Negretto Resaltado

Download as pdf or txt
Download as pdf or txt
You are on page 1of 30

Minority Presidents and

Demomutic Pdormance
in Latin America
Gabriel L. Negretto

ABSTRACT
A widely accepted argument among scholars of Latin American
presidential regimes is that interbranch cooperation is impaired
when the president’s party falls short of a majority of seats in the
legislature. This argument fails to consider three factors that affect
the performance of minority presidents: the policy position of the
president’s party, the president’s capacity to sustain a veto, and the
legislative status of the parties included in the cabinet. This article
argues that the greatest potential for conflict in a presidential regime
occurs when the president’s party lacks the support of both the
median and the veto legislator and no cabinet coalition holding a
majority of legislative seats is formed. This hypothesis is supported
using data on executive-legislative conflicts and on interrupted
presidencies in Latin America during the period 1078-2003.

A rich literature on presidential regimes has repeatedly associated


party fragmentation and its likely effect, minority presidents, with
ineffective government, high levels of executive-legislative conflict, and
even democratic instability. So far, however, single-case studies have
provided no persuasive explanation of the causal mechanisms behind
this association, nor have large-N statistical studies proven beyond a
doubt that the variables are significantly correlated. The absence of con-
clusive results stems from the lack of attention to variables that might
explain the differing performances of presidential regimes, other than
the share of seats of the president’s party in Congress.
This study argues that three factors affect the level of interbranch
cooperation in the event that the president’s party has no independent
control over the assembly. These factors are the policy position of the
president’s party, the president’s capacity to sustain a veto, and the leg-
islative status of the parties included in the cabinet. Based o n an analy-
sis of these factors, this study proposes the hypothesis that the greatest
potential for conflict in a presidential regime occurs when the presi-
dent’s party lacks the support of both the median and the veto legisla-
tors and no cabinet coalition holding a majority of legislative seats is
formed. This hypothesis is supported using data on executive-legislative
conflicts and on the premature termination of presidencies in Latin
America during the period 1978-2003.

63
64 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

The argument is presented by proposing a new typology of minor-


ity situations in presidential regimes and a series of hypotheses about
levels of interbranch cooperation, based o n whether the president’s
party controls the pivotal legislators necessary to pass a bill or sustain a
veto. These hypotheses are then explored through analysis of how often
majority cabinet coalitions are formed in minority situations and how
they would be expected to affect executive-legislative relations. Prelim-
inary support for the hypotheses is developed using an index of exec-
utive-legislative conflicts and cases of premature termination of consti-
tutional terms in Latin America during the period 1978-2003. The
discussion begins with a review of the literature on minority presidents
and their impact o n democratic government.

MINORITY PRESIDENTS AND


DEMOCRATIC
PERFORMANCE
Scholars of presidential regimes have traditionally focused almost exclu-
sively on the share of legislative seats of the president’s party to classlfy
forms of government in presidential regimes. From this perspective, the
main dichotomy is whether the government is unified or divided. A pres-
idential government is unified when the president’s party has a majority,
and divided when the president’s party has a minority of seats in a uni-
cameral or bicameral legislature. A divided government refers to two dif-
ferent forms of separation between presidents and assemblies: “divided
control,” when a party other than the president’s has control over a leg-
islative majority; and “no majority” situations, when neither the presi-
dent’s nor any other party holds a majority of seats in the assembly.’
Under any of these conditions, presidential regimes are generally
assumed to increase the number of actors whose agreement is neces-
sary to approve legislation and make political decisions. Unlike students
of the U.S. presidential regime, students of political institutions in Latin
America generally agree that divided government is almost always asso-
ciated with gridlock; unconstitutional, unilateral actions; and interbranch
strife.2 Given the typical frailty of democracies in Latin America, more-
over, minority presidents have also been considered a major factor lead-
ing to government instability and democratic breakdown. Mark Jones
(1995, 38) summarizes this view well when he notes that even those
w ho find merit in the maintenance of presidential regimes in the region
usually agree that presidential systems that consistently fail to provide
the president with sufficient legislative support are unlikely to thrive.
Because the probability of the president’s party enjoying majority
support in Congress tends to decrease as the number of parties sharing
legislative seats increases, students of Latin American presidential
regimes consider the level of party fractionalization in Congress, as
NEGRETTO: MINORITY PRESIDENTS 65

measured by the effective number of legislative parties, to be a crucial


determinant of the performance of presidential regime^.^ Mainwaring
(1993) and Mainwaring and Shugart (1997a, 1997b, 4661, for instance,
argue that presidential regimes cannot work well under conditions of
extreme multipartism. The underlying assumption is that party system
fragmentation weakens the partisan power of the president, making leg-
islative bargaining more complex and increasing the likelihood of exec-
utive-legislative deadlock. Although one way to overcome this situation
is to rely on cabinet or legislative coalitions, these authors argue that the
separate survival of presidents and assemblies provides parties with
weak incentives for building and maintaining interparty coalitions.
The idea that only a majority president can secure a stable, effective
democratic regime found empirical support in the experience of Latin
American democracies during the 1960s and 1970s. The cases of Colom-
bia, Costa Rica, and Venezuela were the only examples of stable presi-
dential regimes during that period. All three had two- or two-and-a-half-
party systems, which often provided presidents with regular support
from legislative majorities. Since 1978, however, the large majority of
presidential regimes in the region coexist with multiparty systems, under
which presidents are most often minority presidents. From 1978 to 2002,
80 presidents have been democratically elected in Latin America, of
which only 26 (32 percent) were majority presidents. The largest group,
54 presidents (68 percent), were minority presidents. Several of these
presidents faced important constitutional crises. Yet in spite of wide vari-
ations in quality and performance, minority presidents did not lead to
the massive failure of democracies in the region. Moreover, most of
these presidents managed to govern and to implement costly economic
reforms in the midst of dire social conditions.
The relative stability of presidential regimes with minority presi-
dents during the last two decades seems consistent with the findings of
recent statistical studies on the relationship between institutional and
partisan variables and the survival of presidential democracies. Cheibub
(2002a, b), for instance, shows that minority presidents, even when they
govern in deadlock situations (the president’s party holds less than a
legislative majority but can sustain a veto), are not related to the col-
lapse of presidential democracies. He also argues (2002b, 124) that it is
not the level of party fractionalization per se that negatively affects the
survival of presidential democracies but the existence of three large par-
ties of relatively equal strength, a situation that tends to occur under
conditions of moderate (rather than extreme) party pluralism.
The contribution of these recent works, however, is limited for
understanding variations in the performance of contemporary presiden-
tial regimes. In the first place, outright democratic breakdown, so
common in earlier decades of the twentieth century, became an increas-
66 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

ingly rare event in the 1980s and 1990s. For this reason, a comparative
explanation of the differing performances of presidential regimes in the
last two decades requires the use of less dramatic measures of perform-
ance, such as the frequency and type of executive-legislative conflicts,
the capacity of presidents to pass significant legislation, or the premature
termination of constitutional terms. Another limitation of recent studies is
that they focus on democratic breakdown but have not been able to pro-
vide a positive explanation of the factors that affect the performance of
presidential regimes. All they have shown to date is a series of negative
findings, which should lead us to rethink some of the accepted views on
the perils of minority presidents and divided governments.*
In spite of significant progress made during the last decade, the
main problem with the traditional research and part of the contempo-
rary work on presidential regimes is that they rely on rather blunt indi-
cators to conceptualize different types of presidential governments and
predict their performance. The share of seats of the president’s party or
the level of party fractionalization in Congress provides only a quantita-
tive measure, not a causal mechanism to explain the performance of a
presidential democracy.
From an analytic point of view, it is not apparent why the minority
legislative status of the president’s party should, by itself, have a nega-
tive impact on the level of interbranch cooperation. Even without major-
ity support in Congress, presidents often have legislative powers to
influence policymaking and obtain cooperation from legislators. A pres-
idential veto, for instance, may be skillfully used to extract concessions
from legislators by inducing either the introduction of a new bill or
amendments to the original bill (Cameron 2000, 20). Presidents also
have a wide array of agenda powers that they can use to force legisla-
tors to approve legislation that they would have not initiated or passed
on their own (Carey and Shugart 1998; Negretto 2004).
Minority presidents also rely on informal legislative coalitions to
obtain support in Congress. To forge these coalitions, the relative bar-
gaining power of the president’s party may be more important than its
share of seats or the level of party fragmentation. If the president’s party
holds, say, 40 percent of legislative seats, and four opposition parties hold
15 percent each, the latter are individually weaker than the former. Under
this distribution of partisan power, while the president is capable of form-
ing four different coalitions, one with each opposition party, there is only
one coalition (of all opposition parties) that can exclude the president’s
party. With the same share of seats, however, the president’s party would
be in a more vulnerable position if the opposition were concentrated in,
say, only two parties sharing 35 and 25 percent of the seats each. In this
situation, three possible coalitions could be formed, and the president’s
party, like the other two parties, could be included in two of them.5
NEGRETTO: MINORITY PRESIDENTS 67

