Strong presidents, robust democracies? Separation of powers and rule of law in Latin America^*

Marcus André Melo

Federal University of Pernambuco (UFPE), Brazil

ABSTRACT

The received wisdom on Latin America in the 1980s and 1990s was that countries where presidents enjoyed strong constitutional powers and where multiparty coalitions prevailed would be doomed to instability and institutional crises, while countries boasting weak presidents and strong parties were expected to consolidate democratic rule. After almost two decades, it is now widely acknowledged that this prediction failed. Recent re-conceptualizations of presidentialism have partly corrected the flaws in the established diagnosis but left unexplained the role of checks and balances and of the rule of law in containing presidential abuse and guaranteeing governability. The paper argues that the key to solving the paradox of strong presidents and robust democracies is that democratic stability in Latin American countries is a function of an extended system of checks and balances. These are ultimately generated by power fragmentation at the time of the constitutional choices over their institutional design and political competition sustaining their effective functioning.

Keywords: Checks and balances; Rule of law; Latin America.

Introduction¹

There has been renewed unease among pundits and experts about the imbalance of powers in Latin America's presidencialismo. This is reminiscent of concerns that were pervasive in the 1970s and 1980s about this system of government. Nonetheless, while recent presidential abuse of power and inter-branch conflict in Venezuela, Bolivia and Ecuador have attracted the bulk of the attention,² in some other countries — including Chile and Brazil — there has been much praise for their being on the road to good governance (Stein and Tommasi 2008). In the latter set of countries, which in the early 1990s were thought to be doomed to failure because of some alleged flaws in their constitutional design — "exaggerated presidentialism", large number of effective parties and corresponding multiparty coalition government and open list proportional representation, among others³ —, there has been, paradoxically, stability or even an increase in the constitutional powers wielded by presidents.⁴ By contrast, the countries with the weakest executives in the region in the 1980s — such as Venezuela, Bolivia and Peru — have been the ones that have experienced greater instability and governability problems. Presidents enjoying few constitutional powers have imposed their preferences resorting to an array of informal authoritarian practices and unconstitutional means, thereby creating great institutional instability. This puzzle begs an explanation.

Indeed, the governability landscape in Latin America is markedly heterogeneous. In the Andean region and in Venezuela, there is widespread recognition that inter-branch tensions have grown and that presidents have increasingly abused power. The concentration of authority in the hands of the executive is the chief concern but other governance problems have also come to the fore, including violations of press freedom and systemic corruption, as well as liaisons with drug cartels in the justice system and legislative branch. In several countries, presidents have meddled with the internal functioning of high courts and of legislatures, going far beyond their constitutional attributes. Some of the new rulers in the region have typically adopted anti-system rhetoric and appealed directly to the populace. The Andean region and Venezuela have indeed been the privileged loci for these developments. They have witnessed the collapse of the party system and experienced a dramatic crisis of political representation, but, while less dramatic, sustained abuse of power has also been the rule in other countries such as Argentina.

From a broader perspective, presidential abuse of power is neither novel nor widespread, as the contrasting cases across Latin America demonstrate. Indeed, the travail of democracy in Latin America in the 20^th century has involved a relentless struggle to rein in presidential powers, and abuse of power by authoritarian rulers has been the rule rather than the exception. The reason for concern is that for many observers such practices suggest a reversion of expectations regarding the consolidation of democracy in the region. Concerns about the prospects for democracy in the wake of the third wave of democratization in the Latin American region acquired an academic expression in the 1990s in the debate about delegative democracy (O'Donnel 1994) — a form of democracy where plebiscitarian leaders, oftentimes outsiders with no previous experience in politics, adopted authoritarian practices and a discourse against existing political institutions but were legitimized through the electoral process. This debate is still ongoing and ushered in a recent concern about the quality of democracy, as measured by social indicators, and quality assessments of political institutions.

The overarching theme in all these discussions is the question of governability and its institutional determinants. It seems to be appropriate to re-evaluate, in the light of new research findings and recent public debates, what is wrong — if anything — with Latin American political institutions. Are the political institutions adopted in the region to blame for the underperformance of democracy in the region? The divergent paths of Latin American countries warn us against a generic problem with Latin American institutions. Many conflicts impairing inter-branch cooperation can be avoided when coalitional strategies are successfully implemented and the rule of law is robust. This may be the reason why countries such as Chile and Brazil have outperformed others in the region in terms of the functioning of political institutions.

