Professional Documents
Culture Documents
(Cambridge Studies in Comparative Politics) Anibal Perez-Linan - Presidential Impeachment and The New Political Instability in Latin America-Cambridge University Press (2007)
(Cambridge Studies in Comparative Politics) Anibal Perez-Linan - Presidential Impeachment and The New Political Instability in Latin America-Cambridge University Press (2007)
General Editor
Margaret Levi University of Washington, Seattle
Associate Editors
Robert H. Bates Harvard University
Helen Milner Princeton University
Frances Rosenbluth Yale University
Susan Stokes Yale University
Sidney Tarrow Cornell University
Kathleen Thelen Northwestern University
Erik Wibbels University of Washington, Seattle
Cambridge University Press has no responsibility for the persistence or accuracy of urls
for external or third-party internet websites referred to in this publication, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
P1: JZP
0521869423pre cuny809-perez 0 521 86942 0 May 8, 2007 19:47
Contents
References 215
Index 235
ix
P1: JZP
0521869423pre cuny809-perez 0 521 86942 0 May 8, 2007 19:47
P1: JZP
0521869423pre cuny809-perez 0 521 86942 0 May 8, 2007 19:47
Tables
2.1. Impeachment crises in Latin America, 1992–2004 page 38
3.1. Incidence of military coups in Latin America, 1950–2004 43
3.2. Presidential crises in Latin America, by type of regime,
1950–2004 46
3.3. Outcomes of presidential crises, 1950–2004 61
3.4. Impact of presidential crises before and after the wave of
democratization 62
4.1. Entries under “Bribery and Corruption,” FBIS Index,
1983 and 1993 66
4.2. Students of communication and documentation per
thousand university students, 1980–ca. 1990 80
5.1. Summary of the six administrations 94
5.2. Fixed-effects models of presidential approval 117
5.3. Instrumental variable model of presidential approval 121
6.1. Bicameralism and models of impeachment 136
6.2. Constitutional rules on impeachment and declaration of
incapacity 140
6.3. Presidential strategies and support for the executive in six
crises 168
6.4. Individual support for the president (logit models) 173
6.5. Observed and predicted support for the president 174
7.1. Presidential crises in Latin America, 1990–2004 192
7.2. Conditions for impeachment and removal from office
(qualitative comparative analysis) 194
7.3. Predictors of presidential crises, impeachments, removals,
and mass protests, 1990–2004 200
xi
P1: JZP
0521869423pre cuny809-perez 0 521 86942 0 May 8, 2007 19:47
Figures
3.1. Competitive regimes in Latin America, 1950–2004 42
3.2. A typology of outcomes of presidential crises 49
4.1. Political scandals, 1980–2004 67
4.2. Television sets per thousand inhabitants, 1975–2001 76
4.3. Communication students per thousand college students,
1971–1993 79
5.1. Presidential approval in six impeachment crises 89
6.1. Support for the president in six impeachment crises 145
xii
P1: JZP
0521869423pre cuny809-perez 0 521 86942 0 May 8, 2007 19:47
The 1990s were an era of great hopes for Latin America. After the demise
of authoritarian regimes in the 1980s and the early 1990s, major economic
reforms were undertaken in most Latin American countries in order to
reduce chronic inflation and promote sustained growth. For many contem-
porary observers, the confluence of democracy and free markets signaled a
break with the past, the dawn of a new era of civil liberties, prosperity, and
political stability.
More than a decade later, it is hard to look back at this period without
a mixture of nostalgia and sarcasm. The legacies of the 1990s varied from
country to country, but they can be generally described as notable achieve-
ments overshadowed by missed opportunities. In the economic realm,
hyperinflation was eventually defeated, but economic growth remained
elusive and poverty resilient. In the political arena, the military eventually
withdrew from politics (not a minor feat), but elected governments, sur-
prisingly, continued to collapse. Starting in the early 1990s, presidents were
removed from office in Brazil, Venezuela, Guatemala, Ecuador, Paraguay,
Peru, Argentina, and Bolivia – in some countries recurrently. This outcome
frequently represented the triumph of an indignant society over a corrupt
or abusive executive, but it seldom prevented the occurrence of new abuses
in later administrations. By the early years of the twenty-first century, it
was clear that the particular circumstances of each crisis represented only
parts of a broader puzzle – a new pattern of political instability emerging
in the region.
This book explores the origins and the consequences of this novel pat-
tern of instability, emphasizing the critical events that defined the new
trend between 1992 and 2004. During this period, civilian elites realized
that traditional military coups had become for the most part unfeasible and
xiii
P1: JZP
0521869423pre cuny809-perez 0 521 86942 0 May 8, 2007 19:47
Pittsburgh, Pennsylvania
August 2006
xvi
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
The president’s seat represents the dream job for most politicians. Presi-
dents are power brokers, party leaders, role models, the daily focus of public
opinion. Presidents speak for the nation, they are primi inter pares among
national political figures. “They say,” former Chilean president Patricio
Aylwin once joked, “that the most difficult task after being president is get-
ting used to not being president.” Presidents, however, are not free from
failure. And the completion of their terms, particularly in Latin America,
is never guaranteed.
This book deals with an extreme form of political failure: presidential
impeachment. Impeachment transforms the luck of the most successful
politician in the country into a model of defeat. Presidents are deprived
of honor and power, deserted by former allies and voters, prosecuted as
ordinary citizens, and many times incarcerated or forced into exile.
In the 1990s, an unprecedented wave of impeachments swept Latin
America. Dwellers of presidential palaces, from Carondelet to Miraflo-
res and from Planalto to the House of Nariño, unexpectedly confronted
this threat. In just over a decade, between 1992 and 2004, six presidents
faced an impeachment process, and four of them were removed from office.
Brazilian President Fernando Collor de Mello in 1992 and Venezuelan Pres-
ident Carlos Andrés Pérez in 1993 were accused of corruption and ousted
on impeachment charges. In 1996, Colombian President Ernesto Samper
was charged with receiving illegal campaign funds from the Cali drug car-
tel. Congress ultimately acquitted Samper, but his political leverage was
greatly diminished as a consequence of the scandal. The following year, the
Ecuadorian Congress confronted President Abdalá Bucaram and, in order
to avoid the institutional intricacies of impeachment, declared the president
mentally impaired. Paraguayan President Raúl Cubas Grau confronted an
1
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
4
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
5
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
1 For a recent attempt to offer such comparative framework, see Baumgartner and Kada
(2003).
6
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
2 I describe the procedure employed to identify and code presidential crises in Chapter 3.
3 This figure includes only coups that took place in situations in which a presidential crisis
was possible (i.e., when a president and a legislature coexisted at the time). I defined as a
coup any episode in which the military successfuly removed the president from office, closed
Congress, or both. The cases are: Argentina (1955, 1962, 1966, 1976); Bolivia (1951, 1964,
1969, 1979, 1980); Brazil (1954 [an ambiguous case because President Vargas committed
suicide in anticipation of the coup], 1955, 1964); Chile (1973); Colombia (1953); Cuba
(1952); the Dominican Republic (1963); Ecuador (1961, 1963, 1970, 2000); El Salvador
(1960, 1979); Guatemala (1954, 1957, 1963, 1982, 1993); Honduras (1954, 1957, 1963,
1972); Panama (1951, 1968, 1985, 1988); Paraguay (1954, 1989); Peru (1962, 1968, 1992);
8
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
hand, presidential crises are not sufficient for democratic instability. About
half (45 percent) of the presidential crises taking place in Latin America
during this period led to some sort of military intervention, and 36 percent
ended with a full-blown military coup; but this book is inspired by the fact
that a vast majority of presidential crises over the last decade and a half have
not been related to any form of regime breakdown.
Episodes of impeachment are thus presented in this book as a subset
of the universe of presidential crises, in turn an extremely hostile form of
executive-legislative interaction. This perspective suggests that impeach-
ment is not just a legal recourse to remove presidents who are proven
guilty of high crimes; it is often an institutional weapon employed against
presidents who confront a belligerent legislature. This may be true even
in well-established democracies like the United States (Perkins 2003, 21).
The question about the proper use of impeachment – that is, its use for the
purpose of punishing actual misdemeanors rather than for merely partisan
reasons – is often difficult to confront, both for the contemporary observer
and for the analyst claiming the advantage of historical hindsight.
In spite of this approach, the reader should not be tempted to assume
that the only relevant actors in the drama are the executive and the legisla-
ture. Impeachment crises often involve subtle negotiations to appease the
military, media investigations, popular protests, and attempts to manipulate
the judiciary (Whitehead 2002, 104). In fledgling democracies, the military
may play a crucial role in shaping the outcome of a crisis – even when
the survival of the regime is not threatened, for instance, generals may act
as mediators. In well-established democracies, other institutions such as
the press and the judiciary may play powerful roles in the confrontation.
Executive-legislative conflicts may implicate various third parties according
to the circumstances, but they usually involve a complex structure in which
both institutions seek the support of a third party against the other branch
(on the role of third party players in different contexts, see Caplow 1968).
Consider, for instance, the strategies adopted by legislators seeking to
oust the president. In some cases, they investigate presidential misdeeds
originally exposed by the press. The Watergate scandal in the United
States and the Brazilian crisis of 1992 illustrate this situation. In other
Uruguay (1973); and Venezuela (1958). Twenty other coups taking place in countries without
an operating legislature (e.g., Ecuador in 1972) were excluded from the count. These events
were identified using The New York Times Index and other historical sources (Fossum 1967;
Needler 1966).
9
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
emergence of popular outrage against the president. I also argue that the
coverage of scandals may be endogenous to presidential approval: not only
did presidential approval decline with media exposés, media outlets were
also more likely to pursue new stories as presidents became less popular
among their readers and audiences. This dynamic created a characteristic
downward spiral leading to impeachment.
With media scandals hurting the administration and mounting public
outrage against the president, even loyal legislators have found it difficult to
resist public pressures for impeachment. Chapter 6 approaches this problem
from a neo-institutional perspective. To what extent can constitutional rules
facilitate or constrain the impeachment process? I argue that constitutional
rules interact with the partisan configuration of Congress, the president’s
ability to form coalitions, and the political context (the nature of the scandals
and the timing of the electoral calendar) to create a legislative shield to
protect the president.
Chapter 7 documents an emerging pattern of political instability in Latin
America. I contrast the episodes discussed in Chapter 2 with three alterna-
tive situations: presidential crises that forced the resignation of the pres-
ident without an impeachment process, presidential crises that never led
to the resignation of the president, and administrations that never faced a
presidential crisis. I use the Argentine crisis of 2001 to illustrate a char-
acteristic situation in which a political debacle outpaced the capacity of
politicians to respond to events. These situations share some important
commonalties with the core cases presented in Chapter 2, but they dif-
fer substantially in the outcome because the president resigned before an
impeachment was feasible. A qualitative comparative analysis of these cases
along with other presidential crises in which the president survived con-
firms the insight developed in earlier chapters: while scandals and legislative
politics are key to explaining the impeachment process, mass mobilization
constitutes the main factor driving the actual removal of presidents from
office – irrespective of the specific procedure employed to achieve this goal.
Further statistical analysis examining a large number of administrations that
never faced a presidential crisis confirms this hypothesis, and shows that a
distinctive form of political turmoil emerged in Latin America in the 1990s.
The last chapter explores the theoretical relevance of those findings for
the study of presidential accountability and democracy in Latin America. A
radical form of “social” accountability seems to be on the rise throughout
the region (Smulovitz and Peruzzotti 2000). Impeachment may be the best
institutional mechanism to channel the outbursts of public indignation, but
12
P1: JZP
0521869423c01 cuny809-perez 0 521 86942 0 May 8, 2007 14:54
13
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
14
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
Brazil, 1992
In 1989, after twenty-one years of military rule (1964–85) and a transitional
government led by José Sarney (1985–89), Fernando Collor de Mello won
the first direct presidential election that Brazil had seen in twenty-nine
years. The governor of the poor state of Alagoas had campaigned against
the political establishment, presenting himself as a political hero backed
by a minuscule organization, the National Reconstruction Party (Partido
da Reconstrução Nacional, PRN). Supported by the largest national TV
network and perceived by many voters as the only alternative to Luiz Inácio
Lula da Silva – the candidate of the leftist Worker’s Party (Partido dos
Trabalhadores, PT) – Collor obtained only 30 percent of the vote in the
first round of the presidential election but ultimately prevailed with 53
percent in the runoff (de Lima 1993).
The presidential election was a spark of hope for a country in trouble.
José Sarney left office haunted by an image of corruption and economic
mismanagement, and the new president inherited an economy on the brink
of hyperinflation – in March of 1990, the month Collor was sworn into
office, the consumer price index rose by 84 percent. In an attempt to dis-
tance himself from the discredited Brazilian elite and in order to show his
willingness to address the problems of the country, Collor formed a cabinet
of unknown politicians and immediately launched his “Collor Plan.”
The Collor Plan introduced a new currency and froze around 80 percent
of the country’s bank savings in order to drain liquidity (Bresser Pereira
1991, 18). But in spite of these extreme measures, government control
of inflation proved elusive. Unable to produce a sustained decline in the
consumer price index, the plan moved on to more orthodox policies in
15
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
May, giving way to the “Collor II” plan in January of 1991. Still lacking
visible success, the minister of the economy resigned five months later.
Chapter 5 will show to what extent the failure of Collor’s economic policy
bred popular dissatisfaction and weakened the position of the president in
the following months.
While the government struggled to control inflation, the first signs
of rent-seeking behavior began to surface. Scandals involving the Collor
administration (although not the president personally) began to emerge
in June of 1990 – just three months after the inauguration – when a
decree allowed the Ministry of Infrastructure to hire contractors for a large
highway maintenance project without public auction. Over the next two
years, the administration suffered media scandals at a rate of one every
two months. The accusations progressively involved high-ranking officials,
members of the cabinet, some of the president’s friends, the first lady – and
finally Collor himself.
On May 13, 1992, the president’s younger brother, Pedro, accused
Collor’s campaign manager, Paulo César (P. C.) Farias, of funneling cor-
ruption money into ghost offshore companies. Two weeks later, Pedro told
the news magazine Veja that Farias managed a vast corruption network for
the president (Collor de Mello 1993). Fernando Collor denied the accusa-
tions, and the police started to investigate Farias’s business. In early June,
sensing political opportunity, Congress created a bicameral committee to
examine the accusations. With just 8 percent of the seats in the Chamber
of Deputies and less than 4 percent of the seats in the Senate, Collor’s party
was unable to control the investigative process. Soon the committee dis-
covered that Farias had routed six and a half million dollars into Collor’s
(and his cronies’) bank accounts. In August, as public pressure against the
administration mounted, the cabinet issued a public statement virtually
withdrawing its support for the president.
By the end of August, the congressional committee had finished its
report. Hundreds of thousands of demonstrators in São Paulo, Rio de
Janeiro, Brası́lia, and other major cities marched calling for the resigna-
tion of the president. As mobilization escalated, various civic leaders for-
mally requested Collor’s impeachment. On September 29, the Chamber of
Deputies approved – by a vote of 441–38 – the impeachment by the Senate
and suspended Collor from office for a period of six months (Lins da Silva
1993b, 126). Three months later, the Senate voted 73–8 to oust Collor
and authorized his prosecution on twenty-two charges of corruption. The
president, anticipating the decision, presented his resignation.
16
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
come to reject the idea that the president could be above the law. With
the 1992 municipal elections approaching (and more than 15 percent of
the deputies running for mayor), massive popular demonstrations against
Collor convinced most legislators that the president should be removed
from power (Weyland 1993, 20–25).
Venezuela, 1993
The “Collorgate” affair served as a model for the removal of Venezuelan
President Carlos Andrés Pérez a few months later. But Pérez’s personal
trajectory and the institutional foundations of his administration differed
substantially from the factors identified as the sources of the crisis in Collor’s
Brazil. Pérez won his second term in 1988 with a clear majority (53 percent
of the vote). In contrast to Collor, he was the most seasoned leader of
the largest party in the country, Democratic Action (Acción Democrática,
AD). He had already been president of Venezuela in the 1970s, an age of
prosperity, and was a mythical figure throughout Latin America when he
was elected for the second time.
During Pérez’s first term in office (1974–79), the oil boom had boosted
living conditions in Venezuela and his administration had embraced
“national-popular” policies, nationalizing the production of iron and oil,
expanding the public sector, and denouncing multinational corporations
and international financial institutions. Pérez’s program for his second term,
however, was radically different. The new administration inherited a critical
situation from the Jaime Lusinchi government (1984–89), one that com-
bined trade deficits, generalized price controls, scarcity of basic products,
and unpaid international obligations.
Immediately after taking office in February of 1989, Pérez announced an
economic reform package known as the Great Turnaround (El Gran Viraje).
The program initially focused on price and interest rate liberalization and
triggered an abrupt rise in inflation, from 7 percent in December of 1988 to
23 percent in March of 1989 (Corrales 2002a, 47). As a result, the presiden-
tial “honeymoon” soon ended: on February 27, 1989, riots against increases
in public transportation fares and a scarcity of basic foodstuffs erupted in the
city of Caracas and spread throughout the country. The ensuing action of
the security forces caused more than 300 deaths (Kornblith 1998, Chapter 1;
López Maya 1999; López Maya 2005, Chapter 3). The Caracazo and the
human rights violations commited during those days left a deep wound in
the Venezuelan political system. In May, the Venezuelan Labor Federation
18
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
19
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
The Notables complained that the Pact for Reform had been ignored despite
the “increasing moral and material decay of the country.” The Perecistas, fol-
lowers of Carlos A. Pérez within the ruling party, lost the internal elections
of Democratic Action in September. The so-called orthodox faction, which
opposed the administration’s policies, retained control of the party’s direc-
torate (the powerful National Executive Committee, CEN). On October
7, two days after the election of the new CEN, AD’s president, Humberto
Celli, publicly demanded – the first sign of the battle to come – the resig-
nation of the cabinet members in charge of the economy.
This was just the beginning. On February 4, 1992, a group of young army
officers led by paratrooper Hugo Chávez Frı́as attempted a coup d’etat.
After the situation was controlled, former president (and respected founder
of the Copei Party) Rafael Caldera gave a speech in Congress condemning
the military action but blaming Pérez for his unpopular policies. In the
absence of a clear sign of rejection on the part of political elites, a second
coup attempt was carried out by navy and air force officers in November
of that year. The coup failed again, but the Pérez administration was now
under siege. The Notables openly campaigned in favor of the resignation of
the president, and although Pérez’s AD controlled 48 percent of the seats
in the lower chamber and in the Senate, partisan support for the president
began to vanish as political leaders explored alternative ways to force an
“institutional” exit of the chief executive (Rey 1993, 101–112; Rodrı́guez-
Valdés 1993). In July, as part of a legislative debate on constitutional reform,
the Chamber of Deputies contemplated the possibility of holding a recall
election in order to shorten the presidential term. The topic moved to the
Senate in late August, but the debate on constitutional reform was stalled
after the second coup attempt (Kornblith 1998, 89–97).
In November of 1992, nineteen days before the coup attempt, the press
began to publish stories about the unknown destiny of more than seven-
teen million dollars earmarked as “secret homeland security funds” for the
Ministry of the Interior. According to the reports, the money (250 mil-
lion bolı́vares) had been converted to dollars using the official Preferen-
tial Exchange Regime (Régimen de Cambio Diferencial, Recadi) and routed
from Interior to the president’s office. The Office of the President refused
to provide any information on the issue, arguing that national security
reasons prevented the disclosure, but the oversight committee (Comisión
de Contralorı́a) of the Chamber of Deputies created a special subcommittee
to investigate the trajectory of the funds (Chitty La Roche 1993). As the
scandal grew, the prosecutor general (Fiscal General ) asked the Supreme
20
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
Court to evaluate the merits of the case in order to prosecute the president.
On May 20, the Court ruled that the case merited further investigation,
giving anxious politicians the excuse they needed to act. The following day,
the Senate unanimously suspended Pérez from office on a temporary basis
and authorized the judiciary to prosecute the president. In late August, after
President Pérez had been on “temporary leave” for three months, a joint
session of Congress declared his leave “permanent,” removing Pérez from
office and appointing Ramón J. Velásquez to serve out the rest of the term
(Kada 2003a, 126–127).
Observers trying to make sense of the Venezuelan crisis at the time
focused on the collapse of the “petro-state” and the resulting decline in
living standards during the 1980s as the underlying forces driving the dis-
solution of the system of elite conciliation that had been in place since
1958. For them, the ousting of Carlos Andrés Pérez was an indication –
confirmed by later events – that the Punto Fijo regime was crumbling.3
Shortly after the events, Juan Carlos Rey argued that the impeachment had
resulted from a deeper “legitimacy crisis of the social and political order”
(Rey 1993, 72; see also Crisp, Levine, and Rey 1995). In Rey’s view, the
decline of the Venezuelan economy had led voters to support Pérez in the
1988 election because he represented the good old times of prosperity. But
the president had “betrayed” them by unilaterally imposing an economic
adjustment program. The political parties (AD in particular) had failed not
only to mobilize mass support for the president’s program, but also to pro-
vide a viable alternative, “abdicating” their leadership roles. Military rebels
had attempted to fill this vacuum in February and November of 1992, until
corruption charges gave Congress an excuse to oust Pérez from office.
One year later, Michael Coppedge (1994a) presented a similar picture
of economic decline, presidential isolation, and institutional failure. For
Coppedge, the ousting of Pérez indicated the erosion of the Venezuelan
governability formula. Since the 1960s, the regime had been structured
as a partyarchy, a model that assigned a central role to the AD and Copei
establishment and that operated through inclusive patterns of representa-
tion, strong party discipline, consensus policy making, and solid relation-
ships between parties and interest groups (Coppedge 1994b). By the 1980s,
3 The pact of Punto Fijo established the ground rules for the nascent Venezuelan democracy
in 1958 (including power-sharing agreements that lasted until the late 1960s). On the pact,
see Karl (1987). For a discussion of the rise and decline of the Venezuelan petro-state, see
Karl (1997).
21
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
however, this formula had started to crumble under the pressure of eco-
nomic decline, which eroded the capacity of parties to control civil society.
In a context of increasing corruption, ossified political structures and dated
political platforms had proven unable to meet new challenges. The saga
leading to Pérez’s impeachment ilustrated the inability of the Venezuelan
institutions to adapt while they desperatetly attempted to search for an
alternative governability formula (Coppedge 1994a, 51).
The crisis of the Venezuelan governability formula would last for at least
another decade. The removal of Carlos A. Pérez was followed by the elec-
tion of Rafael Caldera in 1993, by a late attempt to adopt neoliberal reforms
in 1996, by the election of Hugo Chávez in 1998, by a heroic attempt to
recreate the country under a “Bolivarian” constitution in 1999, and by a
long period of political turmoil leading to another failed military coup in
2002 (López Maya 2005, Chapters 9–12). In the meantime, however, the
hurricane of impeachments had veered toward neighboring Colombia.
Colombia, 1996
As Ramón J. Velásquez finished his interim term and handed the govern-
ment to the new Venezuelan president, Rafael Caldera, another administra-
tion was taking over in neighboring Colombia. President Ernesto Samper
had narrowly defeated the Conservative candidate, Andrés Pastrana, with
50.6 percent of the vote in the runoff election of June 19, 1994. In contrast
to Fernando Collor and Carlos A. Pérez, Samper had presented himself as
a counter-reformist, a social democratic leader of the Liberal Party who
was willing to reduce the pace of the economic reforms imposed by his
predecessor and co-partisan, César Gaviria.
In the week following the election, the press began to publish transcripts
of a telephone conversation in which one of the heads of the Cali drug cartel
suggested that they had contributed some 3.5 million dollars to Samper’s
campaign (Dugas 2001; Hinojosa and Pérez-Liñán 2003). In late June,
another tape linked Samper’s campaign treasurer, Santiago Medina, to the
Cali cartel. The origin of the tapes is unknown, although most people
believed at the time that the U.S. Drug Enforcement Administration (DEA)
was using them to put pressure on the Colombian government (on the
origins of the tapes, see Vargas, Lesmes, and Téllez 1996, 32–36).
In April of 1995, Prosecutor General Alfonso Valdivieso opened an inves-
tigation (known as “Proceso 8000” for its case number) on the infiltration of
narco-financing into the political campaigns. On July 26, Santiago Medina
22
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
Ecuador, 1997
A few weeks after Ernesto Samper was acquitted by the Colombian
Congress, a new president was elected south of the border in Ecuador.
In July of 1996, Abdalá Bucaram, leader of the Roldosista Party (Partido
24
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
4 “Roldosista” refers to the followers of former President Jaime Roldós Aguilera (Bucaram’s
brother-in-law), who died in a plane crash in 1981.
5 The campaign motto (“De un solo toque”) was a soccer metaphor implying that the can-
didate would score a goal with just one shot. On the campaign, see de la Torre (1996) and
Freidemberg (2003, Chapter 9).
25
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
was perceived by the Quito elite as too flamboyant, even rude – maybe
appropriate for the coast but not for the capital. The Quito press had reser-
vations about the new president from the outset and supported the admin-
istration only when major issues, like peace talks between Bucaram and
Peruvian President Alberto Fujimori, were at stake. After mid-November,
the press in Guayaquil – the PRE’s stronghold – also began to criticize the
government.
Criticism of the administration focused on three different themes. The
first one was the “undignified” behavior of the chief executive. For instance,
the president would shave his mustache on television in exchange for charity
contributions, call former president Rodrigo Borja a “donkey,” and mock
his adversaries and critics. The Roldosistas still contend that much of that
criticism was nothing but an elitist reaction against a president who felt
quite comfortable among, and acting like, the Ecuadorian masses. A second
theme was the administration’s political stubbornness. Early in the term,
Bucaram twisted some arms in Congress to obtain legislative approval for
his unpopular nominee for comptroller general (see Chapter 6). Later on,
he made a point of supporting ministers who were involved in scandals
and were being investigated by the legislature – the minister of education,
charged with plagiarizing her doctoral dissertation, and the minister of
energy, accused of violent outbursts and of harassing union members. Last
but not least, corruption became a major issue after Bucaram’s third month
in office. Public officials were said to collect a “party tax” from business peo-
ple, and the customs service was depicted as highly corrupt. Even though
investigative journalists had difficulties getting stories on the record –
the sources were afraid to confront the government – the administration
suffered, on average, a new scandal every two weeks. For the most part,
Bucaram dismissed the accusations as upper-class hypocrisy and assumed
that the scandals would not hurt his popularity.6
On the policy front, the administration was accused of producing a lot
of controversy with few tangible results. Bucaram made several controver-
sial proposals early in his term – the distribution of subsidized milk under
the brand “Abdalact,” the death penalty for anyone convicted of rape, and
the creation of an “Ethnicity Ministry” against the will of the indigenous
movement, among others. At the same time, however, he made an explicit
effort to show an orthodox profile in terms of economic policy. Facing an
economy in shambles and a country burdened by foreign debt (an estimated
26
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
45 percent of the budget was commited to servicing the debt in 1997), the
administration insisted on pegging the Ecuadorian sucre to the U.S. dol-
lar – a strategy modeled after the Argentine “convertibility” program of
1991. Indications of this new policy were clear from the outset, as Bucaram
hired the former Argentine minister of the economy Domingo Cavallo. The
blueprint of the program, however, was delayed for several months and only
presented to the public on December 1, 1996, when presidential approval
was already in decline (support was around 40 percent in Guayaquil and 20
percent in Quito) and opposition forces had prepared for the battle to come.
In response to the economic measures, which included an increase in the
costs of transportation and natural gas, some demonstrators began to protest
in early January. The Patriotic Front (Frente Patriótico), an alliance of trade
unions and social movements, called a general strike for February 5, 1997. In
the meantime, well-known politicians and former presidents like Rodrigo
Borja and Osvaldo Hurtado began to call for the resignation of the president
(Carrión 1997, 139). This climate of revolt significantly eroded the public
standing of the president – survey data collected by Informe Confidencial sug-
gest a fall of some 20 percentage points in both Quito and Guayaquil. With
an increasing number of civic organizations announcing that they would
join the strike, U.S. Ambassador Leslie Alexander denounced “rampant
corruption” in the customs service on January 29, 1997. The declaration
was read by many as a tacit support for the antigovernment movement.
