a Unified Europe, pp. 347-368 (North-Holland, Amsterdam).
Banister, D. and J. Berechman (1993), "Transport policies and challenges in a unified Europe: introduction",
in: D. Banister and J. Berechman, eds., Transport in a Unified Europe, pp. 1-25 (North-Holland,
Amsterdam).
Beesley, M.E. (1990), "Collusive, predation, and merger in the U.K. bus industry", Journal of Transport
Economics and Policy 24:295-310.
Ben-Akiva, M., A. de Palma and I. Kaysi (1991), "Dynamic network models and driver information systems",
Transportation Research 25A:251-266.
Berechman, J. (1993), Public Transit Economies and Deregulation Policy (North-Holland, Amsterdam).
Berechman, J. and G. Giuliano (1985), "Economics of scale in bus transit: a review of concepts and evidence,
Transportation 12:313-332.
Bergstrom, T.C. (1982), "When is a man's life worth more than his human capital?", in: M.W. Jones-Lee, ed.,
The Value of Life and Safety (North-Holland, Amsterdam).
Bly, P.H., F.V. Webster and S. Pounds (1980), "Effects of subsidies on urban public transport", Transportation
9:311-331.
Boyer, M. and G. Dionne (1987), "The economics of road safety", Transportation Research 21B:413-431.
Brown, J. (1973), '"Toward an economic theory of liability", Journal of Legal Studies 2:323-350.
Brownstone, D. and T.F. Golob (1992), "The effectiveness of ridesharing incentives: discrete-choice models of
commuting in southern California", Regional Science and Urban Economics 22:5-24.
Button, K., (1993), Transport, the Environment and Economic Policy (Edward Elgar, Aldershot, UK).
Calabresi, G. (1970), The Costs of Accidents: A Legal and Economic Analysis (Yale University Press, New
Haven, CT).
Calfee, J.E. and C. Winston (1993), "The consumer welfare effects of liability for pain and suffering: an
exploratory analysis", Brookings Papers on Economic Activity: Microeconomics, NSI:133-196.
Calfee, J.E. and C. Winston (1998), "The value of automobile travel time: implications for congestion policy",
Journal of Public Economics 69:83-102.
Calvert, J.G., J.B. Heywood, R.E. Sawyer and J.H. Seinfeld (1993), "Achieving acceptable air quality: some
reflections on controlling vehicle emissions", Science 261:37-45.
Cervero, R. (1984), "Cost and performance impacts of transit subsidy programs", Transportation Research
18A:407-413.
Cervero, R. (1989a), "Jobs-housing balance and regional mobility", Journal of the American Planning
Association 55:136-150.
Cervero, R. (1989b), America's Suburban Centers: The Land Use-Transportation Link (Unwin Hyman,
Boston).
Cervero, R. (1994), "Making transit work in suburbs", Transportation Research Record, No. 1451, pp. 3-11.
Cervero, R. (1997), Paratransit in America: Redefining Mass Transportation (Praeger, Westport, CT).
Chan, Y. and F.L. Ou (1978), "Tabulating demand elasticities for urban travel forecasting", Transportation
Research Record 673:40-46.
Chirinko, R.S. and E.P. Harper, Jr. (1993), "Buckle up or slow down? New estimates of offsetting behavior and
their implications for automobile safety regulation", Journal of Policy Analysis and Management 12:270-
296.
1992 K.A. Small and J.A. Gdrnez-lbdAiez
Cline, W.R. (1991), "Scientific basis for the greenhouse effect". Economic Journal 101:904-919.
Crampton, G.R. (1997), "Labour-market search and urban residential structure", Environment and Planning A
29:989-1002.
Dahlgren, J.W. (1998), "High occupancy vehicle lanes: not always more effective than general purpose lanes".
Transportation Research-A.32:99-114.
Daniere. A.G. (1995), "Transportation planning and implementation in cities of the third world: the case of
Bangkok", Environment and Planning C: Government and Policy 13:25-45.
Davis, S.C. (1997), Transportation Energy Data Book: Edition 17, Report ORNL-6919 (Oak Ridge National
Laboratory, Oak Ridge, TN).
Dodgson. J.S. (1986). "Benefits of changes in urban public transport subsidies in the major Australian cities",
The Economic Record 62:224-235.
Dodgson, J.S., C.RP Newton and Y. Katsoulacos (1992), "A modelling framework for the empirical analysis of
predatory behaviour in the bus services industry", Regional Science and Urban Economics 22:51-70.
Downs. A. (1962), "Thelaw of peak-hour expressway congestion'. Traffic Quarterly 16:393-409.
Downs. A. (1992), Stuck in Traffic: Coping with Peak-hour Traffic Congestion (The Brookings Institution.
Washington, DC).
Duve, J.L. (1994). "How congestion pricing came to be proposed in the San Diego region: a case history", in:
National Research Council, Vol. 2, pp. 318-333.
Economic Report of the President (1995), Together with the Annual Report of the Council of Economic
Advisors (Washington: US Government Printing Office).
Emmerink, R.H.M., E.T. Verhoef, P. Nijkamp and P. Rietveld (1996), "Information provision in road transport
with elastic demand"'. Journal of Transport Economics and Policy 30:117-136.
Engel, E. and R. Fischer and A. Glatovic (1997), "Highway franchising: pitfalls and opportunities', American
Economic Review 2:68-72.
EC (European Communities-Commission) (1994), Socioeconomic Cost of Road Accidents, Report EUR
15464 (Office for Official Publications of the European Communities, Luxembourg).
ECMT (European Conference of Ministers of Transport) (1993), Statistical Trends in Transport 1965-1989
(OECD Publications Service, Paris).
ECMT (European Conference of Ministers of Transport) (1995). European Transport Trends and Infrastruc-
tural Needs (OECD Publications Service, Paris).
Europa Publications (1992), The Europa World Year Book: 1992 (Europa Publications, London).
Evans, A.W. (1990), "Competition and the structure of local bus markets", Journal of Transport Economics
and Policy 24:255-281.
Evans, A.W. (1991), "Are urban bus services natural monopolies?' Transportation 18:131-150.
Fielding, G.J. (1994), "Private toll roads: acceptability of congestion pricing in southern California', in:
National Research Council, Vol. 2, pp. 380-404.
Fielding, G.J. and D.C. Johnston (1992), "Restructuring land transport in New Zealand". Transport Reviews
12:271-289.
Fielding, G.J. and D.B. Klein (1993), "How to franchise highways", Journal of Transport Economics and
Policy 27:113-130.
Fouracre, P.R., R.J. Allport and J.M. Thompson (1990), The Performance and Impact of Rail Transit in Devel-
oping Countries, Research Report No. 278 (Transport and Road Research Laboratory, UK Department of
Transport, Crawthome, Berkshire).
Fridstr0m, L. and S. Ingebrigtsen (1991), "An aggregate accident model based on pooled, regional time-series
data". Accident Analysis and Prevention 23:363-378.
Fuhs, C.A. (1993), Preferential Lane Treatments for High-occupancy Vehicles, National Cooperative Highway
Research Program, Synthesis of Highway Practice 185 (National Academy Press, Washington. DC).
Giuliano, G. (1991), "Is jobs-housing balance a transportation issue?", Transportation Research Record
1305:305-312.
Giuliano, G. (1992), "An Assessment of the Political Acceptability of Congestion Pricing", Transportation
19:335-358.
Ch. 46: Urban Transportation 1993
Giuliano, G. and K.A. Small (1993), "Is the journey to work explained by urban structure?", Urban Studies
30:1485-1500.
Giuliano, G. and K.A. Small (1995), "Alternative strategies for coping with traffic congestion", in: H. Giersch,
ed., Urban Agglomeration and Economic Growth, pp. 199-225 (Springer-Verlag, Berlin).
Giuliano, G. and M. Wachs (1997), "An employer panel for evaluating the effectiveness of trip reduction
incentives", in: T. Golob, R. Kitamura and L. Long, eds., Panels for Transportation Planning: Methods and
Applications, pp. 129-151 (Kluwer Academic Publishers, Dordrecht).
Giuliano, G., D.W. Levine and R.F. Teal (1990), "Impact of high occupancy vehicle lanes on carpooling
behavior", Transportation 17:159-177.
Glaister, S. (1993), "Bus deregulation in Britain", in: J. Carbajo, ed., Regulatory Reform in Transport: Some
Recent Experiences (World Bank, Washington, DC), pp. 50-70.
Glazer, A., D.B. Klein and C.A. Lave (1995), "Clean on paper, dirty on the road: troubles with California's
smog check", Journal of Transport Economics and Policy 29:85-92.
G6mez-Ibfiez, J.A. (1985), "Transportation policy as a tool for shaping metropolitan development", Research
in Transportation Economics 2:55-81.
G6mez-Ibfiez, J.A. (1989), "A global view of auto dependance", Journal of the American Planning
Association 57:376-379.
G6mez-Ibfiez, J.A. (1992), "The political economy of highway tolls and congestion pricing", Transportation
Quarterly 46:343-360.
G6mez-Ibdfiez, J. A. (1996), "Big city transit ridership, deficits and politics: avoiding reality in Boston",
Journal of the American Planning Association 62:30-50.
G6mez-Ibiez, J. A. (1997), "Estimating whether transport users pay their way: the state of the art", in: D.A.
Greene, D.W. Jones and M.A. Deluchi, eds., The Full Costs and Benefits of Transportation: Contributions
to Theory, Method, and Measurement, pp. 149-172 (Springer-Verlag, Berlin).
G6mez-Ibhiez, J.A. and J.R. Meyer (1993), Going Private: The International Experience with Transport
Privatization (The Brookings Institution, Washington, DC).
G6mez-Ibfiez, J.A. and K.A. Small (1994), Road Pricing for Congestion Management: A Survey of Inter-
national Practice, National Cooperative Highway Research Program, Synthesis of Highway Practice 210
(National Academy Press, Washington, DC).
Goodwin, P. B. (1992), "A review of new demand elasticities with special reference to short and long run
effects of price changes", Journal of Transport Economics and Policy 26:155-169.
Goodwin, P.B. (1996), "Empirical evidence on induced traffic: a review and synthesis", Transportation 23:35-
54.
Gordon, P. and R. Willson (1984), "The determinants of light-rail transit demand: an international cross-
sectional comparison", Transportation Research 18A:135-140.
Gordon, P. and H.W. Richardson (1989), "Gasoline consumption and cities: a reply", Journal of the American
Planning Association 55(3):342-346.
Gordon, P. and H.W. Richardson (1994), "Congestion trends in metropolitan areas", in: National Research
Council, Vol. 2, pp. 1-31.
Gordon, P., H.W. Richardson and M.-J. Jun (1991), "The commuting paradox: evidence from the top twenty",
Journal of the American Planning Association 57:416-420.
Gramlich, E.M. (1994), "Infrastructure investment: a review essay", Journal of Economic Literature 32:1176-
1196.
Greene, D.L. (1992), "Vehicle use and fuel economy: how big is the "rebound" effect?", Energy Journal
13:117-143.
Guensler, R. and D. Sperling (1994), "Congestion pricing and motor vehicle emissions: an initial review", in:
National Research Council, Vol. 2, pp. 356-379.
Hahn, R.W. (1989), "Economic prescriptions for environmental problems: how the patient followed the
doctor's orders", Journal of Economic Perspectives 3:95-114.
Halcrow Fox and Associates (1989), The study of mass transit in developing countries, report prepared for the
UK Department of Transport, Transport and Road Research Laboratory, Overseas Unit.
Hall, J.V. (1995), "The role of transport control measures in jointly reducing congestion and air pollution",
Journal of Transport Economics and Policy 29:93-103.
1994 K.A. Small and J.A. Gn6mez-lbdiez
Hanks, J.W., Jr. and T.J. Lomax (1991), "Roadway congestion in major urban areas: 1982 to 1988",
Transportation Research Record 1305:177-189.
Harrington, W. and A. Krupnick (1996), "Public support for congestion and pollution fee policies for motor
vehicles: survey results", working paper, Resources for the Future, Washington, DC.
Harrington, W., M.A. Walls and V.McConnell (1995), "Shifting gears: new directions for cars and clean air",
discussion paper 94-26-REV, Resources for the Future, Washington, DC.
Harvey, G.W. (1994), "Transportation pricing and travel behavior", in: National Research Council, Vol. 2, pp.
89-114.
Hass-Klau, C. (1990), The Pedestrian and City Traffic (Belhaven, London).
Hass-Klau, C. (1993), "Impact of pedestrianization and traffic calming on retailing: a review of the evidence
from Germany and the UK", Transport Policy 1:21-31.
Hau, T.D. (1994), "Estimation of marginal congestion costs, congestion tolls and revenues for urban road use
in Indonesia", in: Proceedings of the International Conference on Advanced Technologies in Transportation
and Traffic Management, pp. 77-88 (Centre for Transportation Studies, Nanyang Technological University,
Singapore).
Hendrickson, C. (1986), "A note on trends in transit commuting in the United States relating to employment in
the central business district", Transportation Research 20A:33-37.
Hensher, D.A. (1993), "Socially and environmentally appropriate urban futures for the motor car", Trans-
portation 20:1-19.
Hilton, G.W. (1974), Federal Transit Subsidies: The Urban Mass Transportation Assistance Program (Ameri-
can Enterprise Institute, Washington, DC).
Holden, D.J. (1989), "Wardrop's third principle: urban traffic congestion and traffic policy", Journal of
Transport Economics and Policy 23:239-262.
International Road Federation (1980), World Road Statistics 1975-1979 (International Road Federation,
Geneva and Washington, DC).
International Road Federation (1990), World Road Statistics 1985-1989 (International Road Federation,
Geneva and Washington, DC).
Jansson, J.O. (1994), "Accident externality charges", Journal of Transport Economics and Policy 28:31-43.
Jansson, K. (1993), "Optimal public transport price and service frequency", Journal of Transport Economics
and Policy 33-50.
Jansson, K. and B. Wallin (1991), "Deregulation of public transport in Sweden", Journal of Transport
Economics and Policy 25:97-107.
Johansen, F. (1989), "Toll road characteristics and toll road experience in selected south east Asia countries",
Transportation Research 23A:463-466.
Jones, P. (1991), "Gaining public support for road pricing through a package approach". Traffic Engineering
and Control 32:194-196.
Jones-Lee, M.W. (1990), "The value of transport safety", Oxford Review of Economic Policy 6:39-60.
Jones-Lee, M.W., G. Loomes and P.R. Philips (1995), "Valuing the prevention of non-fatal road injuries:
contingent valuation vs. standard gambles", Oxford Economic Papers 47:676-695.
Kahn, S. (1986), "Economic estimates of the value of life", IEEE Technology and Society Magazine 5:24-31.
Kain, J.E (1970), "How to improve urban transportation at practically no cost", Public Policy 20:335-358.
Kain, J.F. (1990), "Deception in Dallas: strategic misrepresentation in rail transit promotion and evaluation",
Journal of the American Planning Association 56:184-196.
Kain, John F. (1997), "Cost-effective alternatives to Atlanta's rail rapid transit system", Journal of Transport
Economics and Policy 31:25-49.
Kain, J.F., with R. Gittell, A. Daniere, S. Daniel, T. Somerville and Z. Liu (1992), Increasing the Productivity
of the Nation's Urban Transportation Infrastructure, US Department of Transportation report DOT-T-92-17
(US DOT, Technology Sharing Program, Washington, DC).
Kawada, K., S. Lawrence and G. Gastelum (1997), "I1-15 congestion pricing project" (San Diego Association
of Governments, San Diego, CA).
Kazimi. C. (1997), "Valuing alternative-fuel vehicles in southern California", American Economic Review,
Papers and Proceedings 87:265-271.
Ch. 46: Urban Transportation 1995
Keeler, T.E. and K.A. Small (1977), "Optimal peak-load pricing, investment, and service levels on urban
expressways", Journal of Political Economy 85:1-25.
Kemp, M.A. (1973), "Some evidence of transit demand elasticities", Transportation 2:25-52.
Kennedy, D. (1995), "London bus tendering: an overview", Transport Reviews 15:341-352.
Kim, S. (1995), "Excess commuting for two-worker households in the Los Angeles metropolitan area", Journal
of Urban Economics 38:166-182.
Klein, D.B., A.T. Moore and B. Reja (1997), Curb Rights: A Foundation for Free Enterprise in Urban Transit
(Brookings Institution, Washington, DC).
Koopman, G.J. (1995), "Policies to reduce CO2 emissions from cars in Europe: a partial equilibrium analysis",
Journal of Transport Economics and Policy 29:53-70.
Kyte, M., J. Stoner and J. Cryer (1988), "A time-savings analysis of public transit ridership in Portland,
Oregon", Transportation Research A 22A (5):345-359.
Lave, C.A. (1985), "Speeding, coordination, and the 55 mph limit", American Economic Review 75:1159-
1164.
Lave, C.A. (1991), "Measuring the decline in transit productivity in the U.S"., Transportation Planning and
Technology 15:115-124.
Lave, C.A. (1992), "Cars and demographics"', Access, 1:4-11 (University of California Transportation Center,
Berkeley).
Lave, C.A. (1996), "What really is the growth of vehicle usage?", Transportation Research Record 1520:117-
121.
Lave, L.B., C.T. Hendrickson and F.C. McMichael (1995), "Environmental implications of electric cars",
Science 268:993-995.
Levy, H. (1994), "China: highway development and management issues, options, and strategies", Report No.
13555-CHA (The World Bank, Washington, DC).
Lindley, J. (1987), "Urban freeway congestion: quantification of the problem and effectiveness of potential
solutions", Institute of Traffic Engineers Journal 57:27-32.
Lindley, J. (1989), "Urban freeway congestion problems and solutions: an update", Institute of Traffic
Engineers Journal 59:21-23.
Liu, Z. (1994), "What caused rising transit deficits and escalating operating expenses?", paper presented at the
annual conference of the Eastern Economic Association, Boston.
Mackie, P., J. Preston and C. Nash (1995), "Bus deregulation: ten years on", Transport Reviews 15:317-339.
Mahmassani, H.S. and R. Jayakrishnan (1991), "System performance and user response under real-time
information in a congested traffic corridor", Transportation Research 25A:293-307.
Manne, A.S. and R.G. Richels (1992), Buying Greenhouse Insurance: The Economic Costs of Carbon Dioxide
Emission Limit (MIT Press, Cambridge, MA).
Mannering, F. and C. Winston (1995), "Automobile air bags in the 1990s: market failure or market efficiency?",
Journal of Law and Economics 38:265-279.
McCarthy, P.S. (1990), "Consumer demand for vehicle safety: an empirical study", Economic Inquiry 27:530-
543.
McCubbin, D.R. and M.A. Delucchi (1996), The social cost of the health effects of motor-vehicle air pollution,
report #11 in the series: The annualized social cost of motor-vehicle use in the United States, based on
1990-1991 data, report UCD-ITS-RR-96-3 (Institute of Transportation Studies, University of California,
Davis).
Merriman, D., T. Ohkawara and T. Suzuki (1995), "Excess commuting in the Tokyo metropolitan area:
measurement and policy simulations", Urban Studies 32:69-85.
Meyer, J.R. and J.A. G6mez-Ibafiez (1981), Autos, Transit, and Cities (Harvard University Press, Cambridge,
MA).
Meyer, J.R., J.F. Kain and M. Wohl (1965), The Urban Transportation Problem (Harvard University Press,
Cambridge, MA).