By exploring the contrasts within Latin America, this paper seeks to establish some preliminary answers to these questions and to the puzzle posed by the contrasting governance outcomes of presidential systems in the region. The main claim developed in the present article is that the key to effective governance and democratic stability in Latin America is the combination of strong presidents with robust checks and balances, and the rule of law. Indeed, while in the early 1990s the institutional design of countries like Bolivia and Venezuela was seen as more conducive to democratic stability and good governance than Chile and Brazil, in the late 2000s, the opposite is true. Bolivia's and Venezuela's institutional design in the early 1990s combined narrow presidential powers with strong party leadership, whereas the constitutional structure of Brazil and Chile rested on strong constitutional powers and weak party leadership.

The main claim in the analysis developed in this article is that the key to promoting sustainable democracy in the region is success in establishing a robust system of checks and balances and rule of law. The latter involves media pluralism, the judicial system and horizontal accountability bodies such as ministérios públicos, tribunais de contas and contralorías, as well as robust mechanisms of parliamentary oversight. In other words, governability also requires that the three branches of government be strong. By exclusively focusing on executive-legislative relations, the extant literature fails to embed them in models of strategic interaction with the latter institutions.⁵ Similarly, the recent burgeoning literature on judicial independence and rule of law in Latin America have uncovered many of the causal mechanisms leading to more effective checks on executives, but have failed to ultimately link this discussion with the discussion on governability and system of government. The analysis presented in this paper is admittedly conceptual and exploratory, but aims to generate testable hypotheses about the determinants of good governance in Latin America.

The paper is organized as follows. The first section reviews the debate on presidentialism and executive-legislative relations, showing the relative inconclusiveness of the debate and, more importantly, that the diagnosis of the ungovernability effects of strong presidents have failed. Strong presidents are not only able to implement their agenda but, more significantly, may be necessary for governability, provided that the institutional checks are in place. The second section discusses institutional and political checks on executive power and the role of the Judiciary, autonomous institutions and the media in reining in executive misbehaviour. Lastly, the final section focuses on the determinants of the rule of law and effectiveness of autonomous institutions. This section sets the stage for a research agenda on the role of checks and balance in containing presidential abuse. There follow some conclusions.

From Regime Types to Separation of Powers and Rule of Law

Over the last thirty years, academics have sought to establish the underlying causes of ungovernability in the region. Much scholarship has centred on the institutional determinants of ungovernability and on institutional crisis. The usual suspects have been the institutions of presidentialism. Indeed now it is possible to trace back the evolution of the intellectual history of this debate and assess to what extent the extant governance problems in the region — particularly as they relate to separation of powers and governability — are associated with this system of government.⁶

In the 1980s and early 1990s, the debate hinged on the nature of presidentialism, taken as a system of government — irrespective of other institutional features — and its alleged inbuilt instability. This view, closely associated with the name of Juan Linz, was proposed as an explanation for the breakdown of democracy and the emergence of military rule in Latin America, but was extended to include presidential systems elsewhere (Linz 1990; Linz and Valenzuela 1994). Indeed this view was very influential in academic circles but criticism of Latin American presidencialismo — a term with strong negative overtones — can be traced back to the beginning of the century. The bottom line of the argument is that presidential systems differ from parliamentary systems in key aspects that are directly related to governability. Presidential systems, according to Linz, create a system of mutual independence — as opposed to the mutual dependence of parliamentary systems — between the executive and legislative branches, allowing for the emergence of executives that do not enjoy majority support. Presidents and prime ministers differ inasmuch as the former are elected separately from the legislative branch and have a fixed term of office. In the language proposed by Shugart and Carey (1992), under presidentialism there is separate origin (a separate popular election for the chief executive) and separate survival (neither the executive nor the legislature may shorten the other's term). Because presidents have fixed terms of office, if a president loses support, he or she remains in office and gridlock ensues.