Up to that point, the story of an unpopular president confronting mass
protests was hardly a reason for headlines in Ecuador. The demonstrations
of February 5, however, took both the government and the opposition by
surprise. The streets of Quito – and, to a lesser extent, other cities in the
country – displayed a colorful coalition of upper-middle-class protesters
upset with Bucaram’s style, indigenous demonstrators led by the Conaie
(Confederation of Indigenous Nationalities of Ecuador), and trade union-
ists opposing the neoliberal reforms. Their only point of agreement was
their contempt for the Bucaram administration.
To my knowledge, the mythical figure of two million protesters taking
to the streets on February 5 and 6 has never been verified. But it is true that
the massive attendance at the rally created both a challenge and an oppor-
tunity for Bucaram’s adversaries in Congress. The protest leaders issued a
“mandate” for Congress to remove the president from office. After some
negotiations, the legislature called for an urgent session the following day.
In a turbulent meeting, the opposition voted to declare Bucaram “men-
tally incapacitated” – a way of avoiding the supermajority required for a
27
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
much popular support (Luna Tamayo 1997; Acosta 1997). In this context,
the mobilization of early February emerged as a unique opportunity to
terminate the discredited government. Elites and masses supported the
demise of the administration, and both cheered as their legislators declared
the president insane.
Abdalá Bucaram would not return to Ecuador for another eight years,
and even then he would be allowed to stay in the country for less than a
month. By then, the removal of elected presidents had become a chronic dis-
ease of Ecuadorian politics. The president elected in 1998, Jamil Mahuad,
was ousted in early 2000 by the combined action of a military revolt led by
Col. Lucio Gutiérrez and a mass mobilization organized by the indigenous
movement. Lucio Gutiérrez in turn won the 2002 presidential election, only
to confront protests against his unconstitutional dismissal of the Supreme
Court – which allowed Bucaram’s short-lived return to the country – in
April of 2005. Challenged by popular mobilizations and a congressional
action to remove him from office, Gutiérrez abandoned the country in late
April (Pallares 2006). I will discuss the events that took place between 1998
and 2005 more systematically in Chapter 7. During that period, the wave
of impeachments shifted toward the southern cone.
Paraguay, 1999
The Paraguayan crises of March 1999, September 2001, and December
2002 cannot be discussed without reference to the atomization of the
Colorado Party (formally, the National Republican Association, ANR) that
followed the fall of dictator Alfredo Stroessner in 1989. Gen. Stroessner
took over in 1954 and ruled Paraguay for thirty-four years with the sup-
port of the ANR and the army. By the late 1980s, the question of suc-
cession was already in the air, and the attempt of the loyalist faction (the
“Militantes”) to gain control of the party and to place Stroessner’s son in
control of the armed forces ultimately ignited a military coup that trig-
gered the transition to democracy in 1989. After displacing the dictator,
Gen. Andrés Rodrı́guez charged a few politicians with corruption (Nickson
1997, 31–32), but Stroessner went into exile and his Militantes regrouped
to support Luis Marı́a Argaña, a traditional Colorado politician and former
chief justice of the Supreme Court.
Unable to seek reelection, Gen. Rodrı́gez left the country in 1993.
Businessman Juan Carlos Wasmosy became the new Colorado president
(1993–98), and Gen. Lino Oviedo, a close collaborator of Rodrı́gez in the
29
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
cavalry, became the strongman of the regime. The relations between Lino
Oviedo and President Wasmosy were friendly at first, but increasing ten-
sions between them led to a clash in April of 1996, when Wasmosy ordered
Oviedo’s retirement. The general initially resisted the decision, triggering
a military crisis that for the first time endangered the Paraguayan transition
to democracy. The Argañistas observed the showdown in silence and won
the internal election for party officials a few days later, when an infuriated
Oviedo ordered his followers to vote for Argaña’s faction (Ayala Bogarı́n
and Costa 1996, 121–122; interview with Angel Seifart, June 2002).
Removed from military command, Lino Oviedo devoted his energy to
Colorado politics. In the following months, while his lawyers maneuvered
to block a trial for military rebellion, the general organized his political
machine. He campaigned to win the Colorado nomination and chose Raúl
Cubas Grau, a businessman with little political experience, as his running
mate for the presidential primary in September of 1997. Afraid of Oviedo’s
popularity among the ANR rank-and-file, President Wasmosy convened a
military court to charge the general with sedition. Oviedo won the primary
against Luis M. Argaña only to be sentenced to ten years in prison; he was
arrested few weeks before the presidential race.
Since General Oviedo, now the official Colorado candidate, was barred
from running, legal provisions mandated that his vice president, Raúl Cubas
Grau, become the new presidential candidate and that Argaña, the runner-
up in the primary, be his running mate. Oviedo’s arrest thus united the
two Colorado factions on the same ticket. In this overall climate of tension
and legal chicanery, Cubas toured the country promising to pardon Oviedo
were he elected. As insurance against this peril, five weeks after the May
election the lame duck Congress passed a law restricting the president’s
pardon authority to convicts who had completed at least half of their prison
terms.
The odd Cubas–Argaña ticket won the 1998 election with 55 percent of
the vote (Turner 1999). Although the Colorado Party won 56 percent of
the seats in the lower chamber and 53 percent of the Senate, the president’s
faction did not control a majority in Congress. With the exit of Wasmosy,
the competition for power within the party now centered on the dispute
between the Oviedo–Cubas faction (named the National Union of Ethical
Colorados, Unace), controlling 34 percent of the seats in the lower chamber,
and Argaña’s Movement for Colorado Reconciliation (MRC), in control of
the vice presidency and 19 percent of the seats.
30
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
31
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
deadly. Seven demonstrators were killed and some seven hundred were
wounded when paramilitary snipers shot at the mob from a nearby build-
ing (Paredes 2001b, 156). Fearing an escalation of the conflict, members of
the government, the opposition, the church, and foreign diplomats nego-
tiated an “honorable solution” to the crisis. Anticipating a defeat in the
Senate trial, President Cubas resigned two days later and sought refuge at
the Brazilian embassy. Luis González Macchi, the speaker of the Senate
and a member of the Argañista faction, took office and formed a cabinet
of “national unity” with members of the Colorado Party (ANR), the Lib-
eral Party (Partido Liberal Radical Auténtico, PLRA), and the National
Encounter Party (Partido Encuentro Nacional, PEN). In the meantime,
Gen. Oviedo abandoned the country.
In an early analysis of the events, Diego Abente Brun (1999), a polit-
ical scientist and later a minister in the González Macchi administration,
interpreted the crisis of March as the culmination of the Paraguayan transi-
tion to democracy. According to Abente, the politics of the transition since
1989 had remained under the control of the Colorado elite and had been
marked by increasing factionalism within the party. Oviedo’s popularity had
grown because Paraguayan voters were tired of traditional politicians and
poor economic performance, but it had quickly faded in the face of Argaña’s
assassination. A “strong citizenry,” formerly unknown in Paraguayan pol-
itics, had unexpectedly taken to the streets to demand the impeachment
of the president, and the mass media had provided live coverage of the
events, igniting public outrage and encouraging mobilization. In this view,
the events of March had shown that public opinion was becoming an impor-
tant force in Paraguay, and that political outcomes would no longer be just
the product of factional disputes.
Paraguay, 2002
In spite of the hopes created by the “Paraguayan March” (as the tragic
events of 1999 eventually became known), the government of national
unity was nothing but a short-lived experiment in consensus democracy
(Pangrazio 2000, 324–325). The alliance formed by Colorados, Liberals,
and Encuentristas, which initially comprised about two-thirds of the seats
in the lower chamber, had few incentives to remain united. Although the
Supreme Court confirmed González Macchi in office until the end of
the Cubas term in 2003, the president’s lack of electoral legitimacy and
the effects of a five-year-long economic recession eroded his favorability
32
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
ratings from 68 percent in April of 1999 to 33 percent one year later (data
from Francisco Capli).
The erosion of the coalition was accelerated by a sequence of media
exposés beginning in November of 1999. Among them was the accusa-
tion, in March of 2001, that the armored car used by the president had
been stolen in Brazil and purchased by the Paraguayan government with
the approval of the finance minister. Two months later, the executive was
accused of transferring sixteen million dollars from the central bank to a
private account in New York.
With a national vice presidential election scheduled for August of 2000,
the coalition soon faced what David Altman has dubbed the “tyranny of the
electoral calendar” (Altman 2000). The PLRA demanded 40 percent of all
positions in the public administration and the nomination of a Liberal as
the coalition’s vice presidential candidate. When González Macchi refused
to meet the demands, the Liberals quickly withdrew from the coalition and
nominated their own candidate in early 2000.
In contrast to the PLRA, the middle-class National Encounter Party
(PEN) remained in the cabinet. The decision created some disappointment
in the ranks, particularly when PEN leaders backed the Colorado candidate
for the vice presidency. In September of 2000, a group of dissidents broke
with the party and formed a new organization called Paı́s Solidario (Solidary
Country). In the end, the PEN preserved seven seats in the lower chamber
and its splinter only two.
The electoral climate and the scandals combined with unpopular policies
to accelerate public unrest. Overwhelmed by the fiscal deficit, the adminis-
tration proposed privatizing telecommunications, water and sewerage, and
the national railroads. In March of 2000, thousands of peasants marched to
Asunción and joined forces with trade union workers against the proposal.
In late June, the national union federation (Central Nacional de Traba-
jadores, CNT) called for a national strike. In October, the CNT marched
to demand additional measures against unemployment.
The administration’s weakness became evident on May 18, 2000, when
cavalry officers linked to Gen. Oviedo attempted a coup d’état. The rebels
surrounded the Congress building and occupied military barracks, the
police headquarters, and a few TV and radio stations. The navy and air
force refused to join them, and eventually the officers surrendered. But
according to one analyst, “it became clear that the citizens were unwilling
to raise one finger to sustain González Macchi in power” (Paredes 2001a,
178).
33
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
Macchi signed the bill immediately, averting an impeachment for the sec-
ond time.
In October, Julio C. Franco resigned as vice president in order to run as
a presidential candidate in the elections of April 2003. Although Congress
was unable to agree on the appointment of a surrogate, Franco’s resignation
freed the dissident Colorados from the problem of voting a Liberal into
office. Aware of this opportunity, the opposition tried to unseat the president
for the third time. On December 5, while President González Macchi was
attending an international meeting in Brazil, a Liberal deputy unexpectedly
asked to bring the impeachment charges (this time based on the case of
the sixteen million dollars, the stolen car, and the president’s support for
corrupt officials) to the floor of the House. The disoriented Argañistas
found themselves unable to block the debate and unwilling to back the
president in the midst of a presidential campaign. The impeachment was
approved with fifty-two votes by the PLRA, Unace, the PEN, Paı́s Solidario,
and some dissident Colorados against only eleven abstentions.
In mid-January, an electoral court banned the nomination of Gen.
Oviedo for the April election, averting the possibility of a split in the
Colorado electorate. Thus, when the Senate began the hearings in late
January, President González Macchi ironically stated that the trial would
be a “nice show.” On February 11, after a ten-hour session, twenty-five
senators voted in favor of removing the president from office, eighteen
voted against, and one abstained. Unable to achieve a two-thirds majority,
the trial failed. On April 27, the Colorado candidate, Nicanor Duarte, won
the presidential race with 37 percent of the vote, and President González
Macchi transferred power to his co-partisan four months later.
in Brazil; the collapse of the petro-state was the source of the problem in
Venezuela; the populist tradition was the underlying issue in Ecuador; the
success of a reformist conspiracy was at stake in Colombia; and the atom-
ization of the Colorado Party triggered repeated crises in Paraguay. These
interpretations are in fact historically accurate, but they fail to explain why
congressional action became a new standard procedure to oust presidents
throughout the region.
A few commonalties shared by the cases shed light on this question. First
and foremost, the use of congressional procedures to remove presidents
from office indicated the unwillingness (or the incapacity) of the military
to do the job. In some cases, as in Brazil and Colombia, civilian politicians
refused to involve the military in the crisis. In others, as in Ecuador in 1997,
the military refused to intervene except as mediators. In yet others, such
as Venezuela in 1992 and Paraguay in 2000, the military attempted to take
over but simply failed. In an international context that discouraged military
intervention in politics, political elites were forced to find constitutional
(or at least pseudo-constitutional) ways to solve their disputes. It is hard
to overstate how much this factor has altered the dynamics of executive-
legislative crises in recent years. The magnitude of this change will be
documented in the following chapter.
The absence of military coups alone is not enough to explain why some
presidents confronted the threat of impeachment and others did not – or
why some presidents survived this threat while others were removed from
office. The succinct case studies presented here suggest that a simple nar-
rative of corruption and punishment usually does not suffice to account for
an impeachment crisis. Beyond the merits of the legal case against the pres-
ident, a complex configuration of social, partisan, and institutional forces
has always determined the executive’s fate.
First, there is the role of the press. In a context of increasing democrati-
zation in the 1990s, the mass media became more willing to unveil scandals
about the president, the first family, and top public officials (Waisbord
2000). The exposés provided the ammunition that opposition legislators
needed to initiate impeachment charges. Second, there is the president’s
ability to bring together a legislative coalition against impeachment. Pres-
idents with the skills to build a strong legislative shield – Samper being the
most visible example – were more likely to prevent the collapse of their
administrations when scandals enveloped them (Hinojosa and Pérez-Liñán
2003). Last but not least, there is popular protest. Increasing freedom of
expression and organization created the possibility of mass demonstrations
36
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
7 “Selection bias” is used with slightly different meanings in econometrics (e.g., in the discus-
sion of Heckman selection models) and in the study of qualitative methods (Lustick 1996).
For simplicity, I keep the discussion restricted to the standard problem of truncation in the
dependent variable.
37
P1: JZP
0521869423c02
38
cuny809-perez
Samper (Colombia, 1996) No Yes Large party and coalition Narrow Avoided trial
Bucaram (Ecuador, 1997) No Yes Small party Broad Declared mentally unfit
Cubas Grau (Paraguay, 1999) No Yes Large party was divided Broad Impeachment (resigned)
15:50
González Macchi (Paraguay, 2002) Failed (2000) Yes Large party was divided Intermediate Impeachment (survived)
P1: JZP
0521869423c02 cuny809-perez 0 521 86942 0 May 8, 2007 15:50
39
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
40
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
41
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
19
18
17
16
15
14
Number of Countries
13
12
11
10
9
8
7
6
5
4
3
2
1
0
1950 1960 1970 1980 1990 2000
Year
and members of the legislature were elected in free and fair elections;
(2) voting rights were granted to a vast majority of the adult population; (3)
civil liberties were respected; and (4) the military did not interfere in civilian
affairs. If any of the four conditions was clearly absent during a particular
year, the country was coded as nondemocratic (or authoritarian). If any
condition was “partially” violated (for instance, if there were accusations of
electoral fraud in certain regions of the country, but they were not thought
to alter the overall result of elections), the country was coded as semidemo-
cratic. For simplicity, I will refer to the set of democratic and semidemo-
cratic countries as “competitive regimes” (Mainwaring and Pérez-Liñán
2005).
Figure 3.1 shows the unprecedented level of democratization achieved
by the region in the 1990s. By 1995, eighteen countries (Argentina, Bolivia,
Brazil, Chile, Colombia, Costa Rica, the Dominican Republic, Ecuador, El
Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay,
Peru, Uruguay, and Venezuela) could be considered competitive regimes,
42
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
and only one (Cuba) was fully authoritarian.1 Although the ultimate reasons
for this historical transformation lie beyond the scope of this study, changes
in the international context, as well as the learning process triggered among
civilian and military elites by the dictatorships of the 1970s, help explain
this evolution.
The transformation of Latin America during the 1980s resulted not just
from the demise of several authoritarian regimes within a few years. It was
also the product of a declining capacity (or willingness) of military officers
to intervene in politics over the long run (Fitch 1998; Mainwaring and
Pérez-Liñán 2005). Once established, the new democratic systems were
less vulnerable to military conspiracies than their predecessors.
Table 3.1 documents this trend by comparing the incidence of military
coups in Latin America over five and a half decades (1950–2004). The unit
of analysis in the table is countries during particular years (e.g., Argentina
in 1950). The right-hand panel shows the incidence of coups among com-
petitive regimes, while the left-hand panel includes all cases (even authori-
tarian regimes, with or without legislatures). Each panel displays the total
number of cases, followed by the percentage of cases that experienced a
43
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
military rebellion (defined as any military action directed against the pres-
ident or Congress) and by the percentage of successful coups (instances in
which the military was able to force the exit of the president or the closure
of the legislature).2
The return of military officers to the barracks in the 1980s is reflected in
the decline in the number of insurrections. In the 1960s, 11 percent of all
competitive regimes experienced a successful military coup, and 19 percent
of them confronted some form of military rebellion. In the 1990s, the rates
were 1 percent and 4 percent, respectively. For the historian concerned with
the long durée, the last decade and a half may look like a brief flash of civilian
stability in a long history of political turmoil. But for politicians operating
in the 1990s, the new context represented a significant break with the past.
How did this change affect the balance of power between the executive and
legislative branches throughout the region? And more importantly, how
did it alter the outcome of extreme interbranch confrontations?
Presidential Crises
In most presidential countries, executives and assemblies regularly confront
each other on policy issues, bargain with each other, defy each other, and
eventually agree on a common policy (or not). This is normal politics,
and certainly not the focus of this book. I will focus on a particular type
of situation labeled “presidential crisis” (Pérez Liñán 2003b). Presidential
crises are episodes characterized by extreme levels of conflict and by the
decision of one elected branch to dissolve the other in order to reshape
its composition. This stance, which would be normal in a parliamentary
system, unleashes the threat of constitutional breakdown in a presidential
regime.
The operational definition of “presidential crisis” employed in this book
includes any episode in which the chief executive threatens to dissolve
Congress or supports a constitutional reform having that purpose, attempts
a military coup against Congress, or “suspends” the term of the legislature
(even if no decree proclaims its “dissolution”) until the next election. It also
2 Military interventions (particularly failed military coups) are not always easy to pinpoint.
The events coded in Table 3.1 were identified using The New York Times Index, Latin American
Weekly Report, and other historical sources (Fossum 1967; Needler 1966). I am indebted to
Annabella España-Nájera and Scott Mainwaring for collecting and sharing important data
on coups. Datasets and computer codes to generate the components of the table are available
at http://www.pitt.edu/∼asp27/Presidential/Impeachment.html.
44
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
3 Note that this definition encompasses both situations in which constitutional mechanisms
are activated and others in which nonconstitutional actions are unleashed. It is the willingness
of one elected branch to reshape the composition of the other, not the outcome of the process,
that defines a presidential crisis. This definition allows us to compare recent confrontations
ending with impeachments to past crises leading to military coups.
4 The qualitative database is available at http://www.pitt.edu/∼asp27/Presidential/
Impeachment.html.
45
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
Competitive Regimes
Democratic Semidemocratic Authoritarian Regimes
Argentina (1989) Argentina (1962) Brazil (1966)
Argentina (2001, twice) Argentina (1976) Brazil (1968)
Bolivia (1983) Bolivia (1979) Brazil (1977)
Bolivia (1985) Brazil (1955, twice) Guatemala (1957)
Bolivia (1990) Colombia (1991) Honduras (1954)
Bolivia (2003) Colombia (1996) Panama (1951)
Brazil (1954) Ecuador (1963) Panama (1955)
Brazil (1964) Ecuador (1970) Panama (1988)
Brazil (1992) El Salvador (1987) Paraguay (1954)
Chile (1954) Guatemala (1993) Paraguay (1959)
Chile (1973) Guatemala (1994)
Colombia (1977) Honduras (1985)
Dominican Republic (1994) Nicaragua (1992)
Ecuador (1961) Panama (1968)
Ecuador (1984) Paraguay (1999)
Ecuador (1987) Paraguay (2001)
Ecuador (1990) Paraguay (2002)
Ecuador (1997) Peru (1991)
Ecuador (2000) Peru (1992)
Ecuador (2004) Peru (2000)
Nicaragua (2004)
Uruguay (1969)
Uruguay (1971)
Uruguay (1973)
Venezuela (1993)
Venezuela (1999)
N 27 21 10
Source: Database on presidential crises, based on Keesing’s Contemporary Archives (1950–86),
Keesing’s Record of World Events (1987–2000) (London: Longman), and country-specific sources.
46
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
6 I use the term to refer to episodes of early reequilibration. “Early reequilibration” refers
to the process by which breakdown is prevented without a major institutional disruption.
By contrast, late reequilibration refers to situations in which an interruption of the existing
regime occurs – as in the transition from the Fourth to the Fifth Republic in France – but
where changes in the political order ultimately take place within the democratic framework.
Linz (1978, 90) was aware of this distinction but never developed the idea.
47
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
48
P1: JZP
0521869423c03
(23)
IV. Legislative Coup Brazil, 1955
Spillover
(26) Short-term
V. Self-Coup Peru, 1992
Presidential (17)
Crises Reequilibration
VI. Short-Term Dictatorship Ecuador, 1963
May 8, 2007
(N = 58) (3)
VII. Removal of the President Panama, 1955
No Spillover Crises without
16:12
(32) Breakdown
VIII. Dissolution of Congress Venezuela, 1999
(35)
IX. Stabilization with Mutual Survival Paraguay, 2002
49
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
Legislative Coups Despite their name, legislative coups against the pres-
ident are not always conducted by the legislators themselves. In most his-
torical circumstances, the members of Congress have simply offered con-
gressional support for a military conspiracy. This leads to an important
distinction between proactive legislators, those who initiate and control the
confrontation with the president, and reactive legislators, those who jump
on the bandwagon of a confrontation driven by the military or by other
social actors. As I will show later, this distinction is relevant for praetorian
as well as for institutional outcomes, for past as well as for contemporary
presidential crises.
The Brazilian coups of 1955 illustrate this point. After the suicide
of Brazilian President Getúlio Vargas in August of 1954,10 Juscelino
50
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
51
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
12 In the end, Velasco’s disruption of the regime led to a long-term breakdown, because the
military later took over in order to prevent Assad Bucaram from winning the presidential
election.
13 The National Council of Administration was a nine-member independent branch of the
executive in charge of domestic policy implementation. Authors concur in pointing out
that it operated in practice as a “third chamber.” On this topic, see Fizgibbon (1952).
14 Although the 1940 constitution allowed the Paraguayan president to dissolve Congress, the
repression and exile of party dissidents suggests that this was more than a mere constitutional
procedure.
52
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
15 The ARENA party (pro-military) had been formed in late 1965 and already controlled 68
percent of the Brazilian Lower Chamber.
53
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
54
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
Long-Term Breakdown
In contemporary Latin America, the confrontations between the execu-
tive and Congress sometimes occurred in a context of broader social and
political turmoil that ultimately led to the demise of the presidential regime
(Bermeo 2003). The best example of this pattern is the establishment of what
Guillermo O’Donnell has called Bureaucratic-Authoritarian (BA) regimes
after the Brazilian coup of 1964, the Chilean and Uruguayan coups of 1973,
and the Argentine coup in 1976 (Collier 1979; O’Donnell 1988).16 Long-
term breakdowns have produced three main outcomes: the ousting of the
president and the establishment of a subservient assembly, the elimination
of Congress and the installation of a puppet president, and the imposition
of a military junta with a long time horizon.
16 The Argentine coup of 1966 also inaugurated a BA regime, but it was not preceded by a
presidential crisis.
55
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
greater autonomy from civilian authorities, and in early 1973 they openly
opposed the appointment of a new minister of defense. President Juan
M. Bordaberry accepted greater military participation in government
affairs through a National Security Council (Lerin and Torres 1987, 9–20;
Weinstein 1988, 44–45). Congressional leaders from almost every quar-
ter refused to back the government on this issue, expecting that President
Bordaberry would be forced to resign (González and Gillespie 1994, 163).
Bordaberry, however, turned to the military for help, and supported the
officers when they demanded that Congress lift the immunity of a senator
presumably linked to the Tupamaro guerrillas (Bermeo 2003, 128–130).
The legislators not only refused to comply, but also warned the presi-
dent that he could be impeached if the senator was arrested (Kaufman
1979, 114). Between April and June, the executive coalition in Congress
collapsed. On June 27, the president dissolved the National Assembly and
appointed a Council of State composed of civilians and military officers. The
armed forces took control of state companies and the central bank, allowing
Bordaberry to remain in office until 1976, when he was replaced by Presi-
dent Aparicio Méndez. Military rule lasted until 1984.
151), but the coup opened the way for seventeen years of military rule under
the aegis of Gen. Augusto Pinochet (Bermeo 2003, Chapter 5).
The examples just presented suggest that long-term breakdowns solve
presidential crises in a Hobbesian way, by expropriating power from the two
conflicting branches in favor of a third player – the army or an individual
dictator. Even when the president is kept in office, as in 1973 in Uruguay,
or when Congress is not shut down, as in 1964 in Brazil, those institutions
are progressively deprived of power because they have to perform new,
diminished functions under the logic of bureaucratic authoritarianism.
Although bureaucratic authoritarianism represented the main pattern of
long-term breakdowns during the period under study, presidential crises
have opened the way for other forms of nondemocratic rule, such as state
corporatism or neo-patrimonial rule. In 1937, President Getúlio Vargas of
Brazil dissolved Congress to impose his Estado Nôvo, a corporatist regime.
In 1954, the Paraguayan assembly accepted the resignation of President
Federico Cháves when a military coup forced him out of office, opening
the way for the ensuing election of Gen. Alfredo Stroessner as president
(Seiferheld 1987). By 1954, the Paraguayan regime was hardly democratic
(opposition to the Colorado Party had been banned since 1947), but it
lacked the strong elements of personalism that Stroessner imposed on the
new regime.
17 In his study of the 1962 military coup in Argentina, Kvaternik identified four possible
outcomes of a democractic crisis: breakdown, reequilibration, failed reequilibration, and
no breakdown. The first two categories reflected Linz’s thinking on the issue, while the
other two were Kvaternik’s own contributions. Failed reequilibration was at the core of
Kvaternik’s study; he related this type of failure to the moderating military interventions
described by Stepan (1971) in the Brazilian context and to the 1962 coup in Argentina. The
fourth category, crisis without breakdown, was not systematically explored.
57
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
the president from office, the legal dissolution of Congress, or some form
of institutional stabilization.
58
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
presidents have made use of self-coups in the past and have promoted con-
stitutional reforms in recent years. The Colombian Constitutional Assem-
bly dissolved Congress in 1991, and a constitutional referendum allowed
for the dissolution of the Guatemalan Assembly and the Supreme Court in
1994. President Hugo Chávez of Venezuela employed a loyal constitutional
assembly to disband an adversarial Congress in 1999 (Crisp 2000, 230–234).
In all cases, a new election followed, and Congress was reshuffled in a way
that resembled parliamentarism.
Peru, Uruguay, and Venezuela have explicit dissolution clauses in their
constitutions, but presidents have been unable to make them work to their
advantage. The Peruvian constitution of 1979 allowed the president to
dissolve the Chamber of Deputies if the latter censured three or more
cabinets (articles 227–229). Although the 1993 constitution reduced the
requirement to the censure of two cabinets, a congressional dissolution
has never been enforced. Similarly, the Venezuelan constitution of 1999
(articles 236 and 240) empowered the executive to dissolve the unicameral
assembly if the latter censured the vice president three times before the last
year of the legislative term (Crisp 2000, 233). In Uruguay, the 1934 charter
(article 141) established the president’s power to dissolve Congress if a
minister was censured by a simple majority of the legislative votes and two-
thirds of the members failed to support the censure in a second vote. With
minor modifications – the threshold for the second vote was reduced to
three-fifths – this rule was preserved in the 1942 and 1966 constitutions. In
1969, Uruguayan President Jorge Pacheco threatened to dissolve Congress
when his minister of industry and commerce was censured by the legislature
(Shugart and Carey 1992, 115). The conflict was decided when a faction of
Pacheco’s party defected to the opposition in the second vote. Support for
the censure reached three-fifths of the votes, and the president was unable
to invoke the dissolution clause (González and Gillespie 1994, 158–162).
20 The idea of stabilization with survival should be distinguished from the Linzean concept
of reequilibration, discussed earlier. Stabilization refers only to the preservation of current
government officials, while reequilibration refers to the preservation of the regime as a
whole.