The location of the president’s party in the policy space may also
have an important effect on the formation of legislative coalitions. Sup-
pose that the president’s party is a centrist party with 20 percent of the
legislative seats, while opposition parties are divided between a leftist
party holding 40 percent of the seats and two rightist parties with 20
percent of the seats each. If parties care about policy, they prefer to
form legislative coalitions that are connected in the policy space.(‘ In this
scenario, there are two coalitions that fulfill this requirement, and the
president’s party is the only party that is a necessary member of both.
In other words, if parties take policy considerations into account when
passing legislation, then a small but relatively centrist presidential party
may be more powerful than larger but more extreme opposition parties,
regardless of the level of party fragmentation.
Besides legislative coalitions, minority presidents can build cabinet
coalitions to seal a formal alliance with other parties in Congress. Recent
work on coalition formation in presidential regimes has shown that
presidents, just like prime ministers, often build multiparty cabinet coali-
tions to overcome the weakness of their parties in Congress. If the cab-
inet is composed of representatives of political parties that collectively
control a majority of seats in the legislature, then government proposals
may obtain majority support in Congress even if the president’s party
alone has a minority legislative status.
This analysis indicates that the classification of presidential regimes
and the construction of hypotheses about their potential performance
cannot be based o n quantitative measures and mechanical associations
alone. To build a more adequate understanding of the nature and per-
formance of different forms of presidential government, we need to start
with a better sense of the distribution of preferences and institutional
powers among the relevant actors in a system of separation of powers.

PRESIDENTS IN MINORITY SITUA~ONS


A simple spatial model of executive-legislative relations can be used to
conceptualize the different minority situations that presidents may face.
From these models, we may derive expectations about levels of inter-
branch cooperation. Let us assume that there are three decisive actors
in separation-of-powers systems: the president’s party, the veto party,
and the median party in Congress.’ The president’s party is the party
that controls the executive office, the veto party controls the legislator
whose support allows the president to sustain a veto, and the median
party controls the median legislator; that is, the legislator whose pro-
posals cannot be defeated by a legislative majority. The model assumes
that each policymaker has a single most-preferred policy along a left-
right dimension and that the utility derived from other policy alterna-
68 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Figure 1. Presidents in Minority Situations

la. Median
Outcome
I
1 I, -
President’s party
--
President Veto legislator Median legislator

Median party
President’s party
- -

tives decreases monotonically as they move farther away from the ideal
point. It also assumes that parties are cohesive, unitary actors.
To isolate the relative position of the president’s party in Congress,
coalitions are excluded from the analysis. What matters in this model is the
share of legislative seats of the president’s party, its location on the policy
space, and the override requirement to sustain a veto. Following this logic,
the model suggests that presidents may face three basic types of minority
situations, which we may call median, divided, and congressional.
A median minority situation occurs whenever the president’s party,
holding less than 50 percent of the legislative seats, is appropriately
located around the center of the distribution, so that n o party or coali-
tion of parties can defeat its proposals by majority vote. In this situation,
it is the location of the president’s party in the policy space, rather than
the size of its legislative representation, that makes possible the support
of both the veto and the median legislator in Congress, as shown in
figure la.HNote that this is logically the same configuration as in “uni-
fied” majority situations, when the president’s party holds more than 50
percent of the seats in the assembly. The difference, of course, is that in
this case the president’s party would always include both the veto and
the median legislator, regardless of its location in the policy space.
Other things being equal, a president whose party controls a major-
ity of legislators should have greater bargaining power than a minority
NEGRE’ITO: MINORITY PRESIDENTS 60

president to implement most of the legislative agenda with relatively


little cost in terms of interbranch negotiation. This is the main reason
why so many scholars argue that separation-of-powers systems cannot
work well when different partisan interests control the presidency and
the congress. Nevertheless, we should expect a median minority presi-
dent to be more effective and face less executive-legislative conflict than
any other minority president. Median presidents may have the support
of a relatively small contingent in the congress but, as long as policy
preferences matter, their parties are pivotal in the formation of any leg-
islative coalition able to reach a majority.
A median minority president may, in this sense, be equivalent to a
single-party minority government in a parliamentary regime. A single-
party minority government can be a viable government in a parliamen-
tary regime if there is no coalition of opposition parties that can defeat
its proposals by majority vote (Strom 1990; Laver and Schofield 1990).
Using the same logic, we could argue that a single-party minority gov-
ernment in a presidential regime could achieve a good record of leg-
islative success and interbranch cooperation if the president’s party con-
trols the median legislator in the congress.
Presidents face what, in a strict sense, can be called a divided situ-
ation when the president’s party controls the veto legislator but a party
other than the president’s has the support of the median legislator, as
shown in figure l b . The threshold required to override a presidential
veto determines the occurrence of this minority situation. In particular,
the probability of a divided situation is higher, the higher the threshold
required to override a presidential veto in Congress, so, for instance, a
divided situation is more likely with a two-thirds than with a one-half
override rule.9 Given the override rule, a divided situation is also more
likely when the threshold applies to each of the separate chambers of a
bicameral assembly than when it applies to the joint vote of a bicameral
assembly or to unicameral assemblies.
This type of minority situation is more closely associated with the
notion of deadlock used by some authors (see, for instance, Cheibub
2002a, 288-89). This is because in this case, the congress can pass leg-
islation and still be unable to change the status quo if the president
prefers the latter to the new policy and the legislative majority lacks the
necessary votes to override the president’s veto. In other words, neither
the president (who lacks a majority) nor the opposition parties (which
lack the votes to override a veto) are able to rule. The risk of a stale-
mate, of course, is the reason that presidential regimes in this situation
are often portrayed as indecisive, costly, and conflictive.
The occurrence of gridlock in a divided situation, however, depends
on the location of the legislative status quo and the policy preferences
of the pivotal actors (Krehbiel 1998). If the initial status q u o policy is
70 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

located between the preferences of the median party and the president
with effective veto, policy changes may, in effect, be impossible. The
president will veto any change that brings the outcome too close to the
median party’s preference and moves it away from the president’s pref-
erence. It is for this reason that the policy space between the two deci-
sive actors defines the “gridlock interval”; that is, the set of policy deci-
sions that will be stable in spite of a legislative majority favoring policy
change, as illustrated in figure lb.
Policy change, however, is possible if the legislative status quo is
outside the gridlock interval. When the legislative status quo is located
at a relatively extreme position with respect to the ideal points of the
decisive actors, for instance, significant policy changes may occur pre-
cisely at the ideal point of the median party. This is because the presi-
dent will not veto proposals made by the median party if they are closer
to the president’s preferences than the status quo is. Policy change,
though only incremental, may also occur if the status quo is outside the
gridlock interval but close to the ideal policy of either the president’s or
the median party (Krehbiel 1998,3639;Brady and Volden 1998, 11-38).
In other words, policy immobilism is not a structural trait of presi-
dential regimes in divided situations. We can assume only that gridlock
may occur from time to time, but not how often, or when it will lead to
interbranch conflict.’O This does not mean, however, that we should
expect the same level of legislative effectiveness and interbranch coop-
eration as in median minority situations. Keeping other factors constant,
a president whose party is located at a median position in the congress
should be able to obtain more support from the congress than a minor-
ity president invested with merely a veto.
A third possible minority situation exists when a party or parties
other than the president’s controls the median and the veto legislator.”
One can call this situation congressional, as illustrated in figure lc,
because in this case, an assembly under the control of opposition par-
ties dominates the legislative process.
This is potentially the most conflictive situation in a presidential
regime. Chosen in a popular election, the president may believe that he
or she has a popular mandate to implement a particular political plat-
form, only to find that he has no institutional power to influence poli-
cymaking. This situation may invite policymakers to engage in what
Juan Linz (1994) rightly identifies as conflicts of “dual legitimacy.”
A nonmedian minority president without a veto has two basic
options: to accept or to challenge congressional rule. Acquiescence to
congressional rule would usually take the form of a coalition govern-
ment in which the president tries to overcome the disadvantage in Con-
gress by distributing cabinet positions among parties whose votes are
needed for a legislative majority. The other option, defiance of con-
NEGRE'ITO: MINORITY PRESIDENTS 71