Where presidents enjoy substantial legislative powers to issue decrees with the force of laws, they would be tempted to bypass congress by issuing decree-laws unilaterally; where they do not, they would attempt to unilaterally change constitutions by mobilizing the direct support of the populace. These attempts exacerbate conflicts, and in many developing countries prompt the military to intervene. A less radical scenario is when conflicts stymie policy-making and the political system remains incapable of producing collective decisions. These deadlocks are less likely to emerge under parliamentarism, according to the well-known Linzian view, because of the latter's flexibility: should a prime minister lose support, due to a motion of censure, the government falls. Presidentialism therefore is supposed to be more conducive either to inter-branch immobilism and/or to crisis (and, by extension, to military coups).

In this stylized Linzian view, parliamentary and presidential systems differ in another crucial dimension: while the former has supposedly inbuilt incentives for party discipline, presidentialism fosters party irresponsibility and individualism. The key element that ensures discipline under parliamentarism, in this view, is the "escape valve" of no-confidence votes, whereby prime ministers threaten legislators with the dissolution of parliament and the call for new elections. The upshot of this threat is that parties become stronger and there are more incentives for coalition formation under parliamentary systems. The lack of a similar tool under presidentialism, in this line of reasoning, explains why in this system parties are weaker and there is less propensity to bargain and to cooperate. Thus, in the Linzian, view presidentialism has a winner-takes-all overall nature that produces confrontation rather than cooperation. Separate origins for presidents and legislatures are part and parcel of this problem because presidents receive a mandate directly from citizens and are symbolically the embodiment of the nation. When they do not enjoy the majority support of elected legislators there emerges a problem of dual legitimacy. Both branches are representative of the citizenry but they may clash, particularly where presidents enjoy minority support, a very likely scenario under presidentialism, according to this analytical model.

This stylized view is well know by Latin American scholars and was the received wisdom in the late 1980s and early 1990s. It provided an explanation for the breakdown of democracy in Latin America as well as predictions about the problems facing the consolidation of the newly established Latin American and Iberian democracies. Linz's mode of explanation provided the analytical key for the institutional malaise affecting these countries. In turn, the success of the Spanish transition provided a model for Latin America and underscored the superiority of parliamentarism in terms of prospects for the survival of democracy. The superior qualities of parliamentary systems were also the object of numerous large sample studies.

A new wave of scholarship in the 1990s led to a much more nuanced notion of presidentialism. Several contributors showed that there is much more variation within presidential democracies than between presidential and parliamentary democracies. The landmark study is Shugart and Carey (1992), who argued that presidential systems differed in crucial dimensions. Presidentialism, in this hugely popular analysis, is only associated with governance crisis where the party system is fragmented and presidents have substantial proactive (exclusive introduction of legislation in specific issue areas, decree authority, agenda power) and reactive powers (veto).

In Latin America there is indeed great variation in the legislative powers of presidents, both overall and in terms of the balance between proactive and reactive powers. Figure 1 provides data compiled by the United Nations Development Program (UNDP), which is primarily based on the typology suggested by Shugart and Carey. The data is normalized from zero to one, and allows us to compare the countries on the two relevant dimensions (reactive and proactive powers). Interestingly, the presidents of Brazil, Chile and Colombia are located at the top of the list of countries with strong legislative powers (and therefore more prone to governability crises), whereas Bolivia, Nicaragua, Paraguay and Honduras are at the bottom. Venezuela was the outlier in terms of reactive powers — its presidents in the pre-Chavez era had virtually no veto power. Mexico and Venezuela also appear at the lower end, along with Costa Rica.⁷ Mexico and Paraguay, which had authoritarian regimes in the 1990s, have presidents who are weak in terms of legislative powers.⁸

In this new wave of scholarship, the key to understanding the nature of party systems was the type of electoral institutions adopted in a constitution or similar statute of a country. By generating a large number of effective political parties, proportional representation was argued to weaken the party system. Conversely, majoritarian systems, by decreasing the number of effective political parties to a small set, were expected to produce a robust party system, which in turn would be instrumental for the support of presidents.

Moreover, Carey and Shugart argued that additional features were viewed as potentially decisive. These include the mechanisms of candidate selection (and the degree of control wielded by party leaders), the ballot structure, and the timing of executive and legislative elections. By exploring the number of micro-institutional features that affect the performance of presidential systems, this new wave of scholarship provided a richer picture of the effects of institutional design on governance outcomes.⁹

The bottom line of the analytical perspective inaugurated by Shugart and Carey (1992) is that not all presidential regimes are prone to institutional crises. Crises tend to occur in contexts of multipartism — particularly those with a large effective number of (undisciplined) parties — where presidents tend not to count on the support of a stable majority. This is a fortiori true if presidents enjoy significant constitutional powers. Therefore, in this line of analysis, it is the combination of the degree of partisan powers of presidents — the extent to which they can count on a stable majority — and their constitutional powers that determine the propensity to governance crises. Where presidents have few constitutional powers but strong partisan powers, this propensity is significantly lowered.