59
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
(e.g., Presidents Andrew Johnson and Bill Clinton in the United States, or
Luis González Macchi in Paraguay). In some cases, stabilization involves an
additional component of regime reequilibration. For instance, to prevent
Arnulfo Arias from winning the presidential election, in 1948 the Panama-
nian legislature passed a resolution “recovering” its status as Constitutional
Assembly (a role exercised in 1945), dismissed President Enrique Jiménez,
appointed the comptroller general as president for four years, and nullified
the recent election. But the Supreme Court rejected the constitutionality of
the resolution, and the national police backed President Jiménez. The pres-
ident ignored the decision, the legislative coup folded, and the president
and the legislators completed their terms (Pippin 1964, 20–28; Pizzurno
Gelós and Aráuz 1996, 341–344).21
21 Arias did not take office until 1951, however. The election results were rigged, and the
Liberal Doctrinario candidate, Domingo Dı́az, was proclaimed the victor.
60
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
61
P1: JZP
0521869423c03 cuny809-perez 0 521 86942 0 May 8, 2007 16:12
Table 3.4. Impact of presidential crises before and after the wave of democratization
63
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
65
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
Table 4.1. Entries under “Bribery and Corruption,” FBIS Index, 1983 and 1993
Year
Country 1983 1993 Difference
Argentina 2 5 3
Bolivia 1 14 13
Brazil 0 69 69
Chile 1 2 1
Colombia 1 16 15
Costa Rica 0 0 0
Cuba 0 1 1
Dominican Republic 0 1 1
Ecuador 0 1 1
El Salvador 0 0 0
Guatemala 1 6 5
Honduras 0 1 1
Mexico 2 3 1
Nicaragua 2 8 6
Panama 1 13 12
Paraguay 0 13 13
Peru 0 1 1
Uruguay 0 0 0
Venezuela 0 46 46
Average 0.6 10.5 +9.9
Total news items on corruption 11 200 +189
FBIS stories on Latin America 24,612 16,310
Items as percentage of all stories 0.04 1.23
Source: Index to FBIS Daily Reports (computer file) (New Canaan, CT: NewsBank/Readex).
media accusations had become more frequent.1 Table 4.1 reflects to some
extent the impact of the impeachment processes in Brazil and Venezuela –
which made corruption a dominant topic during 1993 – but it also suggests
that revelations were on the rise in most countries.
FBIS news stories under “Bribery and Corruption” reflect exposés that
do not necessarily involve the president and his or her entourage. However,
scandals are relevant for the unfolding of an impeachment process when
1 The increase in the number of entries on bribery and corruption is not the result of an
increase in the sheer number of news stories reported by FBIS. As shown in Table 4.1, the
total number of stories on Latin America was smaller in 1993 than in 1983. Items on bribery
and corruption represented only 0.08 percent of all news items on Latin America and the
Caribbean indexed between 1980 and 1985, and 1.31 percent of those indexed between 1990
and 1995, a general pattern consistent with the two years selected as benchmarks.
66
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
30
25
Number of Scandals
20
15
10
0
1980 1985 1990 1995 2000 2005
Year
Figure 4.1 Political scandals, 1980–2004. Source: Latin America Weekly Report
(LAWR).
Even if corruption were on the rise, however, this fact would not be
enough to explain the surge in scandals. To a large extent, scandals may
take place without corruption, and vice versa (Jiménez Sánchez 1994, 14).
Given the definition presented here, in order for a scandal to take place,
the mass media must uncover corrupt behavior. Therefore, the study of
scandals must explain why malfeasance is exposed rather than the causes of
corruption per se (Waisbord 1994, 21).
In fact, social scientists and media analysts have pointed out that the
Latin American press has become increasingly willing to dig into dishonest
politics. In the mid-1990s, a report by the Freedom Forum Media Studies
Center noted “[w]hat has changed is the media’s determination to uncover
malfeasance and to hold government officials accountable to the public.
The combat between the press and the political powers is intensifying in
Latin America” (Vanden Heuvel and Dennis 1995, 14). Alves (1997) simi-
larly documented the emergence of a “vanguard press” – a new generation
of Latin American newspapers providing more aggressive coverage and
displaying greater professionalism and independence.
If this is the key to understanding the rise of scandals, how can we explain
this major transformation in the press? Over the last three decades, most
Latin American countries have undergone four major changes that have
enhanced the capacity of the media to publicize political wrongdoing and
created incentives for politicians to disclose it. Political democratization has
meant more freedom of the press throughout the region. Economic reforms
have deprived governments of regulatory tools traditionally used to control
the press and allowed for the emergence of a new and more powerful breed
of media corporation. The growth of television has made news available to
larger segments of the population and put pressure on newspapers to run
investigative stories. Last but not least, over the last few decades journalists
have become, in the words of a Brazilian journalist, “more professional
and less emotional.”2 This trend has fostered new professional values and
expectations regarding the role of the press in the new Latin American
democracies (Waisbord 1996).
The new state of affairs in which the press has greater opportunities to
investigate scandals has not been without pitfalls. Latin American journal-
ists have many times pursued scandals for their own sake; the effects of
deregulation on media markets are contested; the commitment of TV net-
works to quality news is, at best, doubtful; and investigative journalism has
68
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
often been confused with publishing unfounded accusations. Like the other
major institutions in Latin America, the press still has important lessons to
learn.
Robert Buckman pointed out, “with the coup all communication media ini-
tially came under military censorship. Only four newspapers still appeared
in the capital: the three Edwards newspapers – El Mercurio, La Segunda, and
Ultimas Noticias – and Picó’s La Tercera. All Marxist-oriented media were
closed permanently, and many of their editors and writers were among
those jailed, exiled, or disappeared” (Buckman 1996, 169). Overt censor-
ship was soon replaced by self-censorship and by more selective forms of
pressure, but the press remained largely controlled. During the 1980s,
a slow process of liberalization allowed for the emergence of moderate
opposition dailies like La Epoca and Fortı́n Mapocho; later in the decade,
the mainstream press began to challenge the official spin on critical issues.
Television news excluded the views of the opposition during the Pinochet
regime, until the Concertación was granted limited airtime for the 1988
plebiscite (Hirmas 1993). With the return to democracy in 1990, the media
gained greater independence (Sunkel 1997), although the Chilean main-
stream press remains on average less aggressive than its Argentine and
Brazilian counterparts, and the Chilean constitution inexplicably holds cit-
izens liable “for crimes and abuses that may be committed in the exercise
of [press] freedom” (article 19, section 12).
The protection of free expression and other civil rights is a necessary
(but not sufficient) condition for the proliferation of political muckraking
( Jiménez Sánchez 1994, 19–20). Although in a few countries, like Costa Rica
and Venezuela, this precondition was already present by the late 1970s,
in the rest of Latin America the third wave of democratization changed
the conditions under which journalism was practiced. What military rulers
could formerly hide from the public is nowadays easily exposed by the media
(Weyland 1998a, 110).
The process of political liberalization fostered the creation of “vanguard”
newspapers like La Prensa in Panama (founded in 1980, closed in 1988, and
reopened in 1990), El Norte or Reforma (1993) in Mexico, Siglo Veintiuno
(1990) in Guatemala, Página/12 (1987) in Argentina, and the now-extinct
La Epoca (1987) in Chile. By developing a more aggressive and professional
brand of journalism, those newspapers changed the rules of the game in their
countries and forced the mainstream newspapers (El Universal in Mexico,
Prensa Libre in Guatemala, and Cları́n in Argentina) to improve their own
coverage (Alves 1997; Hughes 2006; Waisbord 1998, 54).
Free speech also allowed the press to play political roles that other institu-
tions had left unattended. In democracies where reciprocal controls among
political institutions are weak – where, in O’Donnell’s terms, there is a lack
70
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
market, media corporations need to secure their share of official perks. The
power to unleash scandal – the power to put investigative stories on hold
or to defect in favor of full disclosure if agreements between media owners
and public officials are not honored – is one of the main strengths that
media corporations bring to their negotiations with governments. For rea-
sons of credibility (in the game-theoretic sense of players being capable of
enforcing their threats), the practice of investigative journalism may be the
best strategy for media corporations willing to negotiate with government
officials.
75
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
400
350
300
250
TV Sets PTI
200
150
100
50
0
1975 1980 1985 1990 1995 2000
Year
Figure 4.2 Television sets per thousand inhabitants, 1975–2001. Source: Interna-
tional Telecommunication Union – World Development Indicators (World Bank
2005). Average based on nineteen Latin American countries.
With satellite, with microwaves, with a more agile television, now the evening news-
casts set the agenda for next-day newspapers. . . . It is of course true that the print
press also sets the agenda for us many times – sometimes it breaks news, stories that
we have not offered the night before. But on average, television shows the night
before, with color and movement, what newspapers will offer the following day.
76
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
Thus, the role of newspapers had to be recycled; they had to offer something else to
their readers, something different from whatever newscasts had offered the night
before. This was a contest in which the winner was the reader, the viewer. . . . After
freedom returned with the end of the dictatorship, that something else in Brazil hap-
pened to be scandal, the frantic chase of thieves. . . . But after a while, that stopped.
Nowadays that something else that newspapers provide is better information with
greater analysis. (interview with Carlos Chagas, August 7, 1998)
This process was less marked in other countries such as Argentina and
Venezuela, where newspapers and morning radio still set most of the jour-
nalistic agenda (Zuleta Puceiro 1993, 74), but in every country newspaper
editors felt pressures to modernize and to provide “something else” in order
to give their products an edge.
Second, broadcast radio and television allowed larger segments of the
population access to daily news, particularly in countries where illiteracy had
precluded massive consumption of print journalism (Pérez-Liñán 2002a).
In Brazil, for instance, 18.3 percent of the population over the age of four-
teen was illiterate in 1990 (ECLAC 1997, 41). In practice, this meant that
one-third of the electorate was illiterate and that two-thirds had not fin-
ished grade school (Folha de São Paulo, August 4, 1998, I-2). UNESCO
estimated that in 1990 illiteracy affected 45 percent of the population age
fifteen and over in Guatemala, 27 percent in El Salvador and Honduras, and
21 percent in Bolivia. Figures were lower but still alarming in Mexico
(13 percent) and Venezuela (10 percent), and fell under 4 percent in coun-
tries like Argentina and Uruguay (see ECLAC 1997, 41). These figures
suggest that the expansion of the broadcast media was more critical for the
diffusion of news in some countries than in others.
Because of its entertaining power, television has become a widespread
medium and a focus of attention for many people.4 Miller and Darling,
for instance, noted that in the late 1990s Mexico had three times more
television sets than telephones (1997, 61). In countries with low newspaper
readership, television has a unique power to multiply the political impact
of scandals, because accusations and investigative stories that would have
limited diffusion through the pages of newspapers may reach every home in
the country once they are exposed on national television. When newscast
producers are interested in scandals – and I have argued that in a con-
text of democratization and growing market competition they increasingly
4 In Venezuela, evening news became late night news, with newscasts placed in the 11 P.M.
slot. In the 1990s, prime time was mostly devoted to soap operas and news magazines.
77
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
5 Interview, July of 1998. Grooscors was Director Nacional de Información under Betancourt,
director of the Oficina Central de Información under Leoni, and founding minister of the
OCI (Oficina Central de Información) in the first Pérez administration. He later pursued a
diplomatic career.
78
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
35.0
30.0
Students per Thousand Enrolled
25.0
20.0
15.0
10.0
5.0
0.0
1970 1975 1980 1985 1990 1995
Year
Brazil Venezuela
79
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
The number of students in social communication has grown faster than the
total number of students in higher education, indicating the rising social
status of the profession.
This trend is not restricted to Brazil and Venezuela. Table 4.2 compares
the evolution of communication careers in seventeen Latin American coun-
tries between the 1980s and the 1990s. Although UNESCO aggregated into
a single category students in social communication and in documentation
majors, most of these students were in journalism programs. The figures do
not allow a clear distinction between graduate and undergraduate programs,
but most students were enrolled at the undergraduate level. Among the
twelve countries with comparable data, eleven showed absolute increases
in the number of students in communication programs over the decade,
and nine displayed significant increases relative to the total enrollment.
Only three countries appeared to suffer some relative decline. One of them
was Venezuela, but the decline was not significant (p = .26, one-tailed
80
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
z-test), and the series from the Consejo Nacional de Universidades dis-
played earlier suggest a sharp increase until the mideighties, then some
decline, followed by a recovery in the 1990s. The only two countries with a
decline in journalism careers were Peru and, most visibly, Chile. Although
the UNESCO figures should be read with some caution, they suggest
that there is a trend toward a more professional press corps in most countries
throughout the region.
This rising number of journalism majors was exposed to teachers for
whom the Watergate scandal in the United States had been a distant but
crucial experience in their own youth. According to the Brazilian journalist
Ricardo Noblat, “After Watergate, the dream of every journalist somehow
became: I will bring my president down if I am capable enough.” Noblat recalled
that during the Collor affair, the image of Watergate was “hanging as a
ghost” over Brazilian journalists, and that many of his colleagues felt proud
when the president was ousted (interview, August of 1998).
Many teachers of this generation had also been trained in the idea that
journalism should confront power. Waisbord (1996, 346) has described a
Latin American tradition that deemed “journalists as partisans equipped
with typewriters.” Resistance against censorship and coercion was a forma-
tive experience for this generation. In countries like Argentina, Brazil, and
Chile, many journalists had fought against – and suffered the repression
of – military regimes in the 1970s. In Venezuela, where civilian govern-
ments ruled, the press corps had viewed with horror a long advertising
boycott (1961–63) against El Nacional, when American and local corpora-
tions condemned the newspaper for being too sympathetic to Fidel Castro
and forced publisher Miguel Otero Silva to leave his own newspaper (Dı́az
Rangel 1994, 96–99; interview with Gloria Cuenca, July 21, 1998). The mil-
itant tradition among Latin American journalists was passed to the younger
generations in the form of a “muckraking ethos” (Waisbord 1994, 29).
Miguel H. Otero, publisher of El Nacional, pointed out that “thirty years
ago, our journalists were on the left, so they had a philosophy [of contest-
ing power]. Journalists nowadays are trained in that culture at journalism
schools” (interview, July of 1998).
In this new context, journalists viewed the exposure of politicians’ mis-
deeds as part of their professional role. Investigation became a professional
value and the disclosure of scandals the path to a successful career. This pro-
fessional culture brought greater incentives for journalists to investigate the
president and his or her collaborators, strengthening the watchdog role of
the press but also creating important institutional problems.
81
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
The [Brazilian] press of the late eighties was too irritated, too exaggerated. That
[behavior] was natural after twenty-one years of military rule. With freedom recov-
ered, the press overstated a little: every politician was a thief, every official was
corrupt. They wanted to denounce every irregularity, because that used to be for-
bidden. . . . Therefore, well, the press exaggerated. (interview, August, 1998)
The new investigative drive has not always been backed by proper
professional skills, and resources. Waisbord (1996, 351) reported that
the investigative teams created by Peruvian newspapers failed to produce
many remarkable stories. A journalist who shall remain anonymous com-
plained about the Venezuelan press lacking a real investigative tradition,
and claimed that most scandal stories are not investigated in depth: “Issues
are born and die within few days. . . . This is a servile press, a press for
hire, a, a . . . very strange press” (interview, July of 1998). Professional values
were eventually set back to nineteenth-century standards in the early 2000s,
when the Venezuelan mainstream press vehemently embraced the task of
82
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
In the 1980s the editorial line combined with a more interactive [sic] relationship
with politicians. . . . The editor had information about corruption, so he collected
statements by politicians who pointed in that direction. The lack of depth in the
investigation – because there was no way of finding documents – was compensated
by statements, by other means. . . . The ones in charge [of the newsroom] had the
intuition that someone was corrupt and produced declarations of politicians saying
so. . . . Here [at El Nacional] they used to say: “Minister So and So is corrupt. Let’s
get that guy who is his enemy to tell us so.” They would interview the guy, run the
story, and then the whole thing relied on the statement from a guy who was the
minister’s enemy saying that the minister was corrupt. They used to do that in good
faith, because they had information but lacked the proof. But we can make mistakes
that way. . . . (interview, July of 1998)
I would not say that this is a malformation of all Brazilian journalism, but we do have
in Brazil too much journalism of declaration. . . . Declarative journalism is [practiced]
in Brazil not only to avoid investigating accusations – although we investigate less
7 The other side of the coin was Chávez’s ambiguous position regarding free speech, his
administration’s hostile reaction to any form of press criticism, and its politicizacion of
state-owned media (see Tanner Hawkins 2003).
83
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
than we could and should do. I think that this is more of a collective state of laziness.
It is easier for a journalist to go to Congress and come back with some statements
than to go after the facts.
According to Noblat, other factors also account for the practice of declar-
ative journalism: a professional culture that has a fascination with statements
and with the spoken word, and the need to have people on record to back
up news stories (interview, August of 1998).
Waisbord (2000, 103–110) called this practice denuncismo (“quick and
easy reporting” based on few sources) and portrayed it as a common prob-
lem in South America (see also Palermo 2002, 307). Describing the situation
of the press in Mexico, Riva Palacio (1997, 29) also noted that “most pub-
lishers and editors were trained in the old school of journalism, in which
statements weight more heavily than actions, and rhetoric is more impor-
tant than information. Newspapers are full of speeches, statements and
press releases.” Declarative journalism is not unique to Latin America; it is
also well known in American newsrooms. Patterson (1994, 245) called this
style of reporting “attack journalism,” and Boorstin (1987, 24–26) noted
that professional pressures had made this practice common long before
Watergate.
8 I do not imply that journalists unilaterally control the agenda. The press relies on govern-
ment officials to set the agenda (Oviedo 1997), and a complex set of formal and informal
rules typically regulates the interaction between politicians and the press corps (e.g., see
Keenan 1997; Riva Palacio 1997).
84
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
9 Most of the journalists interviewed for this project agreed that sources – in politics, just
as in show business – use disclosure as a weapon against their rivals. Francisco Figueroa
(EFE, Caracas) summarized this view: “No one has ever unveiled a scandal unless there was
a rivalry involved. [Scandals serve] to attack your enemy, to knock off your rivals in a public
contract, in an election, or in other situations.” Miguel Otero (El Nacional, Caracas) pointed
out that “sources speak due to the struggle for power: internal factions seek to destroy others;
someone was left out of the loop; someone wants to take over and needs to bring down the
guy at the top.” Carlos Chagas (TV Manchete, Brası́lia): “Sources generally speak in pursuit
of their interests – interests, revenge, vendetta. . . . If the guy was betrayed [he threatens]:
I am going to implicate you all. But, well, some of them speak with good intentions, too. . . . ”
85
P1: JZP
0521869423c04 cuny809-perez 0 521 86942 0 May 8, 2007 16:48
86
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
To what extent are scandals capable of compromising public support for the
president? Although the six presidents discussed in Chapter 2 took office
with the support of significant majorities, they suffered considerable losses
in public approval throughout the process leading to the impeachment
crisis. The average president in this group started his term with an approval
rating of 64 percent but left office with a rating of just 23 percent. An overall
declining trend in approval ratings is not uncommon among presidents in
office, but these cases are notable for the level of deterioration and for the
ultimate consequences of the process. The escalation of public discontent
fueled mass protests that ultimately encouraged impeachment proceedings
against the president.
Although impeachment charges in each case resulted from specific accu-
sations against the president – with the exception of Bucaram, who was
declared mentally unfit after a long series of exposés – the conventional
wisdom suggests that unpopular economic reforms and poor economic per-
formance were important factors explaining the emergence of public unrest
in those countries. On the one hand, the historical narratives presented in
the second chapter suggest that at least in three cases (Bucaram, González
Macchi, and Pérez) popular protest was in part a direct reaction to the
administration’s attempts to impose neoliberal reforms. On the other hand,
students of advanced industrial democracies have consistently shown that
macroeconomic performance is one of the best predictors of public sup-
port for the government. This argument is particularly relevant for Latin
America during this period: Kurt Weyland has convincingly argued that the
impeachments of Fernando Collor and Carlos Andrés Pérez resulted not
only from political isolation but also from their incapacity to “stop further
economic deterioration” (Weyland 2002, 158).
87
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
1 Unfortunately, the reelection of Presidents Carlos Menem of Argentina (1995) and Alberto
Fujimori of Perú (1995) lends credibility to these doubts. On the problem of type I and type
II errors in the impeachment process, see Kada (2003a).
88
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
90
80
70
60
50
40
30
20
10
0
0 5 10 15 20 25 30 35 40 45 50 55
Months in Office
Figure 5.1 Presidential approval in six impeachment crises. Sources: Datafolha and
IBOPE (Brazil), Napoleón Franco y Asociados (Colombia), Market (Ecuador), First
(Paraguay), and Consultores 21 (Venezuela).
popular support fell significantly. Luis González Macchi lost the support of
58 percent of the population; Carlos Andrés Pérez, of 50 percent; Fernando
Collor, of 48 percent; Ernesto Samper, of 40 percent; and Raúl Cubas
Grau, of 26 percent. Abdalá Bucaram’s approval rating declined between
17 and 46 percent points, depending on whether polls were conducted in his
stronghold, the city of Guayaquil, or in the Ecuadorian highlands. Those
figures suggest that the politics of scandal discussed in the previous chapter
may have weakened those presidents significantly.
Figure 5.1 traces the erosion of approval ratings for the six presidents
over the course of their terms. On average, they lost between one and one
and a half points of popularity every month. The exceptions were Abdalá
Bucaram and Raúl Cubas, who lost ground at a much faster pace – about
five and three points per month, respectively. Not surprisingly, Bucaram
and Cubas were not able to finish their first year in office.
89
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Economic Reforms
Throughout the 1990s, several Latin American presidents adopted (or
attempted to adopt) harsh economic reforms. In spite of their campaign
promises, presidents told voters that their bank savings would be frozen;
that gas rates, electricity rates, and bus fares would be raised; that subsidies
for the poor would be cut; and that public employees would be laid off
(Stokes 2001; Weyland 2002). A quick look at the data suggests that these
announcements had significant political costs. Fernando Collor’s approval
rating (71 percent in March of 1990) fell to 36 percent by June, after all bank
accounts were frozen for eighteen months. Carlos Andrés Pérez’s approval
rating (70 percent in January of 1989) dropped to 46 percent by March, after
increases in bus fares triggered violent riots. Confirming this perception,
one of his ministers noted that “Pérez displayed an unusual willingness to
incur the political costs that inevitably accompany major reforms” (Naı́m
1993, 46).
Multivariate analyses tend to support the “political cost” hypothesis.
Buendı́a (1996) showed that basic price increases led to declining presiden-
tial approval in Mexico, and Weyland (1998b) documented a link between
support for economic reforms and presidential approval in Venezuela.
However, the relationship between economic adjustment and presidential
approval deserves a more careful examination. On the one hand, because
presidents typically launch economic reforms early in their terms (when
they are stronger, but when their approval rates are volatile), it is unclear
how much of the subsequent drop in public support is the result of their
policies and how much is explained by the end of the “honeymoon” period
(Gronke and Brehm 2002). On the other hand, voters may support reformist
governments despite the costs of adjustment, either in the hope that ensu-
ing economic recovery will compensate for the losses (Stokes 1996a) or to
90
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Macroeconomic Performance
Studies of presidential approval and voting behavior in industrial democ-
racies have consistently shown that economic performance is a major pre-
dictor of public support for the government (Beck 1991; Lewis-Beck and
Paldam 2000; Lewis-Beck and Stegmaier 2000). Although there is general
agreement that people hold the president accountable for economic perfor-
mance, the literature has shown some disagreement on the specifics. It has
been argued that the electorate is more sensitive to economic downturns
than to prosperity (Bloom and Price 1975), that voters react more strongly
to trends in the national economy than to their individual (pocketbook)
situation (Kinder and Kiewiet 1981), that their evaluations of the econ-
omy are prospective rather than retrospective (Chappell and Keech 1985;
MacKuen, Erikson, and Stimson 1992), and that the impact of retrospective
evaluations is greater when an incumbent is running for reelection (Nadeau
and Lewis-Beck 2001).
Given the need to achieve comparability across cases and the existing lim-
itations of information, I will bypass some of these debates and concentrate
instead on the more general question of how macroeconomic performance
(inflation and unemployment) shaped the decline in public support for the
six administrations under study. Susan Stokes (2001) has shown that dur-
ing this period Latin American voters based their judgment of incumbent
governments on actual macroeconomic results rather than on the ideologi-
cal content of the policies implemented. In turn, Weyland (1998b) showed
that pocketbook expectations of the economy dominated the evaluations of
Carlos Andrés Pérez and his economic program between 1989 and 1993.
There is also clear evidence of economic voting in new democracies outside
Latin America (e.g., Hesli and Bashkirova 2001).
colleagues (2002) concluded that the framing of media scandals may have a
significant impact on presidential approval even when the president’s popu-
larity appears to be impregnable to the accusations. In a comparative study
of public opinion trends during the Colombian and Venezuelan crises, Jaime
Bermúdez (1999) similarly showed that media coverage was an important
predictor of Samper’s and Pérez’s approval ratings.
Unfortunately, the effects of media scandals on presidential approval
may be hard to pinpoint. For instance, in a study of the Lewinsky episode,
Fischle (2000) found that the impact of subjective considerations (e.g., the
importance given to the allegations) on individual reactions to the scandal
was conditional on the respondent’s prior support for President Clinton.
Stokes (2001, 135–136, 149) claimed that in the case of Carlos A. Pérez,
poor economic performance encouraged citizens to believe that govern-
ment corruption was to blame.2 These findings suggest that the effect of
scandals on aggregate approval rates at time t may be mediated by the
president’s political capital at time t – 1.
At the same time, the previous chapter suggested that strategic politi-
cians, editors, and publishers may be more likely to engage in the politics
of scandal when the president is weak and when the public already distrusts
the chief executive. Like the general public, journalists may lend little cre-
dence to the first accusations reaching the newsrooms, but they may actively
pursue exposés if prior scandals have enveloped the executive. A Brazilian
journalist discussing the Collor case recalled that “[w]e heard the first cor-
ruption story, then the second, then the third. . . . In the end we realized
that the country was ruled by the Mob” (Carlos Chagas, interview August
1998). The news media may therefore be reluctant to investigate a popular
president, but they may easily join the “feeding frenzy” once the execu-
tive is discredited (Sabato 1993). If this hypothesis is correct, scandals and
presidential approval may have a reciprocal causal relationship.3
In the following section, I reconstruct the interactions among neoliberal
reforms, economic performance, media scandals, and presidential approval
in the countries under study. Approval data (presented in Figure 5.1) is con-
fronted with data on inflation and unemployment for Brazil (sources: IBGE
2 Similarly, Elimar Pinheiro claimed that in Brazil “for corruption to be identified, investi-
gated, and made public . . . it was necessary to make the President symbolically responsible
for the situation of economic difficulty that most people were suffering” (Pinheiro 1992,
32).
3 I am indebted to Susan Stokes for her comments on this issue.
92
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Case Studies
Table 5.1 summarizes the situation of the six administrations under study.
The top panel compares the administrations in terms of their adoption of
neoliberal reforms and the average monthly and yearly levels of inflation
and unemployment. As a reference point, I have also included performance
indicators for the administrations that immediately preceded them.5 It is
not clear that those administrations performed much worse than their pre-
decessors (although some deterioration seemed to occur under Collor and
Pérez regarding inflation and under Samper and Bucaram regarding unem-
ployment), but they were clearly unable to implement successful policies.
4 Sources used to identify scandals included the following newspapers and magazines: Folha de
São Paulo (Brazil), El Nacional (Venezuela), El Comercio (Ecuador), Ultima Hora (Paraguay),
Semana (Colombia), and Latin American Weekly Report (for several countries). For the
sequence of events in the Samper case, I also relied on Betancourt Pulecio (1996).
5 The administrations of José Sarney in Brazil, César Gaviria in Colombia, Sixto Durán Ballén
in Ecuador, Juan Carlos Wasmosy in Paraguay, and Jaime Lusinchi in Venezuela.
93
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Administration’s González
Average Performance Collor Pérez Samper Bucaram Cubas Macchi
Economy
Adjustment policy Yes Yes No Yes No Yes?