gressional rule, would usually take the form of unilateral presidential


government, either by the use of powers granted by the constitution,
such as decrees, or by extraconstitutional actions, such as usurping leg-
islative powers or dissolving Congress. The risk of a congressional situ-
ation is that the latter option may be particularly tempting in the event
of a crisis that demands of the president a rapid and effective solution.12
Table 1 lists the majority and minority situations under which pres-
idents governed in 18 Latin American countries from 1778 to 2003, clas-
sified according to the categories proposed in this model. For each
country, the presidents' situations are classified according to the presi-
dential party's share of seats and its policy position relative to the loca-
tion of the median and the veto party on a left-right scale. Parties were
located on this scale on the basis of expert judgments.'3
The table includes all the presidents who completed a full presi-
dential term or whose terms were prematurely terminated between 1778
and 2003. Whenever a nonconcurrent congressional election changed
the situation of the president's party, the corresponding years of the
presidency were counted as a different case. Following this methodol-
ogy, table 1 includes 75 cases, of which 24 represent majority and 51
minority presidents. Within the group of minority presidents, 15 gov-
erned in median, 19 in divided, and 17 in congressional situations.
Because this classification is based on the legislative status of the
president's party, the party's policy position, and the president's institu-
tional powers to sustain a veto, it does not take into account the legisla-
tive status of the whole cabinet, which may include parties other than
the president's. But this factor may be crucial if we want to move from
a classification of presidents to a classification of presidential govern-
ments. This is the next step in considering the full range of possibilities.

MINORl"Y PRESIDENTS AND


COALITION
GOVERNMENT
This study has argued that we cannot predict the impact of minority
presidents on interbranch cooperation without knowing whether the
president's party has the support of the median and the veto legislator
in Congress. But we also need to consider the impact of cabinet coali-
tions. Students of cabinet coalitions in presidential regimes have bor-
rowed the distinction between single-party and coalition government
from the analysis of parliamentary regimes. Each of these is classified
into majority or minority government, depending on whether the par-
ties included in the cabinet can muster a majority of seats in the con-
gress (Deheza 1998, 152). The case of coalition majority government is
particularly relevant here.
(continued on p . 74)
72 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Table 1. Presidents in Majority and Minority Situations


in Latin America, 1978-2003

Legislative
President Term Party Status Situation
Argentina
Alfonsin 1983-89 UCR Minority Divided
Menem I 1989-95 PJ Minority Median
Menem I1 1995-97 PJ Majority Unified
Menem I1 1997-99 PJ Minority Median
De la Rba 1999-2001 UCR Minority Divided
Bolivia
Siles Zuazo 1980-85 UDP Minority Divided
Paz Estenssoro 1985-89 MNR Minority Median
Paz Zamora 1989-93 MIR Minority Congressional
Sanchez de Lozada I 1993-97 MNR Minority Median
Banzer 1997-2001 ADN Minority Congressional
Sanchez de Lozada I1 2002-3 MNR Minority Congressional
Brazil
Sarney 1985-89 PFL Minority Congressional
Collor de Mello 1990-92 PRN Minority Congressional
Cardoso I 1994-98 PSDB Minority Congressional
Cardoso I1 1998-2002 PSDB Minority Congressional
Chile
Aylwin 1989-94 PDC Minority Divided
Frei 1994-2000 PDC Minority Divided
Colombia
Turbay 1978-82 PL Majority Unified
Betancur 198246 PC Minority Congressional
Barco 1986-90 PL Majority Unified
Gaviria 1990-94 PL Majority Unified
Samper 1994-98 PL Majority Unified
Pastrana 1998-2002 PC Minority Congressional
Costa Rica
Carazo 1978-82 PU Minority Divided
Monge 1982-86 PLN Majority Unified
Arias 1986-90 PLN Majority Unified
Calder6n 1990-94 PUSC Majority Unified
Figueres 1994-98 PLN Minority Median
Rodriguez 1998-2002 PUSC Minority Median
Dominican Republic
Balaguer [I1 198690 PRSC Minority Divided
Balaguer [II] 1990-94 PRSC Minority Divided
NEGRETI’O: MINORITY PRESIDENTS 73

Table 1 (continued)

Legislative
President Term Party Status Situation
Balaguer [IIII 1994-96 PRSC Minority Divided
Fernindez Reynd 1996-98 PLD Minority Median
Fernindez Reyna 1998-2000 PLD Minority Divided
Ecuador
Rold6s 1979-81 CFP Minority Divided
Febres Corder0 1984-88 PSC Minority Divided
Borja 1988-90 ID Minority Median
Borja 1990-92 ID Minority Divided
Duran Ballkn 1992-96 PRU Minority Divided
Bucaram 1996-97 PRE Minority Divided
Mahuad 1998-2000 DP Minority Divided
El Salvador
Duarte 1985-88 PDC Majority Unified
Duarte 1988-89 PDC Minority Divided
Cristiani 1989-9 1 ARENA Majority Unified
Cristiani 1991-94 ARENA Minority Divided
Calder6n 1994-99 ARENA Minority Divided
Guatemala
Cerezo 1986-90 PDCG Majority Unified
Serrano 1991-93 MAS Minority Congressional
Arzu 1996-99 PAN Majority Unified
Portillo 2000-2003 FRG Majority Unified
Honduras
Suazo C6rdoba 1982-86 PLH Majority Unified
Azcona 198690 PLH Majority Unified
Callejas 1989-93 PNH Majority Unified
Reina 1993-97 PLH Majority Unified
Flores 1997-2001 PLH Majority Unified
Mexico
Zedillo 1997-2000 PRI Minority Median
Nicaragua
Barrios 1990-96 UNO Majority Unified
Aleman 19962001 AL/PLC Minority Median
Panama
Perez Balladares 1994-99 PRD Minority Median
Paraguay
Rodriguez 1989-93 ANR Majority Unified
Wasmosy 1993-98 ANR Minority Congressional
Cubas 1998-2002 ANR Majority Unified
74 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Table 1 (continued)
Legislative
President Term Party Status Situation
Peru
Belahde 1980-85 AP Minority Median
Garcia 1985-90 PAP Majority Unified
(APRA)
Fujimori I 1990-92 C90 Minority Congressional
Fujimori I1 1995-2000 C90 Majority Unified
Fujimori I11 2000 C90 Minority Congressional
UWPaY
Sanguinetti I 1984-89 PC Minority Median
Lacalle 1989-94 PN Minority Congressional
Sanguinetti I1 1994-99 PC Minority Median
Venezuela
Herrera Campins 1978-83 COPE1 Minority Congressional
Lusinchi 1983-88 AD Majority Unified
Perez 1988-93 AD Minority Median
Caldera 11 1994-99 CN Minority Congressional
Chiivez I 1998-2000 MVR Minority Congressional

Source: Based on Payne et al. 2002, Elections Around the World, and Nohlen 2003
for electoral data; Coppedge 1997, Mainwaring and Scully 1995, Alckntara and Frei-
denberg 2001 for party policy positions.