Meyer, M.D. (1994), "Alternative methods for measuring congestion levels", in: National Research Council,
Vol. 2, pp. 32-61.
Miller, TR. (1993), "Costs and functional consequences of U.S. roadway crashes", Accident Analysis and
Prevention 25:593-607.
1996 K.A. Small and J.A. Gdmez-lbdhiez
Mills, G. (1995), "Welfare and profit divergence for a tolled link in a road network", Journal of Transport
Economics and Policy 29:137-146
Mogridge, M.J.H., D.J. Holden, J. Bird and G.C. Terzis (1987), "The Downs/Thomson paradox and the
transportation planning process", International Journal of Transport Economics (Revista Internazionale di
Economia dei Trasporti) 14:283-311.
Mohring, H. (1972), "Optimization and scale economies in urban bus transportation", American Economic
Review 62:591-604.
Mohring. H. (1979), 'The benefits of reserved bus lanes, mass transit subsidies, and marginal cost pricing
in alleviating traffic congestion", in: P. Mieszkowski and M. Straszheim, eds., Current Issues in Urban
Economics, pp. 165-195 (The Johns Hopkins University Press, Baltimore, MD).
Morlok, E.K. and P.A. Viton (1985). "Recent experience with successful private transit in large U.S. cities", in:
C.A. Lave, ed., Private Transit: The Challenge to Public Transportation, pp. 121-149 (Ballinger, Cambridge,
MA).
Nash, C.A. (1988). "Integration of public transport: an economic assessment", in: J.S. Dodgson and N. Topham,
eds., Bus Deregulation and Privatisation, pp. 97-118 (Gower, Aldershot. UK).
National Research Council (1991), Rethinking the Ozone Problem in Urban and Regional Air Pollution
(National Academy Press, Washington, DC).
National Research Council (1994), Curbing Gridlock: Peak-period Fees to Relieve Traffic Congestion,
Transportation Research Board Special Report 242 (National Academy Press, Washington, DC).
Nelessen. A. and L.K. Howe (1995). "Flexible, friendly neighborhood transit: a solution for the suburban
transportation dilemma", Working Paper No. 86 (Center for Urban Policy Research, Rutgers University,
Piscataway, NJ).
Newbery. D.M. (1988), "Road user charges in Britain", Economic Journal 98: 161-176.
Newbery, D.M. (1990), "Pricing and congestion: economic principles relevant to pricing roads", Oxford
Review of Economic Policy 6:22-38.
Newman. P.W.G. and J.R. Kenworthy (1989), Cities and Auto Dependence: A Sourcebook (Gower, Aldershot,
UK).
Newman. P.W.G. and J.R. Kenworthy (1991), "Transport and urban form in thirty-two of the world's principal
cities", Transport Reviews 11:249-272.
Nowlan, D.M. and G. Stewart (1991), "Downtown population growth and commuting trips: recent experience
in Toronto", Journal of the American Planning Association 57:165-182.
OECD (Organisation for Economic Co-Operation and Development) (1994), Congestion Control and Demand
Management (OECD, Paris).
Oum, T.H., W.G. Waters and J.-S. Yong (1992), "Concepts of price elasticities of demand and recent empirical
estimates", Journal of Transport Economics and Policy 26:139-154.
Peltzman, S. (1975), "The effects of automobile safety regulations". Journal of Political Economy 83:677-725.
Perl, A. and J. Pucher (1995), "Transit in trouble?: the policy challenge posed by Canada's changing urban
mobility", Canadian Public Policy 21:261-283.
Perry, J.L. and T.T. Babitsky (1986), "Comparative performance in urban bus transit: assessing privatization
strategies", Public Administration Review 46:57-66.
Phang, S-Y and M. Asher (1997), "Recent developments in Singapore's motor vehicle policies", Journal of
Transport Economics and Policy 31:211-220.
Phang, S-Y and R. Toh (1997), "From manual to electronic road congestion pricing: the Singapore experience
and experiment", Transportation Research E 33:97-106.
Pickrell, D.H. (1985a), "Rising deficits and the uses of transport subsidies in the United States", Journal of
Transport Economics and Policy 24:281-298.
Pickrell, D.H. (1985b), "Urban rail in America: a review of the regional planning association study", US
Department of Transportation Volpe Center Report No. SS-64V.5-3.
Pickrell, D.H. (1992), "A desire named streetcar: fantasy and fact in rail transit planning", Journal of the
American Planning Association 58:158-176.
Prud'homme, R. (1997), "On transport-related extemalities", Working Paper No. 97-15 (Observatoire de
'Economie et des Institutions Locales, University of Paris).
Ch. 46: Urban Transportation 1997
Pucher, J. (1988), "Urban travel behavior as the outcome of public policy", Journal of the American Planning
Association 54:509-519.
Pucher, J. (1995a), "Urban passenger transport in the United States and Europe: a comparative analysis of
public policies", Part 1: "Travel behavior, urban development, and automobile use", Transport Reviews
15:99-117.
Pucher, J. (1995b), "Urban passenger transport in the United States and Europe: a comparative analysis of pub-
lic policies", Part 2: "Public transport, overall cost comparisons and recommendations", Transport Reviews
15:211-227.
Pucher, J., D. Hendrickson and S. McNeil (1981), "Socioeconomic characteristics of transit riders: some recent
evidence", Traffic Quarterly 35:461-483.
Pucher, J. and S. Kurth (1995), "Making transit irresistible: lessons from Europe", Transportation Quarterly
49:117-128.
Pucher, J. and C. Lefevre (1996), The Urban Transport Crisis in Europe and North America (Macmillan,
Basingstoke, UK).
Pucher, J., A. Markstedt and I. Hirschman (1983), "Impacts of subsidies on the costs of urban public transport",
Journal of Transport Economics and Policy 17:156-176.
Quinet, E. (1994), "The social costs of transport: evaluation and links with intemalisation policies", in: Euro-
pean Conference of Ministers of Transport (ECMT), Internalising the Social Costs of Transport, Publications
Service, pp. 31-75 (OECD, Paris).
Rom, M. (1994), "The politics of congestion pricing", in: National Research Council, Vol. 2, pp. 280-299.
Rossetti, M.A. and B.S. Eversole (1993), Journal to Work Trends in the United States and Its Major Metropol-
itan Areas, 1960-1980. Available from the National Technical Information Service at Springfield, Virginia
as Publication No. FHWA-PL-94-012 (Volpe National Transportation Systems Center, US Department of
Transportation, Cambridge, MA)..
Rouwendal, J. and P. Rietveld (1994), "Changes in commuting distances of Dutch households", Urban Studies
31:1545-1557.
Ruster, J. (1997), "A retrospective on the Mexican toll road program (1989-1994)", in: The Private Sector in
Infrastructure: Strategy, Regulation, and Risk, pp. 117-124 (World Bank, Washington, DC).
SACTRA (1994). Trunk Roads and the Generation of Traffic (Her Majesty's Stationery Office, London).
Schrank, D.L., S.M. Turner and T.J. Lomax (1993), Estimates of Urban Roadway Congestion-1990, Research
Report 1131-5 (Texas Transportation Institute, College Station).
Shah, A.M. (1990), "Optimal pricing of traffic extemalities: theory and measurement", International Journal
of Transport Economics 17:3-19.
Sharman, K. and M. Dasgupta (1993), Urban Travel and Sustainable Development: An OECD/ECMT Study
of 132 Cities (Transport Research Laboratory, Department of Transport, Crawthorne, UK).
Shavell, S. (1987), An Analysis of Accident Law (Harvard University Press, Cambridge, MA).
Sherret, A. (1975), Immediate Travel Impacts of Transbay BART. Peat, Marwick, Mitchell &Co. Report TM
15-3-75 to US Department of Transportation (National Technical Information Service, Springfield, VA).
Shimazahi, T. and M. Rahman (1995), "Operational characteristics of paratransit in developing countries of
Asia", Transportation Research Record, No. 1503, pp. 49-56.
Shoup, D. (1994), "Cashing out employer-paid parking: a precedent for congestion pricing?", in: National
Research Council, Vol. 2, pp. 152-199.
Small, K.A. (1983), "Bus priority and congestion pricing on urban expressways", in: T.E. Keeler, ed., Research
in Transportation Economics, Vol. 1, pp. 27-74 (JAI Press, Greenwich, CT).
Small, K.A. (1985), "Transportation and urban change", in: P.E. Peterson, ed., The New Urban Reality, pp.
197-223 (Brookings Institution, Washington, DC).
Small, K.A. (1 992a), Urban Transportation Economics (Harwood Academic Publishers, Chur, Switzerland).
Small, K.A. (1992b), "Using the revenues from congestion pricing", Transportation 19:359-381.
Small, K.A. and J.A. G6mez-Ibbfiez (1998), "Road pricing for congestion management: the transition from the-
ory to policy", in: K.J. Button and E.T. Verhoef, eds., Road Pricing, Traffic Congestion and the Environment
(Edward Elgar, Aldershot, UK).
Small, K.A. and C. Kazimi (1995), "On the costs of air pollution from motor vehicles", Journal of Transport
Economics and Policy 29:7-32.
1998 K.A. Small and J.A. Gdmrnez-bdiez
Small, K.A. and S. Song (1992),' "'Wasteful' commuting: a resolution", Journal of Political Economy 100:888-
898.
Small, K.A., C. Winston and C.A. Evans (1989), Road Work: A New Highway Pricing and Investment Policy
(Brookings Institution, Washington, DC).
Stone, J.R., A. Nalevanko and J. Tsai (1993), "Assessment of software for computerized paratransit opera-
tions", Transportation Research Record, No. 1378, pp. 1-9.
Teal, R.F. (1988), "Public transit service contracting: a state report", Transportation Quarterly 42:207-222.
Teal, R.F. (1991), "Issues raised by competitive contracting of bus transit service in the USA", Transportation
Planning and Technology 15:391-403.
Teal, R.F. (1993), "Implications of technological developments for demand responsive transit", Transportation
Research Record, No. 1390, pp. 33-42.
Thomson, J.M. (1977), Great Cities and Their Traffic (Gollancz, Peregrine Edition, London).
Traffic speeds in central and outer London: 1993-94 (1995), Traffic Engineering and Control 36:101.
Turner, R.P. and P.R. White (1991), "Overall impacts of bus deregulation in Britain", Transport Planning and
Technology 15:203-229.
Turvey, R. and H. Molring (1975), "Optimal bus fares", Journal of Transport Economics and Policy 9:1-7.
Tyson, W.J. (1989), A Review of the Second Year of Bus Deregulation (Association of Metropolitan Counties,
London).
UK Department of Transport (1993), Bus and Coach Statistics in Great Britain 1992/93 (Her Majesty's
Stationary Office, London).
United Nations (1975, 1985, 1992), Demographic Yearbook (New York: UN).
US Census Bureau (1994, 1995, 1997, 1998), Statistical Abstract of the United States (US Government Printing
Office, Washington, DC).
US FHWA (US Federal Highway Administration) (1987, 1988, 1990, 1995), Highway Statistics (US
Government Printing Office, Washington, DC).
Vickrey, W.S. (1955), "Some implications of marginal cost pricing for public utilities", American Economic
Review, Papers and Proceedings 45:605-620.
Vickrey, W.S. (1968), "Automobile accidents, tort law, externalities, and insurance: an economist's critique",
Law and Contemporary Problems 33:464-487.
Vickrey, W.S. (1969), "Congestion theory and transport investment", American Economic Review, Papers and
Proceedings 59:251-260.
Viscusi, V.K. (1993), "The value of risks to life and health", Journal of Economic Literature 31:1912-1946.
Viton, P.A. (1980), "The possibility of profitable bus service", Journal of Transport Economics and Policy
14:295-314.
Viton, P.A. (1983), "Pareto-optimal urban transportation equilibria", in: T.E. Keeler, ed., Research in
Transportation Economics, Vol. 1, pp. 75-101 (JAI Press, Greenwich, CT).
Viton, P.A. (1992), "Consolidations of scale and scope in urban transit", Regional Science and Urban
Economics 22:25-49.
Viton, P.A. (1995), "Private roads", Journal of Urban Economics 37:260-289.
Wachs, M. (1989), "U.S. transit subsidy policy: in need of reform", Science 244:1545-1549.
Walters, A.A. (1961), "The theory and measurement of private and social cost of highway congestion",
Econometrica 29:676-699.
Webster, EV., P. Bly, R.H. Johnston, N. Pauley and M. Dasgupta (1986), "Changing patterns of urban travel",
Transport Reviews 6:49-86, 129-172.
Wheaton, W. (1982), "The long run structure of transportation and gasoline demand", Bell Journal of
Economics 13:439-455.
White, P.R. (1990), "Bus deregulation: a welfare balance sheet", Journal of Transport Economics and Policy
24:311-332.
White, P.R. (1995), "Deregulation of local bus services in Great Britain: an introductory review", Transport
Reviews 15:65-89.
Whitelegg, J. (1993), Transport for a Sustainable Future: The Case for Europe (Belhaven Press, London).
Ch. 46: Urban Transportation 1999
Winston, C. and B. Bosworth (1992), "Public infrastructure", in: H.J. Aaron and C.L. Schultze, eds.,
Setting Domestic Priorities: What Can Government Do?, pp. 267-293 (Brookings Institution, Washington,
DC).
Winston, C. and E Mannering (1984), "Consumer demand for automobile safety", American Economic
Review, Papers and Proceedings 74:316-319.
Zho, M. and V. Sisiopiku (1997), "Relationship between volume-to-capacity ratios and accident rates",
Transportation Research Record 1581:47-52.
Chapter47
STEPHEN ROSS*
University of Connecticut
JOHN YINGER
Syracuse University
Contents
1. Introduction 2003
2. Positive theory 2003
2.1. The impact of local fiscal variables on housing prices and community choices 2004
2.2. The impact of bidding and sorting on public service levels and property tax rates 2019
2.2.1. Key questions in models of public service determination 2019
2.2.2. Examples of voting models 2023
2.3. Models in which both housing and fiscal variables are endogenous 2026
2.3.1. Epple, Filimon and Romer 2026
2.3.2. Other equilibrium proofs 2029
3. Empirical research 2030
3.1. The impact of public services on housing market outcomes 2030
3.1.1. Property tax capitalization 2030
3.1.2. Public service capitalization 2033
3.1.3. Heterogeneity 2035
3.1.4. Zoning 2036
3.2. The impact of bidding and sorting on public service outcomes 2037
3.2.1. Income and price elasticities 2037
3.2.2. Population and community characteristics 2038
3.2.3. Tiebout bias 2040
4. Normative analysis 2042
4.1. The efficiency of a system of local government 2042
* The authors are grateful to Liz Bernhard, Alex Stricker and Jamila Thompson for research assistance and
to Graham Crampton, Dennis Epple, Hydeo Konishi, Thomas Nechyba, David Sjoquist and Myma Wooders
for helpful comments.
2001
2002 S. Ross and J. Yinger
Abstract
This chapter reviews the literature on the boundary between urban economics and
local public finance, defined as research that considers both a housing market
and the market for local public services. The first part of the chapter consid-
ers positive theories. This part presents the consensus model of the allocation
of households to jurisdictions, which is built on bid functions and household
sorting, as well as alternative approaches to this issue. It also examines models
of local tax and spending decisions, which exhibit no consensus, and reviews
research in which both housing and local fiscal variables are endogenous. The
second part of the chapter considers empirical research, with a focus on tax and
service capitalization, on household heterogeneity within jurisdictions, and on
the impact of zoning. The third part considers normative theories about a de-
centralized system of local governments. This part examines the extent to which
such a system leads to an efficient allocation of households to communities or
efficient local public service levels, and it discusses the fairness of local public
spending. This review shows that the bidding/sorting framework is strongly sup-
ported by the evidence and has wide applicability in countries with decentralized
governmental systems. In contrast, models of local public service determination
depend on institutional detail, and their connections with housing markets have
been largely unexplored in empirical work. Ever since Tiebout (1956), many
scholars have argued that decentralized local governments have efficiency advan-
tages over centralized forms. However, a general treatment of this issue identifies
four key sources of inefficiency even in a decentralized system: misallocation of
households to communities, the property tax, public service capitalization and
Ch. 47: Sorting and Voting 2003
heterogeneity. Few policies to eliminate these sources of inefficiency have yet
been identified. Finally, this review explores the equity implications of household
sorting and other features of a decentralized system.
1. Introduction
The literature on urban public finance revolves around two questions: How do
households select a community in which to live? What determines the level of
local public services? To answer these questions, researchers must analyze both
the housing market and the market for local public services-and how these
two markets influence each other-and therefore must operate on the bound-
ary between urban economics and local public finance. This chapter surveys the
literature on this boundary. It covers theoretical work, both positive and norma-
tive, that considers both housing and local public services, along with related
empirical work. Section 2 presents positive theory, Section 3 reviews empirical
research, and Section 4 considers normative analysis. Most of the research on
these topics concerns the US, but this review explores both the applicability of
existing research to and availability of evidence for other countries.
2. Positive theory
The literature focuses on three central conceptual problems: (1) how does the
housing market allocate households to communities when local public services
and taxes vary from one community to the next?; (2) how do communities select
the level of local public services and tax rates? These two problems obviously
interact as the people who move into a community become voters. Hence, the
third problem is: (3) under what conditions are solutions to the first two problems
compatible, that is, when does an urban equilibrium exist? This section addresses
the conceptual literature on each of these problems.
2004 S. Ross and J. Yinger
A broad consensus has emerged concerning the appropriate way to model the first
conceptual problem, namely, the allocation of households to communities. This
consensus has general applicability to any country with active housing markets,
reasonably mobile households, and multiple local governments that exhibit some
variation in public service levels or tax rates. It evolved from the central positive
insight of Tiebout (1956), namely, that households care about local public ser-
vices and local taxes and compete for entry into the most desirable communities,
and builds on the analytical framework developed in the pioneering work of
Ellickson (1971).
The current consensus is contained in similar models presented by Epple
et al. (1983, 1984, 1993), Epple and Romer (1989), Henderson (1977, 1991),
Pogodzinski and Sjoquist (1985, 1991, 1993), Wheaton (1993) and Yinger (1982,
1985, 1995). It has two components: bidding and sorting.
Maximize P= H ( 1
H(1+ t*)
H, Z (2.1)
Subject to U(Z, H, S) = U°(Y),
Us/Uz MB
H(1 +t*) H(1 + t*)
and
P P/r
Pt = - - 1 - (2.3)
r+t 1+ t*
2 Ellickson (1971) and Epple et al. (1983, 1984, 1993) talk about "indifference curves" in [P(l + t*), S]
space, which are the same thing as "bid functions".
3 For a bid-function approach that adds differences in preferences, see Epple and Platt (1998).
2006 S. Ross and J. Yinger
H PH _PH/r
V= =- = . (2.4)
r r+t Il+t*
Sorting. These results describe the housing bids for a single household but do not
reveal how different types of households are sorted into jurisdictions. The key to
sorting is to recognize that Eq. (2.2), the slope of the bid function with respect to
S, has a different value for different household types. This leads to a well-known
diagram, which is in Ellickson (1971), Henderson (1977), Yinger (1982), Epple
et al. (1983, 1984) and Wheaton (1993) and presented here as Fig. 1.5 Housing
suppliers sell to the highest bidder, so observed prices are the envelope of the bid
functions, and the group with the steepest bid function wins the competition for
housing at the highest value of S. The exact boundaries between income classes
depend on the heights of the bid functions, not just the slopes. These heights
are related in turn to each group's utility level. 6 Nevertheless, the ordering of
household types along the S-axis depends only on the relative slopes of their bid
functions.