Monthly inflation (1) 21.3 3.4 1.5 2.7 0.5 0.9
Annual inflation (2) 1690.3 47.7 20.9 24.4 11.5 8.4
(Prior administration) 532.3 18.4 27.3 37.4 13.8 11.5
Unemployment (1) 5.4 9.0 11.1 10.1 12.7 9.5
Unemployment (2) 3.7 9.4 10.1 10.4 5.4 6.8
(Prior administration) 3.2 10.7 9.3 7.8 5.3 5.4
Scandals
Scandal index 43 33 31 256 48 16
President involved? Yes Yes Yes No Yes? No?
Approval (3) Quito Coast
Maximum approval 63 70 81 53 54 57 77
Minimum approval 11 19 33 4 11 27 8
Average rate of change −1.4 −2.8 −0.9 −2.8 0.3 −3.1 −1.6
Protest coalition Broad Broad Narrow Broad Medium Broad Medium
Months in office 32 54 49 7 8 38
Sources: Economy: (1) Average monthly (or quarterly) rates compiled from IBGE and SPES
(Brazil), DANE (Colombia), INEC (Ecuador), Departamento de Cuentas Nacionales and DGEEC
(Paraguay), BCV and OCEI (Venezuela). These figures are used in the pooled time-series models
to be discussed later. (2) Yearly rates reported by the World Development Indicators (World Bank,
2005). Comparisons to previous administrations are based on yearly figures.
Scandals: Folha de São Paulo (Brazil), El Comercio (Ecuador), Betancourt Pulecio (1996) and Semana
(Colombia), Ultima Hora (Paraguay), El Nacional (Venezuela), and Latin American Weekly Report (for
several countries). See the appendix table to this chapter for a complete list of scandals.
Approval: Datafolha and IBOPE (Brazil), Napoleón Franco (Colombia), Informe Confidencial
(Ecuador), First (Paraguay), and Consultores 21 (Venezuela). (3) Surveys in Ecuador were conducted
separately in Quito and Guayaquil by Informe Confidencial.
The table also presents monthly averages for the scandal index intro-
duced in the previous section, and indicates whether the president was
directly affected by the accusations. Episodes of direct involvement include
Pedro Collor’s accusations against his brother Fernando, the investigation
of the 250 Million case in Venezuela, and the accusations against Ernesto
Samper by Santiago Medina and Fernando Botero. No such accusations
were ever explicitly made against Abdalá Bucaram, and the evidence is
mixed regarding Paraguayan Presidents Raúl Cubas and Luis González
Macchi. Cubas was accused of being in contempt of the Supreme Court,
but his reluctance to imprison Oviedo (as requested by the Court) was never
94
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
a secret. By contrast, it was always assumed that González Macchi was per-
sonally involved in several corruption scandals, but the opposition failed to
create a high-profile investigation linking the president to the cases.
The bottom panel summarizes the reaction of the mass public to each
administration’s performance. In addition to the rate of change in presi-
dential approval ratings, I included a qualitative assessment of the scope of
the social movements calling for the resignation of the president. This
assessment is based on historical analysis of the cases to be presented
later.
Table 5.1 suggests that the combination of scandals and (failed) eco-
nomic reforms can erode presidential popularity and encourage multiclass
protests against the president. There are, however, partial exceptions to this
pattern. In Guayaquil, where Bucaram’s party preserved historical patron-
age networks among the popular sectors, the combined effects of scandals
and economic reforms were less harmful for the administration. In another
case, the willingness of the González Macchi administration to compromise
on policy issues hindered the formation of a multiclass coalition in favor of
impeachment. As I will show, Paraguayan peasants mobilized against spe-
cific policies, but the administration was able to deactivate peasant protests
and prevented a mass movement in favor of impeachment.
The evidence presented in the table fails to establish an indisputable
causal pattern, suggesting instead that these factors can provide only partial
explanations for the emergence of public unrest. In order to address this
problem, I follow two opposite but complementary strategies. In the fol-
lowing sections, I engage in process tracing to establish the precise sequence
of events leading to the impeachment crises. In contrast to my approach
in Chapter 2, these narratives emphasize the evolution of the key variables
presented in Table 5.1 (see George and Bennett 2005, Chapter 10). Build-
ing on the insights of these case studies, in the final part of the chapter I
conduct a quantitative analysis of the interrelationships among neoliberal
policies, economic performance, media scandals, and public outrage.
96
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
97
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Venezuelans seemed astonished when the new government disclosed in early 1989
that foreign currency reserves were severely depleted, that the fiscal deficit for 1988
exceeded 9 percent of the GDP, that the current account had registered its largest
deficit in history, and that all prices, from interest rates to eggs, from medicine to
bus fares, were artificially low and impossible to sustain. (Naı́m 1993, 28)
But President Pérez still had his charisma: he was a living myth among
Latin American politicians, and 70 percent of the voters had a favorable
image of him in January of 1989, right after the election. He decided to
invest this political capital in a bold gamble: economic recovery would
require a “great turnaround” of the oil-dependent, import-substitution
model. In February of 1989, the administration announced an economic
program that included price deregulation, liberalization of exchange and
interest rates, and adjustments in public sector prices (gasoline; public trans-
portation; water, telephone, and electricity rates; and other public services).
The long-term components of the reform involved plans for the adoption
of a value-added tax (a shocking novelty for Venezuelans), trade liberaliza-
tion, and privatizations (Naı́m 1993; Stokes 2001; Weyland 2002; interview
with Miguel Rodrı́guez, July of 1998).
As an immediate consequence of the measures, the bolı́var suffered a
sharp devaluation (170 percent); interest rates climbed by 30 percent; and
99
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
The President has admitted . . . that he is alone in his efforts to push the adjustment
program forward. And indeed, he is. His party does not support him . . . and nei-
ther do the trade unions, the social and professional associations, the students,
the Church. Industrialists, construction companies, and agricultural producers
are all disappointed because the “opening” has hit them badly and high interest
rates are driving them to bankruptcy. Outrage among marginal sectors, blue-collar
workers, and the middle class is now almost explosive. (El Nacional, February 10,
1990, D2)
Although the economy stabilized after the initial shock, and Venezuela
began to grow at a fast pace (reaching a 10 percent growth rate in 1991,
thanks to the first Gulf War), popular protests against the administration’s
policies continued, and a wave of scandals compromised the credibility of
the government.7 After the first gubernatorial election in the history of
modern Venezuela took place in 1989, in April of 1990 an opposition gov-
ernor announced that the minister of transportation had demanded kick-
backs on all public works conducted in his state. The same minister was
6 Supporters of the “great turnaround” argued in later years that the key failure lay in the
communications strategy – the assumption being that proper salesmanship would have led
the population to accept the program. The behavior of Venezuelan voters in later years casts
some doubt on this theory.
7 Pérez’s first year in office (1989) was dominated by scandals related to the previous admin-
istration – the manipulation of the preferential exchange regime by Lusinchi’s secretary and
mistress, Blanca Ibáñez, and the distribution of twenty-five Jeeps bought with public funds
among loyal (i.e., Lusinchista) AD leaders during the campaign leading to the primaries.
Although such exposés were not counted as part of the exposure index, one should won-
der to what extent they hurt Pérez in the long run. The accusations pointed to Lusinchi,
his main rival within the party, but they reinforced public perceptions of AD as a corrupt
party and initiated a battle for public opinion that in the end worked against Pérez and his
collaborators (Bermúdez 1999).
100
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
104
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
The first accusations against Samper came [in June of 1994] from the defeated
presidential candidate, Andrés Pastrana, from the newspaper owned by his fam-
ily, La Prensa, and from other public figures like candidate Enrique Parejo. Many
105
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Restrepo’s analysis suggests that public outrage emerged only after 1995
and was concentrated in the Colombian middle and upper classes, as well
as in some organized sectors of civil society (see Bermúdez 1999, 110–
115). Describing the demonstrations in favor of Samper’s impeachment,
Cepeda Ulloa also noted that the “demonstrations of students and ladies8
never acquired a popular flavor, reinforcing the unpalatable class overtones
that characterized the most boisterous and notorious confrontation with a
president [ever]” (Cepeda Ulloa 1996, 80).
After the Chamber of Representatives dismissed the impeachment
charges in mid-1996, the incidence of scandals declined. In a futile attempt
to convey its own verdict against Samper, the Clinton administration can-
celled the Colombian president’s visa. But it was not until early 1997 that
new accusations against specific ministers surfaced. In March, the minister
of defense resigned after being linked to a drug-related scandal. In May,
the minister of transportation was accused of allowing Liberal politicians
to use a public helicopter for campaign purposes. Three months later, the
ministers of energy and communications resigned after Semana revealed
that they had granted FM radio frequencies to political friends and split the
benefits “half-and-half.” By then, the measure of exposure to scandals was
around thirty points – a pace of one exposé every three months – and any
possibility of impeachment had been forestalled. The president’s approval
rating remained above the 30 percent mark until the end of the term, while
the unemployment rate rosc to 15 percent. The Samper administration had
become a case of surviving without governing.
8 The translation is mine. Cepeda employs the term “señoras” (ladies) with an overt class
connotation to refer to the upper-class women leading some demonstrations.
106
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
9 The Roldosistas sensed the opportunity to introduce important political reforms as well.
They dreamed of a constitutional reform that would allow presidential reelection, establish
a five-tear term, and schedule legislative elections concurrent with the presidential runoff.
108
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
110
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
The public’s intimate conviction that Gen. Oviedo had organized the
slaughter of the vice president triggered mass protests and drove the
general’s popularity down to 15 percent in April. President Cubas, respon-
sible for Oviedo’s freedom, also saw his approval rating drop to 20 percent.
In spite of the sharp decline, those figures indicate the presence of a hard-
core group of Unace voters willing to back the Oviedo–Cubas team under
any circumstances. As mass protests unfolded, pro- and anti-impeachment
demonstrators fought to gain control of the square in front of the Congress
building. The first group was composed mostly of middle-class students,
outraged at the slaughter of Argaña, and peasant unions, mobilized to
demand policies for agricultural debt relief. The second group was made
up of Unace militants. The debacle of the administration took place after
March 26, when paramilitary snipers shot the anti-Oviedo demonstrators
in the plaza. Confronting an impeachment in the Senate and major interna-
tional pressures, and unwilling to see himself associated with the outbreak
of violence, President Cubas boarded an airplane to Brazil two days later.
As explained in previous chapters, the Paraguayan elite responded to the
crisis by forming a coalition government headed by the little-known speaker
of the senate, Luis González Macchi. Initially, the collegial political style
of González Macchi constituted an asset. Although not a popularly elected
president, González Macchi started his term with a 68 percent approval
rating, reaching 77 percent by July of 1999. However, public support began
to fade afterward, and the coalition quickly broke up. One year later, in July
of 2000, the president’s approval rating was already at 44 percent. Two years
later, in July of 2001, his popularity had plummeted to 10 percent. By then,
Gen. Oviedo, bête noire in exile, had rebuilt his popularity to levels ranging
around 40 percent.
The context of the decline in González Macchi’s popular support was
one of dismal economic performance. Unemployment, already at 13 per-
cent during the Cubas period, reached 16 percent in 2000. The average
growth rate between 1996 and 2000 was −1.9 percent (World Bank 2005).
Confronted with mounting deficits, the administration embarked on a pro-
gram to privatize telecommunication services, water and sewerage, and
the national railroads. The policy met with suspicion among orthodox
Colorados and was overtly opposed by the trade unions and the peas-
ant movement. In March of 2000, fifteen thousand peasants marched
to Asunción and joined forces with the trade unions against privatiza-
tions. In June, the National Workers Confederation (Central Nacional de
111
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Trabajadores, CNT) called for a national strike. Even though many Lib-
erals agreed with the privatization program, electoral incentives – the vice
presidential election was scheduled for August – drove them away from
supporting the administration’s policies.
The decline in approval ratings was also encouraged by a sequence of
media exposés beginning in November of 1999, when the comptroller gen-
eral was accused of embezzlement, money laundering, and corruption. As
the case received ample coverage in February and March of 2000, the media
disclosed additional stories about irregular contracts in the Social Security
Institute (IPS). The scandal index rose from zero in October of 1999 to
seventeen points in February of 2000 – a scandal every six months – and
approval ratings declined from 51 percent in November of 1999 to 33 per-
cent following the peasant protests in March of 2000.
Some military officers saw this drop in public support as an opportu-
nity and attempted to seize power in May of 2000, fourteen months after
González Macchi had taken office. The wave of accusations receded for
a few months in the aftermath of this failed coup, but resurged later in
2000 when the Ministry of Public Works was charged with diverting funds
collected at toll roads for purposes other than road maintenance. In Febru-
ary of 2001, the press published stories revealing that the president had
instructed the Social Security Institute to invest almost half of its reserves
to rescue a public bank. Around that time, a Japanese development bank
accused Minister of Public Works Walter Bower of manipulating the rules
for the allocation of road maintenance contracts.
In March, press investigations confirmed ongoing stories that the BMW
regularly used by President González Macchi had been stolen in Brazil, and
that the Finance Ministry had authorized the purchase of the stolen car.
Two months later, the executive was accused of transferring $16 million
from the coffers of two banks intervened by the Central Bank to a private
account in New York. The head of the central bank was forced to resign,
and an opposition front of Liberals and Oviedistas in Congress called for
the impeachment of the president.
Thirty thousand peasants again marched to Asunción to demand access
to credit and better conditions for cotton production in March of 2001. Fol-
lowing new protests against privatizations in August, the Catholic Church
organized a round table of national dialogue. By then, presidential approval
ratings were around 8 percent and the scandal index had reached twenty-
six points – or a scandal every four months. Although the opposition talked
112
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
113
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
10 Luis González Macchi was charged with instigating the illegal transfer of $16 million
from the Central Bank to a private account and sentenced to six years in prison by a
Paraguayan court in June of 2006. In December, he was sentenced to eight years in prison
for embezzlement. At the time of this writing the former president was appealing the
sentences.
114
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Multivariate Analysis
The key question, of course, is how to disentangle the simultaneous effects
of those causal factors. In order to address this problem, I developed a
multivariate model of presidential approval pooling the data for the six
administrations. The dependent variable is the percentage of respondents
supporting the president at any point in time (month). The independent
variables reflect, on a monthly basis, the evolution of the key factors pre-
sented in Table 5.1. The introduction of neoliberal economic reforms was
captured by an intervention variable that adopts a value of zero before the
policy was launched and of one afterward. Macroeconomic conditions were
measured using monthly changes in the consumer price index and the rate
of unemployment. The incidence of scandals was captured using the scan-
dal index presented early in the chapter (the number of media exposés over
the elapsed months in office). I also introduced two intervention variables
to reflect the “honeymoon” (the first three months in office) and any period
when presidents were directly implicated by the investigations.11
11 Collor was considered “directly involved” in the accusations between May and September
of 1992; Pérez between November 1992 and June 1993; Samper between July 1995 and
June 1996; and Cubas Grau between December 1998 and March 1999. No explicit charges
were ever made against Bucaram or González Macchi (accusations in these latter cases
remained diffuse, compromising the president through the deeds of collaborators and family
members).
115
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
12 With a few recent exceptions (Cima 1998) there are no comparative surveys of presidential
approval in Latin America. The survey instruments used by Consultores 21, Napoleón
Franco y Asociados, Informe Confidencial, and First captured presidential favorability (i.e.,
the percentage of respondents with a positive image of the president), while Datafolha and
IBOPE measured presidential job approval (i.e., the percentage of respondents ranking the
president’s performance as excellent or good).
13 In Brazil, Datafolha and IBOPE’s surveys (N ≈ 5,000) were typically conducted in all “major
cities” (usually defined as São Paulo, Rio de Janeiro, Belo Horizonte, Curitiba, Porto Alegre,
Salvador, Recife, Fortaleza, Belem, and Brası́lia or Goiania), although in one case the sample
covered only São Paulo. Consultores 21 (Venezuela) used national samples for urban areas
covering approximately 77 percent of the adult population (N ≈ 1,500). Napoleón Franco
y Asociados (Colombia) developed national polls for the Samper administration covering
thirty-six urban and rural municipalities (N ≈ 1,500), as well as surveys in the four major
cities only (Bogotá, Barranquilla, Cali, Medellı́n; N ≈ 850). Informe Confidencial (Ecuador)
measured favorability in the two largest cities, Quito and Guayaquil (N ≈ 400 for each city).
116
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
117
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
Analysis Given the limitations of the data, the most conservative strategy
was to perform the analysis using fixed-effects regression models.14 This
approach was able to accommodate the irregular structure of the dataset
while at the same time offering a rigorous way to deal with unit effects (to
capture, for instance, the overall level of support commanded by each pres-
ident, or the unique nature of the instrument employed by each pollster).
This approach does not assume that unit effects are uncorrelated with other
independent variables or that they follow any particular distribution. Each
administration was treated as a unit, with the exception of the Bucaram
case, in which parallel surveys were conducted in Quito and Guayaquil
independently.
Table 5.2 presents the results of the analysis. In addition to the stan-
dard fixed-effects estimator (5.2.1), I verified the results in several ways: the
dependent variable (a percentage naturally bounded between 0 and 100) was
“linearized” by taking the natural logarithm of the odds of approval (5.2.2);
standard errors were corrected to allow for the possibility of different resid-
ual variance and residual covariance across panels (5.2.3); observations were
weighted (as discussed earlier) to correct for the irregular frequency of sur-
veys (5.2.4); and estimates were obtained using a Tobit model (5.2.5). The
results in four of the five models suggest that the adoption of neoliberal poli-
cies hurt public support for these presidents by sixteen points on average,
even after controlling for the honeymoon period. Rising unemployment
may also have eroded presidential approval (an increase from, say, 5 to
10 percent unemployment predicted, on average, a nine-point decline in
approval ratings).
Although the effects of inflation are not significant, this result must be
interpreted with caution. In this sample, inflation data are biased because
of the extreme values during the first months of the Collor administra-
tion (prices rose by 82 percent in March of 1990). Attempts to correct
for this problem by using a log transformation of inflation still yielded
insignificant coefficients for this variable. This finding is consistent with
Kurt Weyland’s claim that extreme inflation may in fact boost public
14 The alternative to pooling was treating each administration separately, conducting individ-
ual time-series analyses. Unfortunately, this strategy generated inconclusive results because
some administrations lasted in office for very short periods, because some important predic-
tors vary across administrations more than within them, and because some variables turned
out to be significant for one case but not for others (see Pérez-Liñán 2001, Chapter 6).
Moreover, the problems of endogeneity addressed later in this chapter would be virtually
intractable if dealing with the series individually.
118
P1: JZP
0521869423c05 cuny809-perez 0 521 86942 0 May 8, 2007 17:31
support for presidents who are perceived as brave risk takers confronting
a crisis (Weyland 2002). Lourdes Sola similarly argued that the legitimacy
provided by Collor’s direct election and the gravity of the economic cri-
sis explained “the surprising responses of public opinion to the reforms,
such as the responses of workers and employees in state-owned enter-
prises, many of whom became supporters of privatization” (Sola 1994,
158).
The results also show that media scandals were a significant force behind
the erosion of presidential approval. Given the coefficients in Table 5.2, we
would expect that a president confronting a scandal every two months (that
is, with an average score of fifty points in the index) would see his or her
approval rates decline between six and eight percentage points. A president
directly implicated by those investigations of wrongdoing would suffer, on
average, an additional loss of ten percentage points.
Feeding Frenzy?
The previous models assume that media scandals take place independent of
presidential approval. However, the earlier discussion suggests that media
informants, editors, and publishers do not operate in isolation from the
broader political context. Editors may be more willing to risk the publi-
cation of controversial stories when the president is weak and when their
audience is avid for criticism. Journalists may be encouraged to pursue
scandals when newsroom managers want to make headlines with them.
And politicians may be tempted to leak sensitive information about the
president when they are dealing with a receptive press corps.
This idea suggests that the emergence of media scandals may not be
random or solely determined by the level of corruption in the executive
branch. In fact, there are good reasons to suspect that presidential approval
and executive performance at any given time t may in part drive the pro-
duction of scandals at time t + 1. If this is the case, the analysis presented in
Table 5.2 may violate some fundamental assumptions of regression analysis:
to the extent that the scandal index is sensitive to prior levels of presidential
approval, any omitted exogenous factor driving up approval at time t could
indirectly reduce the likelihood of scandals at time t + 1. This pattern of
reverse causation would make the error term covary with two independent
variables, the scandal index and the involvement of the president’s image in
the investigations, and therefore could bias the estimates in models 5.2.1
through 5.2.5.
119
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
121
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
(If this is the case, a new problem of endogeneity would arise, which needs
to be corrected at this stage). Although none of the variables proved to be
statistically significant, the residuals of the logistic model were also saved
as a “purged” instrument; they are highly correlated with the involvement
dummy, but not with the scandal index (r = −.01) or with presidential
approval (r = −.03).
Using those instruments, I conducted the two-stage least-square anal-
ysis. The first stage is reflected in Models 5.3.3 and 5.3.4, which estimate
the proxies for the scandal index and the presidential involvement variable.
Scandals are less frequent during the honeymoon period and more fre-
quent when presidents adopt unpopular policies, but there is no evidence
that scandals are more likely when the economy is doing poorly. In turn,
presidents are more likely to be directly affected by the accusations when
recurrent scandals have eroded their credibility and when unemployment
is high. (I have limited the number of decimal points in the table for reasons
of space, but the bold entries indicate that the coefficients, albeit small, are
significant at the .05 level.)15 Taken together, these models suggest that
there is some evidence in favor of the argument about strategic behavior
on the part of the press, although further research needs to be conducted
on this issue.
The second stage (5.3.5) replicates the general model of presidential
approval (5.2.1) using the instrumental variables instead of the original indi-
cators (standard errors are estimated based on the variance of the original
predictors). Both variables have significant coefficients, suggesting that –
irrespective of whether exposés are in fact driven by the strategic calcula-
tions of the media – presidential approval is hurt both by the frequency of
scandals and by their proximity to the chief executive. The coefficient for
the scandal index is stronger in this estimation: a president with a scandal
score of fifty points would see his or her popularity decline tweleve points
on average (about twenty points if he or she is directly involved in the
accusations).
15 The coefficient for unemployment in 5.3.4 was .004 (s.e. = .001), indicating that a 10
percent increase in unemployment would yield a small increase of 4 percent in the prob-
ability of scandals directly involving the president. The coefficient for the scandal index
was .00019 (s.e. = .00005), indicating that a president facing a scandal every month (S =
100) would see the risk of being personally involved grow by just 2 percent. Note that 5.3.4
relies on a crude linear probability assumption, which may partly explain the small size of
the coefficients. The results should therefore be taken as merely indicative of underlying
effects.
122
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
Conclusions
Were calls for impeachment a legitimate response to corruption scandals,
the product of media manipulation, or just an indication of policy failure
in the context of weakly institutionalized democracies? Maybe the answer
is: all of the above. The combination of case studies and quantitative anal-
ysis has presented a complex picture of the sources of popular discontent.
High levels of unemployment, the imposition of neoliberal reforms, and
media scandals (particularly when they involved the president directly)
eroded presidential popularity and encouraged popular uprisings against
these administrations.
The analysis also suggests three conclusions. First, it seems that any
single scandal in isolation is unlikely to result in a public opinion crisis.
In most of the cases discussed here, media exposés accumulated over time,
eroding the president’s reputation and making every new accusation more
credible than the previous one. Second, the timing and frequency of scandals
is not a product of chance. Although limited by problems of comparability,
the quantitative evidence indicates that scandals are likely to multiply when
an administration is weak. The results of the two-stage least-squares analysis
suggest a very distinctive form of feeding frenzy: if an exogenous factor hurts
the popularity of an administration, weakness may encourage subsequent
leaks and investigations, which will further erode the public standing of
the president and in turn encourage new exposés. Mass support for the
president may discourage the production of media scandals, while political
weakness may initiate a downward spiral of accusations and declining public
trust.
Shall we conclude from this finding that the fourth estate is just an
opportunistic actor? The answer to this question will vary from case to
case, but a more general point has been advanced in Chapter 4: members of
the press should be considered strategic players, just as members of other
institutions (courts and legislatures, for instance) are treated in this way by
contemporary political science. Paradoxically, a modern press corps with
autonomous career goals and advanced investigative skills is more likely to
engage in this sort of strategic behavior than a traditional press reliant on
official sources or driven by partisan goals.
Third, the decline in presidential approval is likely to force the removal
of the president only if it translates into significant public mobilization
(Hochstetler 2006). But not every form of mobilization constitutes a serious
challenge to presidential survival. Protests can be destructive when they
123
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
124
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
Appendix Table to Chapter 5. Main scandals involving the president, close collaborators, or the
first family
(continued)
125
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
126
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
(continued)
127
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
128
P1: JZP
0521869423c05a cuny809-perez 0 521 86942 0 May 8, 2007 17:20
(continued)
129
P1: JZP
0521869423c05b cuny809-perez 0 521 86942 0 May 8, 2007 17:20
131
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
Even when they are besieged by accusations and protests, presidents may
avoid an impeachment if they can rely on loyal legislators. In this chap-
ter, I explore how scandals and public outrage were translated into leg-
islative action against the executive. Under ideal conditions, members of
Congress would impeach the president only if there were sufficient proof
of a “high crime,” and would refrain from doing so if accusations were
merely grounded in partisan or personal motivations. In reality, however,
legislators are hardly able to detach themselves from the broader social
and political context in which a presidential crisis takes place (Kada 2000;
2003b, 148–149). On the one hand, a partisan Congress may protect the
chief executive even when – as in the case of Colombia – there are important
reasons to pursue an in-depth investigation. On the other hand, legislators
may press charges against the president even when there is no real proof
or public sentiment in favor of impeachment – as in the cases of Panama in
1955 and the United States in 1998.
This issue is crucial for the credibility of Congress as a democratic insti-
tution. As a collective body, Congress should act in ways that strengthen
its credibility and public standing. Individual legislators, however, may
encounter personal incentives that run counter to this collective course
of action (see Mayhew 1974).1 Concerned about their own careers,
1 Legislators are aware of this challenge. Colombian Representative Pablo Agámez Agámez
noted that “[t]he trial against the President [Samper] has simultaneously become, for the
national and the international public opinion, a trial against Congress” (Gaceta del Con-
greso, June 19, 1996, 466–467). During the critical session of May 21, 1993, in which the
Venezuelan Senate authorized the Supreme Court to initiate the trial of President Pérez,
Senator Rafael Caldera claimed that “Congress can save its prestige vis-à-vis a people that
has recently questioned it, by giving [them] a provisional president, a new government that
opens new horizons . . . ” (Diario de Debate del Senado, May 21, 1993, 701).
132
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
members of the president’s party may seek to shield the chief executive
from any accusations, while opposition legislators may pursue an impeach-
ment even if the evidence against the president is weak (Kada 2003a).
Given the presence of partisan incentives, what determines the congres-
sional decision to protect or to impeach the president? I contend that this
is the result of four factors: constitutional rules, the party system, the pres-
ident’s relation with Congress, and the overall political context (the nature
of scandals and the electoral calendar, for instance). In the first section of
this chapter, I compare different constitutional traditions and show that in
every presidential system Congress has at least the power to block the
impeachment process. It follows that the partisan composition of Congress
interacts with key constitutional rules to facilitate the formation of a “leg-
islative shield” to protect the president. In the second section, I present the
stylized facts of this process. The model formalizes the role of legislative
bodies as “veto players” with the capacity to block an impeachment. The
third section takes a long historical detour in order to explore the ability
of presidents to use these institutions to their advantage. The six case stud-
ies indicate that presidents who isolated themselves from Congress or who
confronted Congress early in their terms became likely targets once media
scandals enveloped their administrations. In the final section, I employ a
statistical model to assess how different contextual factors encourage indi-
vidual legislators to shield the president from impeachment. Although par-
tisan affiliation is part of the story, the results indicate that presidential
leadership truly matters: isolated presidents and, to a lesser extent, presi-
dents who adopt a confrontational stance vis-à-vis Congress are more likely
to be impeached, while presidents who build extensive legislative coalitions
early in their terms are likely to be shielded (Coslovsky 2002). At the same
time, electoral considerations also weigh in legislators’ decisions: the more
distant the next election is, and the higher the presidential approval rating,
the more willing legislators are to protect the chief executive.