From a conceptual point of view, when a minority president


appoints a cabinet that includes representatives of parties that collec-
tively hold a majority of legislative seats, it is more appropriate to clas-
sify the case as a particular form of majority government. In addition,
and in terms of the performance of presidential regimes, a coalition
majority government might significantly improve the president’s bar-
gaining position in Congress. Even if opposition parties receive a dis-
proportionately small share of cabinet positions in relation to their leg-
islative strength, we would expect executive-legislative relations to be
smoother than when the president’s party governs alone or when a cab-
inet coalition is formed but falls short of a majority in the legislature.
The study of coalition formation is relatively recent in the compar-
ative analysis of presidential regimes, partly because the conventional
wisdom was that coalitions were not a significant phenomenon outside
parliamentary regimes (see Deheza 1998; Amorim N e t 0 1998, 2002;
Altman 2002; Chasquetti 2001; Cheibub et al. 2004; Martinez-Gallardo
2005). Just as in multiparty parliamientary regimes, however, coalition-
making is an important activity in multiparty presidential regimes. The
NEGRE'ITO: MINORITY PRESIDENTS 75

main differences between cabinet coalitions in presidential and parlia-


mentary regimes stem from the different conditions that determine the
formation and duration of governments in each regime.
Whereas in a presidential regime, the president has the power to
appoint the government regardless of the size of the presidential party in
Congress and its policy location, in a parliamentary regime, the forma-
teur is usually the largest or the median legislative party (Laver and
Schofield 1990). In addition, although the autonomy of presidents in the
formation and maintenance of cabinets may be subject to restrictions, par-
liamentary cabinets are strictly dependent on legislative support to obtain
investiture and win confidence votes.I4 This means that while minority
governments (either single-party or coalition) exist in a parliamentary
regime, they must always count, at least implicitly, on the support of a
legislative majority (Strom 1990). A minority presidential government,
however, can form and survive even with a hostile majority in Congress.
These structural differences should and do have an impact on the fre-
quency and nature of coalitions in presidential and parliamentary regimes
(Samuels and Eaton 2002). Cheibub et al. (20041, for instance, have found
that cabinet coalitions in general, and majority cabinet coalitions in partic-
ular, are relatively more frequent under parliamentarism than under pres-
identialism. Amorim Net0 (2006),on the other hand, has shown that exec-
utive coalitions in presidential systems also differ in kind from cabinet
coalitions in parliamentary systems, in that in the former, presidents often
include nonpartisan members and do not observe the norm of propor-
tionality in the distribution of portfolios among coalition allies.
To account for the process of coalitionmaking among presidents in
minority situations, this study examined the partisan composition of
cabinets to determine whether these presidents built cabinet coalitions,
and whether these coalitions had a majority status in Congress. Table 2
lists the cabinet coalitions formed with presidents in minority situations
and the legislative status of the parties they included. A cabinet or exec-
utive coalition is defined here as the set of representatives of parties that
hold cabinet positions. This coalition is considered to be a majority
coalition only if the parties holding cabinet posts gather a majority of
seats in a unicameral assembly or in both chambers of a bicameral
a~semb1y.l~ Unless there was evidence that an appointment was made
on individual rather than on partisan terms, it was assumed that the
inclusion of a single cabinet member from a party other than the presi-
dent's was sufficient to obtain support from that party in Congress. It
was also assumed, again, that parties are cohesive, unitary actors.
Table 3 (p. 78) shows the frequency and type of cabinet coalitions
formed. Such formation clearly is a relatively frequent phenomenon
among presidents in minority situations. Of the 51 minority presidents
included in the database, 27 (53 percent) formed cabinet coalitions.
76 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Table 2 . cdbinet Coalitions Among Presidents in Minority Situations,


1978-2003
President Legislative Status of Coalition

Argentina
De la Rha Minority

Bolivia
Siles Zuazo Minority
Paz Zamora Majority
SiinchezdeLozadaI Majority
Biinzer Majority
Siinchez de Lozada I1 Majority

B d
Sarney Majority
Collor de Mello Minority
Cardoso I Majority
Cardoso 11 Majority

Chile
Alwyn Minority
Frei Minority

Colombia
Betancur Majority

Dominican Republic
Ferniindez Reyna Majority
Ferniindez Reyna Minority

Ecuador
Rold6s Minority
Febres Corder0 Minority
Durin Ballen Minority
Bucaram Minority

Panama
Perez Balladares Minority

Peru
Belacnde Majority
Fujimori I Minority

Uruguay
Sanguinetti I Minority
Lacalle Majority
Sanguinetti I1 Majority
NEGRETTO: MINORITY PRESIDENTS 77

Table 2 (continued)

President Legislative Status of Coalition


Venezuela
Caldera I1 Minority
Chive2 I Minority

Sources: Based o n LAW?; Keesing’s Record of World Events On Line; Funddcibn


CIDOB; and database on cabinet coalitions provided by Cecilia Martinez-Gallardo,
Data on Ecuador collected from Burbano de Lara and Garcia 1998. Data on Panama
and Uruguay provided by David Altman.

Within each category, it seems that the weaker the bargaining position
of the president, the more frequent the inclusion of representatives from
parties other than the president’s. Six (40 percent) of the presidents in
median situations, 9 (47 percent) in divided situations, and 12 (71 per-
cent) in congressional situations formed cabinet coalitions.
The situation is different with majority cabinet coalitions. Relatively
few cabinet coalitions have a majority status in Congress. Of the 27
coalitions formed, 15 (29 percent) were minority and 12 (24 percent)
were majority coalitions. Within each category, 4 (27 percent) of the
presidents in median situations, none (0 percent) of the presidents in
divided situations, and 8 (47 percent) in congressional situations formed
majority cabinet coalitions. Here again, the largest percentage of major-
ity coalitions in congressional situations is also consistent with the
weakness of presidents in these cases.
With this information at hand, this study proposes a new classifica-
tion. Instead of a simple division between majority and minority presi-
dents, we may differentiate between majority and minority presidential
governments. Majority governments will be divided into single-party
and coalition. Single-party majority governments are all those cases
labeled “unified” in table 1. Coalition majority governments include all
the cases in which minority presidents built a multiparty cabinet coali-
tion holding a majority of seats in Congress. Minority governments,
whether single party or coalition, are divided into median, divided, and
congressional, according to the typology in table 1. In other words, we
assume that in the absence of a majority cabinet, what matters for the
performance of presidential regimes is the relative position of the pres-
ident’s party in regard to the median and veto parties.
With this revision, we are left with 36 cases of majority government
(24 single-party and 12 coalition) and 39 cases of minority government
(11 median, 19 divided, and 9 congressional). This classification will be
used to test the main hypotheses of this study about the expected levels
of interbranch cooperation in presidential regimes.
78 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Table 3. Frequency and Type of Coalitions Among Presidents in


Minority Situations
~~ ~

Minority Cabinet
Situation N Coalitions (%) Minority (%) Majority (%)
Median 15 6 (40) 2 (13) 4 (27)
Divided 19 9 (47) 9 (47) 0 (0)
Congressional 17 12 (71) 4 (24) 8 (47)
Total 51 27 (53) 15 (29) 12 (24)

TESTING
THE PERFORMANCE
OF
MINORITY PRESIDENTIAL GOVERNMENTS

The conventional view of minority presidents, as noted, is that their per-


formance is associated with high levels of executive-legislative conflict and
democratic instability. This study critically examines that view by looking
at three other variables: the policy position of the president’s party, the
president’sveto-sustaining capacity, and the legislative status of the parties
included in the cabinet. Two indicators were used to perform a prelimi-
nary test of the impact of these factors on the performance of minority
presidential governments. These indicators were the frequency of execu-
tive-legislative conflicts and the number of premature terminations of a
president’sconstitutional term in Latin America from 1978 to 2003.
Following a methodology proposed by Mark Jones (19951, an index
was constructed of executive-legislative conflicts, based on information
in the Latin America Weekly Report from 1978 to 2003. The index
reflects the number of articles per year that reported executive-legisla-
tive conflicts, compared to the total number of articles that had politics
as a main or secondary theme. The conflicts reported were related either
to events of ordinary legislative bargaining, such as the rejection of pres-
idential proposals or presidential vetoes, or to acts of congressional con-
trol, such as the censure of cabinet ministers o r impeachments. The unit
of analysis is each presidency classified according to the typology pro-
posed in the previous section.l6 The presidencies include constitutional
periods duly completed or prematurely terminated by 2003.
Using the same universe of cases, a separate analysis of those pres-
idencies that ended prematurely was also undertaken. A constitutional
term was considered to end prematurely when presidents were ousted
from office after a successful impeachment process; when they decided
to dissolve the congress, either legally or illegally; when they were
forced to resign after mass protests against the government; or when
their government was overthrown by a military coup. Fourteen cases of
premature termination occurred between 1978 and 2003.
NEGRE'ITO: MINORITY PRESIDENTS 79

Table 4. Executive-Legislative Conflict Index

Mean Standard
Type of Government N Frequency Deviation
Majority
Single party 24 0.082 0.130
Coalition 12 0.088 0.073
Minority
Median 11 0.091 0.099
Divided 19 0.114 0.118
Congressional 9 0.147 0.072

Source: Based o n LAW" 1978-2003. Countries and years: Argentina, 1983-2001;


Bolivia, 1982-2003; Brazil, 1985-2002; Chile, 1989-2001; Colombia, 1979-2001; Costa
R i a , 1979-2001; Dominican Republic, 19862001; Ecuador, 1979-99; El Salvador,
1985-2001; Guatemala, 19862003; Honduras, 1982-2001; Mexico, 1997-2000;
Nicaragua, 1990-2001; Panama, 1994-99; Paraguay, 1989-98; Peru, 1980-2000;
Uruguay, 1985-2001; Venezuela, 1979-2001.