This approach implies that heterogeneous jurisdictions are possible. If a ju-
risdiction provides a value of S at which the bid functions of two groups cross,
4 The first formal derivation of Eq. (2.2) is in Henderson (1977). Equation (2.3) is equivalent to the standard
expression for property tax capitalization, which has a long history. See Yinger et al. (1988).
5 Ellickson (1971), Henderson (1977) and Epple et al. (1983, 1984) draw this figure with P(1 + t*) on the
vertical axis, whereas Yinger (1995) and Wheaton (1993) draw it with P. As discussed in later footnotes, this
difference is inconsequential for sorting, but does influence the role that bid functions play in the analysis. For
illuminating sorting, however, drawing the figure with P on the vertical axis makes more sense because P is
the price that housing suppliers receive; suppliers will sell or rent to the household type with the highest P, not
the highest P(1 + t*). Moreover, a focus on P(1 + t*) requires additional analysis, supplied by Epple et al.
(1983, 1984) but not Ellickson (1971), to capture the behavior of housing sellers. Henderson (1977) assumes
that suburbs can be costlessly created so that P is constant across suburbs. We return to this assumption below.
6 Readers familiar with urban models will recognize the logic here. In an open urban model, the height of
a bid function (over location) is fixed by a group's utility in a system of cities; in a closed model the height is
raised (and utility lowered) until there is enough room for all members of the group. See Brueckner (1987).
Ch. 47: Sorting and Voting 2007
P1
P2
P3
Ii
*
! ' I
,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Cn
S1 S S
both groups are content to live there. In several models, including those of Epple
et al. (1983, 1984, 1993), income is modeled as a continuous variable and every
jurisdiction contains a range of income classes. In the bid-function approach,
therefore, jurisdictions are more homogenous than with a random allocation of
households, but may nevertheless be fairly heterogeneous.
Figure 1 also implies that observed differences in housing prices across ju-
risdictions do not necessarily correspond to a household's willingness to pay for
the corresponding service quality differences. When income classes are defined
continuously, as in Epple et al. (1983, 1984, 1993), the observed price difference
between two communities with adjacent values of S equals the willingness to
pay only for "border" or marginal households which are indifferent between
residing in either community.7 In addition, observed price differences between
communities with nonadjacent levels of S may not correspond to any group's bid
differences. 8
7 With discrete income classes, one class may fill a jurisdiction, implying that its bid exceeds the bid of
every other class. In this case, observed price differences between that jurisdiction and jurisdictions inhabited
by other classes, even if they provide adjacent levels of S, may not correspond exactly to the bid differences of
any group.
8 Pogodzinski (1988) criticizes Yinger's (1982) claim that capitalization may not be "complete" in the
sense that observed price differences may not correspond to some group's bid difference. This strikes us
as a semantic disagreement, not a substantive one. From the perspective of the marginal household class,
complete capitalization always exists between communities with adjacent values of S (Pogodzinski's claim),
but estimated price differences across a range of communities need not correspond to the bid differences of
2008 S. Ross and J. Yinger
Most local public services are normal goods, but the standard market mecha-
nisms are not available to allow higher income households to buy more of them.
A natural question to ask, therefore, is whether or not bidding alone is suffi-
cient to sustain "matched" sorting, which is defined to exist when higher income
households sort into jurisdictions with higher S. To answer this question, it will
prove convenient to write problem (1) in a somewhat different form, introduced
by Wheaton (1993). 9 To be specific, let us assume that both S and t are known
functions of some jurisdictional characteristic, say Y. We interpret Y as a juris-
diction's median income, but it could, as in Wheaton, be a more general index
that incorporated preferences or other factors that boost service quality.
Furthermore, we can simplify the problem by introducing a jurisdiction's
budget constraint, ignoring for now nonresidential property and revenue from
sources other than the property tax. 10 Because all households in a jurisdiction
must pay the same price per unit of housing services, Y affects house value only
through its impact on average house size, written H, and the average property
taxbase in the community is PH(Y)/r. On the expenditure side, the cost of pub-
lic services per household is written E[S(Y), Y]. This form recognizes not only
that costs increase with the public service quality, S, but also, as discussed more
fully below, that costs may depend on Y itself. Thus, the jurisdiction's budget
constraint is
any particular group (Yinger's claim). Moreover, as pointed out in the previous footnote, with discrete income
classes a marginal household type may not exist.
9 Although we use Wheaton's (1993) technique, we do not reproduce his problem, which unlike the rest
of the literature, enters land and housing capital as separate arguments in the household utility function. We
return to this issue in footnotes 13 and 60.
10 A budget constraint is not needed to describe how bids change as Y changes--or to look at matched
sorting. Without a budget constraint, we can make S and t in Eq. (2.1) functions of Y and use the envelope
theorem to derive
MB ) dS P/r dt
H( t*) dY +t* dY
Inserting a budget constraint helps to draw out the intuition of the problem by removing ambiguity about the
relationship between Y and t.
Ch. 47: Sorting and Voting 2009
P1
P2
P3
I i
I I
, , I
lI I-
1 Y2 Y3
Fig. 2. Consensus bidding and sorting net of taxes.
Solving Eq. (2.5) for t* and substituting the result into Eq. (2.1) yields:
Maximize y-z
P= _
Y-(Y) _ E[S(Y), Y]
H, Z (2.6)
Subject to U(Z, H, S(Y)] = U°(Y).
Another application of the envelope theorem now reveals that the slope of the bid
function with respect to Y is
( -) (MB- MC-)
CH) + (- ) - (2.7)
The second pair of parentheses marks the net benefit from additional services,
that is, the excess of a household's marginal benefit from S over its tax price,
which equals MC multiplied by the voters tax share, H/H = V/V. As illus-
trated in Fig. 2, this term, and hence P?, could be positive or negative.
The third pair of parentheses indicates the percentage increase in taxbase that
accompanies an increase in Y weighted by spending per unit of taxbase. This
term measures a household's cost savings from moving to a jurisdiction where
Y is larger and hence its tax price is lower.ll Finally, the fourth pair of paren-
11 Ellickson (1971) was the first to identify this effect and to point out that it gives high income jurisdictions
an incentive to keep out low income households. See also Hamilton (1975) and Henderson (1977).
2010 S. Ross and J. Yinger
theses marks the link between community characteristics and the cost of S. The
evidence, discussed below, indicates that aE/aY is negative, so the cost savings
from moving to a higher income jurisdiction serves to boost housing bids there.
Now consider two income classes that both live in the same jurisdiction. As
shown in Figs. 1 and 2, this implies that they both have the same bid there or, to
put it another way, pay the same P. However, the slopes of their bid functions
with respect to S, described by Eq. (2.2) and Fig. 1, or with respect to Y, de-
scribed by Eq. (2.7) and Fig. 2, may differ. 12 For "matched" sorting to occur, the
class with the higher income must have a bid function with a steeper slope at all
values of S; that is, the derivative of P9 with respect to Y must be positive. From
Eq. (2.7) this derivative is13
aP~
P( MB dS X
aY
By- ( B dSY (
YH d )
)' (2.8)
Maximize P ES(Y),
H Y](2.9) (2.9)
H, Z
Subject to U[Z, H, S(Y)] = U°(Y).
Thi Hesul)
(MB MCY) + (2 Hj - (y . (2.10)
This result is analogous to Eq. (2.7), but its derivative with respect to income
is much more complicated because three terms are now functions of Y: the tax
share, Y/Y; the term in the third pair of parentheses, which reflects the benefits
from moving to a community with a higher taxbase; and the last term, which
reflects the gain from moving to a lower-cost community. Hence, the simple
sorting rule in Eq. (2.8) no longer holds.16
each unit of S costs $g, and this cost is the same in all jurisdictions. The second
part rules out public goods, for which the per-unit cost declines with population.
Second, the supply of housing associated with any given service-tax package is
perfectly elastic. Because the supply of housing cannot be elastic in a jurisdiction
with fixed boundaries, this assumption requires that the boundaries of existing
jurisdictions be flexible or that new jurisdictions can be costlessly created. Third,
a jurisdiction can use zoning restrictions to set the consumption of H at exactly
the optimal level. In our framework, this assumption implies that H is no longer
a choice variable for a household in search of housing but is instead a function of
Y. Note also that the third assumption has two parts: zoning must be exact, in the
sense that it sets the same minimum value of H for every household, and it must
be set at the optimal level of H.
As shown earlier, a household's utility (and hence its housing bid) increases
as a jurisdiction's average taxbase increases, that is, as its own tax share de-
creases. This result reflects the fact that with a property tax, households with
relatively high H (and hence a high tax share) cover some of the cost of S for
households with relatively low H (and hence a low tax share). Hamilton's (1975)
assumptions make it possible for high income households to prevent this type
of subsidization. The assumption of perfectly elastic housing supply makes it
possible for anyone to set up a homogeneous community with housing priced at
P, which is the marginal cost of H when land is priced at the agricultural rental
rate.17 Moreover, exact zoning ensures that everyone consumes the same level of
H (because no household would voluntarily consume more than the minimum
and therefore subsidize its neighbors) and perfect zoning ensures that this level
of H is optimal; thus, the community can keep out anyone who wants to consume
less H than the optimal level for current residents.
In formal terms, the Hamilton (1975) assumptions transform the household's
problem into one in which the household knows that the value of H it selects also
will be the average value of H in the jurisdiction (Epple et al., 1978). Thus, Eq.
(2.1) becomes:
Maximize P= Y - Z - gS
H, Z, S (2.11)
Subject to U(Z, H, S) = U°(Y).
17 Earlier work by Edel and Sclar (1974) predicted, like Hamilton (1975), that housing prices would not
vary in the long run because the supply of housing is elastic. See also Epple et al. (1978) and Henderson (1977,
1985).
Ch. 47: Sorting and Voting 2013
The first-order conditions of this problem imply that a household chooses S and
H so that
Applying the envelope theorem and assuming that, to start with, everyone is in a
jurisdiction where H equals H, we find that
(2.14)
According to Eq. (2.12), the first two terms drop out; households place no value
on the increases in S and H associated with moving to the higher income com-
munity. The third term drops out by the assumption that zoning is exact, so that
everyone's H is at the minimum allowable value. The final term drops out by the
assumption that g is fixed across communities. Thus, under Hamilton's (1975)
assumptions, P = 0 equals zero. This defines the situation first described by
Pauley (1976), in which there is no capitalization because every household lives
in the community where its bid is highest. This situation is illustrated in Fig. 3.
18 This formulation assumes that if a household moves to a community with a higher value of Y, the zoning
constraint there will be binding. A more general, but no more illuminating, version of the problem would allow
H to exceed H.
2014 . Ross and J. Yinger
p ---- 0 -- -
I I
- - - - - - - ~~ ~~I- ~7 --w
J II I
Y Y2 Y3 Y
Fig. 3. Hamilton bidding and sorting.
This discussion not only confirms Hamilton's reasoning, but also reveals that
each of Hamilton's (1975) assumptions is crucial. Without the ability to cost-
lessly create new jurisdictions, the base community may have heterogeneity in
housing consumption, the first part of Eq. (2.12) may not hold, and thus the
first term in Eq. (2.14) may not equal zero.l 9 Without zoning at the optimal
level of housing consumption in the base community, the second part of Eq.
(2.12) may not hold, and thus the second term in Eq. (2.14) may not equal zero.
Without exact zoning, bids reflect the advantages of moving to a jurisdiction with
a higher taxbase, and the third term is not zero. And if the cost of S declines with
income, households will bid up the price in the higher income-and hence lower
cost-jurisdiction so that the last term is not zero.
The assumptions that lead to Hamilton's (1975) equilibrium appear extreme.
First, as pointed out by Yinger (1982) and Rose-Ackerman (1983), the possibility
for new jurisdictions is limited by transportation costs; one cannot bid land away
from agriculture at great distances from employment concentrations. 2 0 Moreover,
no one has provided any evidence that jurisdiction boundaries change to increase
19 Conditions analogous to Eq. (2.12) with heterogenous housing (and a property tax) are derived in Section
4.1.3.
20 This is one of the few points at which the logic of a spatial model, such as those cited in footnote 1, is
brought into a multijurisdiction framework. For a more formal treatment of space with many jurisdictions, see
Crampton (1996).
Ch. 47: Sortingand Voting 2015
the supply of housing where housing bids are high.2 1 Furthermore, the evidence,
reviewed below, indicates that g declines as a jurisdiction's income goes up.
Finally, the assumptions about zoning are very strong. Actual zoning tools,
such as lot-size restrictions and set-back rules, appear far too blunt to control H
precisely, and indeed voters would have to be foresightful to set zoning barriers
at exactly the optimal level of H. For example, if voters are myopic, in the sense
that they take the existing value of H as given, then their perceived price of H is
P(1 + t*) = P + E(S)/H, and they select a level of H that is too low due to
this property tax distortion. To prevent subsidization of incoming residents, they
will set the zoning level equal to their own preferred H, so that it, too, will be
too low. In contrast, if voters' own houses are exempt from the zoning constraint
they select, they have a strong incentive to set the constraint above their own
H so that future entrants will subsidize their taxes (White, 1975; Miceli, 1991,
1992). This incentive suggests that zoning constraints may be too strict, in the
sense that they force people to consume more H than they want, and may reduce
property values.
Fischel (1992) argues that local zoning tools are varied and flexible and
presents evidence that zoning ordinances are driven primarily by fiscal concerns.
According to Fischel, this evidence supports the Hamilton (1975) model. In our
view, however, the existence of fiscal zoning is consistent with many models, not
just with the Hamilton model. Voters who do not understand that matched sorting
may arise naturally may pass zoning ordinances to mimic what the market would
do anyway. If the elasticity condition in Eq. (2.8) is not satisfied, communities
may pass zoning ordinances that ensure matched sorting but do not eliminate
capitalization. 2 2 Moreover, if public service costs decline with income, so that the
last term in Eq. (2.14) is not zero, communities may set zoning above the optimal
level of H in order to keep out low income and, hence, high cost households. 2 3
Finally, communities may use zoning as a form of insurance against upward
shifts in the bid functions of lower income groups due to immigration or natural
population growth.
Thus, a test of Hamilton's model is not whether fiscal zoning exists, but
whether, or not, fiscal zoning is set at exactly the optimal level of H. Moreover,
21 As pointed out by Henderson (1985), jurisdiction boundaries change fairly often, particularly in the
Southwest. However, Epple and Romer (1989) show that these changes do not involve expansions of high-bid
groups into low-bid areas as the Hamilton and Henderson models imply.
22 Henderson (1977), Wheaton (1993) and Yinger (1995) describe zoning rules that serve this function.
23 This point is made by Oates (1977b), who presents survey evidence that cost-related fiscal zoning is
far more important than the taxbase related fiscal zoning in the Hamilton approach. Fischel (1992) does not
distinguish between these two types of fiscal zoning.
2016 S. Ross and J. Yinger
this model implies that the zoning restriction (or level of H) in one jurisdiction
should be independent of the zoning decisions of surrounding jurisdictions; if
every household obtains its optimal H, no resident has an incentive to move, and
no voter needs to consider what other jurisdictions do.
The Hamilton model also predicts no capitalization, so that a house value
regression will obtain zero coefficients for S and t. This model does not imply
that people do not care about S and t, or even that S and t do not affect their
housing bids, only that these effects cannot be observed. If P (or V, holding H
constant) does not vary with S and t (or, equivalently, with factors that influence
S and t, summarized by Y here), then an econometric analysis of P has nothing
to work with and will conclude that S and t have no impact on p. 24 If P does vary
with S and t, so that coefficients for these two variables can be estimated, then
the principal prediction of the Hamilton model, a constant P, can be rejected.
Statistically significant capitalization of S or t therefore serves as a rejection of
the Hamilton model.2 5
This conclusion can be clarified by distinguishing between the bid function of
a particular class and the observed housing price function. As shown in Fig. 3,
the Hamilton bid function for any class reaches a maximum at the S in its own
jurisdiction. This bid function reflects the willingness to pay both for higher S
and for lower t. If that bid function could be observed, its slope would be zero at
that service quality. To put it another way, the Hamilton model predicts that the
marginal impact of higher S (due to higher Y) on that groups's bids is exactly
offset by the marginal impact of a change in t (due to higher Y). Because each
class's bid function is observed at only one point, however, this prediction cannot
be tested. A regression is based on the values of P at different values of Y,which,
according to Hamilton, should not differ.
Several scholars have argued that the Hamilton approach implies zero net
capitalization, not no capitalization at all. 26 Zero net capitalization exists when
the "border" household's willingness to pay to avoid the difference in t between
24 This argument does not require the use of Y. The Hamilton model assumes that any community can be
costlessly replicated, so no community can possess an advantage over other communities in the production of
S-an advantage that might. if it existed, show up in P.
25 The argument in the text applies to interjurisdictional capitalization. A related argument can be made
for intrajurisdictional capitalization: If differences in S or t exist within a jurisdiction (as demonstrated by
their impact on property values) then the homogeneity predicted by the Hamilton model must not exist. Thus,
statistically significant intrajurisdictional capitalization also constitutes evidence against the Hamilton model.
26 See Sonstelie and Portney (1980) and Oates (1994). Hamilton himself originally believed that his ap-
proach involved no capitalization but then changed his mind. As he put it (1983b; p. 87). the precise link
between taxes and public services in his approach "has unfortunately led some observers (see Hamilton, 1975)
to refer to this case as zero capitalization rather than full capitalization, thus creating some confusion in the
literature". As we show below, Hamilton's change of mind depends on another extreme assumption. As noted
Ch. 47: Sorting and Voting 2017
two communities just equals its willingness to pay for the associated difference
in S. In our judgement, this view confuses the bid function of a given household
type with the observed housing price function. As shown above, it is indeed true
that an infinitesimal increase in Y leads to offsetting impacts on P through S and
through t, but Fig. 3 clearly shows that this change in bids is not what is observed
between two communities. Because the second derivative of P with respect to Y
is negative, a class will bid less in a community other than in its own-unless
the values of Y in the two communities differ by only an infinitesimal amount.
Hence, it cannot be said that there is zero net capitalization between jurisdictions
1 and 2 in Fig. 3; a household who lives in jurisdiction 1 is unwilling to pay
enough for the increment in S it would receive in jurisdiction 2 to cover what
it would lose from higher t, so its (unobserved) bid in jurisdiction 2 is below
the market price. With anything less than an infinite number of jurisdictions,
therefore, it is the elastic supply of housing, not a balance between benefits and
costs of S for each household class, that keeps the price of housing constant in
the Hamilton model.
In a later paper, Hamilton (1976) employs an assumption that ensures zero net
capitalization without an infinite number of communities, namely, that the bene-
fits from S can be measured by local public spending. In particular, he examines
a community that contains high and low income housing (designated HIH and
LIH) and assumes that the price of HIH is given by Hi = H' + (X - Hiti)D,
"where Hi = the value of the HIH dwelling unit in community i; H' = the value
of the same structure in a homogeneous HIH community; X = the per household
expenditure on the local public service; ti = community i's property tax rate;
D = the discount factor that converts a constant annual stream into a present
value" (Hamilton, 1976: p. 746). In formal terms, this assumption implies that the
marginal willingness to pay for (or marginal benefit from) $1 of public spending
is $1. Because $1 of property taxes can finance exactly $1 of public spending, this
assumption ensures that every household's willingness to pay for a difference in
S between any two communities is exactly offset by the associated property tax
difference.