Impeachment
Presidential impeachment is the main constitutional procedure allowing
Congress to remove the president from office. In the American consti-
tutional tradition, the term “impeachment” refers to a trial of the presi-
dent initiated by the lower house and conducted by the Senate. Although
133
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
2 The constitutions of Guatemala, Honduras, and Nicaragua are vague about the impeach-
ment procedure. They empower Congress to authorize crimimal charges against the exec-
utive, but there is no explicit provision for the suspension or ousting of the chief executive
by impeachment.
3 The constitutions included in the table were adopted in Argentina (1853, 1949, 1994); Bolivia
(1945, 1961, 1967, 1995); Brasil (1946, 1967, 1969, 1988); Chile (1925, 1980); Colombia
(1886, 1991); Costa Rica (1949); Cuba (1940); the Dominican Republic (1942, 1962, 1966,
1994, 2002); Ecuador (1946, 1967, 1978, 1998); El Salvador (1945, 1950, 1962, 1983);
Guatemala (1945, 1956, 1965, 1985); Honduras (1936, 1957, 1965, 1982); Mexico (1917);
Nicaragua (1948, 1950, 1974, 1987 with reforms in 1995); Panama (1946, 1972); Paraguay
(1940, 1967, 1992); Peru (1933, 1979, 1993); Uruguay (1942, 1952, 1967); and Venezuela
(1945, 1947, 1953, 1961, 1999). Three constitutions (Paraguay 1940, 1967; and Nicaragua
1987 before 1995) were excluded because they did not allow for any procedure against the
president.
135
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
Impeachment Model
Congress Congressional Hybrid Judicial Total
Unicameral 5 0 15 20
(Percentage) (25.0) (0.0) (75.0) (100.0)
Bicameral 20 11 6 37
(Percentage) (54.1) (29.7) (16.2) (100.0)
TOTAL 25 11 21 57
(Percentage) (43.9) (19.3) (36.8) (100.0)
Note: In the congressional model, legislators accuse and judge the president; in the judicial
model, the judiciary always intervenes in the trial phase; in the hybrid model, the judiciary
judges the president only for common crimes.
Source: Grijalva and Pérez-Liñán (2003).
4 Under the 1961 constitution (which ruled the Pérez impeachment), the lower chamber
played no role in the process. The 1999 charter is unclear about the voting threshold. I am
indebted to Naoko Kada for her comments on this issue.
136
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
Declaration of Incapacity
Under some constitutions, Congress may also remove the president from
office after declaring his or her of physical or mental incapacity (e.g., the
5 Article 85 defines crimes de responsabilidade as the actions of the president against the constitu-
tion – particularly those compromising the existence of the federal union, the independence
of the other branches or local governments, political and civil rights, internal security, hon-
esty in administration, the budget, and the law or judicial decisions.
6 During the 1996 crisis, it was argued that the Colombian representatives were the “natural
judges” of the president – implying that only a legislative solution was legitimate, and that
they should act as “judges” and not as politicians. In my view, this thesis (supported by
lawyers, most legislators, and the president himself) only created confusion regarding the
proper function of the lower chamber. Not only was the House the natural prosecutor of the
president (its constitutional role is not that of a jury), but political considerations are also,
by definition, part of the legislative phase in every impeachment process.
137
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
138
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
139
P1: JZP
0521869423c06 cuny809-perez 0 521 86942 0 May 8, 2007 18:7
Table 6.2. Constitutional rules on impeachment and declaration of incapacity in Latin America
and the United States, 1990–2004
140
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
trial.
j The House presumably authorizes trial with simple majority, but two-thirds are necessary
to suspend president.
k Congress intervenes only if the president challenges declaration of incapacity issued by
recommendation.
Sources: Grijalva and Pérez-Liñán (2003), Kada (2000), and specific constitutions.
141
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
7 The concept of “veto players” is part of a broader family of related terms that includes
“veto points” (Birchfield and Crepaz 1998; Immergut 1992) and “veto gates” (Shugart and
Haggard 2001; see Kada 2002 on the impeachment process). Because the term “veto player”
has gained extensive currency (Ames 2001; Hallerberg and Basinger 1998; Kay 1999; Tsebelis
1999, 2002), I adopt it in this study.
8 For instance, urging the vote against President Pérez, Venezuelan Senator Pedro P. Aguilar
argued that “we must complete a foretold act. We had all agreed that if the Supreme Court
found reasons to declare the trial germane, here, in the Senate, our hands would rise to
authorize Court’s prosecution of Carlos Andrés Pérez . . . ” (Diario de Debates del Senado,
Friday, May 21, 1993, 714).
142
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
the president is able to keep control over Congress, his or her constitutional
removal is virtually impossible. The degree of legislative support necessary
to prevent the initiation of an impeachment or the president’s removal from
office is therefore determined by the decision threshold required for each
intervening chamber.
Interestingly enough, in the 1990s restrictive procedures involving two
chambers and requiring supermajorities (like the ones employed in Brazil
and Paraguay) were as lethal as the more flexible procedures required in
Venezuela and Ecuador. Although we may be tempted to conclude that
institutional design does not matter in the context of presidential crises,
this is not necessarily the case. The role of constitutional rules can only
be assessed vis-à-vis the capacity of specific presidents to mobilize sup-
port among members of Congress. Whether a “legislative shield” against
impeachment is created depends upon both the number of loyal legislators
in Congress and the constitutional voting thresholds required to prevent
an impeachment.
143
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
9 Procedural rules allow for a third option – abstention. But an abstention counts as a nonvote
if the rules demand a majority (or supermajority) of the members present, and as a vote
against impeachment if the rules demand a majority of all the members.
144
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
0 v = .34 1
0 v = .5 1
0 v = .5 1
0 v = .34 v2 = .5 1
0 v = .34 1
0 v = .34 1
Figure 6.1 Support for the president in six impeachment crises. P = proportion
of the chamber voting against impeachment; v = threshold required to prevent
an impeachment; v2 = threshold to prevent the declaration of incapacity; L =
distance between the level of legislative support obtained and the level required
by the constitution (L = P − v). Sources: Congressional records for the respective
countries.
10 I use the term “coalition” to refer to the set of legislators who consistently vote with the
administration on key issues – in this case, against impeachment – without assuming the
existence of a previous electoral alliance, participation in the cabinet, or a formal legislative
pact.
146
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
147
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
terms. Even though later media scandals and popular discontent may drive
many legislators “to the left” (i.e., closer to zero), the number of members
of Congress above the support threshold may still be sufficient to prevent
an impeachment process. By contrast, presidents who alienate members of
Congress early in their terms (either by confronting them openly or by
ignoring their needs) may be in an extremely weak position to confront
an impeachment crisis. Frustrated legislators may leak critical information
to the press in order to ignite media scandals; they may use their oversight
powers to investigate and publicize the accusations; and they may ultimately
refuse to act as veto players, triggering the fall of the administration.
This issue is particularly relevant when the president, out of preference or
necessity, seeks to impose a policy agenda that most legislators oppose – for
instance, neoliberal reforms in Venezuela or the liberation of Lino Oviedo
in Paraguay. Faced with a potentially hostile Congress, the executive may
adopt one of three possible strategies. The first one is to implement the
policy, bypassing Congress and ignoring congressional leaders whenever
possible. I refer to this strategy as isolation. For example, the president may
refuse to bargain with a recalcitrant Congress, imposing policies by decree
(Cox and Morgenstern 2001). Or, if his party controls a majority of the
seats, the president may refuse to accommodate the party’s needs, invoking
rigid notions of party discipline (what Corrales dubbed party-neglecting
behavior).
In other cases, presidents prefer to “go public” and openly defy Congress
in the hope that this strategy will mobilize public opinion and induce leg-
islative polarization to their benefit. The strategy of confrontation (close to
what Corrales dubs “hostility”) is less respectful of the institutional order,
since presidents are often forced to denigrate legislators publicly. The third,
and most expensive, strategy is to negotiate with legislative leaders on a
regular basis, even if this leads to repeated concessions in terms of pol-
icy or patronage (or both). This strategy appropriate for presidents who
are professional politicians with solid relationships in Congress (and who
understand the existing room for party or opposition accommodation) or
for presidents who lack strong policy preferences (and thus can afford to be
party- or opposition-yielding). It is, however, a difficult strategy to imple-
ment when presidents claim to be anti-party politicians (Kenney 2004) or
when they have rigid policy preferences. In the following sections, I show
how these three strategies have worked for presidents with both small and
large legislative contingents.
148
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
149
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
a victory without the “it is in giving that we receive,” without the Persian market
style of the Sarney administration. (Veja, April 18, 1990, 23)11
11 The Brazilian press usually uses the phrase “franciscan politics” (“it is in giving that we
receive”) as a caustic allusion to pork barrel. The irony was coined by José Sarney’s minister
of industry and commerce, Roberto Cardoso Alves.
150
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
Octavio Amorim has argued that Collor sustained the strategy of ruling
by decree for more than a year because the cost of making concessions to
the largest parties would have been very high. Hoewever, political isolation
bred increasing dissatisfaction both with the way he handled the relationship
with Congress and with the overall performance of his government. In April
of 1991, Congress debated and nearly approved the Jobim bill, intended to
regulate the use of decrees by the executive (Amorim Neto 2002, 76).
Brazilian politicians, even those who followed President Collor to the
end, confirm this diagnosis of isolation. According to Luiz Moreira (PTB-
BA), one of the president’s allies in Congress, Collor “adopted a rather
authoritarian attitude. And besides that, creating antipathy among legisla-
tors, he made it very difficult to get access to him. He isolated himself and
did not keep an open dialogue.” Zé Gomes, a loyal congressman of the rul-
ing PRN, also acknowledged the need for a “more professional relationship
with Congress” during those years (personal interviews, October of 1999).
Ricardo Fiuza – a PFL leader appointed by Collor as his chief of staff to
reorganize the congressional front in 1992 – reportedly told press secretary
Cláudio Humberto Rosa e Silva: “forging a majority in the Chamber is the
easiest thing in the word, Claudinho. It is just getting a list of the deputies
and check, one by one, their political needs. I can do that in a week and
the President will have a majority. But I need his authorization . . .” (Rosa e
Silva 1993, 83–84).12
By the time Pedro Collor spoke to Veja and the impeachment process had
begun in Congress, the legislative coalition was in shambles. Fiuza knew
it was too late to form an anti-impeachment bloc: “With the accusations
rolling against the president, with the left in the opposition, with an adver-
sarial Congress because Collor had not taken care of the deputies, with-
out [elite] support or any governability pact . . . Collor was not impeached,
12 A few months earlier, while still a deputy, Fiuza had already complained to the press secretary
that the administration was not giving enough attention to its allies:
Listen, son: I am the leader of the government [bloc in Congress], right? Well, I recom-
mended Mrs. Rosane a candidate for the position of local head of the LBA – somebody who
had been nominated by one of my mayors in the state of Pernambuco. She not only ignored
my suggestion, but appointed another person who opposed the President during the cam-
paign. The mayor can think only two things: either this deputy he has been supporting
for 20 years is a moron (bunda-mole) with no prestige in the government whatsoever, or he
made no effort to push his candidate. Whatever he thinks, I am on the losing side. Now,
if that has happened to me, just imagine what is going on. . . . (quoted in Rosa e Silva 1993,
84)
151
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
he was not judged, he was lynched,” explained his former chief of staff.13
Another congressman who voted against impeachment recalls
. . . a fact that clearly exemplifies what happened. . . . Close to the time of the
impeachment, President Collor began to telephone legislators requesting their sup-
port. One deputy from Paraná got the call and thought that it was a practical joke.
He kept the conversation for a while and at the end he realized that, indeed, it
was the president calling. Then he simply said: “Mr. President, how many times I
attempted to talk to you and you never gave me an appointment. It is not now that
I will take care of your request simply because you gave me a call.” And he voted
against [Collor]. This shows the kind of dissatisfaction that Collor had created
among his legislative base. (Interview with Luiz Moreira, October 14, 1999)
13 Interview with Ricardo Fiuza, Brası́lia, October 19, 1999. In his memoirs, Samper (2000,
108) also uses the lynching metaphor.
14 Votes to appoint Alarcón (on August 1, 1996) were distributed as follows: PRE, eighteen;
DP, eleven; ID, four; FRA, two; LP, two; PCE, two; others, five (see Burbano de Lara and
Rowland 1998, 82).
152
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
The agreement that tenuously linked the PRE with the ID and the DP
did not survive the election of Alarcón. As explained in the previous chap-
ter, the administration insisted on nominating Fernando Rosero, a staunch
PRE politician, for the post of comptroller general. The leaders of the ID,
the DP, and the PSC complained that the candidate was too close to the gov-
ernment to become an independent watchdog. The administration refused
to replace the nominee, seeking instead to break party discipline among the
opposition.15 After two failed attempts (on August 20 and 21), the motion
to appoint Rosero passed on August 22 by a vote of 38–5, while thirty-seven
deputies abstained in a vain attempt to prevent the necessary quorum. In
late October, the PRE reached a new agreement with the powerful Par-
tido Social Cristiano (PSC) to appoint twelve Supreme Court justices. But
the PSC leader, Heinz Moeller, made clear that the party would remain
in the opposition and would demand the resignation of Education Minis-
ter Sandra Correa – by then already accused of plagiarizing her doctoral
dissertation. President Bucaram publicly responded that PSC leaders were
just “wasting their time” (El Comercio, October 25, 1996, A2).
By late November – when the first one hundred days of “truce” with the
government were over – most legislators had overtly moved to the opposi-
tion camp. Bucaram had offended ID partisans by calling their leader, for-
mer president Rodrigo Borja, a “donkey.” Bucaram had accused Alexandra
Vela, a key DP representative, of obstructing justice in the investigation
of the “assassination” of former president Jaime Roldós – to which Vela
responded with a libel suit against the president.16 The administration
aroused the anger of the Pachakutik leaders by creating an “ethnic min-
istry” – breaking a previous deal with the indigenous movement, which saw
15 Opposition deputies who voted in favor of Rosero explained to the press that the adminis-
tration had promised investment for their districts (El Comercio, August 23, 1996). In later
months, legislators denounced the presence of a “man with a briefcase” in charge of buy-
ing votes for the administration (Saltos 1997, 124). Politicians on both sides of the aisle
depicted those actions as presidential attacks on the opposition parties’ cohesion and not
as logrolling. However, Paco Moncayo – at the time head of the military Joint Command
and later an ID legislator – noted that the practice of buying votes was widely practiced by
all parties (interview, November 1999).
16 President Jaime Roldós (and first lady Marta Bucaram, Abdalá’s sister) died in a plane
crash in May of 1981. The Bucaram family claims that the crash was the result of sabotage
(interview with Elsa Bucaram, November 17, 1999). Deputy Alexandra Vela believed that
challenging the president in court would be a powerful signal to many people who feared
the abuses of the Bucarams (interview with A. Vela, December 1999).
153
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
17 According to Raúl Baca (ID), only two candidates to replace Bucaram had some consti-
tutional legitimacy: Vice President Rosalı́a Arteaga and the president of Congress, Fabián
Alarcón. The constitutional reform of 1996 had left the issue of succession unclear, and
Arteaga, the “natural” candidate, was widely opposed among the opposition legislators –
apparently because she could not guarantee a “regime change.” Alarcón faced strong oppo-
sition, too, but he was the only candidate who could claim a legitimate right to take office
if Bucaram was deposed (interview, November 1999).
18 In later days, following an agreement with the military, Congress had to “unpack” the
declaration – postponing the appointment of Alarcón and allowing Vice President Arteaga
to take office for few days in order to preserve the constitutionality of the succession.
155
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
Because the constitutionality of the vote was dubious and the outcome of
the crisis was uncertain, PRE legislators were reluctant to abandon the boat
while Bucaram was still in office.19 However, the Roldosistas controlled only
23 percent of the seats and were able to gather limited support from a
few additional Conservative, CFP, and Alfarista deputies. This bloc would
have sufficed to prevent the formation of a two-thirds majority, but it was
unable to stop the simple majority needed for the declaration of incapacity.
Bucaram’s failure originated in the small size of his party and in his con-
frontational style (a hindrance for coalition building), rather than in the
betrayal of his followers.
19 After the Alarcón administration had taken over, however, most of the nineteen PRE
deputies disaffiliated and negotiated with the interim president. According to Burbano
de Lara and Rowland (1998, 36): “The acute political crisis created by the demonstrations
of February 5 and the ousting of President Bucaram had a demolishing effect on the rul-
ing party, the PRE. One year after winning the presidential election, the PRE had lost 12
deputies.”
156
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
secretary Luis Alfaro Ucero and aligned with former President Lusinchi,
and a faction of ‘Renovadores’ led by Héctor Alonso López and aligned
with incumbent President Pérez” (Coppedge 1994b, 103). Pérez won the
presidential election with 53 percent of the vote, while his party obtained
48 percent of the seats both in the lower house and in the Senate, leaving
him with the impression that he had won the popular election and that his
opponents within the party had lost (interview with Carlos A. Pérez, July
10, 1998).
President Pérez had learned early in the presidential campaign of 1988
that the orthodox faction of AD opposed his economic reform program.
After taking office, he appointed independent “technocrats” (well-trained
professionals like Imelda Cisneros, Ricardo Hausmann, Moisés Naı́m, and
Miguel Rodrı́guez) to key economic posts and followed a “party-neglecting
strategy” in the implementation of public policy (Corrales 2000; 2002a).
However, the orthodox faction won the 1991 internal elections and gained
control of the powerful National Executive Committee (CEN) at the
national convention.
According to Corrales (1997, 97), AD responded to Pérez’s party-
neglecting strategy in two ways. First, the party began to act as a “virtual
opposition force.” Second, its leaders relaxed the traditional party discipline
in Congress, allowing a few dissidents to defy the government. After Hugo
Chávez’s coup attempt in February of 1992, the orthodox opposition within
the party became stronger. Given the tradition of strong party loyalty within
Acción Democrática, most criticism of the president had taken place behind
closed doors rather than publicly; but party leaders now began to challenge
the unpopular “technocrats” in an open way. Shortly after the coup attempt,
for instance, AD’s president, Humberto Celli, publicly called for a reshuf-
fling of the cabinet, stating that the departure of the powerful head of the
central bank (and former minister of planning) Miguel Rodrı́guez was a
“point of honor” for the party.
As a result of the first coup attempt, presidential power began to vanish.
Pérez appointed an eight-member advisory council that included opposi-
tion politicians and some “notables.” In March, the president announced a
series of measures in response to the committee’s recommendations. A pro-
posal for constitutional reform was introduced in Congress; price increases
for oil and electricity were suspended; and price controls on some basic
foodstuffs and medicines were reintroduced. As part of the negotiations,
Pérez supported a reshuffling of the Supreme Court in which he lost sev-
eral loyal justices.
157
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
158
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
20 The Conservatives participated in the Samper cabinet until January of 1996, when they
formally withdraw at the peak of the scandal (Samper 2000, 69). But a group of them
(ironically called “los lentejos”) remained in the cabinet and opposed an impeachment. The
most prominent legislators in this group (which constituted almost half of the Conservative
bloc) were Isabel Celis and Carlina Rodrı́guez (interview with Rep. Benjamı́n Higuita,
C-Antioquia, May 24, 2000).
21 According to Ungar Bleier (1995, 87): “The [1994] elections reinforced the atomization
that was already present and the collapse of regional cacicazgos [bosses] which, with few
exceptions, were replaced in Congress by ‘small’ bosses.” On this process, see Pizarro
(2002).
159
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
. . . very positive; [he had] a very good personal relationship with Congress members.
I would say that the last two presidents had an excellent relation with Congress –
Gaviria and Samper, President Barco was too distant. Gaviria and Samper were
presidents who met with Congress members, who explained their bills to them,
who get them involved in their projects. They took care of Congress members,
they returned their calls. They had a fluent relationship with Congress – with both
Liberals and Conservatives. They had majorities in Congress, but besides such
natural majorities, they crafted broader coalitions, political blocs beyond their own
party. And that was useful for them because, since this is a Congress with no party
discipline, there are issues on which even the members of the president’s party may
vote against [him]. (Interview, May 24, 2000)22
22 Rivera, a member of the Gavirista faction, voted in favor of impeachment on June 12, 1996.
160
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
Therefore, it was hard for some dissident Liberals to investigate the presi-
dent. According to the head campaign treasurer, 69 of the 111 votes against
impeachment were cast by legislators “who were treasurers themselves or
received money from the regional treasurers for the Samper campaign in
different districts” (Medina Serna 1997, 223).23
As a result of these factors, the Liberal Party (as well as some Conser-
vatives) consistently shielded President Samper from impeachment.24 The
president invested his political capital to guarantee his survival in office
rather than to push any controversial legislative agenda. In the end, the
investigative committee rejected most of the evidence supplied by the pros-
ecutor general and reported in favor of the president (Comisión Ciudadana
1996, 46–60). On June 12, 1996, the Chamber of Representatives finally
voted 111–43 to close the investigation. The Liberal Party remained cohe-
sive and the opposition divided: eighty out of ninety Liberals voted against
impeachment, while only twenty-seven Conservatives (out of forty-eight)
supported a trial.25
23 Medina Serna (1997, 223) erroneously noted that the total number of votes in favor of
Samper was 113 (instead of 111). Rodrigo Rivera, then president of the Chamber of Rep-
resentatives, believes that this factor was marginal in deciding the Liberal vote. Because
regional treasurers were not responsible for the fund-raising process, most legislators were
not at risk in spite of the scandal (interview, May 2000).
24 As a fourth explanation, Samper has claimed that American pressure (through decertifica-
tion) heightened nationalism among members of Congress and strengthened his position
(interview, May 22, 2000).
25 Six legislators of smaller parties voted for impeachment and ten others against. I follow
my own reconstruction of the roll call; because former Liberals and Conservatives tend to
disguise themselves as “independents,” figures vary slightly according to the source. See,
for instance, Bermúdez (1999, 55), Hinojosa and Pérez-Liñán (2003, 75), and Ungar Bleier
(1997, 77).
161
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
26 As explained in previous chapters, while the Supreme Court deliberated the issue a new
martial court had convened, revisited Oviedo’s case, and declared the general not guilty.
President Cubas thus claimed that there was no legal ground for him to recapture Oviedo.
162
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
logic of the Colorado Party – in the Colorado Party, the logic of building
majorities always prevails. . . . It would have been possible for the president
to work with the vice president, to pact.”
Even the opposition was willing to compromise – with one condition:
On December the 27 [1998], I had the last conversation with [President Cubas]. We
[the opposition] told him that we could guarantee governability and support for the
policies he had outlined, only if he could distance himself from Oviedo. No way. We
had to stop the meeting because he did not give a damn about his administration.
(interview with Gonzalo Quintana, June of 2002)
The tensions escalated in January of 1999, after the chief justice gave
Cubas seventy-two hours to capture Oviedo and reminded the president
that he could be impeached if he remained in contempt. Vice President
Argaña prepared for the takeover: by mid-March, the Committee of Con-
stitutional Affairs of the Chamber of Deputies had drafted the accusation
bill. Within days, the Chamber of Deputies modified its internal proce-
dures to reduce the number of votes required to authorize an impeachment
from 67 percent to 66.5 percent of the members present – a convenient
interpretation of the two-thirds clause that could eventually save a vote or
two. Inaugurating a new concept of filibustering, some Oviedista deputies
cut the cords of the chamber’s sound system in a vain attempt to block the
approval of the new rule (Ultima Hora, March 11, 1999, 2).
At the same time, the national convention of the Colorado Party voted on
March 14 to extend the terms of the Argañista authorities in control of the
party. After a short battle in which Colorado politicians attacked each other
with the convention chairs, the Oviedistas abandoned the meeting, formed
a parallel convention, and occupied the party headquarters. The steering
committee (Junta de Gobierno) of the ANR requested the intervention of
the judiciary to evict the intruders and issued a communiqué stating that
“President Cubas is not a representative of the Colorado Party.”
Roberto Céspedes has explained the rise of the confrontational style
of the Oviedismo in late 1998 and early 1999 as a response to its increasing
political isolation: the Supreme Court turned against Oviedo; the Colorado
convention outmaneuvered Unace; and Congress prepared for an impeach-
ment. Oviedo responded with an escalation of conflict against the Court,
the ANR, and Congress. “Without any citizen or partisan response, this
escalation continued until the assassination of Vice President Luis Marı́a
Argaña” (Céspedes 1999, 148).
163
P1: JZP
0521869423c06a cuny809-perez 0 521 86942 0 May 8, 2007 18:9
On March 17, the Oviedista deputies attempted their last trick to derail
the impeachment by bringing the accusations to the floor at a time when
the opposition lacked the necessary votes, but floor consideration of the
issue was rejected by a vote of 44–27. Instead, the chamber set a date for the
debate, April 7. Although it was not at all clear that Argaña could produce
a two-thirds majority by early April, it became evident that he was paving
his way to the presidency. The brutal response came a few days later, when
Argaña’s car was intercepted by a death squad and he and his bodyguards
were shot.
With the assassination, Cubas was left in a position of irreversible conflict
with the Democratic Front. Public opinion immediately blamed Oviedo for
the crime; the Argañistas feared for their lives; and the opposition parties
feared a democratic breakdown. In the chaotic hours that followed, the
president’s attempts to reestablish links with the Democratic Front were
doomed to failure. Cubas appointed his brother Carlos as the new minister
of the interior, but the congressional leaders did not show up when the
minister invited them to a meeting that night. Instead, they called for an
extraordinary session of the Chamber of Deputies to take place in the early
hours of March 24.
The deputies of the Democratic Front met in the chamber at dusk and
carefully conducted a head count. As they initiated the session at 8:00 A.M.,
the crowds outside Congress cried for impeachment. The speaker ordered
the security guards to ensure that nobody be allowed to enter the chamber
after the session had started. In the meantime, the crowds prevented an
Oviedista leader from reaching the building’s entrance, until he drove his
SUV over the building’s metallic fence. The deputy run into the building
holding a handgun and was intercepted by the guards, who disarmed him
and locked him in a room. Fearing the arrival of more Oviedista deputies,
the Democratic Front closed the debate and moved to a vote. The surprised
Oviedistas saw the impeachment charges approved by a margin of a single
vote (49–24). Within minutes, the Senate had been notified of the charges
against the president (Ultima Hora, March 24, 1999, 2; interviews with
Marcelo Duarte and Aristides González).
pact that drove González Macchi to the López Palace involved, for the
first time since 1946, the formation of a coalition cabinet. The Argañistas
and some former officials of the Stroessner dictatorship (“Jurassic Stron-
istas” in local press parlance) gained control of six cabinet positions (inte-
rior, defense, treasury, education, health, and public works), the Liberals
of two (foreign affairs and agriculture), and the Encuentro Nacional of the
remaining two (justice and industry) plus the office of planning. Part of the
agreement was that the coalition would present a common candidate for
the coming vice presidential election, and that he would be a Liberal.
After eight months of confrontational politics under Cubas Grau,
González Macchi presented himself as a natural negotiator. A pundit noted
“the friendly, parliamentary face of President Lucho. We had not seen cour-
tesy in the López Palace since [President] Rodrı́guez left office. The politics
of friendship is nowadays a topic of praise and study among political sci-
entists; it is not just gossip for the social pages or for the frivolous press”
(Rodrı́guez 1999, 22).27
The ruling coalition of the Democratic Front began to disintegrate later
in 1999, when internal disputes within the PLRA created tensions with the
president regarding the Liberal leader to be appointed as minister of foreign
affairs. González Macchi finally imposed his preferred Liberal candidate,
but the losing factions within the PLRA began to distance themselves from
the government (interview with José Moreno Rufinelli, June of 2002). In
February of 2000, 759 of the 1,091 delegates at the PLRA national con-
vention voted to abandon the government coalition and to nominate an
opposition candidate for the coming vice presidential election to replace the
late Argaña. The gamble paid off, and the Liberal candidate, Julio César
“Yoyito” Franco, won the election in August (interview with Miguel A.
Saguier, June of 2002).
Paradoxically, the Liberal triumph in the August election consolidated
González Macchi’s support among the Argañistas. They filled the two PLRA
cabinet positions and realized that any attempt to oust González Macchi
would drive Franco to the presidency, hurting their access to patronage and
their capacity to win the presidential election in 2003. At the same time,
González Macchi was able to preserve the support of Encuentro Nacional.
But after the national convention of the party decided to remain in the coali-
tion in April of 2000, internal voices of dissent grew louder. In September,
27 For studies of the “politics of friendship” in the Paraguayan context, see Frutos and Vera
(1993) and Caballero Carrizosa (1999).