Executive-LegislativeConflicts

Table 4 shows the mean frequency of conflict and the standard deviation
for each category of government. As might be expected, the mean level
of executive-legislative conflict tends to increase as we move from single-
party and coalition majority government to minority governments.
Among the latter, however, the category with the highest mean level of
executive-legislative conflict is congressional minority government.
A t test of the equality of means shows that interbranch cooperation
is severely impaired only when the president's party lacks the support
of both the median and the veto legislator and no cabinet coalition hold-
ing a majority of legislative seats is formed. The mean levels of inter-
branch conflict of single-party or coalition majority government and
median or divided minority government are not statistically different
from 0. There is, however, a statistically significant difference, at the 0.05
level, between the mean levels of executive-legislative conflict of major-
ity governments and congressional minority government. There is also
a statistically significant difference, at the 0.10 level, between the mean
levels of conflict of single-party or coalition majority government and
congressional minority government.
The crucial factor for predicting the level of executive-legislative
conflict, therefore, is not simply the share of seats of the president's
party. Other factors must concur to account for a significant increase in
the level of executive-legislative conflicts. These factors are related to
the policy position of the president's party, the ability o f the president
80 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

to sustain a veto, and the legislative status of the parties included in the
cabinet.
The main reason for these results is that several categories of gov-
ernment show wide variation in the levels of executive-legislative con-
flict. Among single-party majority governments, for instance, the Cubas
presidency (1998-2002) in Paraguay had the highest frequency of inter-
branch conflict in the entire database (60 percent). A relatively high
level of conflict (29 percent) also occurred during the presidency of Vio-
leta Chamorro (1990-96) in Nicaragua. While these cases may be out-
liers, they suggest the need to find an adequate measure of intraparty
factionalism to determine the level o f interbranch cooperation in single-
party majority governments.
There is also wide variation in the levels of interbranch conflict of
median and divided minority governments. In the case of median gov-
ernments, the assumption that parties are disciplined and legislators care
mostly about policy does not hold equally for all the countries under
consideration. Parties in Chile and Uruguay, for instance, seem to be
more ideology-based than parties in Ecuador or Peru. This variation
should have an impact on how the policy position of the president’s
party matters for interbranch relations. Among divided governments, the
differences in performance may depend on the preferences of policy-
makers regarding the status quo. In the absence of an external shock, a
president armed with only a veto and intending to introduce drastic
policy changes may face intense interbranch conflict, as occurred during
the presidencies of Febres Cordero (1984-88), D u r h Ballen (1992-96),
and Bucdram (1996-97) in Ecuador.
It is interesting to note that while coalition majority government
does not appear to perform differently from median or divided govern-
ments, it does differ when compared to congressional governments. As
argued earlier, presidents who face a congressional situation may be
able to avoid conflicts with Congress if they attempt to distribute cabi-
net positions among parties whose votes are necessary to reach a leg-
islative majority. The presidency of Fernando Collor de Mello in Brazil
serves As a good illustration. Although all presidents in Brazil governed
in a congressional situation, only Collor (1990-92) failed to build and
maintain a majority cabinet coalition during his term.” He faced the
highest rates of interbranch conflict among all Brazilian presidents and
was forced to resign after an impeachment process.

Presidencies Interrupted
The cases of presidencies terminated before the end of the constitu-
tional term merit a separate analysis. In the 1960s and 1970s, a consti-
tutional crisis in Latin America typically involved the participation of the
NEGRETTO: MINORITY PRESIDENTS 81

military, whether to replace the democratic regime with a military dic-


tatorship or provisionally to suspend the constitution. Since 1978, how-
ever, constitutional crises have had a different outcome. With the partial
exception of Peru from 1992 to 1995, constitutional crises in Latin Amer-
ica have ended not in a formal breakdown or suspension of democratic
procedures but in a premature termination of constitutional terms. As
Perez Lidan (2003, 162) points out, contemporary constitutional crises
are more related to government than to regime instability.
Table 5 provides a list of interrupted presidencies from 1978 to
2003, indicating the president, year, and reason for termination. It shows
that a total of 14 presidencies were interrupted from 1978 to 2003. In
most cases, the origin of the crisis involved a complex interaction
between social mobilization, either spontaneous or activated by politi-
cal actors, and institutional conflict.IH
In four cases (Collor, Perez, BUCardm, and Cubas), presidencies were
terminated in the context of an impeachment process, and in one case
(Suazo) in the context of a series of executive-legislative conflicts that
included the threat of impeachment. In five cases, presidents resigned or
abandoned the office, three times (Alfonsin, De la Rua, and De Lozada)
due to mass protests against the government’s policies, once (Mahuad)
due to an attempted military coup, and once (Fujimori 111) because of a
corruption scandal. In one case (Balaguer HI), the president decided to
call early elections because of internal and international pressures. In
one case (Chiivez), the term was terminated by the president’s decision
to convene a popular constituent assembly and reelect authorities under
the new constitution. In two cases (Fujimori I and Serrano), the crisis
consisted of the illegal dissolution of the existing congress, although in
one of these (Serrano) the president failed in this attempt.
In spite of the variations, there is a common element in all these
cases: the premature termination of the presidential term signals, in its
most extreme form, the breakdown of cooperation between presidents
and congresses. Whatever the nature of the crisis that presidents and
legislators had to face, it seems clear that in most cases, either the con-
gress or the president saw the elimination of the other as the solution
to the crisis. Even in cases in which social conflict preceded interbranch
conflict, either the president or the congress decided to withdraw sup-
port for the other branch. From this perspective, the forced resignation
or impeachment of presidents, on the one hand, and the replacement
or dissolution of congresses by presidents, o n the other, can be seen as
symmetrical noncooperative outcomes.
Of the 14 cases of premature termination of terms, one occurred in
single-party majority government, one in coalition majority government,
one in median, six in divided, and five in congressional governments. If
we divide the cases by type of government, we see that the category
82 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Table 5. Premature Termination of Presidents’ Constitutional Terms in


Latin America, 1978-2003

President Date Motive


Siles Suazo November 1984 Forced resignation amid economic
crisis, social upheaval, coup threat,
executive-legislative conflicts,
including threat of impeachment
Alfonsin July 1989 Forced resignation amid economic
crisis, social upheaval
Fujimori I April 1992 Dissolved Congress, with support
of military, after a series of execu-
tive-legislative conflicts over
president’s power to rule by decree
on economic management and
terrorism
Collor August 1992 impeached by Congress on corrup-
tion charges amid economic crisis,
social upheaval, long series of
executive-legislative conflicts
Serrano May 1993 Forced resignation after failed
attempt to dissolve Congress and
rule with support of military
Perez August 1993 Removed from office by Congress
on corruption charges initiated after
violent social upheaval against eco-
nomic policy and failed coup
Balaguer August 1996 Early call for elections due to
internal and international pressures
after allegations of fraud in 1994
presidential elections
Bucaram February 1997 Removed from office by Congress
on charges of mental incapacity ini-
tiated after mass protests against
president’s economic policy and
alleged government corruption
Cubas March 1999 Resigned after impeachment
process initiated by Chamber of
Deputies
Ch5vez I April 1999 Irregular constituent assembly
terminated constitutional period
and replaced existing congress
NEGRE'ITO: MINORITY PRESIDENTS 83

Table 5 (continued)

President Date Motive


Mahuad January 2000 Abandoned office after massive
protests against economic policies
and attempted coup
Fujimori I11 November 2000 Removed from office by Congress
after corruption scandal
De la R6a December 2001 Forced resignation after mass
protests against economic policy,
and after the congressional opposi-
tion in congress rejected national
union coalition to support the gov-
ernment
Sanchez de September 2003 Forced resignation after massive
Lozada I1 protests against economic policy
~ ~~