This assumption is extreme. Whereas a standard demand curve sets MB =
MB(Y, S), this assumption sets MB = $1, which requires both the income and
price elasticities of demand for S to equal zero. This requirement is strongly
contradicted by the evidence (discussed below). In short, without adding another
extreme assumption (either infinite communities or zero income and price elas-
earlier, several other scholars are clear that the Tiebout or Hamilton assumptions imply no capitalization (Edel
and Sclar. 1974; Epple et al., 1978).
2018 S. Ross and J. Yinger
ticities of demand for S), the Hamilton (1975) approach implies that neither t
nor S has any effect on property values. 2 7 Moreover, even with one of these
assumptions, one cannot distinguish empirically between zero net capitalization
and no capitalization. 2 8
This analysis also applies to zoning. Fischel (1992) argues that Hamilton's
(1975) approach predicts a positive relationship between restrictive zoning con-
trols and property values. As shown in Fig. 3, however, Hamilton zoning (along
with his other assumptions) implies that the price of housing services equals
P; any link between zoning and property value implies a deviation from this
prediction.2 9
The Hamilton model also implies homogeneity within a community in both
S and H. As observed by Goldstein and Pauly (1981), however, homogeneity
in public service demands does not imply income homogeneity; a low income
household with a strong taste for public services might demand the same S as a
high income household with a weak taste for public services. As shown below,
these unobserved tastes for S complicate both tests of the Hamilton (1975) model
based on heterogeneity and the estimation of public service demand equations.
27 Hamilton's (1976) equation for HIH and the equivalent one for LIH were designed primarily to introduce
heterogeneity and capitalization into his framework and were one of the first attempts to introduce bidding
notions into the Tiebout literature. Since the consensus bidding model does not require zero income and price
elasticities for S, the Hamilton bidding model is not pursued here.
28 With zero net capitalization, willingness to pay for S is perfectly correlated with willingness to pay for a
reduction in t, so coefficients for S and t cannot be estimated. If these two variables are not perfectly correlated,
then one cannot claim zero net capitalization, even if at some value of S the willingness to pay for an increase
in S exactly cancels the willingness to pay for the associated increase in t.
29 Fischel (1992: p. 175) also argues that effective zoning to offset nuisances will not affect house values.
The failure of some studies "to detect spillover effects can be taken as evidence that zoning works well in
their jurisdictions, contrary to the more usual inference that zoning is not justified". We believe the same point
applies to fiscal zoning; zoning set at the optimal H should not affect housing prices.
Ch. 47: Sorting and Voting 2019
trast, a finding of heterogeneity in demands for S and H leads to a rejection of
Hamilton (1975), but is consistent with both the consensus model and Henderson.
2.2. The impact of bidding and sorting on public service levels and property
tax rates
Unlike the literature on bidding and sorting, the literature on the determinants of
local public service quality shows no signs of a consensus, largely because of the
wide range of issues that must be considered. Many of these issues are complex
and several of them involve questions of voters' perceptions about which we
have little information. As a result, innumerable combinations of assumptions
are possible, and no two scholars use the same model.
30 A sixth question, namely, how public services are distributed within a jurisdiction, has been neglected in
this literature; virtually all studies assume an equal distribution.
31 See Wildasin (1986) fora literature review. Romer and Rosenthal (1979) present a critique of the median-
voter model, Holcombe (1989) evaluates evidence for it, and Aronsson and Wikstrom (1996) provide a recent
test of it using data from Sweden.
2020 S. Ross nd . inger
budget and the other in which they use the savings to reduce property taxes. Kon-
ishi (1997) employs the d-majority voting rule developed by Greenberg (1979),
which makes it possible to apply voting logic to decisions involving many local
public goods. 3 2 Finally, public service levels may be determined by coalition (or
jurisdiction) formation, a process in which all consumers in a coalition agree
in advance to a level of public service provision; in equilibrium no subset of
consumers has an incentive to form a new coalition.3 3 The literature reviewed
here does not consider highly centralized governmental arrangements in which
public service outcomes are imposed by a central government.
(2) Are voters owners or renters? For two reasons, most of the research on
local voting focuses on owners. First, almost two-thirds of households in the US
are owners. Second, when housing bids are included in the analysis (which is a
requirement to be considered in this review!) the median-voter framework breaks
down for renters. Renters' voting incentives depend on the difference between
their willingness to pay for public services (as expressed in housing bids), and
those of renters moving into the commnunity. If resident renters are identical to
entering renters, then any benefit from increased public services is exactly offset
by an increase in rents, and resident renters do not care about the public service
level (Stiglitz, 1983; Yinger, 1985). If resident and entering renters are not iden-
tical, renters care about the level of public services, but one cannot specify their
incentives without a simultaneous examination of voting and sorting, as in Epple
and Romer (1991). 34
(3) Do voters understand that their decisions may affect housing prices? This
question is the first of two key questions about the extent to which voters are
myopic. This question arises primarily because of capitalization; changes in S
and t alter housing prices. Moreover, voters' perceptions about housing prices are
important for at least four reasons. First, an increase in housing prices represents
an increase in the opportunity cost of a voter's housing. Because a homeown-
er's mortgage payments are based on the price of his house at the time it was
purchased, this increase may not show up in any observed market transaction,
32 Greenberg (1979) shows that a necessary and sufficient condition for an equilibrium to exist is that the
maximal set of winning coalitions in a community consists of all coalitions that contain more than d/(d + 1)
of its residents, where d is the number of dimensions over which public services are evaluated; d = I for the
median-voter model.
33 This choice mechanism is consistent with Hamilton's (1975) assumption of costless community forma-
tion. Wooders (1989) develops a model based on this choice mechanism, but the model does not include a land
market.
34 Epple et al. (1983, 1984, 1993) interpret their model as a renter model because households do not perceive
capital gains or losses in their voting decisions, but it can also be interpreted as an owner model in which
households do not perceive the effect of voting outcomes on the value of their property.
Ch. 47: Sorting and Voting 2021
but it still represents a change in the voter's implicit market rent. Second, with
flexible assessments, an increase in housing price leads to an increased property
tax payment. Third, a change in housing price leads to a capital gain or loss for a
homeowner, which represents a change in the voter's permanent income. Fourth,
an increase in housing price alters the market value of houses in the community,
and hence affects the community's budget constraint. 3 5 Voting outcomes may
depend on whether or not voters recognize such effects.
(4) Do voters understand that their decisions may affect housing consump-
tion, both their own and that of other households in the community? Changes in
S and t also can influence housing consumption, primarily through changes in
housing prices. The question is whether or not voters are aware that this might
occur. Do voters know, for example, that they might want to adjust their own H
after a change in S or t? In addition, changes in housing consumption affect the
community budget constraint. If housing prices go up, for example, and other
households cut back their H, then a voter's tax share could increase. Voters'
decisions may be influenced by their perceptions of these possibilities.
(5) What is the technology of public production? Different scholars make
different assumptions about the technology of public production. Three charac-
teristics of this technology are particularly important: returns to service quality,
the degree of publicness, and the impact of community characteristics. 3 6 As in
the simple community budget constraints presented earlier, assumptions about
technology typically appear in a cost function. A cost function that reflects all
three of these characteristics can be written E = E(S, N, A), where N is the
jurisdiction's population or number of pupils, and A is a vector of community
attributes that influence production costs (see Duncombe and Yinger, 1993).
The first characteristic is summarized by the elasticity of cost per unit of S
with respect to S. A zero (negative) [positive] value for this elasticity implies
constant (increasing) [decreasing] returns to quality scale.3 7 The vast majority
of studies assume constant returns, although, as discussed in Section 3.2.2, the
limited available evidence does not support this assumption.
The second characteristic is summarized by the elasticity of cost per capita
with respect to N. A zero (negative) [positive] value for this elasticity implies
35 Changes in the S or t could also alter the price of commercial and industrial property (Carroll and Yinger,
1994), but this effect is not considered here.
36 A fourth aspect is returns to scope due to sharing of inputs across governmental activities. See, for ex-
ample, Gyapong and Gyimah-Brempong (1988), Grosskopf and Yaisawaring (1990) or Duncombe and Yinger
(1993).
37 Duncombe and Yinger (1993) show that returns to quality scale depend on both technical returns to scale,
which are measured by the elasticity of total costs with respect to G, and on the elasticity of G with respect to
S, which is influenced by community characteristics.
2022 S. Ross and J. Yinger
38 Duncombe and Yinger (1993) show that returns to population scale depend on technical returns to scale
(defined in the previous footnote) and on the elasticity of G with respect to N. The role of technical returns to
scale is often missed.
39 In production theory, cost functions are derived from production functions assuming cost minimization.
Although local governments may not minimize costs, this assumption is also used to estimate public cost
functions. A general treatment of cost functions under other assumptions is unavailable.
40 Bradford et al. (1969) assume that community characteristics do not affect factor substitution in the first-
stage production function. A more general form that allows for such an effect is TC = c'[s-1 (S, N, A), W,
A'], where A' is the set of community characteristics that affect factor substitution. See Duncombe and Yinger
(1993).
41 An exception is the bidding model of Epple and Romano (1997), that brings in A by assuming that
education costs depend on student ability. They examine the sorting of public school students by ability
and income, derive conditions under which equilibria stratified by ability are expected to exist, and find that
multiple equilibria exist under these conditions.
Ch. 47: Sorting and Voting 2023
however, some empirical work uses this framework to estimate a comprehensive
cost index that indicates the impact of N and A on TC, holding S constant.
Epple et al.'s myopic voters. One approach to voting, illustrated by the work
of Epple et al. (1983, 1984, 1993), is to assume that voters are myopic in the
sense that they do not know that their decisions about S and t will influence
housing prices or the average housing consumption in their jurisdiction. A voter's
perceptions about her own H are not crucial; Epple et al. (1983) show that the
equilibrium is the same whether, or not, the voter treats her own H as fixed. In
formal terms, the median voter maximizes her utility subject to her own and the
community's budget constraint under the assumption that P is fixed. Substituting
the community budget constraint, E(S) = t*PH, into the household budget
constraint, this problem becomes:
Maximize U(Z, H, S)
Z, H, S (2.15)
Subject to Y= Z+ + () H.
MB = MC- (2.16a)
H
and
UH E(S)
= P+ =P(1 + t*). (2.16b)
UZ H
In this model, therefore, the median voter's decision about S is influenced by her
tax share and her decision about H is influenced by the property tax.
Although this model is simple, it does not necessarily have an equilibrium.
Epple et al. (1984) analyze "viable" within-community or internal equilibria,
defined as ones in which the value of aggregate housing consumption at the
equilibrium t results in exactly the revenue necessary to finance the equilibrium
E. Moreover, an equilibrium requires that an interactive process of voting and
2024 S. Ross and J. Yinger
42 Epple et al. (1984) is based on a general cost function for S, but the proof in Epple et al. (1993) depends
on more restrictive assumptions about technology.
43 Epple and Romer (1991) prove that, even without strong assumptions about preferences, an internal
equilibrium exists in a model of income redistribution when voters anticipate the effect redistribution policy
has on both housing stock and community population. They also demonstrate that an equilibrium is unlikely
to exist in a redistributive public economy when voters are myopic.
Ch. 47: Sorting and Voting 2025
capitalization at all: both possibilities are consistent with coefficients of zero for
all tax and service variables. Equation (2.7) and its analog with zoning, namely
Eq. (2.14), also reveal that one cannot isolate net capitalization without account-
ing for other fiscal factors that vary across jurisdictions and influence bids, in-
cluding the taxbase and the cost of S. At best, therefore, a house-value regression
can provide only a rough guide to the decision rule used by local voters.
condition for the first case holds. The first of these cases covers both the myopic
voters of Epple et al. (1984) and Yinger's (1982, 1985) voters, who are aware of
capitalization. The second case covers the voters in Crane (1988, 1990), who are
aware only of the possibility that changes in S and t affect the opportunity cost
of their housing.
Nevertheless, many plausible cases lead to different decision rules. Suppose,
for example, that voters are aware of property tax capitalization, but are not
aware that S affects property values, and that they are aware of the impact of
capitalization on either the value of their house or the opportunity cost of their
housing, but not both. If they are aware of capital gains, voters perceive an added
cost to increasing S, namely, the capital loss associated with the required higher t.
By Eq. (2.4), the annualized capital loss, d(rV)/dt, is r/(r + t)% of V; to reflect
this cost, the right side of Eq. (2.16a) must be multiplied by [1 + r/(r + t)].
In contrast, the higher t associated with higher S lowers the opportunity cost
of housing, also r V, so an awareness of the opportunity cost, but not of capital
gains, lowers the perceived cost of S, and the right side of Eq. (2.16a) must be
multiplied by [1 - r/(r + t)].
In short, many different assumptions about voter perceptions and capitaliza-
tion lead to the same well-known decision rule, Eq. (2.16a). However, these
perceptions are poorly understood, and reasonable alternative assumptions can
lead to public service quality that is either above or below the level implied by
Eq. (2.16a).
44 An alternative literature, reviewed in Conley and Wooders (1998), examines the existence and nature
of the core in economies with local public goods and no land market. Conley and Wooders (1997) show. for
example, that in large economies with endogenous community formation, price taking equilibria exist and are
nearly optimal.
Ch. 47: Sorting and Voting 2027
45 Pogodzinski and Sjoquist (1985, 1991, 1993) and Henderson (1991) provide informal proofs of equi-
librium with specific functional forms and specific parameter values by showing that their simulation models
have solutions.
46 In principle, power equalizing grants also could cause trouble for equilibrium. These grants involve a
higher matching rate for jurisdictions with lower wealth (and most likely lower income). See Ladd and Yinger
(1994). One study (Feldstein, 1976b) argues that power-equalizing grants would lead to lower spending on
education in higher wealth school districts. Nechyba (1996) introduces power-equalizing grants into a general
equilibrium model with multiple communities and also finds a negative correlation between local spending and
income. However, Feldstein's (1976b) estimated price elasticities are far higher than that obtained in the rest
of the literature (discussed in Sections 3.2.1 and 4.2.1). Moreover, no state has implemented a pure form of a
power-equalizing grant. See Reschovsky (1994). These grants are unlikely to cause disequilibrium!
47 This problem cannot arise with decreasing returns to N and equal land area. In this case, the high income
community will have a higher MC only if it has a larger N. A larger N requires smaller per-household H,
which in turn requires a higher P. A higher P is possible, but only through capitalization, that is, only if S is
higher.
2028 S. Ross and J. Yinger
faces a higher MC could be eliminated simply by having the two income classes
trade communities.4 8
(2) The marginal cost of delivering public services increases with community
income. Because MC is part of the tax price, this case, like the previous one,
could lead to lower demand for S in the high income than in the low income
community. Thus, to take advantage of the low tax price in the low income
community, high income households might outbid low income households there.
As high income households move in, however, MC will rise, thereby preventing
equilibrium. As we will see, the empirical evidence indicates that this case is
unlikely.
(3) The ratio of median to mean housing consumption (the median voter's
tax share) increases with community income. In this case, the median voter in
the low income community faces a more favorable tax price for S than does the
median voter in the high income community, and might, therefore, demand more
S. This case obviously depends on the underlying distribution of income and on
sorting.4 9
Epple et al. (1983, 1984, 1993) prove that these conflicts between sorting
and voting cannot arise when the cost of producing S is linear in the product
of N and S and the income elasticities of demand for H and S are positive.
This assumption about public service technology restricts the extent of returns
to population or quality scale and rules out variation in MC with respect to
community characteristics. 50 Moreover, the tax-share problem is eliminated by
assuming linear costs and positive elasticities. 5 1
We suspect that these assumptions are stronger than necessary to rule out
these conflicts. 52 These assumptions are sufficient because they ensure that high
income voters never face higher tax prices than lower income voters, but they
48 This same switching logic could be used to eliminate situations in which jurisdictions vary by the
technology for producing H or S.
49 As shown by Hanson and Kessler (1997). a similar issue arises when government spending is funded with
an income tax. They examine a model with purely redistributive governments, following Epple and Romer
(1991), but myopic voters, as in Epple et al. (1983, 1984. 1993). They prove that a stratification equilibrium
will not exist unless the ratio of median-to-mean income in high income jurisdictions is less than this ratio in
low income jurisdictions; otherwise, high income households have an incentive to move.
50 Diminishing returns to S cannot lead to this conflict. In this case, jurisdictions with higher S have higher
MC, but all jurisdictions still face the same price profile over values of S. Formally, in a homogeneous commu-
nity, the demand function can be written S = S[MC(S), Y]. Totally differentiating this equation with respect
to S and Y reveals that diminishing returns serve only to scale down the income effect and cannot fully offset
it.
51 Epple et al. (1984) also assume a uniform income distribution, which rules out case three, but the income
distribution assumption is not necessary for existence given their other assumptions.
52 These assumptions may be necessary, however, to assure that empty communities do not arise (Rose-
Ackerman, 1979).
Ch. 47: Sorting and Voting 2029
do not appear to be necessary because higher tax prices for high income voters
do not lead to sorting-voting conflicts unless /, is fairly high. Moreover, these
assumptions restrict some aspects of public service technology, such as decreas-
ing returns to population scale or nonconstant returns to quality scale, that do
not lead to sorting-voting conflicts. Further research on necessary conditions for
equilibrium is warranted.
their tax subsidy will be capitalized into their prices. This creates an incentive
for homebuilders and landowners to change the composition of a community's
housing stock. As a result, Nechyba's stratification may not persist in the long
run unless it is associated with differences in site quality or exogenous location
characteristics.
3. Empirical research
This section reviews empirical work in urban public finance. The focus is on
studies that test hypotheses related to the consensus view of bidding and sorting,
the Hamilton (1975) approach, and aspects of voting that are related to housing
markets. 58
V= - /5 r+,B(3.1)
r r-B(5) r + t
Despite its clear conceptual foundations, however, few studies make use of this
equation's functional form. Many studies follow Oates (1969), for example, and
approximate Eq. (3.1) by including t or the logarithm of t as an explanatory vari-
able. When data on tax and house value changes over time are available, simple
but exact functional forms are possible. Let subscripts indicate time periods, then
Eq. (3.1) leads to
AV -5,Bt -BAT
AV - -pAt or AV =- (3.2)
V1 r+ t2 r
If t2 is roughly constant (as it might be after a revaluation), then both of these
equations reduce to a constant multiplied by the change in a tax variable.6 1 The
second equation can be estimated, for example, when tax payments undergo
a large change (as with Proposition 13, a property tax limit in California; see
Rosen, 1982). The estimating equation also should reflect the role of income tax
deductions for property tax payments, which cannot be claimed by homeowners
who do not itemize; Eisenberg (1996) and deBartolome and Rosenthal (in press)
find that capitalization is significantly higher for itemizers than for nonitemizers.
The second methodological difficulty is that the property tax variable is en-
dogenous. Assessed values, and hence tax payments, are based to some degree
on market values, and effective tax rates are defined with market value as the
denominator. As first pointed out by Oates (1969), capitalization equations there-
fore must be estimated with a simultaneous equations procedure. Moreover, the
choice of instruments is important (see Pollakowski, 1973),62 and the best studies
formally model assessor behavior to identify appropriate instruments.