165
P1: JZP
0521869423c06b cuny809-perez 0 521 86942 0 May 8, 2007 18:14
a group of Encuentristas led by Carlos Filizzola broke with the party and
formed a new organization called Paı́s Solidario (Solidary Country). In the
end, the PEN preserved seven seats in the lower chamber and its splinter
only two.
As the disgruntled Oviedistas regrouped their forces and later approached
the Liberals to form an opposition front, inconclusive “social pacts” became
part of the administration’s style. In June of 1999, the coalition government
and some social groups agreed on a government program that was rapidly
forgotten. In July, a new attempt at “concertación” reached no solid con-
clusions. Four months later, the administration called for a new round of
negotiations after a large peasant demonstration. At the time, the Liberals
were already distancing themselves from González Macchi (Ultima Hora,
June 15–16, 2002, 4).
By the end of 2000, the opposition camp was consolidated with the
votes of the Liberals, the Oviedistas, and the Paı́s Solidario members of
Congress. At the same time, corruption scandals began to erode the image
of the administration. The president was soon accused of using a stolen
BMW and of diverting $16 million from the central bank to private
accounts in the United States. These scandals opened the way for a con-
stitutional accusation led by the PLRA and the PEN dissidents (interviews
with Rafael Filizzola and Miguel A. Saguier, June of 2002). Early in July
of 2001, the government forces in the Chamber of Deputies attempted
to reject the impeachment charges, but they were defeated in a 45–34
vote that postponed the treatment of the issue. With the government
coalition unable to force a vote and the opposition incapable of reaching
the two-thirds majority, González Macchi was now barely shielded from
impeachment.
In late August of 2001, the church offered to coordinate a new round of
social “dialogue” to promote governability. As part of the negotiations, the
opposition in Congress finally agreed to subject the impeachment accusa-
tions to a vote even though that meant their defeat. On September 6, 2001,
the impeachment charges were brought to the Chamber of Deputies and
rejected in a 38–38 vote. The PLRA, the Oviedistas, Paı́s Solidario, and two
Encuentro Nacional dissidents supported the accusation. The Argañistas
and a majority of the PEN opposed the trial.
The dialog promoted by the church began in November of 2001 and
lasted until March of 2002, with few tangible results. The opposition soon
introduced new charges against the president for corruption and human
rights violations. In early June, when peasant demonstrators marched to
166
P1: JZP
0521869423c06b cuny809-perez 0 521 86942 0 May 8, 2007 18:14
Summary
The extensive historical discussion presented here has shown how the pat-
tern of executive-legislative relations established by a president early in his
term may shape the willingness of legislators to shield the executive from
impeachment at a later point. This view is consistent with Salo Coslovsky’s
interpretation of presidential crises in the 1990s. In his study of the Collor
and Bucaram cases, Coslovsky argued that “presidents that did not respect
the consultation and negotiation processes that characterize democracy
provoked a deadlock that was solved through impeachment” (Coslovsky
2002, 8). Table 6.3 summarizes the situation of each president by the time
the impeachment charges reached the floor of the legislature. The table
presents the main presidential strategy adopted with regard to Congress
and the structure of legislative support in terms of the analytical model
described in the previous section.
167
P1: JZP
0521869423c06b cuny809-perez 0 521 86942 0 May 8, 2007 18:14
Table 6.3. Presidential strategies and support for the executive in six crises
Support
Administration Strategy Vote Held in (Date) v d S A P L = P−v
Collor de Mello I Deputies (9/29/92) .34 .28 .06a .10 .12 −.22
Pérez I Senate (5/21/93) .50 .00 .48 .00 .00 −.50
Samper N House (6/12/96) .50 .90 .54 .23 .72 .22
Bucaram C Congress (2/6/97) .50 .95 .23 .24 .46 −.04
Cubas Grau I/C Deputies (3/24/99) .34 .57 .57 .00 .33 −.01
González Macchi N Deputies (9/6/01) .34 .77 .57 .07 .50 .16
Deputies (12/5/02) .34 .00 .57 .00 .00 −.34
Senate (2/11/03) .34 .48 .48 .20 .43 .09
Key: Strategies: I = isolation; N = negotiation; C = confrontation. Conditions: v = veto
threshold; d = support within the president’s party; S = size of the president’s party (at the time of
the vote if count is of members present); A = support provided by coalition members; P = overall
support at the impeachment vote (P = dS + A); L = legislative shield.
a Figure for PRN alone. Including PFL forces, S = .23 (d = .27).
this cohesion was insufficient to shield the president in the midst of the cri-
sis. Taken by surprise, the loyal legislators failed to prevent both Bucaram’s
declaration of incapacity and Cubas Grau’s impeachment.
By contrast, the remaining two presidents adopted a policy of negotiation
with Congress begining in their early days in office. Samper’s negotiating
style was not only the result of his affable personality and his experience
as a seasoned politician, but also the product of informal institutions and
coalition-building practices inherited from the days of the National Front.
President “Lucho” González Macchi also had the right personality and
professional experience for the role of negotiator, but his weak political
skills, the atomization of the Paraguayan party system, and the tyranny of
the electoral calendar – which first encouraged the defection of the Liberals
from his coalition and later placed the Argañistas in an ambiguous position –
increasingly left him isolated. Building on González Macchi’s weaknesses,
the opposition attempted to oust him from office time after time, each
attempt coming closer to success than the previous one. Yet in the end both
Samper and González Macchi bargained their way to the end of the term,
even though this meant surviving without governing.
Political scientists are often inclined to see presidential behavior as an
outcome that deserves explanation rather than as an independent factor
operating in the political process, and therefore may wonder whether pres-
idential styles are in fact a by-product of some other conditions that in turn
drive presidential crises. For instance, presidents may be more willing to
negotiate when they have a small party in Congress (or alternatively, they
may be more likely to isolate themselves when they face a hostile legis-
lature). I will address this problem in the following section, by analyzing
the effect of presidential style on legislative behavior while controlling for
other institutional and contextual factors. (I will also return to this issue in
Chapter 7, by looking at a broad sample of presidential administrations in
order to dispel risks of selection bias). Although the analysis of presidential
strategies often involves processing large amounts of historical information
and inevitably requires some degree of subjective assessment, a systematic
treatment of this independent variable may yield valuable insights for the
study of presidentialism in the future.
historical record suggests that the political environment also plays a critical
role in determining the propensity of legislators to support an impeachment
process. By “environment” I simply mean a set of factors that, although
exogenous to executive-legislative relations, have the capacity to affect leg-
islators’ careers and therefore to alter their incentives to remain loyal to
the president. The previous case studies indicate that three additional fac-
tors may be crucial to shaping the political environment in the midst of a
presidential crisis: constituency pressures, the mass media, and the electoral
calendar.
Constituency pressures, discussed in the previous chapter, are reflected
in the collapse of presidential approval ratings and the emergence of mass
protests calling for the resignation of the chief executive. Naoko Kada has
distinguished between the “preconditioning” and the “resolution” effects
of public opinion in the impeachment process. According to Kada, hos-
tile public opinion may encourage legislators to investigate the president
(preconditioning effect), but an incisive congressional investigation may
also fuel public outrage, which in turn may increase public pressure for
an impeachment (Kada 2003b, 148–49; see also Smulovitz and Peruzzotti
2000). Popular protest, a factor to be discussed in the concluding chapter
in greater detail, represents a stronger form of pressure because legislators
are confronted with a destabilizing escalation of public unrest.
The mass media, the focus of Chapter 4, also mold the environment for
impeachment by investigating and publicizing the accusations against the
president. The important question from the perspective of this chapter is
whether media effects take place only indirectly, by shaping public opinion
and constituency pressures, or also directly, by signaling to legislators the
right course of action. Direct media influence over Congress could take
place if legislators trusted the opinions or the investigations conducted by
the mass media, or if they feared becoming the focus of media criticism for
supporting the president. Meinke and Anderson, for instance, found that
the members of the U.S. House were less likely to support the president on
key bills during the Watergate, Iran–Contra, and Lewinsky scandals, even
after controlling for presidential approval ratings (Meinke and Anderson
2001). On the other hand, it is also possible that politicians take media
scandals seriously only when they ignite constituency pressures, but dismiss
the accusations if their voters fail to react.
The third factor deserving consideration is the electoral calendar. Pre-
sumably, the proximity of elections increases strategic behavior on the
part of politicians and potentially promotes party defection. According to
170
P1: JZP
0521869423c06b cuny809-perez 0 521 86942 0 May 8, 2007 18:14
In Brazil, I counted all members of the PDS, the PFL, the PL, the PRN,
and the PTB as part of the president’s coalition. In Paraguay, I took mem-
bers of the Oviedista faction as the only legislative bloc supporting Raúl
Cubas, and members of the Argañista faction plus the PEN as the coalition
supporting Luis González Macchi. In the other two cases, the president’s
coalition coincided with his party. In Colombia, where information about
factional identification was not available for the individual legislators, I sim-
ply assumed that the Liberals represented the president’s coalition and that
the Conservatives did not. In previous sections, I have shown that some
Conservative factions supported Samper’s policies while some Liberal fac-
tions opposed them, but it was impossible to identify the individual mem-
bers of these factions. In Ecuador, I counted the PRE as the only member
of the president’s coalition because by early 1997 no other party was willing
to support the Bucaram administration openly, even though the smaller
parties may have been willing to secretly form “ghost coalitions” (Mejı́a
Acosta 2003).
The dataset also includes two variables capturing the president’s strat-
egy vis-à-vis Congress. One dichotomous variable indicates the situation
of legislators facing an isolationist president (Collor and Cubas, since votes
for Pérez were not available), while a second dummy variable reflects
the situation of legislators dealing with a confrontational administration
(Bucaram and again Cubas Grau, given Oviedo’s style). Legislators inter-
acting with presidents prone to negotiation (Samper and González Macchi)
were treated as the reference category.
Finally, three additional variables depict the political environment at the
time of the vote. For the most part, these variables reflected conditions that
varied across administrations rather than across individual legislators. The
president’s approval rating for the month in which the vote took place was
used to capture constituency pressures (at the national level, since district-
level approval data was of course not available). The scandal index (described
in Chapter 5) corresponding to the same month was used to assess the role
of the media. And the impact of the electoral calendar was measured by the
number of days pending until the next legislative election. This indicator
ranged from 1,495 days at the time the Paraguayan House charged Raúl
Cubas Grau, to 75 days when the Senate acquitted Luis González Macchi. In
the Brazilian case, because a municipal election took place few days after the
impeachment vote, and because several observers have claimed that national
legislators usually pursue careers at the subnational level (Mainwaring
1999b; Samuels 2003; Samuels and Mainwaring 2004), I identified the
172
P1: JZP
0521869423c06b cuny809-perez 0 521 86942 0 May 8, 2007 18:14
eighty deputies running for mayor in the 1992 election and assigned them
a very short time horizon (4 days to the municipal election), while all
other deputies were assumed to have a longer horizon (734 days to the
1994 race).
Table 6.4 presents the results of three logistic regression models in which
the dependent variable is the individual legislator’s support for the presi-
dent (i.e., a vote against impeachment). The results indicate that presidents
who build extensive legislative coalitions are, not surprisingly, less likely to
confront impeachment charges. They also suggest that strategies based on
presidential isolation and, to a lesser extent, interbranch confrontation can
be lethal once an impeachment crisis explodes.
The coefficients for the electoral calendar and the constituency pres-
sure variables show the expected signs: the longer the time horizon and the
higher the presidential approval rating, the more likely legislators are to
173
P1: JZP
0521869423c06b cuny809-perez 0 521 86942 0 May 8, 2007 18:14
support the executive. However, the sign of the scandal index coefficient
defies the initial expectations – more scandals seem to encourage members
of Congress to rally around the president, suggesting that the media exer-
cise a powerful but indirect influence on the impeachment process. In the
absence of constituency pressures, most legislators presumably interpret
media scandals as political chicanery and close ranks with the executive.
In model 6.4.2, I replaced presidential approval with an alternative mea-
sure of public outrage (a dummy variable indicating whether mass protests
called for the resignation of the president) and obtained equivalent results.
Model 6.4.3 corrected for the undue influence of large legislatures. For
instance, the Brazilian Chamber of Deputies had 503 members, while the
Paraguayan Chamber of Deputies had just 80 members. The model simu-
lated a resizing of all chambers to the equivalent of one hundred seats. The
results remained unchanged.
Table 6.5 compares the observed and the predicted values for the key
parameter in the analytical model presented in this chapter: P, the pro-
portion of legislators favoring the president. The predictions are highly
accurate, with the exception of the ones referring to the final crisis of the
González Macchi administration.
According to model 4.6.1, about 14 percent of the deputies should have
voted against the impeachment of González Macchi in December of 2002
174
P1: JZP
0521869423c06b cuny809-perez 0 521 86942 0 May 8, 2007 18:14
(while in reality no one backed the president), and only 28 percent of the
senators should have supported the president in the trial vote a couple of
months later (as opposed to 43 percent). The discrepancies in this case
suggest that a complete model of legislative impeachment should probably
incorporate additional measures of strategic behavior. In particular, fur-
ther extensions of the model may need to capture the free-riding behavior
of lower house legislators. In cases like Paraguay and the United States,
where the president is not suspended from office before the impeachment
trial, lower house members may allow the accusation to proceed, knowing
that their partisan colleagues in the Senate will pay the political price for
shielding the president.
Although the evidence presented in this chapter indicates that institu-
tional factors play a major role in the impeachment process, it also suggests
that there is a causal relationship between the style of presidential leadership
and the strength of the legislative shield. Presidents who followed an isola-
tionist strategy found themselves in a particularly weak position to prevent
an impeachment process later in their terms. There was little uncertainty
about the outcome of the votes against Fernando Collor and Carlos Andrés
Pérez, because by the time of the crisis a large proportion of legislators were
already detached from the executive. By contrast, presidents who adopted
a confrontational style were in a stronger position even if they ultimately
failed to survive in office. The outcome of the vote was highly uncertain in
the cases of Abdalá Bucaram and Raúl Cubas Grau. Bucaram would have
survived an impeachment, but he did not control enough support to pre-
vent a declaration of incapacity. Cubas Grau’s supporters in the Paraguayan
Chamber of Deputies barely failed to prevent his impeachment by one vote.
A sustained pattern of negotiation with Congress clearly appeared to be the
best insurance against impeachment. The historical record shows that local
observers easily anticipated the outcome in the cases of Ernesto Samper and
(with the exception of the surprise in December of 2002) Luis González
Macchi, because the political environment was more stable and a larger
number of legislators were known to be loyal to the president.
175
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
176
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Argentina, 2001
President Fernando de la Rúa won the 1999 election as the candidate of
the Alianza, a coalition of the traditional Radical Party and the center-left
Frepaso (Novaro 2002; Ollier 2001). The new administration inherited –
and embraced – the so-called convertibility policy from its predecessor.2 In
1991, in response to hyperinflation, the Menem administration had estab-
lished a currency board, pegging the Argentine peso to the U.S. dollar at a
1:1 rate. This policy immediately stabilized the Argentine currency, brought
inflation down, and allowed for the return of credit and some years of eco-
nomic growth. Unable to rely on the inflationary tax, the Argentine gov-
ernment was forced to deal with its deficits by cutting expenditures and by
borrowing, since a comprehensive tax reform proved politically infeasible
and the revenues from privatization were soon exhausted (Corrales 2002b;
Schamis 2002). After 1999 the limitations of this model became evident
when Brazil, the largest partner in the Southern Cone Common Market,
devalued its currency and Argentine leaders realized that the exchange rate
policy was too rigid. As Argentine companies lost competitiveness in inter-
national markets, the unemployment rate skyrocketed to 17 percent by the
end of the first Menem administration (in October of 1995), remaining at
14 percent by the end of his second term (in October of 1999).3
The de la Rúa administration inherited a budget deficit and soon real-
ized that it had walked into a trap: any attempt to cut government spending
or raise taxes would deepen the recession, while any attempt to boost the
economy would worsen the fiscal imbalance. An orthodox strategy would
facilitate access to the financial markets in the short run, but it would com-
pound social tensions and compromise the federal tax base needed to serve
2 Because many people had contracted debts in dollars and because of the harsh memories of
hyperinflation, a large majority of the population opposed the devaluation of the Argentine
currency. A Gallup poll conducted in the greater metropolitan areas of Buenos Aires, Rosario,
Córdoba, Mendoza, and Tucumán on December 20, 2001 (already in the midst of the crisis)
showed that 79 percent of the respondents opposed the devaluation of the Argentine peso; 78
percent rejected a full dollarization of the economy; and only 45 percent wanted to abolish
the existing convertible peso (La Nación, December 23, 2001, 10).
3 See “Tasa de Desocupación Abierta en los Principales Aglomerados Urbanos,” Instituto
Nacional de Estadı́stica y Censos (INDEC), http://www.indec.mecon.ar/default.htm.
177
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
the debt in the long run. A gamble on economic growth, by contrast, would
restrict immediate access to credit because financial operators were antici-
pating a slow recovery and therefore a likely default. Unable either to cut
deficits or to curb unemployment, the administration increasingly projected
an image of ineffective rule (Bonvecchi 2002).
Despite electoral promises of transparent administration and a more
humane approach to the economy, the Alianza government was shaken in
October of 2000 when trade union leaders denounced the minister of labor
and the head of the intelligence service (a close friend of the president) for
bribing twelve senators in order to get an unpopular labor law passed.
As the scandal unfolded, presidential approval ratings dropped below
20 percent, and Vice President Carlos Alvarez, the leader of Frepaso,
resigned to protest the corruption scheme (Weyland 2002, 201–204).
According to Enrique Peruzzotti, the episode had deleterious effects
because the administration had made the fight against corruption one of
its prominent banners. Alvarez’s resignation and the subsequent fracture of
the government coalition around this issue indicated that the problem of
corruption affected all political parties (Peruzzotti 2005, 243).
The president responded to this challenge by increasing his political
isolation, “surrounding himself with ‘friends and family’ and shutting top
Radicals out of his inner circle” (Schamis 2002, 87). By early 2001, it was
clear that fiscal deficits had escalated, that access to the international capital
markets had become too costly, and that a recovery in tax collection was
unlikely. To make things worse, new authorities at the U.S. Treasury and the
International Monetary Fund were increasingly reluctant to offer assistance,
under the principle that unconditional support would encourage private
lenders to behave imprudently (Corrales 2002b). A radical attempt to slash
government expenditures by appointing a new minister of the economy
in early March met political resistance and failed within days (Bonvecchi
2002, 139–147). Unable to pull the country out of a three-year recession,
in March of 2001 the president appointed Domingo Cavallo – a former
minister under Menem and the architect of the convertibility policy – as
his economic czar, reinforcing his image as a weak leader who had virtually
abdicated policy making in favor of Cavallo.
Paradoxically, this appointment created uncertainty about the future of
the convertibility program. The new minister proposed to flexibilize the
fixed exchange rate by pegging the Argentine peso to a basket of currencies
that included the dollar and the euro. In July, facing skepticism in the finan-
cial markets, the administration turned to a more orthodox strategy and
178
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
announced a “zero deficit” plan that slashed public salaries, pensions, and
spending (Corrales 2002b).4 A side effect of this policy may have been the
reduction of local patronage resources and the disruption of the clientelistic
networks that distributed food among the poor – with catastrophic conse-
quences five months later (Auyero 2005). The Alianza lost the midterm
elections in October of 2001, and the Peronist Party (Partido Justicialista,
PJ) regained control of a plurality of the seats (45 percent) in the Chamber
of Deputies and retained a majority (56 percent) in the Senate.5
The outcome of the midterm election led the government to reorient its
strategy once again. In November, the Ministry of the Economy announced
new social programs, minor tax cuts, and a plan to restructure the public
debt. Investors reacted with skepticism, and capital flight accelerated. On
the evening of December 1, 2001, Domingo Cavallo announced a set of
desperate measures intended to prevent a run on banks and to save the
convertibility policy: for a period of ninety days, citizens would be allowed
to extract only 250 pesos (the equivalent of $250) in cash from their bank
accounts every week; all other transactions would be restricted to checks,
debit cards, and electronic transfers; and no person traveling outside of the
country would be allowed to take more than a thousand dollars.
The announcements activated a ticking bomb in a country that had reg-
istered 18 percent unemployment during the previous month. As Christmas
sales plummeted and the recession worsened, trade unions and shopkeepers
mobilized to protest the economic policy. Early demonstrations appeared
to be isolated episodes with little political weight, but in mid-December
food riots erupted in several provincial cities. By December 19, 2001, food
riots had reached the outskirts of Buenos Aires, unleashing a rapid chain of
events that led to the collapse of the administration.
During the afternoon, as riots spread throughout the country, de la Rúa’s
chief of staff requested the resignation of all ministers to open the way for
a coalition government. That night, in a vain attempt to show that he
was still in control of the situation, the president made a televised speech
announcing the declaration of a state of siege. Spontaneously, middle-class
4 Feldstein (2002, 12) erroneously claimed that Argentine wages did not decline during the
late 1990s. By contrast, the government cut public employees’ salaries by 15 percent in May
of 2000 and another 13 percent (including pensions for all retired people) in July of 2001
(Bonvecchi 2002, 125). The private sector followed the same course.
5 The midterm election sent a clear signal of popular discontent to the Argentine elite, as 25
percent of the registered voters did not show up at the polls (even though voting is legally
mandatory) and 21 percent of the voters annulled their ballots.
179
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
demonstrators defied the curfew, took to the streets, and began a pan-
beating protest demanding the resignation of Domingo Cavallo. When
police forces tried to repress acts of vandalism in downtown Buenos Aires
later that night, seven people were killed. A few hours later, the president
accepted the minister’s resignation, but in the meantime, intense nego-
tiations between Radical Party operatives and Peronist leaders to form a
coalition government had reached a deadlock.
The next morning, the Peronist leaders politely excused themselves from
meeting with the president. After two days of riots, demonstrations, and
police repression, more than 30 people had died, more than 400 had been
wounded, and more than 3,000 had been arrested all over the country. At
5 P.M., the leader of the Peronist bloc in the lower house announced that the
party was ready to initiate impeachment proceedings against President de la
Rúa. The president understood the message and submitted his resignation
to Congress.6
6 Pablo Calvo and Rolando Barbano, “Represión en Plaza de Mayo: Una Batalla con Cinco
Muertos,” Cları́n, December 21, 2001. Graciela Mochkofsky, Gabriela Litre, and Mariano
Obarrio, “El Fallido Plan de De la Rúa para Sobrevivir,” La Nación, December 23, 2001, 12.
180
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
As Marı́a Matilde Ollier (2003) has pointed out, the change in govern-
ment had not resolved two major problems driving the crisis – the depen-
dence on foreign credit and the fragmentation of political power. Within
two days, it became clear that President Rodrı́guez Saá was planning to
defy the transition pact and remain in office until the end of the de la Rúa
term in 2003. In an attempt to gain the support of various factions within
the Peronist Party, the new administration appointed some old-guard party
leaders with an unpopular record of corruption. The betrayed PJ governors
waited in silence.
On December 28, a federal judge ruled against restrictions on bank
accounts, authorizing one citizen to withdraw his savings. Later that after-
noon, however, the Supreme Court overruled the decision and upheld the
restrictions by a vote of seven of its nine members – ordering the benefi-
ciary of the initial ruling to return his money to the bank. That night, as
they learned about the judicial decision on the news, upset middle-class
demonstrators marched to the Supreme Court, calling for the resignation
of the justices, and to the government palace, demanding the removal of
public officials with well-known records of corruption. More than fifteen
thousand people camped in the Plaza de Mayo, in front of the presidential
palace. As young demonstrators began to burn tires and paint the doors of
the government palace with graffiti, the police charged with tear gas and
dissolved the demonstration. Some twelve officers were hurt; thirty-three
demonstrators were arrested; and a new wave of looting of banks in the sur-
rounding areas broke out. A few protesters entered the Congress building
and sacked its offices until they were dispersed.
The protests of December 28 showed that public outrage was directed
toward all politicians and institutions, not just against the defunct de la
Rúa administration (Armony and Armony 2005; Peruzzotti 2005, 246).
Even the Supreme Court, widely discredited after being taken over by
Menemista justices in the early 1990s, became a major target of public
outrage (see Helmke 2005; Larkins 1998). President Rodrı́guez Saá called
for an emergency meeting with the Peronist governors on December 30,
and all members of the cabinet submitted their resignations to allow
for a new round of negotiations within the party. However, only six
Peronist governors (of fourteen) showed up for the meeting, and new
demonstrations erupted outside the president’s summer house. Under-
standing that the leaders of the party had withdrawn their support in
retaliation for his attempt to remain in office until 2003, Rodrı́guez Saá
181
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Other Episodes
Between 1990 and 2004, four other presidents were forced to leave office
without the use of an impeachment or a declaration of incapacity. These
episodes took place in Guatemala in 1993, in Ecuador and Peru in 2000, and
in Bolivia in 2003.8 Reasons of space prevent me from discussing these cases
in great detail, but I offer a brief overview of them in order to concentrate
on their comparative aspects in the next section.
7 Mariano Obarrio, “La Fractura que Hizo Caer al Presidente,” La Nación, December 31,
2001, 4
8 Presidents were also ousted in Bolivia and Ecuador in 2005 – in ways that reinforced the
conclusions of this chapter.
182
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Ecuador, 2000 In 1998, Jamil Mahuad became the first elected pres-
ident after the fall of Abdalá Bucaram. But Mahuad’s situation deteri-
orated quickly in the midst of an economic crash that included a col-
lapse of the exchange rate, the freezing of bank accounts, and a mora-
torium on Ecuador’s foreign debt. The Mahuad administration proposed
a “dollarization” plan that resembled Bucaram’s convertibility program. As
in 1997, the Indigenous Federation (Conaie) opposed the measures and
led a popular uprising. This time, however, the social coalition against
the executive was narrower, and the legislative conditions for impeach-
ment were uncertain. Indigenous leaders overcame this obstacle through
an alliance with middle-rank military officers (Pallares 2006). On January
21, 2000, a coalition of indigenous demonstrators and military officers led
by Col. Lucio Gutiérrez occupied Congress and overthrew the president
(Herrera Aráuz 2001). Although unable to remain in power, President
Mahuad refused to submit his resignation, and the international community
conveyed the message that a military junta would not be welcome. Invert-
ing the Guatemalan situation, the top Ecuadorian generals outmaneuvered
the middle ranks and turned to Congress in search of a constitutional out-
come (Lascano Palacios 2001). Congress declared Mahuad out of office, and
183
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Vice President Gustavo Noboa was sworn in to complete the rest of the
term.
Peru, 2000 Peruvian President Alberto Fujimori took office in 1990 and
immediately adopted a strategy of confrontation with Congress. The pres-
ident was well known for his “arrogance towards political leaders, includ-
ing his own advisers, [and his] vitriolic slander of civilian institutions”
(McClintock 1993, 114). On December 6, 1991, six months before the
Collor scandal erupted in Brazil, Fujimori almost became the first case of
congressional removal in the decade. When he suggested that Congress was
under the influence of drug-trafficking lobbyists, legislators met overnight
and almost declared the president “morally unfit” to rule. The Senate
approved the motion unanimously, but the Chamber of Deputies, fearing
that they would incite a military coup, rejected the declaration of incapacity
in a 51–60 vote (Kenney 2004, 177–191).
In April of 1992, taking advantage of the low popularity of Congress and
his support among military officers, Fujimori closed the legislature, called
for the election of a constituent assembly, and rewrote the constitution
(Kenney 2004). (This episode was to inspire Guatemalan President Jorge
Serrano a year later.) By 1995, a significant decline in inflation, the cap-
ture of the leader of the main terrorist organization, and spending targeted
on social programs and public works allowed Fujimori to win reelection
and to keep a majority in Congress (see Stokes 2001, Chapter 5; Weyland
2002, Chapter 7). But by the end of the second term, the administration’s
corruption and authoritarian leanings had eroded its popularity. Fujimori
“won” the 2000 runoff election after the opposition candidate, Alejandro
Toledo, refused to participate, anticipating electoral fraud. The adminis-
tration confronted international criticism and increasing public unrest, and
the situation only worsened when a TV station aired a video showing the
head of the intelligence service and strongman of the regime bribing an
opposition congressman.