Source: Based on Latin American Weekly Report, 1978-2003; Fundacicin CIDOB.

with the highest relative percentage of premature terminations is con-


gressional government. Five out of nine presidencies in this category-
that is, 55 percent-were interrupted before the end of the constitu-
tional term. Divided government, with 6 out of 19, or 32 percent of the
presidencies interrupted, appears as the second most frequent type. At
the other extreme, single-party majority government is the category with
the lowest percentage of interrupted presidencies, followed by coalition
majority governments and median minority governments. These relative
frequencies are shown in table 6.
A rare events logistic regression was used to estimate the impact of
each form of government on the probability of a premature termination
of constitutional terms.19 The dependent variable is the interruption of
constitutional terms, coded as a dummy variable, which equals 1 when
the constitutional term was prematurely terminated and 0 otherwise.
The main independent variable is the type of government. Model 1
traces the impact of majority presidents on the probability of early ter-
mination. The variable MAJORITYPRES is coded as a dummy variable,
which equals 1 when the president's party has a majority status in Con-
gress, and 0 otherwise. The main variable in model 2 is MAJORITYGOV,
which equals 1 when the cabinet has a majority status in Congress, and
0 otherwise. Model 3 measures the impact of each form of minority gov-
ernment on the probability of a premature termination of constitutional
terms. Three dummy variables are used: MEDIAN, DIVIDED, and CON-
GRESSIONAL, coded as 1 when the government is median, divided, or
84 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Table 6. Premature Termination of Constitutional Terms by


Type of Government

Premature
Type of Government N Termination Percent
Majority
Single party 24 1 4
Coalition 12 1 8
Minority
Median 11 1 9
Divided 19 6 32
Congressional 9 5 55

congressional, and 0 otherwise. Majority government, either single-party


or coalition, is used as the implicit comparison group.
Two control variables, DECREE and CENSURE, were added to the
three models. DECREE traces the impact of the president’s power to
enact decrees of legislative content on the probability of a premature
termination of constitutional terms. It is coded as a dummy variable that
equals 1 when the constitution grants the power to issue decrees that
become law immediately, and 0 otherwise. CENSURE measures the
impact of the power of Congress to force the resignation of cabinet min-
isters on the probability of early termination. It is also coded as a
dummy variable that equals 1 when cabinet ministers are subject to a
binding motion of censure, and 0 otherwise.
Table 7 summarizes the results of the regressions. In model 1,
MAJORI’IYPRES had the expected (negative) sign but did not reach sta-
tistical significance. DECREE and CENSURE were not found to be sig-
nificant, either. In model 2, the coefficient of MAJORITYGOV was sig-
nificant at the 0.05 level and negative, indicating that the probability of
a premature termination of constitutional terms decreases from minority
to majority government. The estimates of the coefficients of DECREE
and CENSURE were positive, but neither was statistically significant. In
model 3, the coefficient of DIVIDED was positive and statistically sig-
nificant at the 0.05 level. The coefficient of CONGRESSIONAL was also
positive and statistically significant at the 0.01 level. MEDIAN, DECREE,
and CENSURE were not statistically significant.
A comparison of the three models indicates that whether or not the
president’s party holds a majority of seats is not the crucial variable for
predicting an irregular interruption of the constitutional term. Having the
support of a legislative majority is relevant only when w e include the leg-
islative status of cabinets. But even if minority government increases the
likelihood of a premature termination of constitutional terms, the proba-
NEGRETTO: MINORITY PRESIDENTS 85

Table 7. Premature Termination of Constitutional Terms in


Latin America, 1978-2003 (rare events logistic regression)

Independent Variables Model 1 Modely- Model 3


MAJORITYPRES -1.559
(1.033)
MAJORITYGOV -1.778*’
(.761)
MEDIAN .820
(1.144)
DIVIDED 1.763-
(.810)
CONGRESSIONAL 2.776***
(.923)
DECREE .285 .419 316
C.627) (.656> (.611)
CENSURE -.Of39 .220 ,320
(.607) (.618) (.621)
CONSTANT -1.050** -.990* -2.775***
(.525) (.556) (.801)

-
Notes: Dependent variable: Premature termination of constitutional terms
N for all three models 75
Numbers in parentheses are robust standard errors
* p < .lo, ** p < .05, *** p < .01

bility is only significant in divided and, particularly, in congressional


minority governments. Specifically, while there is a 26 percent probabil-
ity of premature termination of constitutional terms in divided govern-
ments, the probability rises to 50 percent in the case of congressional
governmenta20This means that the relative probability of a n early termi-
nation of constitutional terms in a congressional government is almost
twice the probability of an early termination in a divided government.
These results reinforce the conclusion of the previous analysis of
executive-legislative conflicts. It is not simply the frequency of “minor-
ity presidents,” as recently argued by some authors (Valenzuela 2004,
121, that facilitated the interruption of presidential terms in Latin Amer-
ica during the last two decades. The rate of interrupted presidencies per
type of presidential government indicates that factors more crucial than
the share of seats of the president’s party are the location of this party
in the policy space, the capacity of the president to sustain a veto, and
the legislative status of the parties included in the cabinet.
86 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Looking at the substance of actual interbranch conflicts, other pres-


idencies that did end their terms without sufficient support from coali-
tions could very well have ended prematurely. Those presidencies tend
to fall into the categories of divided and congressional governments.
This was the case for the divided presidential governments of Febres
Cordero, Boja (for the period 1990-92), and Durin BallCn in Ecuador. All
these presidents tried to alleviate Ecuador’s fiscal crisis and debt problems
with austerity policies that were paralyzed or watered down in Congress.
In reaction to the massive protests generated by these policies, opposition
legislators continuously harassed the president by impeaching his ministers
(Conaghan 1995, 452-53). In the last year of Borja’s presidency, moreover,
Congress formally asked the Supreme Court to initiate impeachment pro-
cedures against the president himself. In 1995, after the censure of several
of Dur5n BallCn’s ministers, Congress impeached his vice president (see
LAW 1990, 1995). Durin Ballen responded with an unsuccessful attempt
to call a constituent assembly and reform the constitution to strengthen the
president’s powers in relation to the congress.
The same is true of several presidents in congressional situations.
AndrCs Pastrana of Colombia (2 la Hugo Chivez of Venezuela) tried but
failed to call a referendum on constitutional reform, which included cut-
ting the number of congressional seats by half and calling new congres-
sional elections (LAW 2000). During Juan Carlos Wasmosy’s presidency
in Paraguay, the president and legislators were engaged in a bitter con-
flict that included a congressional threat of impeachment (LAW!? 1994b).
Rafael Caldera’s second presidency in Venezuela, like his first, became a
symbol of divisive government. Caldera decided to face the country’s
deep economic crisis by suspending economic rights and implementing
economic measures by decree (Crisp 2000). When Congress attempted
to restore those rights, it was forced to back down by Caldera’s threat to
call a referendum on a constitutional reform proposal, which would have
enabled him to dissolve Congress (LAW 1994a).
Individual case studies would probably confirm the findings of both
the comparative analysis of the frequency of interbranch conflicts and
the special case of premature termination of constitutional terms. The
probability of observing high levels of executive-legislative conflict or
interruption of constitutional terms in presidential regimes increases
when the president’s party lacks the support of both the median and the
veto legislator and no cabinet coalition holding a majority of legislative
seats is formed.