Because house values are affected by many factors other than t, careful con-
trols are important to insulate capitalization estimates against omitted variable
60 As stated in footnote 9, the form derived by Wheaton (1993) is somewhat different, but it has never been
tested.
61 Intrajurisdictional tax capitalization applies to deviations from the average tax rate. Redefining t as a
deviation from the community average, a constant effective tax rate after revaluation implies that r2 = 0, so that
this form reduces to -Ai/r. See Yinger et al. (1988).
62 Pollakowski (1973) is a comment on Oates (1969); Oates (1973) is the reply.
2032 S. Ross and J. Yinger
bias. Some studies avoid this bias with an extensive list of housing and neighbor-
hood characteristics and public service measures, but other studies do not have
adequate controls. Studies based on tax changes must control for changes in both
housing and neighborhood characteristics.
Finally, the form of Eqs. (3.1) and (3.2) precludes separate estimation of r and
fl. Most studies follow Oates (1969) in assuming a value for r, but in this case the
value of : depends on an untested assumption and different studies use different
values of r. Yinger et al. (1988) point out that the appropriate r is a real discount
rate, defined as a nominal rate minus anticipated inflation; by subtracting a pre-
diction of anticipated inflation from the nominal interest rate, they estimate r to
be about 3% in their sample period. Do and Sirmans (1994) claim to estimate
r, but in fact, they simply reverse the usual procedure by assuming / = 1 and
calculating the implied r.
Virtually all the literature estimates the capitalization of current property tax
differences. Under the assumption that current differences will persist indefi-
nitely, the theory behind Eq. (2.4) implies that these differences should be fully
capitalized, that is, that B = 1. In fact, however, current differences may not be
expected to persist. In this case, Yinger et al. (1988) show that the capitalization
of current tax differences, the /i in Eq. (3.1), is related to the capitalization of the
present value of the expected tax stream, say /3', as follows:
where N' is the length of time current tax differences are expected to persist.
The theory indicates that /' = 1, but the estimated /3 clearly need not equal 1,
and indeed need not equal the same value under all circumstances. In the case
of intrajurisdictional tax differences, for example, if an upcoming revaluation is
expected to eliminate current tax differences in 10 years and r = 0.03, then Eq.
(3.3) implies that the estimated /3 will be only 26% even if /3' = 1.
Virtually every study of property tax capitalization finds a statistically sig-
nificant negative impact of property taxes (or property tax changes) on house
values. 63 The vast majority these studies use data from the US, but a few, includ-
ing Chinloy (1978) and Hamilton (1979), use data from Canada. Estimates of
/3 vary widely, but if r is set at 3%, the estimates of /Bfor the best studies fall
between 15 and 60% (Yinger et al., 1988; deBartolome and Rosenberg, 1997).
No study yet provides a definitive estimate of /3', but Yinger et al. (1988) and
Eisenberg (1996) present some evidence that it is close to 100%. These results are
63 All the studies that find no effect have serious methodological flaws. See Yinger et al. (1988).
Ch. 47: Sorting and Voting 2033
consistent with the consensus bidding model, but contradict both the Hamilton
and Henderson approaches.
64 The general equilibrium nature of the sorting problem causes difficulties for interpreting studies of service
capitalization based on before-and-after data. See Pogodzinski and Sjoquist (1994).
65 Alternative approaches, namely, bidding or random utility models (McFadden, 1974; Ellickson, 1981;
Lerman and Kern, 1983), are used by Quigley (1985), Nechyba and Strauss (1998) and Chattopadhyay (1998)
to estimate the impact of public spending on house values.
2034 S. Ross and J. Yinger
These results all support the key prediction of Eq. (2.2) and contradict the
Hamilton and Henderson approaches. Far more work needs to be done, however,
to obtain comprehensive measures of S and to estimate the implied marginal ben-
efit function in Eq. (2.2). Although the large literature on hedonic prices identifies
several methods for estimating demand or marginal benefit functions (Follain
and Jimenez, 1985; Epple, 1987; Chattopadhyay, 1998), only one study, Boije
(1997), applies such methods to local public services; using data from Sweden
with per capita local spending as a measure of S, Boije estimates income and
price elasticities of demand for S equal to 0.09 and -0.89, respectively.
Several studies have looked at net capitalization, which, as noted earlier, can
provide a rough guide to local spending decisions. Oates (1973) calculates that,
at average values for the explanatory variables, the negative impact on median V
of a property tax increase is roughly offset by the positive impact of the increased
educational spending the resulting revenue could fund. He does not find zero net
capitalization at all values of spending and t, however, so the implications for
local voting are unclear. Brueckner (1979, 1982) solves the community budget
constraint for t, substitutes the result into Eq. (3.1), and argues that to maxi-
mize its property values a community will increase S up to the point where its
marginal net impact on house values is zero, which is the same thing as zero
net capitalization. Using a sample of communities in Massachusetts, Brueckner
(1982) estimates this version of Eq. (3.1), finds that S has a coefficient of zero,
and concludes that communities maximize property values. We do not find this
conclusion persuasive. Brueckner's procedure cannot distinguish between zero
net capitalization everywhere (which is consistent with his argument), and zero
net capitalization on average or no capitalization (which are not). Moreover,
his data have severe limitations (Rubinfeld, 1987); for example, public service
quality is measured by spending per capita, and the regression includes only one
housing characteristic.
A promising new look at net capitalization is provided by Bradbury et al.
(1997). They distinguish between towns that are constrained by Proposition 2,
a property tax limit in Massachusetts, and those that are not, and find that house
values increase with spending only in constrained towns. This result supports the
view that these communities used a decision rule similar to Eq. (2.16a) and that
Proposition 2½ pulled them below their preferred spending level. Because of this
forced drop in spending, the marginal benefit from increased spending exceeded
the cost for a wide range of households, including the marginal household bid-
ding for housing in these communities. Thus, according to Eq. (2.7), increases in
spending resulted in higher house values.
Ch. 47: Sorting and Voting 2035
3.1.3. Heterogeneity
Several scholars have argued that theories about local public finance can be tested
by examining the extent to which communities are homogeneous. As shown
above, the Hamilton model does not require homogeneity in income, but it does
predict that individual jurisdictions should be substantially more homogenous
in income than is a metropolitan area as a whole. Unfortunately, however, this
prediction is not helpful for testing Hamilton's model, because the consensus
bid framework yields the same prediction. In models with a fixed number of
jurisdictions with fixed land areas, Epple et al. (1983, 1984, 1993) and Epple and
Platt (1998) show that jurisdictions will not exhibit perfect income homogeneity,
but will be more homogeneous than the urban area. In the Hamilton model with
the Goldstein and Pauly (1981) amendment, income heterogeneity arises because
of taste differences; in the Epple et al. (1983, 1984, 1993) models it arises because
the population must be spread out among a fixed number of communities; and in
Epple and Platt (1998) it arises for both of these reasons. But these differences in
mechanisms do not lead to differences in predictions about heterogeneity.
As a result, most existing studies of community heterogeneity fail to pro-
vide tests of the Hamilton model. Pack and Pack (1977) find that individual
jurisdictions are not homogenous by income or occupational status, but are con-
siderably more homogeneous than the urban area in which they are located.
Hamilton et al. (1975), Eberts and Gronberg (1981) and Munley (1982) all find
that within-jurisdiction homogeneity increases with the number of jurisdictions
in a metropolitan area. Heikkila (1996) examines heterogeneity over a broad
range of socioeconomic characteristics and finds substantial homogeneity. 6 6 Fi-
nally, Gramlich and Rubinfeld (1982) analyze a survey of household attitudes
toward the level of public services in their jurisdictions and conclude that house-
holds are grouped based on their unobservable tastes for public services. Al-
though several of these authors interpret their results as evidence that Tiebout or
Hamilton sorting is taking place, all of these results are consistent with both the
Tiebout/Hamilton approach and the consensus bidding framework.
One possible way to distinguish between the Hamilton and bidding approaches
is to examine homogeneity in H and S. While Hamilton's model might be a
reasonable representation of the world with something less than perfect homo-
geneity in H and S, it does require that the within-jurisdiction heterogeneity in H
66 Heikkila uses factor analysis to generate indexes of heterogeneity across characteristics and analysis of
variance to test whether or not census tracts of similar types tend to be found in the same jurisdiction. This
approach avoids the problem of statistically generated within-jurisdiction homogeneity that has been discussed
by Dowding et al. (1994).
2036 S. Ross eind J. Yinger
3.1.4. Zoning
According to the Hamilton (1975) model, each jurisdiction's zoning restrictions
are entirely determined by the housing demand of residents and, therefore, should
not be affected by the zoning choices of nearby jurisdictions. Using data from
Connecticut, Lenon et al. (1996) find that each town's zoning, taxing, and spend-
ing policies are significantly and positively related to the corresponding policies
of nearby towns. This result directly contradicts the Hamilton prediction.
Moreover, neither perfect zoning, in the Hamilton sense, nor zoning that sim-
ply ratifies a sorting outcome should have an impact on house values. Without
the perfect sorting of the Hamilton model, however, communities with stricter
zoning might be better protected from the entry of (and resulting subsidies to)
low income households and therefore have higher property values. If this zon-
ing is not very exact, households that do move in may be forced to consume
more housing than they prefer, so zoning also might have a negative impact on a
community's average house value.
The empirical literature often finds that zoning is not neutral, thereby contra-
dicting the Hamilton prediction, but does not clearly indicate the role that zoning
plays. Frech and Lafferty (1984), Katz and Rosen (1987) and Schwartz et al.
(1986) provide evidence that increased growth controls raise property values,
whereas, McMillen and McDonald (1991) find that zoning for single-family
houses (but not apartments) tends to be neutral, that is, to follow the market.
67 Hoyt and Rosenthal (1997) argue that a test of Hamilton's (1975) model can be derived from the premise
that all residents of a jurisdiction should have the same marginal benefit from additional public services. Re-
gardless of the merits of this argument, their empirical implementation does not provide a test of the Hamilton
model because it examines heterogeneity within neighborhoods, which are likely to be far more homogeneous
than jurisdictions. For a formal derivation of this point, see Epple and Romano (1997).
68 Gramlich and Rubinfeld (1982) observe the extent of heterogeneity in demand for S, but do not compare
their results with observed heterogeneity in income.
69 Hamilton et al. (1975) found that a greater number of jurisdictions did not lead to more within-community
homogeneity in house value. This result also contradicts the Hamilton model, which predicts that expanded
opportunities for sorting will lead to more homogeneity in both S and H.
Ch. 47: Sorting and Voting 2037
The most recent study, Pogodzinski and Sass (1994), finds that property values
decline with stricter land use controls.
The enormous literature on local voting includes many studies that determine
whether or not housing market outcomes influence public service choices. Sev-
eral key conceptual issues have not been explored empirically, however, includ-
ing voter perceptions about capitalization and about the property taxbase.
land, including direct democracies, where the median-voter model should work
particularly well, and representative democracies. The tax share term is based
on the local income tax. For direct democracies, a is about 1.3 and /u is about
-0.75. Both elasticities are somewhat smaller in absolute value for represen-
tative democracies. Borge and Ratts0 (1995) examine spending in Norwegian
municipalities. They estimate values of a near 1.0 and of guthat are mostly in
the -0.5 to -0.9 range. Aronsson and Wikstr6m (1996) estimate that a = 0.82
and ut = -0.47 for municipalities in Sweden. Sone (1993) estimates a model of
expenditure determination for cities in Japan. He argues that municipal officials
in Japan attempt to please the mean resident (not the median voter). Moreover,
Sone develops a tax-share term designed to reflect the fact that the property tax
is one of several taxes levied by Japanese cities and does not appear to be the
marginal tax; this term is the mean household's tax payment divided by the city's
total taxbase (with an adjustment for tax exporting). Using data for individual
cities, Sone estimates this model for several types of spending and three different
years. For overall spending, a falls between 0.5 and 0.7, and i between -0.6
and -1.1. Values of a and A for individual services exhibit greater variation.
71 This formulation follows Bradford et al. (1969) in assuming that A only influences the second stage
of public production-the link between G and S. As pointed out earlier, however, A also may influence the
production of G. Duncombe and Yinger (1993) find, for example. that population and building height affect
the tradeoff between firetrucks and firefighters in providing fire-protection services.
Ch. 47: Sorting and Voting 2039
share, and other demand variables on the right side, which is one method used to
estimate a and ut with community-level data.7 2
Despite their analytical connection, production and cost studies have some
important differences. Production studies focus on a single output, whereas cost
studies cover all the outputs for a broad type of public spending, such as educa-
tion, police or fire. Moreover, production and cost studies may apply to different
units of observation. Educational production studies typically focus on the school
or student level, for example, whereas educational cost studies typically focus on
the school district.
Several studies have explored the role of N in production or cost. 73 In the
case of education, many cost studies find a U-shaped relationship between cost
per pupil and the number of pupils in a district. Early studies are reviewed in Fox
(1981); more recent studies include Downes and Pogue (1994) and Duncombe
and Yinger (1997, 1998). A production study of fire services by Brueckner (1981)
finds some degree of publicness; to use the term defined earlier, he finds that
y < 1. A cost study of fire services by Duncombe and Yinger (1993) finds that
y = 1. In the case of police services, both Schwab and Zampelli (1987) and Ladd
and Yinger (1991) find that y > 1. Pommerehene (1978) finds clear evidence of
publicness in aggregate public spending in Swiss municipalities, with a value of
y around 0.6. The study of Japanese cities by Sone (1993) finds that y is close
to one, both for overall spending and for most subcategories.7 4
The importance of community characteristics in educational production and
cost functions has been known since Coleman (1966) found that a student's
achievement (as measured by test scores) depends heavily on the social and
economic characteristics of the other students in her class. This result has been
duplicated by many studies, many of which are reviewed in Hanushek (1986).
A recent production study by Ferguson (1991), for example, finds that, all else
being equal, a student's achievement test scores are higher if her classroom has
a smaller share of students living in poverty, with limited English proficiency, or
from female, single-parent families. These variables also are significant in recent
72 This approach requires assumptions about the forms of the cost and demand functions. Most studies
assume a form that implies constant returns to quality scale. Duncombe and Yinger (1993) use a general form
to estimate the cost function directly controlling for S and find strongly increasing returns to quality scale.
73 As first pointed out by Hayes (1986), community characteristics may influence y. Duncombe and Yinger
(1993: p. 55) argue, for example, that "the cost of assuring a certain quality of fire service for a new household
is likely to be higher in communities with poor building condition than in communities with good building
condition". They find evidence to support this view.
74 All of these results except those of Brueckner (1981) and Pommerehene (1978) contradict a key
assumption of the Henderson (1991) model presented earlier.
2040 S. Ross and J. Yinger
educational cost studies by Downes and Pogue (1994) and Duncombe and Yinger
(1997, 1998).
Community characteristics also affect the production and cost of other public
services.75 Duncombe (1991), for example, finds that building age, poverty, and
the presence of commercial and industrial capital, all raise the cost of fire protec-
tion; Ladd and Yinger (1991) find that the costs of both police and fire services
rise with a city's poverty rate; and Bradbury et al. (1984) find that the overall cost
of local services rises with population density and housing age. Sone (1993) finds
that several community characteristics influence public service costs in Japan; for
example, the cost of fire services increases with population density.
75 Hamilton (1983a) argues that community income is correlated with other community characteristics and
suggests that this correlation explains the observed difference between estimates of the income elasticity of
demand and of the spending elasticity of grants, the so-called flypaper effect. Wyckoff (1991) tests and rejects
this argument; in fact, he finds that the inclusion of additional community characteristics increases the dif-
ference between these two estimated elasticities. Schwab and Zampelli (1987) pursue the Hamilton approach
using functional form restrictions to try to separate the demand and cost roles of income. Other studies in the
literature have focused on other community characteristics, including the poverty rate, as environmental cost
factors and interpreted income as a demand variable.
76 Using data from Sweden, Aronsson and Wikstrbm (1996) test and reject the hypothesis that the resident
with median income is the median voter.
Ch. 47: Sorting and Voting 2041
These issues can be illustrated by considering community and individual de-
mand equations for S. Community demand is
S = bY + ij, (3.5)
where the subscript i indicates a voter; the asterisks indicate that a variable ap-
plies to an individual voter, not the community; and 17 is a random error term.
The Goldstein/Pauly argument is that Y is correlated with v, so that Eq. (3.4)
will lead to a biased estimate of b, but that with perfect Tiebout/Hamilton sorting,
Sj = S, so that Eq. (3.5), which does not lead to bias, can be estimated with
observed levels of S.
If sorting is not perfect, however, the Goldstein/Pauly procedure involves
estimating
The problem of imperfect sorting also affects Rubinfeld et al. (1987). They
examine the difference between community and individual demand for public
service, or the difference between Eqs. (3.4) and (3.5):
They argue that a household will reside in a community that does not deliver their
preferred level of S, only if no other community offers a closer level. Individuals
with extreme values of will not live in communities with others who have
similar incomes, and for the individuals within a community, the distribution of
r will be truncated at both ends. This truncation creates a correlation between
and the explanatory variables. Rubinfeld et al. (1987) show how to correct for this
correlation but also assume that v and the explanatory variables are uncorrelated,
which, as shown earlier, requires no sorting by tastes.
In short, the use of a household sample eliminates the Goldstein/Pauly bias
only if Tiebout/Hamilton sorting is perfect, whereas standard median-voter es-
timation is unbiased only if there are no unobserved taste factors. The work
following Goldstein and Pauly (1981) identifies additional biases that arise with
a household sample when sorting is not perfect. However, none of these studies
eliminates the original Goldstein/Pauly bias, which remains in place with imper-
fect sorting. 78 The only way to eliminate the original bias is to combine a method
that identifies the true median voter within a jurisdiction (Reid, 1991; Aronsson
and Wikstr6m, 1996; Epple and Sieg, 1997) with an analysis of spending across
jurisdictions.
4. Normative analysis
The lasting fame of Tiebout (1956) derives primarily from his claim that "voting
with one's feet" is analogous to shopping for a commodity and leads to efficient
provision of local public services. The literature has struggled with questions of
efficiency ever since.
81 If wages for a household type do not vary across jurisdictions, uniform wage and head taxes across all
residents of all communities also results in efficiency.
82 In fact, Flatters et al. (1974) show that even with pure public goods and no differential costs across types
of households, voters will not select uniform head taxes across different communities.
83 For a description of the British poll tax, see Besley et al. (1997).
84 Under some conditions, intergovernmental grants might do it. See, for example. Schwab and Oates
(1991).
Ch. 47: Sorting and Voting 2045
The most complete formal treatment of this issue is provided by Schwab and
Oates (1991).85 In a model without land or housing, they derive explicit efficiency
conditions and show that one must account not only for the production effects but
also for the deadweight loss associated with heterogeneity (which is discussed
later). They also confirm the Wildasin (1986) result that the only way to achieve
an efficient outcome is to levy different head taxes on different types of people-
with a higher tax on type-i than on type-2 people. No article yet explores these
important issues with a full treatment of housing markets.