In October of 2000, sensing the weakness of the administration, politi-
cians began to realign. The vice president resigned, and at least ten members
of Fujimori’s legislative coalition switched parties. The president under-
stood that his legislative shield was growing thin. In a context of media
revelations, mounting demonstrations, and declining legislative support,
Fujimori opted for a fast exit. On November 17, 2000, he flew to Japan
and announced his resignation from Tokyo. Five days later, the stunned
184
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Peruvian Congress voted to reject his resignation and to impeach the run-
away president on the grounds of “moral incapacity.”
185
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Demilitarization
Chapter 3 documented a sharp decline in military interventions since 1978.
During the 1980s and 1990s, international and domestic factors greatly
reduced the incentives for military intervention in Latin American poli-
tics, and military officers saw their influence in domestic affairs diminished
(Hunter 1997). At the international level, the end of the Cold War encour-
aged changes in U.S. policy toward Central America and facilitated the
attainment of peace accords in El Salvador (1992) and Guatemala (1996).
The Organization of American States closed ranks to protect fledgling
democracies and amended its charter in 1997 to suspend any member
country when its “democratically constituted government is overthrown
by force” (Chapter III, article 9). The United States (at least until the
Venezuelan coup attempt of 2002) and the European Union consistently
signaled low tolerance for military turmoil, threatening to cut economic
and military aid to countries confronting the risk of a coup.
Internally, military intervention in politics was discredited by human
rights violations and the poor economic performance of military govern-
ments in the 1970s. Political learning altered the perceptions of both civilian
9 “The exact mechanism often comes down to a question of timing, such as whether Congress
can put together a formal impeachment process, settles for a hasty vote to remove the
president using other mechanisms, or the president simply resigns before formal proceedings
are initiated” (Hochstetler 2006).
186
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
10 The Southern Cone Common Market is formed by Argentina, Brazil, Paraguay, and
Uruguay, with Bolivia being an associate member.
187
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Watchdog Politics
A new context of greater government respect for freedom of the press,
combined with more powerful media outlets, has promoted the emergence
of what Waisbord (2000) called “watchdog journalism.” I documented in
Chapter 4 how elected presidents, their families, and cabinet members
became increasingly exposed to accusations and scandals. However, scandals
were not present in all episodes of presidential removal, and the evidence
suggests that exposés were not strictly necessary for the emergence of pub-
lic outrage against the president. As shown in Chapter 5 (and as illustrated
188
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
11 The three causes were the “betrayal of the motherland, graft, or crimes affecting national
honor” (article 59 of the 1979 constitution).
189
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
190
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
goal of all sectors. Finally, the presence of a legislative shield (L) was
acknowledged when the president controlled a legislative party or coali-
tion with a proportion of seats significantly greater than the seats required
to block an impeachment process. The dubious cases are those in which
the party had only a marginal advantage, was divided, or relied on an
uncertain coalition. Natural limitations of space prevent me from dis-
cussing the nuances of all coding decisions here, but more detailed case
192
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
13 The Balaguer case resembles what Mustapic (2005) called the “presidential” way of ter-
minating an administration – with the chief executive negotiating an anticipated election.
Mustapic identified the cases of Raúl Alfonsı́n (1989) and Eduardo Duhalde (2003) as exam-
ples of this pattern in Argentina. In the case of Balaguer, Congress approved the required
constitutional reforms; I have therefore counted this episode as a presidential crisis (see
also Marsteintredet and Berntzen 2006).
193
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Table 7.2. Conditions for impeachment and removal from office (qualitative comparative
analysis)
14 The analysis was conducted using fs/QCA 2.0. If every independent variable has j categories
and there are k variables, the number of feasible sufficient configurations (considering the
possibility that each variable may or may not have a significant impact) equals (j + 1)k − 1.
194
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
for the removal of the president (7.2.2). Models 7.2.1 and 7.2.2 treat all
variables as dichotomous factors (without any adjustment for the question
marks in Table 7.1; that is, they rely on conventional crisp sets). Given the
small number of cases, all tests in Table 7.2 are “veristic”: the presence of a
deviant case is enough to eliminate a particular solution (Ragin 2000).
The results of the analysis reinforce the previous conclusions: the absence
of military intervention, the presence of scandals, and a weak legislative
shield are necessary for an impeachment to take place. However, those
conditions do not pass a test of sufficiency because in two cases, Peru in
1991 and 2000, the three elements were present to some extent, but no
impeachment took place (however, note that Congress attempted a decla-
ration of incapacity in 1991 and declared Fujimori unfit for the post after
he had resigned in 2000).
Model 7.2.2 indicates that, during this historical period, popular upris-
ings were the only condition necessary to remove an elected president from
office. Uprisings often interacted with other factors (military action, as
in Ecuador in 2000, or partisan opposition, when the president lacked a
legislative shield) to trigger the collapse of an administration. The table
also reports the simplifying assumptions made to reach those conclusions:
although no case in Table 7.1 displays military intervention combined with
protests and legislative support, or protests without scandals and no leg-
islative shield, it is assumed that such configurations would also trigger the
fall of an administration.
Models 7.2.3 and 7.2.4 accommodate some of the historical nuances
discussed earlier by treating the analytical categories as fuzzy sets (Zadeh
1965). Conventional comparison assumes “crisp” categories: cases belong
to a particular category (set) or not (e.g., Sartori 1970). The analysis of
fuzzy sets relaxes this restriction, allowing for a continuous membership
function ranging between zero (full exclusion from the set) and one (full
inclusion in the category). The assignment of partial membership scores
may rely on various procedures (Verkuilen 2005); for the sake of simplicity,
I will just assume that cases with partial membership in any of the cate-
gories presented in Table 7.1 always lie at the crossover point (0.5). The
goal of this procedure is not just to “split the difference” in cases that involve
difficult coding decisions. By acknowledging the ambiguity of the histori-
cal record explicitly, this strategy forces us to reconsider the results of the
previous comparative analysis and to verify the results under an alterna-
tive coding scheme. Because categories are not strictly dichotomous, the
equivalent of negation (“not a member of ”) for fuzzy sets corresponds to
195
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
1 − m, where m is the membership score for the case in any given set.
Membership in an intersection (e.g., M and S) is determined by the mini-
mum membership score in any of the components, while membership in a
union (M or S) is determined by the maximum membership score in any of
the sets.
The second column in Table 7.2 presents the results of the analysis
based on this approach. Two of the factors previously identified (no mil-
itary intervention and scandals) were confirmed as necessary conditions
for an impeachment, but a weak legislative shield was dropped from the
list. This change was driven by four cases in which an impeachment took
place despite partial legislative support (Paraguay in 1999 and 2002) or in
which an impeachment was attempted even though the president had artic-
ulated a shield (Colombia in 1996 and Paraguay in 2001). The presence
of the four “deviant” cases also prevented the detection of any sufficient
configurations.
Similarly, model 7.2.4 confirmed the role of popular uprisings as nec-
essary conditions for the fall of elected presidents. The sufficient config-
urations detected in the sample suggest that popular protests may lead to
this outcome in at least three different ways: by prompting a military inter-
vention against the president (Ecuador in 2000) or preventing an autogolpe
(Guatemala in 1993);15 by empowering an adversarial majority in Congress
(which may initiate an impeachment); and by destabilizing an unpopular
government, even if no scandals have taken place and therefore, according
to model 7.2.3, an impeachment is not feasible.
The last two models incorporate questions about the relevance of
executive-legislative conflict into the fuzzy set analysis. The assumption
in those models is that the twenty-four cases that constitute the sample
may belong to the underlying population to varying degrees. The results of
this analysis suggest, not surprisingly, that a significant level of executive-
legislative conflict is necessary for an impeachment to take place, and that
mass protests may be sufficient to force the exit of the president even in
the absence of interbranch confrontation. However, given that all cases in
Table 7.1 have at least some degree of membership in the set of presiden-
tial crises, the number of simplifying assumptions in this analysis is quite
large.
15 Note the ambiguity of the M term in Model 7.2.4, which may indicate that the protests
support the military or that they oppose their action.
196
P1: JZP
0521869423c07 cuny809-perez 0 521 86942 0 May 8, 2007 19:13
Quantitative Evidence
The last column in Table 7.2 underscores the need to analyze a broader
sample of cases including administrations that never faced a presidential
crisis. Consider, for instance, the findings about the legislative shield pre-
sented in Chapter 6. The evidence suggests that presidents who are willing
to negotiate with the opposition from the outset are less likely to confront an
impeachment crisis later in the term. At the same time, willingness to logroll
may be correlated with being a minority president. For reasons of per-
sonality or for other reasons, some minority presidents (such as Fernando
Collor, Alberto Fujimori, and Abdalá Bucaram) may not be willing to nego-
tiate, but most minority presidents probably will. If this is the case, weak
partisan powers could be a blessing because they would force presidents to
bargain with the legislature from the first day in office. But this strategic
behavior would create a paradox: presidents with smaller parties (a small
parameter S according to the terminology of Chapter 6) would also work
harder (and earlier) to secure broader coalitions (a larger A) and also to
keep their followers happy (achieving greater cohesion, d). By looking only
at presidential crises, we would not be able to perceive this effect because
only “unreasonable” minority presidents would self-select into the sample.
The obvious solution to this problem is to select a representative sample
of administrations irrespective of whether they confronted a presidential
crisis. Unfortunately, the small number of impeachments and declarations
of incapacity complicates any quantitative analysis. Moreover, presidential
crises are better analyzed as a process than as a discrete outcome. There are
at least three stages in the process leading to a successful impeachment: the
emergence of the presidential crisis, the decision by Congress to authorize
(or block) an impeachment or declaration of incapacity, and the removal
of the president from office. As the previous sections of this chapter have
shown, different variables may play different roles at these stages.
To develop a large-N comparative test, a research team at the Uni-
versity of Pittsburgh collected information on seventy-five presidential
administrations in eighteen Latin American countries between 1990 and
2004.16 The dataset includes all administrations considered democratic or
16 Data collection was supported by the Center for Latin American Studies and by the Faculty
of Arts and Sciences at the University of Pittsburgh. I am indebted to Andrea Castagnola,
Agustı́n Grijalva, Germán Lodola, and Juan Carlos Rodrı́guez Raga for their assistance
with this project.
197
P1: JZP
0521869423c07a cuny809-perez 0 521 86942 0 May 8, 2007 19:11
17 Latin American Weekly Report (London: Latin American Newsletters Ltd., 1990–2001).
18 As shown in Chapter 6, in most countries the lower (or only) house authorized the impeach-
ment process. Exceptions to this pattern were Bolivia (where a joint session authorizes the
impeachment, which is conducted by the Supreme Court) and Venezuela before 1999, where
the Senate authorized a trial conducted by the Supreme Court.
19 For four administrations, I used the percentage of seats controlled by the president’s
coalition rather than the share of his or her party. Three of them corresponded to the
Concertación’s governments in Chile (Aylwin, Frei, and Lagos). I treated the size of the
Concertación as a better indicator of the level of support for the president in Congress
because Carey (2002) has shown that the coalition behaves cohesively as a unified actor.
The other case was the Endara administration in Panama, where I treated the anti-PRD
198
P1: JZP
0521869423c07a cuny809-perez 0 521 86942 0 May 8, 2007 19:11
coalition as a unit, and adjusted its size to account for its decline after a couple of years. (A
similar case was the Chamorro administration in Nicaragua, but the Nicaraguan constitu-
tion did not contemplate an impeachment process until 1995.)
20 If LAWR reported factional disputes in the ruling party, I computed the legislative shield
as L = P/3 − v.
199
P1: JZP
0521869423c07a cuny809-perez 0 521 86942 0 May 8, 2007 19:11
Table 7.3. Predictors of presidential crises, impeachments, removals, and mass protests,
1990–2004
21 Military interventions in Peru (1992) and Guatemala (1993) were not coded because they
were initially directed against Congress and not against the president. Additional tests
showed that including the Guatemalan crisis would not alter the results.
200
P1: JZP
0521869423c07a cuny809-perez 0 521 86942 0 May 8, 2007 19:11
able to play this proactive role because repeated media scandals had eroded
the president’s credibility and mass protests had made the survival of the
administration untenable.
Alternative outcomes occurred when military officers ignored the his-
torical conditions that restricted their role in politics, or when the pres-
ident resigned, anticipating his downfall. The first example of deviation
from impeachment happened when the armed forces disregarded the trend
toward demilitarization and sought to disrupt the democratic process.
Although these coups ultimately failed, they accelerated the fall of pres-
idents in Guatemala (in 1993) and Ecuador (in 2000). The second kind of
deviation happened when the president acted strategically before Congress
could consider his removal. This pattern was characteristic of situations
marked by spiraling social conflict, when presidents feared facing criminal
charges (or furious mobs) unless they could manage their exit from power.
Popular protests outpaced the articulation of a legislative front, and the
president resigned before an impeachment process could be set in motion.
In those cases, Latin American legislators worked reactively, providing
a constitutional framework for the government transition once the fall of
the administration was already in sight. In Guatemala in 1993, in Ecuador
and Peru in 2000, in Argentina in 2001, and in Bolivia in 2003, legislators
contained the disruptive effects of the crisis by legitimizing the president’s
removal – and sometimes by offering a parliamentary mechanism to replace
the ousted leader (Marsteintredet and Berntzen 2006; Mustapic 2005).
202
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
politicians’ careers, and for the formation of public policy.1 From time to
time, country specialists were puzzled by crises without breakdown, but
they tended to analyze those episodes in isolation (see Chapter 2).
What are the theoretical and normative implications of the new pattern
of political instability? Are crises without breakdown mere functional equiv-
alents of traditional military coups, or do they demonstrate the presence
of checks and balances in Latin America? Are they signals of democratic
erosion, or of democratic vitality? Students of comparative politics have
only begun to address those questions, and I suspect that we will see heated
debates in the years to come. Although analysts are naturally inclined to
interpret presidential crises according to their own intellectual frameworks
and ideological propensities, debates will not be settled easily because crises
without breakdown leave multiple and sometimes contradictory legacies.
Among them are the ability of civilian elites to resolve conflicts without mil-
itary intervention, the emergence of unexpected checks and balances, and
the capacity of social movements to exercise vertical accountability through
insurrectional politics. Those legacies – and the paradoxes they embody –
constitute the focus of this concluding chapter. Exploring those paradoxes
may not yield conclusive answers, but it will force us to rethink much of
our current understanding of Latin American presidentialism.
204
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
democratic regime (Lamounier 1994; Linz 1990; Riggs 1988; Shugart and
Mainwaring 1997; Stepan and Skach 1993; Valenzuela 1994).
In the 1990s, new voices questioned this view by pointing out that not
every form of presidentialism is equally dangerous. The argument was
reframed in terms of destabilizing veto players (Tsebelis 1995, 321–322).
According to Cox and McCubbins, the separation of powers characteristic
of presidential constitutions breeds deadlock when the two branches also
display a separation of purpose (Cox and McCubbins 2001). Thus, the pres-
ence of different partisan players in control of the two branches of govern-
ment was seen as a potential source of turmoil. Scott Mainwaring showed
that the interaction between presidentialism and multipartyism constitutes
a difficult combination for democracy (1993, 212–213), and Mark Jones
claimed that stable presidentialism usually requires an executive able to
command a legislative majority (Jones 1995, 160).
As the debate on the perils of presidentialism cast shadows on the
fate of the new Latin American democracies, the multiplication of crises
without breakdown in the 1990s challenged the terms of the debate
(Mainwaring 1999a, 109–110; Pérez-Liñán 2003b). How should these cases
be interpreted? Were they another manifestation of presidential instability,
a challenge to the theory, or just a few irrelevant anomalies?
For Arturo Valenzuela, the collapse of fourteen presidential admin-
istrations2 between 1985 and 2004 simply underscored the well-known
problems of presidentialism: in presidential systems, disputes about spe-
cific grievances may easily escalate into debates about whether the chief
executive should resign, and the related polarization may in turn transform
the governmental crisis into a full-blown crisis of the constitutional order
(Valenzuela 2004, 12).
There is little doubt – and the historical evidence presented in Chapter
3 proved this point – that the underlying propensity of presidential systems
toward interbranch confrontation has not subsided after the third wave of
democratization. But Valenzuela’s interpretation poses an obvious question:
if current presidential crises do not trigger democratic breakdowns, where
exactly do the perils of presidentialism lie? Instability is now installed at the
level of the government, not at the level of the regime. Based on this fact,
most studies of recent presidential crises have drawn an almost diametrically
2 In addition to the ten cases discussed in Chapter 7, Valenzuela discussed the fall of presidents
Hernán Siles Zuazo of Bolivia (1985), Raúl Alfonsı́n of Argentina (1989), and Jean Bertrand
Aristide of Haiti (in 1991 and again in 2004).
205
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
The concern with unconstrained presidents has survived into the present,
sometimes inspiring stereotypical representations. Maybe the most dra-
matic depiction of Third World presidentialism was put forward by Carles
Boix, who suggested that because elected presidents are not fully account-
able to the other branches of government, they may seize most of the
nation’s fixed assets and impose a dictatorship (Boix 2003, 152).
A much more nuanced image of presidential dominance pervaded the
debate about delegative democracy in the 1990s. Guillermo O’Donnell
argued that many new democracies “rest on the premise that whoever wins
election to the presidency is thereby entitled to govern as he or she sees
fit, constrained only by the hard facts of existing power relations and by a
constitutionally limited term of office” (O’Donnell 1994, 59).
O’Donnell’s ideal type emphasized three different but related problems:
the absence of horizontal accountability (i.e., control over the executive
exercised by Congress or other institutions); a pattern of politically isolated
and highly technocratic economic decision making; and a cycle leading from
presidential “omnipotence” early in the president’s term to political “impo-
tence” by the end of the administration. Subsequent work on this topic
emphasized the problems of horizontal accountability (e.g., Larkins 1998;
O’Donnell 1998) and unilateral policy making (e.g., Ghio 2000; Panizza
2000), but for the most part ignored the cyclical nature of presidential
207
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
power that became so evident in most presidential crises (as partial excep-
tions, see Helmke 2002; Schmidt 2000; Weyland 1993).
The multiplication of impeachments in the early 1990s posed an impor-
tant question for this traditional view. In the absence of horizontal account-
ability, how could Congress have challenged presidential dominance, held
the president accountable, and removed him from office time after time?
The first possible solution to this puzzle is that impeachments took place
in countries that were not delegative democracies. But this answer is clearly
unsatisfactory. Brazil, the nation that initiated the wave in 1992, was one
of the cases that O’Donnell identified with the ideal type, and among the
examples of weak horizontal accountability were established polyarchies
such as Colombia and Venezuela (O’Donnell 1998, 112).
A second way to account for this paradox is by discarding the concept
of delegative democracy altogether, claiming that it fails to reflect the true
attributes of Latin American presidentialism in the 1990s. Neo-institutional
studies have revisited the role of legislatures vis-à-vis the executive branch,
noting that delegation is one among several possible strategic interactions
between the executive and Congress (Carey and Shugart 1998; Cox and
Morgenstern 2001; Palanza 2006). Thus, when legislators failed to exercise
their oversight they may have done so for strategic reasons rather than as a
result of institutional impotence. Congressional leaders may have realized
that challenging a popular president would not help their political careers,
or they may have wanted the executive branch to take sole responsibility
for unpopular policies. Yet, when confronted with popular demands to do
so, they were fully capable of conducting an investigation and removing the
president from office.
The problem with this answer is that, paradoxically, it ignores the strate-
gic effects of horizontal accountability. It is true that the multiplication of
impeachments proved that congressional oversight could ultimately be exer-
cised, but these outbursts of congressional action represented an intermit-
tent form of accountability, very different from the institutionalized form
that is (at least in theory) characteristic of well-established democracies.
Like any other form of law enforcement, impeachment is successful only
when its shadow discourages the perpetration of misdeeds. This strategic
effect is imperfect anywhere, but the surprise and skepticism with which
many Latin Americans observed the unfolding of impeachment processes
indicates that horizontal accountability was not an institutionalized trait of
these regimes.
208
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
I suspect that there is a third and more complex answer to this puz-
zle, an answer that fits O’Donnell’s thesis of presidential power cycles as
well as Smulovitz and Peruzzotti’s thesis about the emergence of societal
accountability in Latin America (Peruzzotti 2005; Smulovitz and Peruzzotti
2000). In the 1990s, the multiplication of impeachments signaled the emer-
gence of a model of checks and balances consistent with, and embedded in,
the new pattern of political instability. The search for presidential account-
ability took place in a context marked by a traditional Latin American
dynamic – executive dominance punctuated by presidential failure – but
also by new conditions, namely, the inability of military leaders to take over
and the related emergence of new decisive actors like the mass media and
the protest movements (Smulovitz and Peruzzotti 2000).
The model of horizontal accountability that emerged in this context
can best be characterized as politicized and spasmodic. By “politicized,” I
mean that short-term electoral, partisan, and personal considerations drive
legislators’ decisions about impeachment as much as their concerns about
presidential dominance or institutional maintenance (Mayhew 1974). By
“spasmodic,” I mean that the impeachment mechanism is activated inter-
mittently as a way to handle extreme situations, without a more continu-
ous exercise of congressional oversight during “normal” times (see Siavelis
2000). The elites used impeachments as a way to control presidents who had
become too unpopular, too unpredictable, or too unwilling to compromise.
But institutional checks were activated intermittently in order to dethrone
undesirable presidents rather than to prevent presidential dominance and
misdemeanors at an early stage.
This position lies somewhere between a pessimistic image of Latin
America as a vast institutional wasteland and an optimistic image of soci-
eties governed by well-established “rules of the game.” In a presidential
regime deprived of any effective mechanism of horizontal accountability,
we would expect presidential excesses to be common and impeachment to
be unknown. In a context of strong horizontal accountability, we would
expect high crimes and misdemeanors to be rare – and impeachment to
be equally unusual. By contrast, presidential misbehavior was frequent in
Latin America during the 1990s, but impeachment emerged as an unex-
pected phenomenon. Impeachments proved that legislators were strong
enough to hold the president accountable when media scandals and popu-
lar protest gave them enough leverage, but not strong enough to discourage
the occurrence of presidential abuse on a regular basis.
209
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
210
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
but the officers have no right to speak for the people. Second, a military
revolt by definition violates a democratic taboo – the use of violence to
achieve power. By contrast, an ideal-typical uprising seems to reverse those
assumptions: it is the people who make their voices heard in the streets, and
any form of violence – to the extent that it takes place – is just indicative
of a desperate situation, proof that the elected rulers have not represented
the people’s interests.
Time after time, these two assumptions have informed the understand-
ing of popular uprisings. Unfortunately, the difference between a “popular
impeachment” and a “popular coup” is often blurred, as the example of
January 2000 in Ecuador illustrates (Pallares 2006; see also Munck 2006,
11). For this reason, the two underlying assumptions deserve some careful
consideration.
First, there is the issue of representation. In principle, there is no clear
democratic principle to support the argument that protests should trump
votes. In other words, why should the will of the people currently trying
to oust the president prevail over the will of the people who cast votes
in favor of the same president in the last election? Working in the midst
of crises, politicians have eschewed abstract philosophical debates in favor
of practical solutions to this dilemma. The evidence presented in previous
chapters has shown that the scope of social protests is critical to determining
their political leverage. The presence of a broad street coalition conveys the
message that something resembling a supermajority wants the president out
of office.
Unfortunately, after a broad coalition succeeds in deposing a president,
the task of representation only begins. Who has the right to set the new
agenda? Interpretations of popular protest often assume that uprisings have
unified goals, convey clear mandates, and have foundational consequences
for the political system; but the evidence in this regard is mixed. Just like the
horizontal accountability exercised by legislators, the vertical accountabil-
ity exercised by social movements is often politicized and spasmodic. Two
dynamics account for this outcome: popular movements may divide over the
new agenda and ultimately play into the hands of traditional actors, or they
may displace those actors from power only to create new unaccountable
governments.
Popular uprisings do not occur in isolation, and they are often shaped
by complex interactions between emerging social forces and conventional
political structures (Pallares 2006). The experience of Argentina after the
2001 crisis is illustrative. The middle-class assemblies that proliferated
211
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
but at the same time has legitimized the potential use of popular protest
as a “democratic” form of praetorian politics. As a positive by-product,
this asymmetric situation has encouraged the emergence of a restrained
style of presidential rule. In the early 2000s, presidents as dissimilar as
Eduardo Duhalde and Néstor Kirchner in Argentina or Hugo Banzer and
Carlos Mesa in Bolivia bowed to the proliferation of popular protests, try-
ing to funnel them through a dense web of negotiations, social programs,
and patronage rather than confronting them with security forces (Gamarra
2004).
A debate about the democratic legitimacy of insurrectional politics is still
pending. Defenders of popular uprisings will be tempted to dress them in
epic narratives, but they will have a hard time convincing other democrats
that social movements have an intrinsic right to forceful political action
while other nonstate actors should commit to human rights and abide by
the rule of law. Critics of popular uprisings will be tempted to cast this
problem as a Hungtingtonian tension between social incorporation and
democratic institutions, but they will have a hard time convincing other
democrats that this approach is not a latent justification for social exclusion
and institutional rigidity.3
Maybe the most important lesson taught by recent presidential crises
is that presidents who exercise virtually unconstrained power in a context
of high popularity often become easy targets when their approval ratings
plummet and mass protests consume their political capital. The impeach-
ment process has emerged as the most effective way to manage the downfall
of an elected administration while protecting a democratic constitution, but
it has generally failed to prevent a new cycle of presidential dominance and
government collapse. Military coups may be the drama of the past, but there
are reasons to believe that crises without breakdown will be Latin America’s
drama for years to come.
3 Interestingly, the recent popular uprisings proved that conflict-ridden incorporation is not
simply a by-product of the modernization process. Street coalitions were often formed by
the poor, but they did not represent new social groups created by the process of economic
development. (For the classic work on this topic, see Huntington 1968.)
213
P1: JZP
0521869423c08 cuny809-perez 0 521 86942 0 May 8, 2007 19:22
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Abente Brun, Diego. 1999. “People Power in Paraguay.” Journal of Democracy 10(3):
93–100.
Acosta, Alberto. 1997. “El Bucarmismo en el Poder.” In ¿Y Ahora Qué? Una Con-
tribución al Análisis Polı́tico-Histórico Actual. Quito: Eskeletra Editorial, 47–90.
Acosta Cerón, Ramiro. 1997. “El Levantamiento del 5 de Febrero: Una Visión
Jurı́dica.” In 5 de Febrero: La Revolución de las Conciencias. Quito: CECS–
FETRAPEC–Fundación José Peralta, 85–111.
Acuña, Carlos, and William C. Smith. 1994. “The Political Economy of Structural
Adjustment: The Logic of Support and Opposition to Neoliberal Reform.” In
Latin American Political Economy in the Age of Neoliberal Reform, ed. William C.
Smith, Carlos H. Acuña, and Eduardo A. Gamarra. Miami: North-South Center
Press, 17–66.
Agresti, Alan. 1996. An Introduction to Categorical Data Analysis. New York: John
Wiley and Sons.
Allen, Philip M. 2003. “Madagascar: Impeachment as Parliamentary Coup d’Etat.”
In Checking Executive Power, ed. J. C. Baumgartner and N. Kada. Westport, CT:
Praeger, 81–94.
Altman, David. 2000. “The Politics of Coalition Formation and Survival in Multi-
Party Presidential Democracies: The Case of Uruguay, 1989–1999.” Party Politics
6(3): 259–283.
Alves, Rosental Calmon. 1997. “Democracy’s Vanguard Newspapers in Latin Amer-
ica.” Paper presented at the forty-seventh annual conference of the International
Communication Association, Montreal, Quebec, Canada, May 22–26.
Ames, Barry. 2001. The Deadlock of Democracy in Brazil. Ann Arbor: University of
Michigan Press.
Amorim Neto, Octavio. 2002. “Presidential Cabinets, Electoral Cycles, and Coali-
tion Discipline in Brazil.” In Legislative Politics in Latin America, ed. S. Morgen-
stern and B. Nacif. Cambridge: Cambridge University Press, 48–78.
Armony, Ariel C., and Victor Armony. 2005. “Indictments, Myths, and Citizen
Mobilization in Argentina: A Discourse Analysis.” Latin American Politics and
Society 47(4): 27–54.