CONCLUSIONS
This study has provided a first step toward developing a more theoret-
ically grounded explanation of the causal mechanisms that foster or
NEGRETTO: MINORITY PRESIDENTS 87

hinder cooperation between executives and legislators when the presi-


dent’s party falls short of a legislative majority. It has argued that three
factors, apart from the share of legislative seats of the president’s party,
should affect the performance of presidential regimes: the location of
the president’s party in the policy space, the president’s capacity to sus-
tain a veto, and the legislative status of the parties included in the cab-
inet. Following this logic, this article has proposed that the most con-
flictive situation in a presidential regime occurs when the president’s
party lacks the support of both the median and the veto legislator and
no cabinet coalition holding a majority of legislative seats is formed.
This hypothesis is supported by both the frequency of executive-
legislative conflicts and the premature termination of constitutional
terms in Latin America during the period 1978-2003. Both indicators
show that the category of presidential government most often associated
with high levels of executive-legislative conflict and interrupted consti-
tutional terms is that in which the president’s party has no control over
the median and the veto legislator, and the parties included in the cab-
inet hold a minority of seats in the congress.
This analysis suggests that research on presidential democracies
should shift from its exclusive focus on the electoral variables that deter-
mine the number of parties competing in elections to the electoral and
institutional variables that make cooperation between branches possible
in the absence of a legislative majority supporting the president. In this
respect, a new research agenda should pay attention to electoral formu-
las that make the selection of presidents whose parties occupy a median
position in Congress more likely than others, the influence of presidents’
reactive and proactive legislative powers on policy outcomes, and the
factors that hinder or foster the formation of majority cabinet coalitions.
A new research agenda should also explore more carefully the rela-
tionship between social conflict, interinstitutional conflict, and constitu-
tional crises. Most cases of premature interruption of constitutional
terms occurred in the midst of mass protests, usually in reaction to
unpopular economic policies. If institutions really have had a decisive
influence on these outcomes, then scholars of presidential democracies
face the challenge of showing to what extent o r under what conditions
institutional rules played an independent causal role.

NOTES
The author gratefully acknowledges valuable comments and criticisms on
previous versions of this paper from Javier Aparicio, Josep Colomer, Bill Keech,
Cecilia Martinez-Gallardo, Bingham Powell, Jr., and Arturo Valenzuela; and
research assistance from Luis Eduardo Escatel. The final version was also
improved by the comments of two anonymous reviewers.
IATIN AMERICAN POLITICS AND SOCIETY 48: 3

1. The definition of divided government as divided control is common


among scholars who study the U.S. presidential regime and assume the exis-
tence of a two-party system. See, for instance, Laver and Shepsle 1991, 252. Stu-
dents of Latin American presidential regimes include “no majority” situations in
the definition because this is by far the most common condition in a region
where multiparty systems abound. Shugart (1995, 327) is perhaps the only
author who differentiates between ‘divided government in the strict sense of
divided control and “no majority” situations.
2. For the debate on whether divided government leads to legislative grid-
lock in the context of the U.S. presidential regime, see Kernell 1991; Mayhew
1991; Sundquist 1992; Jones 1994; kehbiel 1998. For the incompatibility
between minority presidencies and stable democracy in Latin America, see Linz
1994; Shugart and Carey 1992; Mainwaring 1993; Jones 1995; Mainwaring and
Shugart 1997a, b.; Kenney 2004.
3. The effective number of legislative parties results from the application
of the Laakso and Taagepera (1979) formula: 1 divided by the sum of the
squared decimal shares of the seats won by each electoral party.
4. The only explanations advanced to date for the instability of presiden-
tial democracy are based on three rather disparate hypotheses: the existence of
three parties of relatively equal strength, the frequency of presidential term
limits, and decentralized policymaking processes. See Cheibub 2002a; Cheibub
and Limongi 2002.
5 . This scenario assumes minimal winning coalitions; that is, coalitions
that turn into a losing coalition by the subtraction of any of their members. See
Laver and Schofield 1990, 92.
6. A coalition is connected when all its members will be adjacent to each
other on a single policy dimension. A coalition is minimally winning and con-
nected when it is n o longer winning or connected by the subtraction of any one
of its members. See Laver and Schofield 1990, 97.
7. This model draws on Krehbiel 1998; see also Colomer and Negretto
2005.
8. Some authors (Jones 1995) have argued that interbranch cooperation in
minority presidencies improves when the percentage of legislative seats of the
president’s party is close to a majority; for instance, equal to or greater than 45
percent. From our perspective here, however, this argument lacks theoretical
foundation. If the president’s party has 45 percent of the seats but is located at
an extreme position on the policy space, executive-legislative relations may
cause more conflict than when the president’s party has, say, 30 percent of leg-
islative seats but is located at a relatively centrist position.
9. In the extreme case, a president would always have a veto, regardless
of the party’s share of seats, if the constitution provided the executive with a
veto that was not subject to override. This is the case in Ecuador, which pro-
vides the president with a veto that cannot be overridden for one year. This is
the reason that most presidents in Ecuador have governed in a divided situa-
tion, even though most of them had only marginal support in the legislature.
10. Interbranch conflict may occur, for instance, when the president sub-
mits a proposal for legislative change and the status quo policy is close to the
ideal point of the median party. Under this configuration of preferences, the
NEGRETTO: MINORITY PRESIDENTS 89

president’s veto is ineffective; the pace and content of legislative change WOLIICI
be entirely in the hands of the median party.
11. In principle, one party may control both, or different parties may con-
trol each of these legislators.
12. According to Amorini Neto and Tafner 2002 and Amorim Neto 2006,
presidents either seek a statutory implementation of policy goals by means of
partisan and proportional majority cabinet coalitions, or attempt to legislate by
decree. It could be argued that the latter strategy (whether or not those powers
are granted by the constitution) is more likely when minority presidencies face
a “congressional” situation.
13. Party policy positions were classified on the basis of a comparison of
the expert opinions gathered by Coppedge (19971, Mainwaring and Scully
(1995), and Alchtara and Freidenberg (2001a, b, c). Whenever the president’s
party was consistently labeled as populist or personalist during a particular
period, the presidency was classified based only on the share of seats of the
president’s party, rather than on its location on the policy space. This was the
case for the Partido de ReconstruGio Nacional (PRN) in Brazil, Concentracion de
Fuerzas Populares (CFP) and Partido Roldosista Ecuatoriano (PRE) in Ecuador,
Cambio 90 (C90) in Peru, and Convergencia Nacional (CN) in Venezuela.
14. One restriction to the autonomy of presidents in the appointment of
cabinet ministers could be the congressional ratification requirement, as in the
United States. A restriction on the autonomy of presidents in maintaining or
changing their cabinets is the capacity of Congress to force the resignation of
cabinet ministers, as in several Latin American countries.
15. A presidency was considered to have formed a majority coalition gov-
ernment when the cabinet maintained a majority legislative status until at least
one year before the end of the presidential term.
16. Unlike Mark Jones (19951, who uses each presidential year as a unit of
analysis, this study did not exclude from the count those years in which the
number of articles was fewer than six. This means that in some cases, the index
of the frequency of conflicts is lower (when no conflicts were reported that
year) and in other cases higher (when conflicts were reported) than Jones’.
17. Collor maintained a cabinet coalition with majority status in both cham-
bers of congress only from March to October 1990. See Amorim Net0 2002, 56.
18. On the role of social conflicts in the interruption of presidencies, see
Hochstetler 2004.
19. A rare events logistic regression is used to obtain an approximately
unbiased estimate of the probability of rare events, such as wars, coups, o r rev-
olutions. The model is described in King and Zeng 1999. The papers and the
software are available online.
20. These probabilities are calculated based on the rare events logistic
regression results, allowing the variable of interest to take a value of 1 while
keeping all the other variables at 0. The stata commands are provided by King
and Zeng 1999.
90 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