The property tax andpublic service capitalization. The first two sources of ef-
ficiency can be illuminated with a simple planner's problem for a homogeneous
community (under the assumption that an efficient allocation of people to com-
munities has been achieved).8 7 The planner maximizes the utility of a representa-
tive resident subject to resource constraints and production technology. Suppose
Z is produced with capital, K, and that H and S are produced with K and land,
L.8 8 Then the planner's problem is to allocate K and L so as to
This problem leads to the following efficiency conditions, where subscripts indi-
cate partial derivatives:
UH ZK US ZK Us HK_ HL (4.2)
and S
(4.2)
UZ HK' UZ SK UH SK L
85 This problem also appears in club models. See Brueckner and Lee (1989).
86 Some other sources of inefficiency have been identified, but they are not central to the models developed
in this paper. See Wildasin (1986) for a review.
87 Adding heterogeneity would not alter the substance of this efficiency analysis.
88 A more general formulation is provided by Yinger (1999).
2046 S. Ross and J. Yinger
ZK ZK HK HL MC
= P, - = MC and - - - (4.3)
HK SK SK SL P
Moreover, regardless of voters' perceptions about capitalization, they respond
to P(1 +t*), not P (see Eq. (2.16b)), so the property tax inserts a wedge between
consumer and producer decisions, and the first condition in Eq. (4.2) is not met.8 9
This is classic distortion from the property tax, which has been emphasized by
Oates (1972) and Mieszkowsi and Zodrow (1989), among others. This distortion
disappears only under the extreme Hamilton assumptions.
If voters have either complete awareness or complete ignorance of public
service capitalization, then according to Eq. (2.16a) when H = H, they treat
MC as if it were the price of S, and the second condition in Eq. (4.2) is satisfied.
Several authors have claimed that decisions about S are efficient because they
meet this condition.9 0 By the theory of the second best, however, this condition
is not a valid guide to efficient provision of S unless all other first-best efficiency
conditions in the overall planning problem are satisfied. As shown earlier, mar-
kets are unlikely to satisfy the efficiency conditions for household allocation to
communities and for housing consumption. The second condition in Eq. (4.2)
illuminates the requirements for efficiency, but cannot by itself be used to show
that efficient service provision is attained.
Furthermore, we do not know whether, or not, voter perceptions lead to the
decision rule in Eq. (2.16a). As discussed earlier, this rule is consistent with many
different assumptions about these perceptions, but some reasonable assumptions
lead to different decision rules-and hence to outcomes that are not consistent
with efficiency (Yinger, 1999). Awareness of capital gains (but not the opportu-
nity cost of housing) and of property tax capitalization (but not of public service
capitalization), for example, results in spending below the level defined by Eq.
(2.16a). Moreover, switching to a head tax does not guarantee efficiency, even
89 Recall, however, that with complete capitalization, P(1 + t*) is constant, so with a fixed housing stock,
the property tax causes no distortion in household choices between H and S. The problem arises because
households and firms respond to different prices.
90 Perhaps the best known example is Brueckner (1979, 1982, 1983), who argues that communities maxi-
mize property values and that this behavior leads to efficiency. For both empirical reasons (discussed earlier)
and the reason in the text, we disagree.
Ch. 47: Sorting and Voting 2047
Moreover, if the ratio of median to mean is greater for MB than for H (or in
Fig. 4 if, at the efficient level of S, the vertical distance between the MB curves
is greater than the vertical distance between the MC curves), the voter selects a
2048 S. Ross and J. Yinger
SACTUAL SOPTIMAL
level of S that is below the efficient level. In effect, the efficiency rule gives more
weight than does the median voter to the strong demand for S from the highest
income households.
Now suppose ~o is the ratio of median to mean income. Then with constant
elasticity demand functions, the ratio of MBm to MB equals ys, where as before,
Eis the income elasticity of MB. Similarly, the ratio of V, to V equals ( , where
0 is the income elasticity of demand for housing. Thus, whenever (p < 1, which
is always is true with a long upper tail for the income distribution, the elasticity
condition that assures matched sorting, namely E > 0, also assures that the level
of S will be below the efficient level in a heterogeneous community. A more
general version of this result is provided by Bergstrom (1973), and matching
grants to achieve efficiency in a heterogenous community are derived by Akin
and Young Day (1976) and Yinger (1985).
4.1.4. Conclusion
In short, even with neutral political institutions, a general equilibrium analysis of
housing and public services reveals four fundamental sources of inefficiency:
misallocation of households to communities, the property tax, public service
capitalization and heterogeneity. The empirical literature supports most of the
key assumptions that lead to these types of inefficiency and leads to the strong
presumption that our system of local governments is inefficient. Moreover, with
Ch. 47: Sorting and Voting 2049
so many possible sources of inefficiency, the theory of the second best implies
that efficiency cannot be inferred from a finding that one of the efficiency con-
ditions is satisfied. The existing literature provides no information, however, on
the magnitude of the excess burdens associated with all this efficiency and hence
sheds no light on the potential gains from efficiency-enhancing policies. 9 1
91 In addition, little work has been done on second-best policies that take some constraints as given. One
exception is Yinger (1985) who builds on the work of Atkinson and Stern (1974) to derive efficient conditions
for service provision when a distortionary property tax is the only available revenue instrument.
92 The impact of commercial and industrial property is quite complex, both because taxes on this property
may be shifted to consumers (Ladd, 1975) and because the presence of this property may raise public service
costs (Ladd and Yinger, 1991).
93 For a contrary view, see Oakland (1994).
2050 S. Rss and J. Yinlger
Fairness issues have long been recognized in the case of education, both by
academics and policy-makers. Moreover, many American states, often in re-
sponse to a court ruling, now provide higher grants per pupil to school districts
with lower property values per pupil. See Downes (1992), Furman et al. (1993),
Reschovsky (1994) and Courant and Loeb (1997). Some grant systems are de-
signed to bring all school districts up to a minimum spending level, and others
are designed to ensure that all districts that levy the same property tax rate will
receive the same spending per pupil. No existing grant system eliminates the
correlation between wealth and school spending, but some of them undoubtedly
lower this correlation significantly.9 4
The role of educational costs is not widely understood, however, and most
courts and grant systems focus on spending per pupil, not public output per pupil.
At best, existing programs contain only limited, ad hoc adjustments for cost-
related factors. This neglect is unfortunate; cost indexes can be estimated and
included in aid formulas (LeGrand, 1975; Bradbury et al., 1984; Ladd and Yinger,
1994; Duncombe and Yinger, 1997, 1998). Because they also ignore variation
in costs and taxbases, property tax limitation measures are likely to undermine
equity by forcing high-cost, low-wealth school districts to make the largest tax
cuts (Ladd and Wilson, 1985).
94 Because grants alter recipient governments' behavior, a grant system that links spending per pupil to a
district's taxbase, called a power-equalizing grant does not necessarily eliminate the positive correlation be-
tween spending and wealth (i.e., achieve wealth neutrality), and might lead to a negative correlation (Feldstein,
1976b; Nechyba. 1996). Recent evidence on the behavioral responses to grants indicates that power-equalizing
grants do not bring this correlation down to zero (Duncombe and Yinger, 1998).
Ch. 47: Sorting and Voting 2051
Consider, for example, a community in which poor assessing practices lead to
assessment/sales ratios, and hence effective property tax rates, that vary widely
from one house to another.9 5 In this situation, an unanticipated revaluation that
brings all houses to the same effective tax rate will result in capital losses for
houses that were previously underassessed and to capital gains for houses that
were previously overassessed. These gains and losses are largely arbitrary, and
hence unfair; a long-time owner whose house was previously underassessed could
be said to be paying back tax breaks in the form of a capital loss, but the same loss
falls on a recent buyer who gained nothing at all from the past underassessment.
One cannot avoid this problem, however, by retaining the poor assessing system,
because such a system hands out regular, small, unannounced effective tax rate
cuts or increases as it allows assessments to diverge from market values. The
resulting incremental gains and losses also are arbitrary, and hence unfair. The
only way out of this dilemma is to pay the one-time fairness cost of revaluation
and then keep assessments up to date in the future (Yinger et al., 1988).
As shown by Wyckoff (1995), capitalization also can offset some or all of the
equalizing impact of an equalizing grant program. Current homeowners and land-
lords in a jurisdiction that receives a higher intergovernmental grant will receive
a capital gain, but anyone who moves into the jurisdiction in the future will have
to pay a higher price or rent-which will offset some or all of the benefit from
the increased public services or lower taxes there. This effect is not important
if one is only concerned with equalizing service levels, but it is central if one
wants to equalize household welfare over the long term. In the case of education,
equalizing service quality for children appears to be a more compelling objective
than equalizing welfare for parents. The issues are not so clear for police and fire.
In any case, more work needs to be done on the appropriate equity objectives and
on the extent to which capitalization undermines equalizing aid programs.
5. Conclusions
The literature reviewed here raises important issues in local public finance for the
US and many other countries, particularly those with decentralized governmental
systems.
Bidding and sorting models have very general applicability. They do not re-
quire local governments to be autonomous but do require mobile households,
a well functioning housing market, and some variation in local service levels
and effective tax rates within a metropolitan area. For example, they could prove
helpful in a country that sets spending levels for local governments without con-
sidering cost differences across jurisdictions. However, the extent of variation in
local service levels and tax rates, and hence, the power of the bidding/sorting
approach, is likely to be positively correlated with decentralization, which varies
widely even among developed nations. Pommerhene (1977) shows, for example,
that the US, Switzerland and Canada are very decentralized, whereas, at the other
extreme, France reserves for the central government many services provided lo-
cally elsewhere, such as primary and secondary education. In addition, Oates
(1972, 1993) finds that developing nations tend to be much less decentralized
than developed nations.
Overall, bidding and sorting models are likely to be applicable in many coun-
tries, but most of the related empirical evidence comes from the US, and the
usefulness of these models in a centralized country has never been tested. More-
over, the importance of these models is likely to increase over time as more
countries move toward decentralized systems. For recent developments in fiscal
decentralization, see Fraschini (1989) on Italy, Kojima (1992a) and Bahl and
Nath (1986) on developing nations, Oh (1991) and Ito (1992) on Korea, Kojima
(1992b) on China, and Leigland (1993) on Indonesia.
The generality of the bidding/sorting framework stands in sharp contrast to the
dependence of collective choice models on institutional detail. The median-voter
model, which is widely used in the literature reviewed here, implicitly assumes
that all residents participate in and influence the collective choice process. This
assumption appears to work well in some circumstances, such as voting in cities
Ch. 47: Sorting and Voting 2053
in Switzerland, especially those that are direct democracies, and in many local
government contexts in the US. Evidence from Sone (1993) suggests that mod-
ified voter models also can be helpful in democracies, such as Japan, that give
public officials considerable latitude.
The median-voter approach has obvious limits with more centralized decision-
making and seems irrelevant in many developing nations where local fiscal
decision-making is dominated by a select few (see Conyers, 1990). Moreover,
Bahl and Linn (1992) document that local governments in developing countries
rely heavily on property taxes, but they account for a small share of public
revenue, and central decision-making greatly limits variation in public services
across local governments. Although trends toward decentralization may increase
their applicability in the future, the voting models explored here appear at present
to be most helpful in developed countries with strong-and varied-local gov-
ernments.
The literature reviewed here is rich and diverse but leaves plenty of room
for future scholars. Many aspects of the voting models remain untested; little
empirical research on bidding and sorting, or their impact on voting has been con-
ducted for countries other than the US; and the implications of local government
inefficiency for public policy remain largely unexplored.
References
Aaron, H.J. (1975), Who Pays the Property Tax: A New View (The Brookings Institution, Washington, DC).
Akin, J.S. and D.J. Young Day (1976), "The efficiency of local school finance", Review of Economics and
Statistics 58:255-258.
Arnott, R.J. and J.G. McKinnon (1977), "The effects of the property tax: A general equilibrium simulation",
Journal of Urban Economics 4:389-407.
Aronsson, T. and M. Wikstr6m (1996), "Local public expenditure in Sweden: a model where the median voter
is not necessarily decisive", European Economic Review 40:1705-1716.
Atkinson, A.B. and N.H. Stem (1974), "Pigou, taxation, and public goods", Review of Economic Studies
41:119-128.
Bahl, R.W. and J.F. Linn (1992), Urban Public Finance in Developing Countries (Oxford University Press,
Oxford).
Bahl, R.W. and S. Nath (1986), "Public expenditure decentralization in developing countries", Environment
and Planning C: Government and Policy 4:405-418.
Barlow, R. (1970), "Efficiency aspects of local school finance", Journal of Political Economy 78:1028-1040.
Barzel, Y. (1973), "Private schools and public school finance", Journal of Political Economy 81:174-186.
Bentick, B.L. (1996), "The differential incidence of an urban land tax depends on the travel intensities of
substitutes for land", Urban Studies 33:1729-1732.
Bergstrom, T. (1973), "A note on efficient taxation", Journal of Political Economy 81:187-191.
Bergstrom, T. and R. Goodman (1973), "Private demands for public goods", American Economic Review
63:280-296.
Besley, T., I. Preston and M. Ridge (1997), "Fiscal anarchy in the UK: modeling poll tax noncompliance",
Journal of Public Economics 64:137-152.
2054 S. Ross and J. Yinger
Black, S.E. (in press), "Do better schools matter? Parental valuation of elementary education", Quarterly
Journal of Economics.
Bogart, W.T. and B. Cromwell (1997), "How much more is a good school district worth'?". National Tax Journal
50:215-232.
Boije, R. (1997), Capitalisation. Efficiency and the Demand for Local Public Services, Economic Studies 33
(Department of Economics, Uppsala University, Uppsala, Sweden).
Borcherding, T.E. and R.T. Deacon (1972), "The demand for services of non-federal governments", American
Economic Review 62:891-901.
Borge, L.-E. and J. Rattso (1995), "Demographic shift, relative costs, and the allocation of local public
consumption in Norway", Regional Science and Urban Economics 25:705-724.
Bradbury, K.L., H.E Ladd, M. Perrault, A. Reschovsky and J. Yinger (1984), "State aid to offset fiscal
disparities across communities", National Tax Journal 37:151-170.
Bradbury, K.L., C.J. Mayer and K.E. Case (1997), "Property tax limits and local fiscal behavior: Did Massa-
chusetts cities and towns spend too little on town services under Proposition 2 ½",Working Paper No. 97-2
(Federal Reserve Bank of Boston, Boston).
Bradford, D.F., R.A. Malt and W.E. Oates (1969), The rising cost of local public services: some evidence and
reflections". National Tax Journal 22:185-202.
Bradford. D.F. and W.E. Oates (1974), "Suburban exploitation of central cities and government structure",
in: H.M. Hochman and G.E. Peterson, eds., Redistribution Through Public Choice, pp. 43-92 (Columbia
University Press. New York).
Brueckner, J.K. (1979), "Property values, local public expenditure, and economic efficiency", Journal of Urban
Economics 11:223-246.
Brueckner, J.K. (1981), "Congested public goods: The case of fire protection", Journal of Public Economics
15:45-58.
Brueckner, J.K. (1982), "A test for allocative efficiency in the local public sector", Joumal of Public Economics
14:311-332.
Brueckner, J.K. (1983), "Property value maximization and public sector efficiency", Journal of Urban
Economics 14:1-16.
Brueckner. J.K. (1987). "The structure of urban equilibria: a unified treatment of the Muth-Mills model", in:
E.S. Mills, ed, Handbook of Regional and Urban Economics, pp. 821-875 (North Holland, Amsterdam).
Brueckner, J.K. and K. Lee (1989), "Club theory with a peer-group effect",. Regional Science and Urban
Economics 19:399-420.
Carroll, R. and J. Yinger (1994), "Is the property tax a benefit tax? The case of rental housing", National Tax
Journal 47:295-316.
Chattopadhyay, S. (1998), "An empirical investigation into the performance of Ellickson's random bidding
model, with an application to air quality valuation". Journal of Urban Economics 43:292-314.
Chinloy, P. (1978), "Effective property taxes and tax capitalization", Canadian Journal of Economics 11:740-
750.
Coleman, J. (1966), Equality of Educational Opportunity (US Government Printing Office, Washington, DC).
Conley, J. and M.H. Wooders (1998), "Anonymous pricing in Tiebout economies and economies with
clubs", in: D. Pines, E. Sadka, and I. Zilcha, eds., Topics in Public Finance (Cambridge University Press,
Cambridge).
Conley, J. and M.H. Wooders (1997), "Equivalence of the core and competitive equilibrium in a Tiebout
economy with crowding types", Journal of Urban Economies 41:421-440.
Conyers, D. (1990), "Centralization and development planning: a comparative perspective", in: P. De Valk and
K. Wekwete, eds., Decentralizing for Participatory Planning (Avebury, Aldershot, UK).
Courant, P.N. and S. Loeb (1997), "Centralization of school finance in Michigan", Journal of Policy Analysis
and Management 16:114-136.
Crampton, G. (1996), "Local government structure and urban residential location", Urban Studies 33:1061-
1076.
Crane. R. (1988). "Second-best property value capitalization", Econonmic Letters 26:175-178.
Ch. 47: Sorting and Voting 2055
Crane, R. (1990), "Price specification and the demand for public goods", Journal of Public Economics 43:93-
106.
deBartolome, C.A. and S.S. Rosenthal (in press), "Property tax capitalization in a model with tax deferred
assets and standard deductions", Review of Economics and Statistics.
Do, A.Q. and C.E Sirmans (1994), "Residential property tax capitalization: discount rate evidence from
California", National Tax Journal 47:341-348.
Dowding, K., J. Peter and S. Biggs (1994), "Tiebout: a survey of empirical literature", Urban Studies 31:767-
797.
Downes, T.A. (1992), "Evaluating the impact of school finance reform on the provision of public education:
the California case", National Tax Journal 45:405-419.
Downes, T.A. and T.E Pogue (1994), "Adjusting school aid formulas for the higher cost of educating
disadvantaged students", National Tax Journal 47:89-110.
Duncombe, W. (1991), "Demand for local public services revisited: the case of fire protection", Public Finance
Quarterly 19:412-436.
Duncombe, W. and J. Yinger (1993), "An analysis of returns to scale in public production, with an application
to fire protection", Journal of Public Economics 52:49-72.
Duncombe, W. and J. Yinger (1997), "Why is it so hard to help central city schools?", Journal of Policy
Analysis and Management 16:85-113.
Duncombe, W. and J. Yinger (1998), "School finance reform: aid formulas and equity objectives", National
Tax Journal 51:239-262.
Dynarski, M., R. Schwab and E. Zampelli (1989), "Local characteristics and local production: the case of
education", Journal of Urban Economics 26:250-263.
Eberts, R.W. and T.J. Gronberg (1981), "Jurisdictional homogeneity and the Tiebout hypothesis", Journal of
Urban Economics 10:227-239.
Edel, M. and E. Sclar (1974), "Taxes, spending and property values: supply adjustments in a Tiebout-Oates
model", Journal of Political Economy 82:941-954.
Edelson, N.M. (1973), "Efficiency aspects of local school finance: comments and extension", Journal of
Political Economy 81:158-173.
Eisenberg, E.E (1996), Intrajurisdictional Property Tax Capitalization Rates, unpublished PhD. Dissertation
(Syracuse University, Syracuse, NY).
Ellickson, B. (1971), "Jurisdictional fragmentation and residential choice", American Economic Review
61:334-339.
Ellickson, B. (1981), "An alternative test of the hedonic theory of housing markets", Journal of Urban
Economics 9:56-63.