215
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
216
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
217
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Castañeda, Jorge G. 1997. “Limits to Apertura: Prospects for Press Freedom in the
New Free-Market Mexico.” In A Culture of Collusion: An Inside Look at the Mexican
Press, ed. William A. Orme, Jr. Miami: North-South Center Press, 131–140.
Cepeda Ulloa, Fernando. 1996. “El Congreso Colombiano ante la Crisis.” In Tras
las Huellas de la Crisis Polı́tica, ed. Francisco Leal Buitrago. Santafé de Bogotá:
TM Editores–FESCOL–IEPRI, 75–97.
Céspedes, Roberto L. 1999. “Los Actores Sociales en el Marzo Paraguayo de 1999.”
In Marzo de 1999: Huellas, Olvido y Urgencias, ed. J. N. Morı́nigo. Asunción:
Universidad Católica, 145–178.
Chappell, Henry W., and William R. Keech. 1985. “A New View of Political
Accountability for Economic Performance.” American Political Science Review
79(1): 10–27.
Chehabi, H. E., and Juan J. Linz. 1998. Sultanistic Regimes. Baltimore: Johns
Hopkins University Press.
Cheibub, José Antonio. 2002. “Minority Governments, Deadlock Situations, and
the Survival of Presidential Democracies.” Comparative Political Studies 35(3):
284–312.
Chernick, Marc. 1997. “The Human Rights Crisis in Colombia: It’s Time to Inter-
nationalize the Peace Process.” LASA Forum 28(3): 20–23.
Chitty La Roche, Nelson. 1993. Doscientos Cincuenta Millones: La Historia Secreta.
Caracas: Pomaire.
Cima. 1998. Informe Anual de Opinión Iberoamericana, 1997–1998. Bogotá: Consorcio
Iberoamericano de Empresas de Investigación de Mercados y Asesoramiento.
Collier, David, ed. 1979. The New Authoritarianism in Latin America. Princeton, NJ:
Princeton University Press.
Collier, David, and James Mahoney. 1996. “Insights and Pitfalls: Selection Bias in
Qualitative Research.” World Politics 49(1): 56–91.
Collier, David, James Mahoney, and Jason Seawright. 2004. “Claiming Too Much:
Warnings about Selection Bias.” In Rethinking Social Inquiry: Diverse Tools, Shared
Standards, ed. H. E. Brady and D. Collier. Lanham, MD: Rowman and Littlefield,
85–101.
Collor de Mello, Pedro. 1993. Passando a Limpo: A Trajetória de Um Farsante. Rio
de Janeiro: Record.
Comisión Ciudadana de Seguimiento (Colombia). 1996. Poder, Justicia e Indignidad:
El Juicio al Presidente de la República Ernesto Samper Pizano. Santafé de Bogotá:
Utópica Ediciones.
Conaghan, Catherine M. 1996. “Public Life in the Time of Alberto Fujimori.”
Working Paper No. 219. Washington, DC: Woodrow Wilson International Cen-
ter for Scholars.
Conaghan, Catherine M. 1998. “The Permanent Coup: Peru’s Road to Presidential
Reelection.” LASA Forum 29(1): 5–9.
Conaghan, Catherine M. 2005. Fujimori’s Peru: Deception in the Public Sphere. Pitts-
burgh: University of Pittsburgh Press.
Coppedge, Michael. 1994a. “Prospects for Democratic Governability in
Venezuela.” Journal of Interamerican Studies and World Affairs 36(2): 39–64.
218
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Coppedge, Michael. 1994b. Strong Parties and Lame Ducks: Presidential Partyarchy
and Factionalism in Venezuela. Stanford, CA: Stanford University Press.
Corrales, Javier. 1997. “El Presidente y Su Gente: Cooperación y Conflicto entre los
Ámbitos Técnicos y Polı́ticos en Venezuela, 1989–1993.” Nueva Sociedad (152):
93–107.
Corrales, Javier. 2000. “Presidents, Ruling Parties, and Party Rules: A Theory on
the Politics of Economic Reform in Latin America.” Comparative Politics 32(2):
127–49.
Corrales, Javier. 2002a. Presidents without Parties: The Politics of Economic Reform in
Argentina and Venezuela in the 1990s. University Park: Pennsylvania State Uni-
versity Press.
Corrales, Javier. 2002b. “The Politics of Argentina’s Meltdown.” World Policy
Journal 19(3): 29–42.
Coslovsky, Salo Vinocur. 2002. Neoliberalism, Populism, and Presidential Impeachments
in Latin America. M.A. thesis, Tufts University.
Cox, Gary W., and Mathew D. McCubbins. 2001. “The Institutional Determi-
nants of Economic Policy Outcomes.” In Presidents, Parliaments, and Policy, ed.
S. Haggard and M. D. McCubbins. Cambridge: Cambridge University Press,
21–63.
Cox, Gary W., and Scott Morgenstern. 2001. “Latin America’s Reactive Assemblies
and Proactive Presidents.” Comparative Politics 33(2): 171–189.
Crisp, Brian F. 2000. Democratic Institutional Design: The Powers and Incentives
of Venezuelan Politicians and Interest Groups. Stanford, CA: Stanford University
Press.
Crisp, Brian, and Rachael E. Ingall. 2002. “Institutional Engineering and the Nature
of Representation: Mapping the Effects of Electoral Reform in Colombia.” Amer-
ican Journal of Political Science 46(4): 733–748.
Crisp, Brian, and Michael J. Kelly. 1999. “The Socioeconomic Impacts of
Structural Adjustment.” International Studies Quarterly 43(3): 533–552.
Crisp, Brian, Daniel Levine, and Juan Carlos Rey. 1995. “The Legitimacy Problem.”
In Venezuelan Democracy under Stress, ed. Jennifer McCoy, Andrés Serbin, William
C. Smith, and Andrés Stambouli. New Brunswick, NJ: Transaction Books, 139–
170.
D’Amico, Margarita. 1992. “TV Años 50: La Epoca de Oro.” Feriado, no. 506
(November 22).
Dassin, Joan. 1984. “The Brazilin Press and the Politics of Abertura.” Journal of
Interamerican Studies and World Affairs 26(3): 385–414.
de la Torre, Carlos. 1996. Un Solo Toque: Populismo y Cultura Polı́tica en Ecuador.
Quito: Centro Andino de Acción Popular.
de la Torre, Carlos. 1997. “Populism and Democracy: Political Discourses and
Cultures in Contemporary Ecuador.” Latin American Perspectives 24(3): 12–24.
de Lima, Venicio A. 1993. “Brazilian Television in the 1989 Presidential Campaign:
Constructing a President.” In Television, Politics, and the Transition to Democracy
in Latin America, ed. Thomas Skidmore. Washington, DC: Woodrow Wilson
Center Press, 97–117.
219
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
220
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Freidemberg, Flavia. 2003. Jama, Caleta y Camello: Las Estrategias de Abdalá Bucaram
y el PRE para Ganar las Elecciones. Quito: Corporación Editora Nacional.
Frutos, Julio César, and Helio Vera. 1993. Pactos Polı́ticos. Asunción: Editorial
Medusa.
Galperı́n, Hernán. 2002. “Transforming Television in Argentina: Market Develop-
ment and Policy Reform in the 1990s.” In Latin Politics, Global Media, ed. E. Fox
and S. Waisbord. Austin: University of Texas Press, 22–37.
Gamarra, Eduardo. 2004. “Carlos Mesa’s Challenges in Bolivia: Will He Succeed
against Overwhelming Odds?” LASA Forum 35(1): 4–5.
Geddes, Barbara. 2003. Paradigms and Sand Castles: Theory Building and Research
Design in Comparative Politics. Ann Arbor: University of Michigan Press.
Geddes, Barbara, and Artur Ribeiro Neto. 1999. “Institutional Sources of Corrup-
tion in Brazil.” In Corruption and Political Reform in Brazil: The Impact of Collor’s
Impeachment, ed. K. S. Rosen and R. Downes. Miami: North-South Center Press,
21–48.
George, Alexander L., and Andrew Bennett. 2005. Case Studies and Theory Develop-
ment in the Social Sciences. Cambridge, MA: MIT Press.
Gerhardt, Michael J. 1996. The Federal Impeachment Process. A Constitutional and
Historical Analysis. Princeton, NJ: Princeton University Press.
Gervasoni, Carlos. 1998. “El Impacto de las Reformas Económicas en la Coalición
Electoral Justicialista (1989–1995).” Boletı́n SAAP 4(6): 67–101.
Ghio, José Marı́a. 2000. “América Latina Después de la Reforma: Incertidumbre
Institucional y Crecimiento Económico.” Reforma y Democracia (16):65–82.
Ginsberg, Benjamin, and Martin Shefter. 1990. Politics by Other Means: The Declining
Importance of Elections in America. New York: Basic Books.
González, Luis Eduardo, and Charles Guy Gillespie. 1994. “Presidentialism and
Democratic Stability in Uruguay.” In The Failure of Presidential Democracy: The
Case of Latin America, ed. Juan J. Linz and Arturo Valenzuela. Vol. 2. Baltimore:
Johns Hopkins University Press, 151–178.
Grijalva, Agustı́n, and Anı́bal Pérez-Liñán. 2003. Presidential Powers Database
[computer file]. Department of Political Science, University of Pittsburgh.
Gronke, Paul, and John Brehm. 2002. “History, Heterogeneity, and Presidential
Approval: A Modified ARCH Approach.” Electoral Studies 21(3): 425–452.
Hagopian, Frances, and Scott Mainwaring, eds. 2005. The Third Wave of Democrati-
zation in Latin America: Advances and Setbacks. Cambridge: Cambridge University
Press.
Hallerberg, Mark, and Scott Basinger. 1998. “Internationalization and Changes in
Tax Policy in OECD Countries: The Importance of Domestic Veto Players.”
Comparative Political Studies 31(3): 321–52.
Hartlyn, Jonathan. 1988. The Politics of Coalition Rule in Colombia. New York:
Cambridge University Press.
Hartlyn, Jonathan. 1994. “Presidentialism and Colombian Politics.” In The Failure
of Presidential Democracy: The Case of Latin America, ed. Juan J. Linz and Arturo
Valenzuela. Vol. 2. Baltimore: Johns Hopkins University Press, 220–253.
Hartlyn, Jonathan. 1998. The Struggle for Democratic Politics in the Dominican
Republic. Chapel Hill: University of North Carolina Press.
221
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
222
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
223
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Kiewe, Amos, ed. 1994. The Modern Presidency and Crisis Rhetoric. Westport, CT:
Praeger.
Kinder, Donald, and Roderick Kiewiet. 1981. “Sociotropic Politics: The American
Case.” British Journal of Political Science 11(2): 129–162.
King, Gary, and Langche Zeng. 2001. “Logistic Regression in Rare Events Data.”
Political Analysis 9(2): 137–163.
Kornblith, Miriam. 1998. Venezuela en los ’90: Las Crisis de la Democracia. Caracas:
Ediciones IESA.
Kuypers, Jim A. 1997. Presidential Crisis Rhetoric and the Press in the Post–Cold War
World. Westport, CT, and London: Praeger.
Kvaternik, Eugenio. 1987. Crisis sin Salvataje: La Crisis Polı́tico-Militar de 1962–63.
Buenos Aires: Ediciones del IDES.
Lamounier, Bolivar. 1994. “Brazil: Toward Parliamentarism?” In The Failure of
Presidential Democracy. Volume 2, The Case of Latin America, ed. Juan J. Linz and
Arturo Valenzuela. Baltimore: Johns Hopkins University Press, 179–219.
Larkins, Christopher. 1998. “The Judiciary and Delegative Democracy in
Argentina.” Comparative Politics 30(4): 423–442.
Lascano Palacios, Mario. 2001. La Noche de los Coroneles: La Rebelión de los Mandos
Medios. Quito: Editorial KESS.
Lavieri, Omar. 1996. “The Media in Argentina: Struggling with the Absence of
a Democratic Tradition.” In Communication in Latin America: Journalism, Mass
Media, and Society, ed. Richard R. Cole. Wilmington, DE: Scholarly Resources,
183–198.
Lerin, François, and Cristina Torres. 1987. Historia Polı́tica de la Dictadura Uruguaya
(1973–1980). Montevideo: Ediciones Nuevo Mundo.
Lewis, Paul H. 2005. Authoritarian Regimes in Latin America: Dictators, Despots, and
Tyrants. Lanham, MD: Rowman and Littlefield.
Lewis-Beck, Michael S., and Martin Paldam. 2000. “Economic Voting: An Intro-
duction.” Electoral Studies 19(2): 113–121.
Lewis-Beck, Michael S., and Mary Stegmaier. 2000. “Economic Determinants of
Electoral Outcomes.” Annual Review of Political Science 3: 183–219.
Lins da Silva, Carlos Eduardo. 1993a. “The Brazilian Case: Manipulation by the
Media?” In Television, Politics, and the Transition to Democracy in Latin America,
ed. Thomas Skidmore. Washington, DC: Woodrow Wilson Center Press, 137–
144.
Lins da Silva, Carlos Eduardo. 1993b. “Brazil’s Struggle with Democracy.” Current
History 92(572): 126–129.
Linz, Juan. 1978. Crisis, Breakdown, and Reequilibration. Vol. 1 of The Breakdown of
Democratic Regimes, ed. Juan J. Linz and Alfred Stepan. Baltimore: Johns Hopkins
University Press.
Linz, Juan. 1990. “The Perils of Presidentialism.” Journal of Democracy 1(1): 51–69.
Linz, Juan. 1994. “Presidential or Parliamentary Democracy: Does It Make a Dif-
ference?” In The Failure of Presidential Democracy: The Case of Latin America, ed.
Juan J. Linz and Arturo Valenzuela. Baltimore: Johns Hopkins University Press,
3–90.
224
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
225
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Markoff, John. 1996. Waves of Democracy: Social Movements and Political Change.
Thousand Oaks, CA: Pine Forge Press.
Marsteintredet, Leiv, and Einar Berntzen. 2006. “Latin American Presidentialism:
Reducing the Perils of Presidentialism through Presidential Interruptions.” Paper
read at the Workshop on Parliamentary Practices in Presidential Systems, ECPR
Joint Sessions of Workshops, Nicosia, Cyprus, April 25–30.
Mata, Aquilino José. 1992. “TV Años 70: La Década que Sacudió a la Televisión.”
Feriado, no. 506 (November 22).
Mayhew, David R. 1974. Congress: The Electoral Connection. New Haven, CT: Yale
University Press.
Mayobre, José Antonio. 2002. “Venezuela and the Media: The New Paradigm.”
In Latin Politics, Global Media, ed. E. Fox and S. Waisbord. Austin: University of
Texas Press, 176–186.
McClintock, Cynthia. 1993. “Peru’s Fujimori: A Caudillo Derails Democracy.” Cur-
rent History 92(572): 112–119.
McConnell, Shelley. 1993. “Rules of the Game: Nicaragua’s Contentious Consti-
tutional Debate.” NACLA Report on the Americas 27(2): 20–25.
McConnell, Shelley. 1997. “Institutional Development.” In Nicaragua without Illu-
sions: Regime Transition and Structural Adjustment in the 1990s, ed. Thomas W.
Walker. Wilmington, DE: Scholarly Resources, 45–63.
Medina Serna, Santiago. 1997. La Verdad Sobre las Mentiras. Santafé de Bogotá:
Planeta.
Meinke, Scott R., and William D. Anderson. 2001. “Influencing from Impaired
Administrations: Presidents, White House Scandals, and Legislative Leadership.”
Legislative Studies Quarterly 26(4): 639–659.
Mejı́a Acosta, Andrés. 1999. “Indisciplina y Deslealtad en el Congreso Ecuatoriano.”
Iconos (6):13–21.
Mejı́a Acosta, Andrés. 2002. Gobernabilidad Democrática: Sistema Electoral, Partidos
Polı́ticos y Pugna de Poderes en Ecuador, 1978–1998. Quito: Fundación Konrad
Adenauer.
Mejı́a Acosta, Andrés. 2003. “Through the Eye of a Needle: Veto Players, Infor-
mal Institutions and Economic Reform in Ecuador.” Paper presented at the
annual meeting of the Latin American Studies Association, Dallas, Texas, March
27–29.
Menendez, Cristina, and Mercedes Kerz. 1993. Autocracia y Democracia: Brasil, Un
Camino al Mercosur. Buenos Aires: Editorial de Belgrano.
Miller, Marjorie, and Juanita Darling. 1997. “The Eye of the Tiger: Emilio
Azcárraga and the Televisa Empire.” In A Culture of Collusion: An Inside Look
at the Mexican Press, ed. William A. Orme, Jr. Miami: North-South Center Press,
59–70.
Morgenstern, Scott. 2004. Patterns of Legislative Politics: Roll-Call Voting in Latin
America and the United States. Cambridge: Cambridge University Press.
Morı́nigo, José N. 1999. “La Disolución del Poder Dual y el Orı́gen de una Nueva
Legitimidad Polı́tica.” In Marzo de 1999: Huellas, Olvido y Urgencias, ed. J. N.
Morı́nigo. Asunción: Universidad Católica, 29–97.
226
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
227
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
228
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
229
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Ribadeneira, Jorge, José Hernández, and Marco Aráuz. 1997. Ecuador Frente al
Vértigo Fatal. Quito: El Comercio.
Riggs, Fred W. 1988. “The Survival of Presidentialism in America: Para-
Constitutional Practices.” International Political Science Review 9(4): 247–278.
Riva Palacio, Raymundo. 1997. “A Culture of Collusion: The Ties That Bind the
Press and the PRI.” In A Culture of Collusion: An Inside Look at the Mexican Press,
ed. William A. Orme, Jr. Miami: North-South Center Press, 21–32.
Rodrı́guez, Juan Carlos. 1999. “Unidad Nacional: Un Desafı́o del Gobierno y de la
Ciudadanı́a.” Última Hora – Correo Semanal (April 3–4): 22.
Rodrı́guez-Valdés, Angel. 1993. La Otra Muerte de CAP. Caracas: Alfadil Ediciones.
Romeu, José Vicente. 2000. Del Caso Remón-Guizado. Panamá: Editorial Mariano
Arosemena.
Rosa e Silva, Cláudio Humberto. 1993. Mil Dias de Solidão: Collor Bateu e Levou. São
Paulo: Geração Editorial.
Saad Herrerı́a, Pedro. 1997. La Caı́da de Abdalá: Un Análisis Actual. Quito: Editorial
El Conejo.
Sabato, Larry. 1993. Feeding Frenzy: How Attack Journalism Has Transformed
American Politics. New York: The Free Press.
Saltos, Napoleón. 1997. “Testigo de Cargo.” In 5 de Febrero, la Revolución de las
Conciencias. Quito: CECS–Fetrapec–Fundación José Peralta, 117–136.
Samper Pizano, Ernesto. 2000. Aquı́ Estoy y Aquı́ me Quedo: Testimonio de un Gobierno.
Santafé de Bogotá: El Áncora Editores.
Samuels, David. 2003. Ambition, Federalism, and Legislative Politics in Brazil.
Cambridge: Cambridge University Press.
Samuels, David, and Scott Mainwaring. 2004. “Strong Federalism, Constraints on
the Central Government, and Economic Reform in Brazil.” In Federalism and
Democracy in Latin America, ed. E. L. Gibson. Baltimore: Johns Hopkins Univer-
sity Press, 85–130.
Sarmiento, Sergio. 1997. “Trial by Fire: The Chiapas Revolt, the Colosio Assassi-
nation and the Mexican Press in 1994.” In A Culture of Collusion: An Inside Look
at the Mexican Press, ed. William A. Orme, Jr. Miami: North-South Center Press,
33–39.
Sartori, Giovanni. 1970. “Concept Misinformation in Comparative Politics.”
American Political Science Review 64(4): 1033–1053.
Saunders, J. V. D. 1964. “A Revolution of Agreement among Friends: The End of
the Vargas Era.” Hispanic American Historical Review 44(2): 197–213.
Schamis, Hector. 2002. “Argentina: Crisis and Democratic Consolidation.” Journal
of Democracy 13(2): 81–94.
Schmidt, Gregory D. 2000. “Delegative Democracy in Perú? Fujimori’s 1995 Land-
slide and the Prospects for 2000.” Journal of Interamerican Studies and World Affairs
42(1): 99–132.
Seiferheld, Alfredo M., ed. 1987. La Caı́da de Federico Cháves: Una Visión Documental
Norteamericana. Asunción: Editorial Histórica.
Serrafero, Mario Daniel. 1997. “Juicio Polı́tico y Derrumbe Institucional en la
Argentina (1976).” Estudios Interdisciplinarios de América Latina y el Caribe 8(2):
41–66.
230
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Shah, Dhavan V., Mark D. Watts, David Domke, and David P. Fan. 2002. “News
Framing and Cueing of Issue Regimes: Explaining Clinton’s Public Approval in
Spite of Scandal.” Public Opinion Quarterly 66(3): 339–370.
Shugart, Matthew S., and John Carey. 1992. Presidents and Assemblies: Constitutional
Design and Electoral Dynamics. Cambridge: Cambridge University Press.
Shugart, Matthew S., and Stephan Haggard. 2001. “Institutions and Public
Policy in Presidential Systems.” In Presidents, Parliaments, and Policy, ed. S.
Haggard and M. D. McCubbins. Cambridge: Cambridge University Press, 64–
102.
Shugart, Matthew S., and Scott Mainwaring. 1997. “Presidentialism and Democracy
in Latin America: Rethinking the Terms of the Debate.” In Presidentialism and
Democracy in Latin America, ed. S. Mainwaring and M. S. Shugart. Cambridge:
Cambridge University Press, 12–54.
Siavelis, Peter. 2000. “Disconnected Fire Alarms and Ineffective Police Patrols: Leg-
islative Oversight in Post-Authoritarian Chile.” Journal of Interamerican Studies
and World Affairs 42(1): 71–98.
Siavelis, Peter M. 2002. “Exaggerated Presidentialism and Moderate Presidents:
Executive-Legislative Relations in Chile.” In Legislative Politics in Latin America,
ed. S. Morgenstern and B. Nacif. Cambridge: Cambridge University Press, 79–
113.
Sigmund, Paul E. 1977. The Overthrow of Allende and the Politics of Chile, 1964–1976.
Pittsburgh: University of Pittsburgh Press.
Skidmore, Thomas. 1999. “Collor’s Downfall in Historical Perspective.” In Cor-
ruption and Political Reform in Brazil: The Impact of Collor’s Impeachment, ed. K. S.
Rosen and R. Downes. Miami: North-South Center Press, 1–19.
Skowronek, Stephen. 1997. The Politics That Presidents Make: Leadership from John
Adams to Bill Clinton. Cambridge, MA: Belknap Press.
Smulovitz, Catalina, and Enrique Peruzzotti. 2000. “Societal Accountability in Latin
America.” Journal of Democracy 11(4): 147–158.
Sola, Lourdes. 1994. “The State, Structural Reform, and Democratization in
Brazil.” In Democracy, Markets, and Structural Reform in Latin America, ed. William
C. Smith, Carlos H. Acuña, and Eduardo A. Gamarra. Miami: North-South Cen-
ter Press, 151–181.
Stepan, Alfred. 1971. The Military in Politics: Changing Patterns in Brazil. Princeton,
NJ: Princeton University Press.
Stepan, Alfred, and Cindy Skach. 1993. “Constitutional Frameworks and Demo-
cratic Consolidation: Parliamentarism vs. Presidentialism.” World Politics 46(1):
1–22.
Stokes, Susan. 1996a. “Public Opinion and Market Reforms: The Limits of Eco-
nomic Voting.” Comparative Political Studies 29(5): 499–519.
Stokes, Susan. 1996b. “Economic Reform and Public Opinion in Peru, 1990–1995.”
Comparative Political Studies 29(5): 544–565.
Stokes, Susan C. 2001. Mandates and Democracy: Neoliberalism by Surprise in Latin
America. Cambridge: Cambridge University Press.
Stokes, Susan, and John Baughman. 1999. “From Policy Change to Preference
Change? Neoliberalism and Public Opinion in Latin America.” Paper presented
231
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
232
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
Vanden Heuvel, Jon, and Everette E. Dennis. 1995. Changing Patterns: Latin
America’s Vital Media. New York: The Freedom Forum Media Studies Center.
Vargas, Mauricio, Jorge Lesmes, and Edgar Téllez. 1996. El Presidente que se
Iba a Caer: Diario Secreto de Tres Periodistas Sobre el 8.000. Santafé de Bogotá:
Planeta.
Velásquez, Osvaldo. 1993. Historia de una Dictadura: De Torrijos a Noriega. Panama:
Litho Editorial Chen.
Verdesoto, Luis. 2004. “¿Hacia Dónde va Bolivia?” Nueva Sociedad (191): 38–49.
Verkuilen, Jay. 2005. “Assigning Membership in a Fuzzy Set Analysis.” Sociological
Methods and Research 33(4): 462–496.
Victoria, Pablo. 1997. Yo Acuso: Un Documentado Pliego de Cargos contra el Presidente
Samper. Bogotá: Ediciones Temas de Hoy.
Villagrán de León, Francisco. 1993. “Thwarting the Guatemalan Coup.” Journal of
Democracy 4(1): 117–124.
Virtue, John, and J. Arthur Heise. 1996. “Controversies over Mass Communication
and Professional Education in the Andean Countries.” In Communication in Latin
America: Journalism, Mass Media, and Society, ed. Richard R. Cole. Wilmington,
DE: Scholarly Resources, 199–216.
Waisbord, Silvio. 1994. “Knocking on Newsroom Doors: Press and Political Scan-
dals in Argentina.” Political Communication 11(1): 19–33.
Waisbord, Silvio. 1996. “Investigative Journalism and Political Accountability in
South American Democracies.” Critical Studies in Mass Communication 13(4): 343–
363.
Waisbord, Silvio. 1998. “The Unfinished Project of Media Democratization in
Argentina.” In Communicating Democracy: The Media and Political Transitions, ed.
Patrick H. O’Neil. Boulder, CO: Lynne Rienner, 41–62.
Waisbord, Silvio. 2000. Watchdog Journalism in South America. New York: Columbia
University Press.
Weinstein, Martin. 1988. Uruguay: Democracy at the Crossroads. Boulder, CO: West-
view Press.
Weyland, Kurt. 1993. “The Rise and Fall of President Collor and Its Impact on
Brazilian Democracy.” Journal of Interamerican Studies and World Affairs 35(1):
1–37.
Weyland, Kurt. 1998a. “The Politics of Corruption in Latin America.” Journal of
Democracy 9(2): 108–121.
Weyland, Kurt. 1998b. “Peasants or Bankers in Venezuela? Presidential Popu-
larity and Economic Reform Approval, 1989–1993.” Political Research Quarterly
51(2): 341–362.
Weyland, Kurt. 1999. “A Paradox of Success? Determinants of Political Support
for President Fujimori.” Paper presented at the fifty-seventh annual meeting of
the Midwest Political Science Association, Chicago, April 15–17.
Weyland, Kurt. 2002. The Politics of Market Reform in Fragile Democracies: Argentina,
Brazil, Peru, and Venezuela. Princeton, NJ: Princeton University Press.
Whitehead, Laurence. 2002. Democratization: Theory and Experience. Oxford: Oxford
University Press.
233
P1: JZP
0521869423ref cuny809-perez 0 521 86942 0 May 8, 2007 19:30
References
234
P1: JZP
0521869423ind cuny809-perez 0 521 86942 0 May 8, 2007 19:59
Index
235
P1: JZP
0521869423ind cuny809-perez 0 521 86942 0 May 8, 2007 19:59
Index
236
P1: JZP
0521869423ind cuny809-perez 0 521 86942 0 May 8, 2007 19:59
Index
237
P1: JZP
0521869423ind cuny809-perez 0 521 86942 0 May 8, 2007 19:59
Index
238
P1: JZP
0521869423ind cuny809-perez 0 521 86942 0 May 8, 2007 19:59
Index
239
P1: JZP
0521869423ind cuny809-perez 0 521 86942 0 May 8, 2007 19:59
Index
240
P1: JZP
0521869423ind cuny809-perez 0 521 86942 0 May 8, 2007 19:59
Index
241
P1: JZP
0521869423ser cuny809-perez 0 521 86942 0 May 8, 2007 19:26