REFFWNCES
Alciintara, Manuel, and Flavia Freidenberg, eds. 2001a. Partidos politicos de
America Latina: Centroamerica, Mexico y Republica Dominicana. Biblioteca
de America Series, 20. Salamanca: Ediciones Universidad de Salamanca.
-. 2001b. Partidos politicos de America Latina: Cono Sur. Biblioteca de
America Series, 19. Salamanca: Ediciones Universidad de Salamanca.
-. 2001c. Partidospoliticos de America Latina: paises andinos. Biblioteca de
America Series, 18. Salamanca: Ediciones Universidad de Salamanca.
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: 259-83.
Amorim Neto, Octavio. 1998. Of Presidents, Parties, and Ministers: Cabinet For-
mation and Legislative Decision-Making Under Separation of Powers. Ph.D.
diss., University of California, San Diego.
-. 2002. Presidential Cabinets, Electoral Cycles and Coalition Discipline in
Brazil. In Legislative Politics in Latin America, ed. Scott Morgenstern and
Benito Nacif. New York: Cambridge University Press. 48-78.
-. 2006. The Presidential Calculus: Executive Policy Making and Cabinet For-
mation in the Americas. Comparative Political Studies 39, 4: 415-40.
Amorim Neto, Octavio, and Paulo Tafner. 2002. Governos de coaliz2o e mecan-
ismos de alarme e incendio n o controle legislativo das medidas provisorias.
Dados 45, 1: 5-38.
Rrady, David, and Craig Volden. 1998. Revolving Gridlock: Politics and Policy
From Carter to Clinton. Boulder: Westview Press.
Burbano de Lara, Felipe, and Michel Rowland Garcia. 1998. Pugna depoderes.
Presidencialismo y partidos en el Ecuador, 197S1997 Quito: Proyecto
CORDES-Gobernabilidad.
Cameron, Charles M. 2000. Veto Bargaining. Presidents and the Politics of Neg-
ative Powers. New York: Cambridge University Press.
Carey, John M., and Matthew Soberg Shugart, eds. 1998. Executive Decree
Authority. New York: Cambridge University Press.
Chasquetti, Daniel. 2001. Democracia, multipartidismo, y coakiones en America
Latina: evaluando la dificil combinacibn. In Tipos de presidencialismo y
coaliciones politicas en America Latina, ed. Jorge Lanzaro. Buenos Aires:
CLACSO. 319-57.
Cheibub, Jose Antonio. 2002a. Minority Governments, Deadlock Situations, and
the Survival of Presidential Democracies. Comparative Political Studies 35:
284-312.
-. 2002b. Presidentialism and Democratic Performance. In m e Architecture
of Democracy: Constitutional Design, Conflict Management, and Democ-
racy, ed. Andrew Reynolds. New York: Oxford University Press. 104-40.
Cheibub, Jose Antonio, and Fernando Limongi. 2002. Democratic Institutions
and Regime Survival: Parliamentary and Presidential Democracies Recon-
sidered. Annual Review of Political Science 5: 151-79.
Cheibub, Jose Antonio, Adam Przeworski, and Sebastian Saiegh. 2004. Govern-
ment Coalitions and Legislative Success Under Presidentialism and Parlia-
mentarism. BritishJournal of Political Science 34: 565-87.
NEGRETTO: MINORITY PRESIDENTS 91

Colomer, Josep, and Gabriel Negretto. 2005. Can Presidentialism Work Like Par-
liamentarism? Government and Opposition 1: 60-89.
Conaghan, Catherine M. 1995. Politicians Against Parties: Discord and Discon-
nection in Ecuador’s Party System. In Mainwaring and Scully 1995. 434-58.
Coppedge, Michael. 1997. A Classification of Latin American Political Parties.
Working Paper 244. Notre Dame: Kellogg Institute, University of Notre
Dame.
Crisp, Brian. 2000. Democratic Institutional Design: 7he Powers and Incentives
of Venezuelan Politicians and Interest Croups. Stanford: Stanford University
Press.
Deheza, Grace Ivana. 1998. Gobiernos de coalici6n en el sistema presidencial:
America del Sur. In El presidencialismo renovado: institucionalismo y
cambio politico en Amkrica Latina, ed. Dieter Nohlen and Mario Fernin-
dez. Caracas: Nueva Sociedad. 151-69.
Centro de Investigation, Docencia, Documentacion y Divulgacion de Relaciones
Internacionales y Desarrollo (CIDOB). 2005. Barcelona: Fundacion CIDOB.
December. <www.cidob.org/bios/Cidob.com>
Elections Around the World. 2004. Ed. Wilfried Derksen. Website database. Ver-
sion 06/23/2004. <www.electionworld.org>
Hochstetler, Kathryn. 2004. Rethinking Presidentialism: Challengers and Presi-
dential Falls in Latin America. Unpublished mss. [Published, Comparative
Politics 38, 4 uuly 20061.1
Jones, Charles 0. 1994. 7be Presidency in a Separated System. Washington, DC:
Brookings Institution Press.
Jones, Mark P. 1995. Electoral Laws and the Survival of Presidential Democra-
cies. Notre Dame: University of Notre Dame Press.
Keesing’s Record of World Events On Line. 2005. <www.keesings.com> December.
Kenney, Charles D. 2004. Fujimori’s Coup and the Breakdown of Democracy in
Latin America. Notre Dame: University of Notre Dame Press.
Kernell, Samuel. 1991, Facing an Opposition Congress: The President’s Strategic
Circumstance. In The Politics of Divided Government, ed. Gary Cox and
Kernell. Boulder: Westview Press. 87-1 12.
King, Gary, and Langche Zeng. 1999. Logistic Regression in Rare Events Data.
Department of Government, Harvard University. Available at <http://gking.
harvard.edu/stats.shtml>Accessed February 2005.
Krehbiel, Keith. 1998. Pivotal Politics: A Theory of U S . Lawmaking. Chicago:
University of Chicago Press.
Laakso, Markku, and Rein Taagepera. 1979. Effective Number of Parties: A Mea-
sure with Application to West Europe. Comparative Political Studies, 12, l :
3-27.
Latin America Weekly Report (LAWR). 1990. November 1.
-’ 1994a. December 8.
-. 1994b. November 8.
-. 1995. Congress Confirmed Its Decision to Impeach Ecuador’s Vice Presi-
dent. September 21.
-. 2000. May 16.
Laver, Michael, and Kenneth Shepsle. 1991. Divided Government: America Is
Not Exceptional. Governance 4, 3: 250-69.
92 LATIN AMERICAN POLITICS AND SOCIETY 48: 3

Laver, Michael, and Norman Schofield. 1990. Multiparty Government: 7be Poli-
tics of Coalition in Europe. Ann Arbor: University of Michigan Press.
Linz, Juan. 1994. Presidential or Parliamentary Democracy: Does It Make a Dif-
ference? In The Failure of Presidential Democracy, ed. Linz and Arturo
Valenzuela. Baltimore: Johns Hopkins University Press. 3-87.
Mainwaring, Scott. 1993. Presidentialisin, Multipartism, and Democracy: The Dif-
ficult Combination. Comparative Political Studies 26: 19S228.
Mainwaring, Scott, and Matthew Soberg Shugart, eds. 1997a. Presidentialism
and Democracy in Latin America. New York: Cambridge University Press.
-. 1997b. Juan Linz, Presidentialism, and Democracy: A Critical Appraisal.
Comparative Politics 29, 4: 449-71.
Mainwaring, Scott, and Timothy Scully, eds. 1995. Building Democratic Institu-
tions: Party Systems in Latin America. Stanford: Stanford University Press.
Martinez-Gallardo, Cecilia. 2005. Presidents, Posts, and Policy: Ministerial
Appointments and Political Strategy in Presidential Systems. Ph. D. diss. I

Columbia University.
Mayhew, David. 1991. Divided We Govern. New Haven: Yale University Press.
Negretto, Gabriel. 2004. Government Capacities and Policy-Making by Decree in
Latin America: The Cases of Brazil and Argentina. Comparative Political
Studies 37: 531-62.
Nohlen, Dieter, ed. 1993. Enciclopedia electoral latinoamericana y del Caribe.
San Jos6, Costa Rica: Instituto Interamericano de Derechos Humanos.
Payne, J. Mark, Daniel Zovatto G., Fernando Carrillo FlBrez, and Andrks Alla-
mand Zavala. 2002. Democracies in Development: Politics and Reform in
Latin America. Washington, DC: Inter-American Development Bank and
the International Institute for Democracy and Electoral Assistance/Johns
Hopkins University Press.
Perez LiAin, Anibal. 2003. Pugna de poderes y crisis d e gobernabilidad: ihacia
un nuevo presidencialismo? Latin American Research Review 38,3: 149-64.
Sarnuels, David, and Kent Eaton. 2002. Presidentialism And, Or, and Versus Par-
liamentarism: The State of the Literature and an Agenda for Future
Research. Paper presented at the Conference “Consequences of Political
Institutions in Democracy,” Duke University, April 5-7.
Shugart, Matthew Soberg. 1995. The Electoral Cycle and Institutional Sources of
Divided Presidential Government. American Political Science Review 89, 2:
327-43.
Shugart, Matthew Soberg, and John M. Carey. 1992. Presidents and Assemblies:
Constitutional Design and Electoral Dynamics. New York: Cambridge Uni-
versity Press.
Strom, Kaare. 1990. Minority Government and Majority Rule. Cambridge: Cam-
bridge University Press.
Sundquist, James L. 1992. Constitutional Reform and Effective Government.
Washington, DC: Brookings Institution Press.
Valenzuela, Arturo. 2004. Latin American Presidencies Interrupted. Journal of
Democracy 15, 4: 5-19.

You might also like