Epple, D. (1987), "Hedonic prices and implicit markets: Estimating demand and supply functions for
differential products", Journal of Political Economy 95:59-80.
Epple, D., R. Filimon, and T. Romer (1983), "Housing, voting, and moving: equilibrium in a model of local
public goods with multiple jurisdictions", in: J.V. Henderson, ed., Research in Urban Economics, Vol. m,
pp. 59-90 (JAI Press, Greenwich, CT).
Epple, D., R. Filimon, and T. Romer (1984), "Equilibrium among local jurisdictions: toward and integrated
treatment of voting and residential choice", Journal of Public Economics 24:281-308.
Epple, D., R. Filimon, and T. Romer (1993), "Existence of voting and housing equilibrium in a system of
communities with property taxes", Regional Science and Urban Economics 23:585-610.
Epple, D. and G.J. Platt (1998), "Equilibrium among jurisdictions when households differ by preferences and
income", Journal of Urban Economics 43:23-51.
Epple, D. and R.E. Romano (1997), "School choice policies, household location, and property values",
Department of Economics Working Paper No. 1997-93 (University of Florida, Gainesville, FL).
Epple, D. and T. Romer (1989), "On flexible municipal boundaries", Journal of Urban Economics 26:307-319.
Epple, D. and T. Romer (1991), "Mobility and redistribution", Journal of Political Economy 99:828-859.
Epple, D. and H. Sieg (1997), "Estimating equilibrium models of local jurisdictions", Graduate School of
Industrial Administration Working Paper No. 1997-94 (Carnegie Mellon University, Pittsburgh, PA).
Epple, D., A. Zelenitz and M. Visscher (1978), "A search for testable implications of the Tiebout hypothesis",
Journal of Political Economy 86:405-425.
2056 S. Ross and J. Yinger
Feldstein, M.S. (1976a), "On the theory of tax reform", Journal of Public Economics 6:77-104.
Feldstein, M.S. (1976b), "Wealth neutrality and public choice in local education", American Economic Review
65:75-89.
Ferguson, R. (1991). "Paying for public education: new evidence on how and why money matters", Harvard
Journal on Legislation 28:465-498.
Fischel, W.A. (1992), "Property taxation and the Tiebout model: evidence for the benefit view from voting and
zoning", Journal of Economic Literature 30:171-177.
Flatters, F., J.V. Henderson and P. Mieszkowski (1974), "Public goods, efficiency and regional fiscal
equalization". Journal of Public Economics 3:99-112.
Follain, J.R. and E. Jimenez (1985), "Estimating the demand for housing characteristics: a survey and critique",
Regional Science and Urban Economics 15:77-107.
Fox, W. (1981), "Reviewing economies of size in education". Journal of Education Finance 6:273-296.
Fraschini. A. (1989), "Local autonomy, accountability, and a new local tax: the Italian debate". Policy and
Politics 17:155-163.
Frech, H.E., III and R.N. Lafferty (1984), "The effect of the California coastal commission on housing prices".
Journal of Urban Economics 16:105-123.
Fuhrmann, S., R. Elmore and D. Massell (1993), "School reform in the United States: putting it into context",
in S. Jacobson and R. Bume, eds., Reforming Education: The Emerging Systematic Approach, pp. 3-27
(Crown Press, Thousand Oakes, CA).
Goldstein, G.S. and M.V. Pauly (1981), "Tiebout bias on the demand for local public goods", Journal of Public
Economics 16:131-144.
Goodspeed, T.J. (1989), "A re-examination of the use of ability to pay taxes by local governments", Journal of
Public Economics 38:319-342.
Gramlich, E.M. and D.L. Rubinfeld (1982), "Micro estimates of public spending demand functions and tests
of the Tiebout and median-voter hypotheses", Journal of Political Economy 90:536-560.
Greenberg, J. (1979), "Consistent majority rules over compact sets of alternatives". Econometrica 47:627-636.
Grosskopf, S. and S. Yaisawaring (1990), "Economics of scope in the provision of local public services",
National Tax Journal 63:61-74.
Gyapong, A. and K. Gyimah-Brempong (1988). "Factor substitution, price elasticity of factor demand and
returns to scale in police protection: Evidence from Michigan". Southern Economic Journal 54:863-878.
Hamilton, B.W. (1975), "Zoning and property taxation in a system of local governments", Urban Studies
12:205-211.
Hamilton, B.W. (1976), "Capitalization of intrajurisdictional differences in local tax prices", American
Economic Review 22:743-753.
HIamilton. B.W. (1979), "Capitalization and the regressivity of the property tax: empirical evidence", National
Tax Journal 32:169-180.
Hamilton, B.W. (1983a), "The flypaper effect and other anomalies", Journal of Public Economics 22:347-361.
Hamilton, B.W. (1983b). "A review: is the property tax a benefit tax?", in: G.R. Zodrow, ed.. Local Provisions
of Public Services: The Tiebout Model after Twenty-Five Years, pp. 85-108 (Academic Press, New York).
Hamilton, B.W., E.S. Mills and D. Puryear (1975), "The Tiebout hypothesis and residential income segrega-
tion", in E.S. Mills and W.E. Oates, eds., Fiscal Zoning and Land Use Controls, pp. 101-118 (Lexington
Books. Lexington).
Hansen. N.A. and A.S. Kessler (1997), "The political geography of tax h(e)avens and tax hells", Discussion
Paper No. A-541 (University of Bonn. Bonn).
Hanushek, E. (1986), "The economics of schooling: production and efficiency in public schools", Journal of
Economic Literature 24:1141-1177.
Harmon, O.R. (1988), "The income elasticity of demand for single-family owner-occupied housing: an
empirical reconciliation", Journal of Urban Economics 24:173-185.
Harrison, D. and D.L. Rubinfeld (1978), 'Tiebout bias on the demand for local public goods", Journal of
Environmental Economics and Management 5:81-102.
Haurin, D.R. and D. Brasington (1996), "School quality and real house prices: inter- and intrametropolitan
effects", Journal of Housing Economics 5:351-368.
Haurin, D.R. (1980). "The effect of property taxes on urban areas", Journal of Urban Economics 7:384-396.
Ch. 47: Sorting and Voting 2057
Hayes, K. (1986), "Local public good demands and demographic effects", Applied Economics 18:1039-1045,
Heikkila, E.J. (1996), "Are municipalities Tieboutian clubs?", Regional Science and Urban Economics
26:203-
226.
Henderson, J.V. (1977), Economic Theory and the Cities (Academic Press, New York).
Henderson, J.V. (1985), "The Tiebout model: bring back the entrepreneurs", Journal of Political
Economy
93:248-264.
Henderson, J.V. (1991), "Separating Tiebout equilibrium", Journal of Urban Economics 29:128-152.
Henderson, J.V., P. Mieszkowski and Y. Sauvageau (1978), "Peer group effects and
education production
functions", Journal of Public Economics 10:97-106.
Holcombe, R. (1989), "The median voter model in public choice theory", Public Choice 61:115-125.
Hoyt, W.H. and S.S. Rosenthal (1997), "Household location and Tiebout: Do families
sort according to
preferences for locational amenities?", Journal of Urban Economics 42:159-178.
Inman, R.P. (1979), "The fiscal performance of local governments: An interpretive review",
in: P. Mieszkowski
and M. Straszheim, eds., Current Issues in Urban Economics, pp. 270-321 (Johns Hopkins University
Press,
Baltimore).
Inman, R.P. (1987), "Markets, governments, and the 'new' political economy", in: A.J. Auerbach
and M.
Feldstein, eds., Handbook of Public Economics, Vol. II, pp. 647-778 (North-Holland, Amsterdam).
Ito, K. (1992), "The reinstatement of local fiscal autonomy in the Republic of Korea", Developing
Economies
30:404 429.
Judd, G.D. and J.M. Watts (1981), "Schools and housing values", Land Economics 57:459-470.
Katz, L. and K.T. Rosen (1987), "The interjurisdictional effects of growth controls on housing
prices", Journal
of Law and Economics 30:149-160.
King, A.T (1973), Property Taxes, Amenities, and Residential Land Values (Ballinger, Cambridge,
MA).
Kojima, R. (1992a), "Central and local finances in developing countries-The trend toward
decentralization:
introduction", Developing Economies 30:311-314
Kojima, R. (1992b), "The growing fiscal authority of provincial-level governments in China",
Developing
Economies 30:315-346.
Konishi, H. (1997), "Voting with ballots and feet: existence of equilibrium in a local public good
economy",
Journal of Economic Theory 68:480-509.
Ladd, H.F.E (1975), "Local education expenditures, fiscal capacity, and the composition
of the property tax
base", National Tax Journal 28:145-158.
Ladd, H.F. and J. Boatright Wilson (1985), "Education and tax limitations: evidence from
Massachusetts"
Journal of Education Finance 10:281-296.
Ladd, H.F. and J. Yinger (1991), America's Ailing Cities: Fiscal Health and the Design of Urban
Policy,
updated edition (Johns Hopkins University Press, Baltimore).
Ladd, H.F. and J. Yinger (1994), "The case for equalizing aid", National Tax Journal 47:211-224.
LeGrand, J. (1975), "Fiscal equity and grants to local authorities", Economic Journal 85:531-547.
Leigland, J. (1993), "Decentralizing the development budget process in Indonesia: Progress and
prospects",
Public Budgeting and Finance 13:85-101.
Lenon, M., S.K. Chattopadhyay and D.R. Heffley (1996), "Zoning and fiscal interdependencies",
Journal of
Real Estate Finance and Economics 12:221-234.
Lerman, S.R. and C.R. Kern (1983), "Hedonic theory, bid rents, and willingness-to-pay: some
extensions of
Ellickson's results", Journal of Urban Economics 13:358-363.
Lipsey, R.G. and K. Lancaster (1956), "The general theory of the second best", Review of Economic
Studies
24:11-32.
Mayo, S.K. (1981), "Theory and estimation in the economics of housing demand", Journal of Urban
Economics
10:95-116.
McDougall, G.S. (1976), "Local public goods and residential property values: some insights and extensions",
National Tax Journal 29:436-437.
McFadden, D. (1974), "Conditional logit analysis of qualitative choice", in: P. Zarembka, ed., Frontiers
in
Econometrics (Academic Press, New York).
McMillen, D.P. and J.F. McDonald (1991), "A simultaneous equations model of zoning and land
values",
Regional Science and Urban Economics 21:55-72.
2058 S. Ross and J. Yinger
Miceli, TJ. (1991), "Free riders and distortionary zoning by local communities", Journal of Urban Economics
30:112-122.
Miceli, T.J. (1992), "Optimal fiscal zoning when the local government is a discriminating monopolist",
Regional Science and Urban Economics 22:579-596.
Mieszkowski, P. and G.R. Zodrow (1989), "Taxation and the Tiebout model", Journal of Economic Literature
27:1098-1146.
Munley, V.G. (1982), "An alternative test of the Tiebout hypothesis", Public Choice 38:211-217.
Nechyba, T.J. (1996), "A computable general equilibrium model of intergovernmental aid", Journal of Public
Economics 62:363-398.
Nechyba, T.J. (1997), "Existence of equilibrium and stratification in local and hierarchical Tiebout economies
with property taxes and voting", Economic Theory 10:277-304.
Nechyba, T.J. and R.P. Strauss (1998). "Community choice and local public services: a discrete choice
approach", Regional Science and Urban Economics 28:51-74.
Noto, N.A. (1976), "The impact of the public sector on residential property values", Proceedings of the
National Tax Association, pp. 192-200.
Oakland, W.H. (1994), "Fiscal equalization: an empty box?" National Tax Journal 47:211-224.
Oates, W.E. (1969), "The effects of property taxes and local public services on property values: an empirical
study of tax capitalization and the Tiebout hypothesis", Journal of Political Economy 77:957-971.
Oates, W.E. (1972), Fiscal Federalism (Harcourt Brace, New York).
Oates, W.E. (1973), "The effects of property taxes and local spending on property values: a reply and yet
further results", Journal of Political Economy 81:1004-1008.
Oates, W.E. (1977a), "An economist's perspective on fiscal federalism", in: W.E. Oates, ed., The Political
Economy of Fiscal Federalism (Lexington Books, Lexington).
Oates, W.E. (1977b), "The use of local zoning ordinances to regulate population flows and the quality of local
services", in: O.C. Ashenfelter and W.E. Oates, eds., Essays in Labor Market Analysis, pp. 201-220 (John
Wiley and Sons, New York).
Oates, W.E. (1985), "Searching for Leviathan: An empirical study", American Economic Review 75:748-757.
Oates, W.E. (1993), "Fiscal decentralization and economic development", National Tax Journal 46:237-244.
Oates, W.E. (1994), "Federalism and government finance", in: J.S. Quigley and E. Smolensky, eds., Modem
Public Finance, pp. 126-161 (Harvard University Press, Cambridge, MA).
Oh, Y.-C. (1991), "The allocation of tax bases between central and local governments in Korea", in: T.
Prud'homme, ed., Public Finance with Several Levels of Government (Foundation Journal Public Finance,
The Hague).
O'Sullivan, A., T.A. Sexton and S.M. Sheffrin (1995), Property Taxes and Tax Revolts: The Legacy of
Proposition 13 (Cambridge University Press, New York).
Pack, H. and J.R. Pack (1977), "Metropolitan fragmentation and suburban homogeneity", Urban Studies
14:191-201.
Pasha, H.A. (1990), "The differential incidence of land taxation", Urban Studies 27:591-595.
Pasha, H.A. and A.F. Aisha Ghaus (1995), "General equilibrium effects of local taxes", Journal of Urban
Economics 38:253-271.
Pauly, M.V. (1976). "A model of local government expenditure and tax capitalization", Journal of Public
Economics 6:231-242.
Pogodzinski, J.M. (1988), "Amenities in an urban general equilibrium model", Journal of Urban Economics
24:260-278.
Pogodzinski, J.M. and T.R. Sass (1994), "The theory and estimation of endogenous zoning", Regional Science
and Urban Economics 24:601-630.
Pogodzinski, J.M. and D.L. Sjoquist (1985), "A numerical approach to comparative statics in a Tiebout-median
voter model", in: J.M. Quigley, ed., Perspectives on Local Public Finance and Public Policy, Vol. 2, pp. 31-58
(JAI Press, Greenwich, CT).
Pogodzinski, J.M. and D.L. Sjoquist (1991), "The effects of the social choice rule on local fiscal variables: a
general equilibrium approach", Regional Science and Urban Economics 21:31-54.
Pogodzinski, J.M. and D.L. Sjoquist (1993), "Alternative tax regimes in a local public good model", Journal of
Public Economics 50:115-141.
Ch. 47: Sorting and Voting 2059
Pogodzinski, J.M. and D.L. Sjoquist (1994), "A note on the estimation of capitalization", Public Finance
Quarterly 21:434-448.
Polinsky, A.M. and D.L. Rubinfeld (1978), "The long run effects of a residential property tax
and local public
services", Journal of Urban Economics 5:241-262.
Pollakowski, H.O. (1973), "The effects of property taxes and local public spending on property
values: a
comment and further results", Journal of Political Economy 5:119-129.
Pommerhene, W. (1977), "Quantitative aspects of federalism: a study of six countries", in: W.E.
Oates, ed.,
The Political Economy of Fiscal Federalism (Lexington Books, Lexington).
Pommerhene, W. (1978), "Institutional approaches to public expenditure: empirical evidence
from Swiss
municipalities", Journal of Public Economics 9:255-280.
Preston, I. and M. Ridge (1995), "Demand for local public spending: evidence from the
British social attitudes
survey", The Economic Journal 105:644-660.
Quigley, J.M. (1985), "Consumer choice of dwelling, neighborhood, and public services",
Journal of Urban
Economics 15:41-63.
Reid, G. (1990), "The many faces of Tiebout bias in local education demand estimates",
Journal of Urban
Economics 27:232-254.
Reid, G. (1991), "Tests of institutional vs. non-institutional models of local public expenditure
determination",
Public Choice 70:315-333.
Reschovsky, A. (1979), "Residential choice and the local public sector: an alternative test
of the Tiebout
hypothesis", Journal of Urban Economics 4:501-520.
Reschovsky, A. (1980), "An evaluation of metropolitan area tax base sharing", National Tax Journal
33:55-66.
Reschovsky, A. (1994), "Fiscal equalization and school finance", National Tax Journal 47:185-198.
Roberts, J. (1992), "A comment on the many faces of Tiebout bias", Journal of Urban Economics 32:45-51.
Romer, T. and H. Rosenthal (1979), "The elusive median voter", Journal of Public
Economics 12:143-170.
Rose-Ackerman, S. (1979), "Market models of local government: exit, voting, and the
land market", Journal
of Urban Economics 6:319-337.
Rose-Ackerman, S. (1983), "Beyond Tiebout: modeling the political economy of local government,
in: G.R.
Zodrow, ed., Local Provision of Public Services: The Tiebout Model after Twenty-Five Years,
pp. 55-83
(Academic Press, New York).
Rosen, H.S. and D.J. Fullerton (1977), "A note on local tax rates, public benefit levels and
property values",
Journal of Political Economy 85:433-440.
Rosen, K.T. (1982), "The impact of Proposition 13 on house prices in northern California:
a test of the
intejurisdictional capitalization hypothesis", Journal of Political Economy 90:191-200.
Rubinfeld, D.L. (1987), "The economics of the local public sector", in: A.J. Auerbach and
M. Feldstein, eds.,
Handbook of Public Economics, Vol. II, pp. 571-646 (North-Holland, New York).
Rubinfeld, D.L., P. Shapiro and J. Roberts (1987), "Tiebout bias and the demand for local
public schooling",
Review of Economics and Statistics 69:426-437.
Samuelson, P.A. (1954), "The pure theory of public expenditures", Review of Economics
and Statistics
36:387-389.
Schmidt, A. (1992), "Private school enrollment in metropolitan areas", Public Finance Quarterly
20:298-320.
Schwab, R. and W.E. Oates (1991), "Community composition and the provision of local public goods: a
normative analysis", Journal of Public Economics 44:217-237.
Schwab, R. and E. Zampelli (1987), "Untangling the demand function from the production function
for local
public services: the case of public safety", Journal of Public Economics 33:245-260.
Schwartz, S., P.M. Zorn and D.E. Hansen (1986), "Research design issues and pitfalls in
growth control
studies", Land Economics 62:223-233.
Sone, A. (1993), "An analysis of municipal government expenditures in Japan", unpublished PhD.
Dissertation
(Syracuse University, Syracuse, NY).
Sonstelie, J.C. and P.R. Portney (1980), "Gross rents and market values: testing the implications
of Tiebout's
hypothesis" Journal of Urban Economics 7:102-118.
Stiglitz, J.E. (1983), "The theory of local public goods twenty-five years after Tiebout: a perspective",
in: G.S.
Zodrow, ed., Local Provision of Public Services: The Tiebout Model after Twenty-Five Years, pp.
17-54
(Academic Press, New York).
2060 S. Ross and J. Yinger
Stull, W.J. and J.C. Stull (1991), "Capitalization of local income taxes", Journal of Urban Economics 29:182-
190.
Sullivan, A.M. (1985). The general equilibrium effects of the residential property tax: incidence and excess
burden", Journal of Urban Economics 18:235-250.
Tiebout, C.M. (1956), "A pure theory of local expenditures", Journal of Political Economy 64:416-424.
Westoff, F. (1977). "Existence of equilibrium in economics with a local public good", Journal of Economic
Theory 14:84-112.