This diagnosis implicitly contained the recipe for institutional therapy and several reforms informed by this emerging consensus were in fact carried out in the region. Bolivia, Mexico and Venezuela introduced mixed-member districts modelled at the German system, while other countries eliminated term limits for presidential elections. How have predictions made in the mid 1990s on the basis of these hypotheses fared in the light of developments in Latin America over the last decade? In fact, some of the predictions turned out not to be accurate. Analysts writing in the influential volume edited by Shugart and Mainwaring, which had enormous influence in the region and beyond, expected two countries to have serious governance crises — Chile and Brazil — and two others, which experienced comprehensive institutional reform — Venezuela and Bolivia —, to be on the road to good governance. With the benefit of hindsight, the predictions for these two sets of countries could be reversed because the former group of countries have outperformed the latter by a significant margin.

The predictions were based on the assumption that these countries adopted open lists with a large number of effective parties and the presidential parties' share of the vote was small. In addition it was also assumed that presidentialism did not generate incentives for coalition formation.¹⁰

In that volume, Siavelis (1997, 349) argued that

The most important consequence of the combination of multipartism and exaggerated presidentialism in the Chilean case is that the problem of doble minoria presidents has not been solved".

In Chile, despite the binomial electoral system, encouraging joint lists among the parties, presidents were expected to receive a small share of votes and, consequently, were bound to encounter governability crises. According to Siavelis (1997, 349) in Chile

Similarly, the Brazilian political system lacked stable majorities as a result of party system fragmentation, caused by the combination of open list proportional representation and robust federalism (wherein regional and state loyalties also contributed to undermine party structures) (Mainwaring and Shugart 1997a, 1997b). Brazil and Chile were the object of great concern due to the fact that the presidential party share of the vote was minimal. During Cardoso's first term of office, the Partido da Social Democracia Brasileira (PSDB) took 12% of seats, while Lagos's Partido Socialista (PS) got 10% in 2000. In a context of lack of incentives for coalition formation, Brazil and Chile were anticipated to be crisis-ridden in the late 1990s and 2000s.

Contrary to expectations, a few countries that have been unstable and performed poorly in terms of democratic governance in the last decade were expected to be successful. Thus Venezuela and Bolivia, following their electoral reforms, were expected to improve:

Likewise, Argentina was argued to have the correct ingredients for the "successful functioning" of its presidential system: a president with strong partisan powers and a strong party system. According to Jones (1997) in the same volume, this success was the product of an array of adequate micro-institutional incentives, including party leaders' control over candidate selection, closed list proportional representation, weak federalism, among other factors. Admittedly, the authors were cautious and did not argue that these institutions would unilaterally produce good governance. Nevertheless, governability was viewed primarily in terms of the ability to form majority governments, and coalition governments were seen as intrinsically problematic.

In the last decade or so, a new wave of scholarship has strongly contested the received wisdom of the 1990s (Cheibub, Przeworski and Saiegh 2004; Colomer and Negretto, 2005). Presidentialism is no longer seen as impairing interbranch cooperation or as leading to ungovernability, and has definitively stopped being conceptualized in a negative vein. The new received wisdom postulates that presidentialism and parliamentarism differ in several important dimensions. While recognizing that constitutional design matters, there has been an emerging consensus that presidentialism does not cause governability problems.¹¹ Some of the alleged important differences still hold: all else being equal, a) government coalitions should occur more frequently under parliamentarism than under presidentialism;¹² and b) parliamentary governments should be more successful legislatively than presidential ones.¹³ However, three key findings contradict earlier views. Firstly, in both systems, minority governments are not associated with less success in terms of the approval of the government's parliamentary agenda.¹⁴ Secondly, minority governments are not associated with deadlocks that lead to the demise of democracy. Thirdly, party system fragmentation does not affect the likelihood of impasses and interbranch conflict. Quite the opposite is true.¹⁵ These claims have been supported by extensive empirical research.¹⁶