Wheaton, W.C. (1993). "Land capitalization, Tiebout mobility, and the role of zoning regulations", Journal of
Urban Economics 34:102-117.
White, M.J. (1975), "Fiscal zoning in fragmented metropolitan areas", in: E.S. Mills and W. Oates, eds., Fiscal
Zoning and Land Use Controls, pp. 31-100 (Lexington Books, Lexington, MA).
Wildasin, D.E. (1986). Urban Public Finance (Harwood Academic Publishers, London).
Wooders, M.H. (1989), "A Tiebout theorem", Mathematical Social Sciences 18:33-35.
Wyckoff, P.G. (1991), "The elusive flypaper effect", Journal of Urban Economics 30:310-328.
Wyckoff, P.G. (1995), "Capitalization, equalization, and intergovernmental aid", Public Finance Quarterly
23:484-508.
Y'nger, J. (1982), "Capitalization and the theory of local public finance", Journal of Political Economy 90:917-
943.
Yinger, J. (1985), "Inefficiency and the median voter: Property taxes, capitalization, heterogeneity, and the
theory of the second best", in: J.M. Quigley, ed., Perspectives on Local Public Finance and Public Policy,
Vol. 2, pp. 3-30 (JAI Press, Greenwich, CT).
Yinger, J. (1995), "Capitalization and sorting: a revision", Public Finance Quarterly 23:217-225.
Yinger, J. (1999), "Capitalization and efficiency", in: A. Panagariya, P. Portney, and R. Schwab, eds., Envi-
ronmental and Public Economics: Essays in Honor of W. E. Oates, pp. 310-26 (Edward Elgar Publishing
Company, Cheltenham, UK).
Yinger, J., H.S. Bloom, A. Boersch-Supan and H.F. Ladd (1988), Property Taxes and House Values: The Theory
and Estimation of Intrajurisdictional Property Tax Capitalization (Academic Press, New York).
AUTHOR INDEX
I-1
I-2 Author Index
2018, 2020, 2030, 2032, 2035, 2036, 1764, 1765, 2004, 2006, 2009, 2010,
2040, 2042, 2043 2012, 2015, 2018, 2027, 2039, 2044
Hamilton, F.E.I., 1823 Hendry, D.F., 1570, 1646
Hamilton, J., 1442 Henning, J.A., 1614
Hamley, D., 1724, 1726 Hensher, D.A., 1958
Hancock, K., 1577, 1580 Herzog, H., 1766, 1767
Hanks, J.W., Jr., 1940 Hills, J., 1566, 1580
Hannah, L., 1738, 1808, 1825, 1830, 1835, Hilton, G.W., 1973
1836, 1843 Hirono, K., 1574
Hansen, J.L., 1566 Hoehn, J.P., 1415, 1416, 1429
Hansen, N.A., 2028 Hoek-Smit, M., 1842
Hanson, G.H., 1468, 1762 Hoff, K., 1921
Hanson, S., 1525, 1526 Hoffman, M., 1793
Hansson, B., 1489 Holcombe, R., 2019
Hanushek, E., 1566, 1575, 2039 Holden, D.J., 1946
Harberger, A., 1564 Holland, S.A., 1574
Hardoy, J.E., 1828 Holmans, A.E., 1580
Harmon, O.R., 1617, 2011 Holt, D., 1426
Harper, Jr., E.P., 1968 Holtz-Eakin, D., 1489
Harrington, D.E., 1567 Holzer, H.J., 1514, 1515, 1520, 1532, 1533
Harrington, W., 1955, 1958, 1959, 1962, 1964 Honghai, D., 1727, 1728
Harris, B., 1891 Hoover, E.M., Jr., 1460, 1469
Harris, J., 1755 Horowitz, J.L., 1628, 1631
Harrison, B., 1520, 1533 Hossain, M., 1882
Harrison, D.J., 1440, 1614, 2033 Hotchkiss, D., 1392, 1522
Harsman, B., 1583, 1843 Howe, H., 1798
Harvey, G.W., 1954, 1960, 1963, 1975 Howe, L.K., 1988
Hass-Klau, C., 1948 Howenstine, E.J., 1806
Hau, T.D., 1942 Hoy, M., 1735, 1736, 1812
Haughwout, A.F, 1429 Hoyt, H., 1575
Haurin, D.R., 1568, 1575, 2004, 2033 Hoyt, W.H., 2036
Hausman, J.A., 1427, 1443 Hsing, Y., 1347
Hay, D.G., 1359, 1360, 1366, 1416, 1429, 1446 Hu, S.C., 1574
Hayami, Y., 1901 Huang, Ju Chin, 1438
Hayes, K., 2039 Hughes, G.A., 1581
Hebbert, M., 1664, 1665 Hughes, J.A., 1747
Hegedus, J., 1798, 1810, 1831, 1832, 1841 Hughes, M.A., 1536
Heikkila, E.J., 2035 Hugo, G.J., 1715
Hekman, J., 1523, 1524 Hulten, C.R., 1365
Helsley, R.W., 1386, 1467, 1468, 1470-1472 Hunt, G.L., 1448
Hendershott, P.H., 1568, 1574, 1575 Huppi, M., 1913
Henderson, J.V., 1349, 1350, 1352, 1353,
1356, 1382, 1386, 1387, 1441, 1465- lhlanfeldt, K.R., 1399, 1400, 1515, 1516, 1537-
1468, 1474, 1477-1481, 1568, 1575, 1539, 1546, 1566, 1644
1729, 1732, 1739, 1756, 1760, 1762, Ingebrigtsen, S., 1969
1-8 Author Index
Ingram, G.K., 1562, 1565, 1571, 1738, 1740, Kasarda, J.D., 1531, 1532
1799, 1800, 1802, 1803, 1842 Kaserman, D.L., 1575
Inman, R.P., 1448, 2019, 2037, 2045 Katsura, H.M., 1815
International Labour Office, 1845 Katz, E., 1712-1714
International Road Federation, 1942, 1943 Katz, L., 1543, 2036
loannides, Y.M., 1568 Katzman, M., 1828
Ito, K., 2052 Kaufman, D., 1798
Ito, T., 1574 Kawada, K., 1955
Kawai, M., 1566
Jackson, J.R., 1616 Kawashima, T., 1476, 1479, 1482
Jacobs, J.. 1462, 1485 Kazimi, C., 1959, 1961, 1962
Jaffe, A.J., 1795, 1796, 1841 Keare, D., 1567, 1799, 1800
Jamison, D., 1918 Kearl, J.R., 1568
Jansson, J.O., 1965 Keeler, T.E., 1954
Jansson, K., 1976, 1982 Kelejian, H., 1362
Jayakrishnan, R., 1948 Kelley, A.C., 1703-1707
Jencks, C., 1519, 1532 Kemp, M.A., 1977
Jensen, M.J., 1367, 1368 Kemp, P.A., 1582
Jimenez, E., 1567, 1598, 1607, 1735, 1736, Kempner, J.L., 1582
1797-1800, 1809, 1812, 1823, 1828, Kendrick, J.W., 1365
1918-1920, 2034 Kennedy, D., 1982
Johansen, F., 1956 Kenworthy, J.R., 1949, 1950. 1975
Johnson, T.E. Jr., 1809, 1810 Kern, C.R., 1631, 2033
Johnston, D.C., 1983 Kessides, C., 1827
Johnston, I., 1525 Kessler, A.S., 2028
Johnston-Anumonwo, I., 1525, 1526 Khan, A.R., 1901
Jones, G., 1843 Khandker, S., 1897
Jones, L.E., 1603, 1616 Kiamba, M., 1796
Jones, P., 1955 Kiel, K., 1439, 1440
Jones-Lee, M.W., 1966, 1967, 1969 Kikuchi, M., 1901
Jovanovic, B., 1485 Kim, C., 1557
Judd, G.D., 2033 Kim, I., 1410
Kim, J., 1830
Kaganova, O.Z., 1832 Kim, K.-H., 1649, 1650, 1656, 1808, 1818,
Kahn, M., 1426, 1441, 1444 1820, 1825, 1826, 1836, 1841
Kahn, S., 1966 Kim, S., 1387, 1468, 1471, 1949
Kain,J.F., 1505, 1509, 1528, 1529, 1565-1568, Kim, S.-J., 1813
1571, 1614, 1947, 1975, 1979, 1980 King, A., 1830
Kakwani, N.C., 1878, 1880 King, A.T., 1567, 2033
Kalachek, E.D., 1529 King, M.A., 1567, 1568
Kalbermatten, J., 1828 King, M.L., 1426
Kaldor, N., 1481 Kingsley, G.T., 1748, 1749
Kamecke, U., 1346 Kirwan, R., 1567, 1573
Kanbur, R., 1721, 1722. 1874, 1891, 1924 Kish, L., 1425
Kanemoto, Y., 1627, 1630 Kiyotaki, N., 1570
Kannappan, S., 1752 Klaassen, L.H., 1359
Author Index I-9
Oates, W.E., 1581, 2015, 2016, 2030-2034, Parsons, G.R., 1567, 1607, 1617, 1628-1630
2043-2046, 2051, 2052 Pasha, H.A., 1738, 1801, 1802, 1841, 2004
Oaxaca, R., 1522 Pauly, M.V., 2013, 2018, 2035, 2040-2042
Oberai, A.S., 1759 Payne, G., 1822, 1829, 1841
Odling-Smee, J., 1576 Pejovich, S., 1796
OECD (Organisation for Economic Co- Peltzman, S., 1968
Operation, and Development), 1951, Peng, R., 1825
1980 Penny, D.H., 1901
Offner, P., 1529 Perl, A., 1973, 1974
Ogawa, H., 1381, 1382, 1385, 1386 Pernia, E.M., 1881, 1882, 1887, 1890, 1898,
Oh, Y.-C., 2052 1901-1903
O'Hara, J.D., 1381 Perry, J.L., 1978
Ohlin, B., 1460 Peterson, P.E., 1519
Ohsfeldt, R.L., 1610, 1611, 1624, 1625 Peterson, W., 1580
Oi, J.C., 1729, 1731, 1732 Pfefferman, D., 1426
Oizumi, E., 1649 Phang, S-Y., 1953
Okpala, D.C.I., 1818 Phelps, M., 1703
Okun, D.A., 1828 Phillips, K.E., 1534
Olsen, E.O., 1562, 1569, 1576, 1582, 1793, Phipps, T.T., 1610, 1624, 1627, 1630
1830, 1833, 1837, 1844 Phongpaichit, P., 1717
Ondiege, P., 1830, 1834 Pickrell, D.H., 1978, 1979
Ondrich, J., 1801, 1802 Piggott, J., 1833
Openshaw, S., 1503 Piore, M., 1518
O'Regan. K.M., 1511, 1539 Pitt, M., 1891, 1897
Orshansky, M., 1872 Platt, G.J., 2005, 2035
Osada, S., 1345, 1348, 1350 Pogodzinski, J.M., 1641, 1643, 1649, 1666,
Ostro, B.D., 1437 1804, 2004, 2007, 2011, 2019, 2027,
O'Sullivan, A.M. 1457, 1462, 2051 2033, 2037
Otsuka, K., 1916, 1917 Pogue, TF., 2039, 2040
Ou, F.L., 1977 Polinsky, A.M., 1566, 1567, 1803, 2004
Oum, T.H., 1977 Pollak, R., 1714
Ozanne, L., 1569, 1570, 1572, 1646, 1836 Pollakowski, H.O., 1568, 1615, 2031
Pommerhene, W., 2037, 2039, 2052
Pace, R.K., 1618, 1619 Pope, C., 1437
Pack, H., 2035, 2036 Population Reference Bureau. 1723
Pack, J.R., 2035, 2036 Porter, R., 1739
Paderanga, C., 1882 Portes, R., 1719, 1754
Page, D., 1843 Portney. P.R., 1437. 1443, 1444, 2016
Pain, N., 1568 Potepan, M.J., 1572
Palivos, T., 1468 Poterba, J.M., 1562, 1570, 1574
Palmquist, R.B., 1598 Potter, J.E., 1898
Palumbo, G., 1362 Pratt, G., 1525, 1526
Panzar, J.C., 1469 Preston, I., 2037
Parikh, K., 1751 Price, R., 1402, 1535
Parr, J.B., 1344 Prud'homme, R., 1940, 1959
Parry, D.L., 1820 Pryce. G., 1571, 1652
Author Index I-13
Psacharopoulos, G., 1868, 1895, 1906, 1918 Rietveld, P., 1512, 1949
Pucher, J., 1942, 1945, 1955, 1973-1975, 1977 Rob, B., 1485
Pudney, S., 1834 Roback, J., 1414, 1417, 1429, 1436, 1437, 1485
Puente, C.A., 1716 Roberts, J., 2041
Punpuing, S., 1743 Robinson, M., 1922
Robinson, S., 1706, 1707, 1711
Qabeer, M., 1654 Rodgers, G., 1898, 1900
Quandt, R., 1804 Rogers, A., 1705
Quibria, M.G., 1869, 1881-1883, 1886, 1890, Rogers, C.L., 1516
1907, 1911, 1912, 1916 Rogerson, P., 1513
Quigley, J.M., 1507, 1511, 1530, 1539, 1561, Rojas, E., 1842
1565-1569, 1575, 1583, 1584, 1607, Rom, M., 1954, 1960
1614, 1627, 1630, 1798, 1843, 2033 Romano, R.E., 2022, 2036
Quinet, E., 1940, 1958, 1959, 1965, 1966 Romer, T., 2004, 2015, 2019, 2020, 2024, 2028
Rose-Ackerman, S., 2014, 2024, 2028, 2029,
Rabianski, J., 1563, 1569 2043
Rahman, M., 1987 Rosen, H.S., 1623-1625, 1630, 1813, 2033
Rahman, O., 1887 Rosen, K.T., 1344-1348, 1357, 1565, 1575,
Rakodi, C., 1734, 1812 1763, 2031, 2036
Ranson, M., 1437 Rosen, S., 1414, 1438, 1442, 1570, 1605, 1620,
Raphael, S., 1534 1621
Rashid, S., 1916 Rosenthal, H., 2019
Ratts0, J., 2038 Rosenthal, L., 1567
Rauch, J.E., 1485 Rosenthal, S.S., 2032, 2036
Ravallion, M., 1544, 1879-1881, 1904, 1913, Rosenzweig, M.R., 1716, 1893, 1896
1834 Ross, S., 1390
Rawls, J., 1870 Rossetti, M.A., 1941, 1973
Rawski, T.G., 1730 Rothenberg, J., 1563, 1568, 1569, 1572, 1573,
Read, C., 1575, 1577 1576, 1577, 1580
Reece, B.F., 1576 Roumasset, J.R., 1899
Rees, A.E., 1398, 1545 Rouwendal, J., 1512, 1565, 1949
Reid, G., 2040, 2041, 2042 Rowntree, B.S., 1911
Reid, M., 1566 Rubenstein, A., 1714
Renaud, B., 1583, 1666, 1726, 1727, 1805, Rubinfeld, D.L., 1440, 1614, 2004, 2033-2037,
1806, 1815, 1817-1821, 1823, 1826, 2040, 2042
1832, 1835, 1841, 1842, 1882 Russo, G., 1512
Reschovsky, A., 2027, 2049, 2050 Ruster, J., 1956
Resnick, M., 1344-1348, 1357, 1763 Rutherford, T.F., 1761
Reza, A.M., 1542 Rutman, G.L., 1796, 1842
Richardson, H.W., 1352, 1464, 1486, 1759, Rydell, C.P., 1569, 1578, 1843
1941, 1950
Richardson, R., 1540, 1546 Sa-Aadu, J., 1842
Richels, R.G., 1959, 1963 Saba, R., 1582
Richmond, J., 1489 Sabatini, F., 1824
Ridge, M., 2037 Sabel, C.F., 1518
Ridker, R.G., 1614 Sabot, R.H., 1755
I-14 Author Index
Turner, B., 1579, 1580, 1582, 1829, 1831 Wachter, S.M., 1568
Turner, R.P., 1983 Wadhva, K., 1823
Turok, I., 1824, 1842 Waldman, D.M., 1447
Turvey, R., 1828, 1976 Wallace, N.E., 1613, 1640
Tyson, W.J., 1985 Wallin, B., 1982
Walter, S., 1458
UK Department of Transport, 1985 Walters, A.A., 1829, 1953
UN Centre for Human Settlements, and the Walz, U., 1468
World Bank, 1793, 1829 Wan, H., 1754
United Nations Development Programme Wang, L., 1834
(UNDP), 1724, 1867, 1909, 1927 Wang, P., 1468, 1471
United Nations, 1715, 1717, 1759, 1881, 1887 Wang, Y., 1729, 1731, 1732
Upton, C., 1356 Ward, P., 1822, 1824, 1825, 1843
Urban Edge, The, 1829 Warford, J., 1828
US Census Bureau, 1942, 1944, 1945, 1969 Watson, C.J., 1576
US FHWA (US Federal Highway Administra- Watts, H.W., 1876
tion), 1944, 1945, 1965 Watts, J.M., 2033
Waugh, F.V., 1597, 1613, 1614
Valenzona, R., 1902 Wegelin, E., 1890
Van den Berg, L., 1359, 1360 Weicher, J., 1561, 1576, 1578, 1584
Van der Heijden, H., 1579, 1580, 1582 Weil, D.N., 1574
Van der Linden, J., 1822 Weinberg, D.A., 1578
van der Veen, A., 1504, 1506 Weinberg, D.H., 1507, 1567, 1575, 1578, 1815
Van Ommeren, J., 1513 Weinstein, M., 1534
Vandell, K.D., 1576, 1795, 1797 Welch, E., 1531
Vapnarsky, C.A., 1344, 1345 Westaway, P., 1568, 1570, 1571
Vashistha, P., 1876 Westoff, F., 2026
Vaupel, J.W., 1728 Wheaton, W.C., 1345, 1392, 1523, 1568, 1570,
Vedder, R., 1767 1571, 1574, 1575, 1581, 1764, 1806,
Verma, S., 1760 1807, 1825, 1975, 2004-2006, 2008,
Vernez, G., 1822 2010, 2015, 2031
Versacaj, J.P., 1796 Wheeler, J.O., 1517
Vickerman, R.W., 1505, 1508 Whinston, A., 1573
Vickrey, W.S., 1953, 1968, 1969 Whitby, M.C., 1644
Vining, D.R., 1351, 1363, 1366 White, J.R., 1571
Vining, D.R., Jr., 1458 White, M.J., 1380, 1389, 1391, 1393, 1396,
Vining, R., 1344 1406, 1408, 1521, 1522, 1573, 1730,
Vipond, J., 1540, 1547 1816, 2015
Viscusi, V.K., 1966 White, P.R., 1983-1986
Viton, P.A., 1955, 1976, 1981 Whitehead, C.M.E., 1570, 1573, 1576, 1578-
Voith, R., 1429 1580, 1582, 1793
von Boventer, E., 1486 Whitelegg, J., 1958
Vrooman, J., 1534 Whittington, D., 1828
Wieand, K., 1380, 1382, 1390, 1391, 1403
Wachs, M., 1952, 1972, 1977 Wikstr6m, M., 2019, 2039, 2040, 2042
Wachter, M.L., 1531, 1545 Wildasin, D.E., 2019, 2043, 2045
Author Index I-17
1-19
I-20 Subject Index