Nonetheless, some minority governments occur under presidentialism and they transpire for the same reasons that they do under parliamentarism: the status quo is preferred by a majority (Cheibub, Przeworski and Saiegh 2004). This intuition is definitional and comes from the very notion that under parliamentarism any existing government is the result of majority preferences because a new government could be formed if the majority prefers it to the status quo. This argument may be extended to presidentialism with a few caveats. Minority presidents may represent a stable equilibrium whereby potential coalitions partners (the non-presidential parties) would opt to stay out of government while preferring the status quo policies enacted by the government. They might wish to stay out of the coalition because participation in government might incur in electoral costs and these parties have better electoral chances if they do not join the government. The key difference in the logic of cabinet formation under the two regimes is that under presidentialism the president must necessarily be a part of the government coalition, thereby reducing the number of potential coalitions. Two distinct logics take place depending on the location of the status quo (vis-à-vis the preferences of the president and the non-presidential parties) and on the president's constitutional powers (agenda powers, veto and decree authority, among others). If the president does not control the legislative process, the logic of coalition formation is identical under presidentialism and parliamentarism (Cheibub, Przeworski and Saiegh 2004).

Deadlock situations under presidentialism emerge in a rare class of cases: where the preferences of a minority president with strong constitutional powers and the other potential coalition parties are wide apart and further from the president's preferences than the status quo. In other words, the non-presidential parties' ideal points and the status quo are close but far from the president's ideal point. In this set up, presidents are unable to offer portfolios to potential partners because in order to attract a party that is close to the president's ideal point in the policy space, he or she must offer it a quantum of ministerial portfolios that is unacceptably high for him or her.¹⁷ On the hand, the president is expected to implement an agenda that is very far from his or her preferences to appease the non-presidential parties, and therefore a stalemate emerges.¹⁸

Although the role of the minority status of governments appears not to play a role in democracy's survival, presidential regimes are found to be less stable than parliamentary regimes. Therefore other factors might be contributing to the higher propensity of democratic breakdowns under presidentialism.¹⁹

The dynamics of coalition formation and their role in the effectiveness of presidential regimes have replaced electoral rules as explanatory variables in much of the debate on institutional design. Coalition management, in particular, appears as a crucial factor in the explanation of presidential success in implementing his or her agenda in Latin America. Cabinets help presidents implement their policy-making strategies (Amorim Neto 2006a; 2006b).²⁰ Presidents may decide to bargain with coalition partners ("the strategy of governing by statute"), in which case they allocate ministerial portfolios to non-presidential parties. Instead, they may opt to utilize their constitutional prerogatives ("strategy based on executive prerogatives"). The former strategy involves the appointment of a majority cabinet, the selection of more partisan ministers, and more proportionality in the distribution of portfolios to coalition parties.

Amorim Neto found that the president's choice depends on the president's preferences, as well as on institutional incentives and economic conditions.²¹ The president is more likely to appoint a majority cabinet, select more partisan ministers and distribute portfolios to parties on a proportional basis. The empirical analysis shows that the determinants of cabinet legislative status are the size of the president's party, "extremist presidents" — presidents whose ideal point is wide apart from potential coalition partners — and economic crises. If the share of votes of the president's party is small and he or she is an extremist holding large constitutional powers, he or she would have more incentives to use his or her presidential prerogatives and would tend not to appoint partisan ministers nor allocate portfolios proportionally. Rather than attempting to establish whether or not presidentialism leads to deadlock situations or presidential success rates in approving his or her agenda (or even democratic survival), this analysis shows that the distribution of preferences matters — confirming the findings discussed previously — and, more importantly, that coalition management is influenced by presidential prerogatives such as veto and decree powers.

Roughly similar findings are reached by Negretto (2006). He argues that the greatest potential for conflict also occurs depending on the president's veto powers and policy preferences, as well as on the partisan distribution of portfolios.²² These recent findings provide a much more nuanced picture of presidentialism. The main lesson to be drawn is that simplistic unilateral diagnoses can be discarded and that presidentialism is not doomed to failure or crisis-ridden. Despite the recent expansion in the legislative powers of presidents in Latin America,²³ there is nothing to fear as to the future of democracy (Negretto 2009).²⁴ Strong presidents have been able to implement their agenda without crises and unilateralism in Chile and Brazil as argued, respectively, by Alemán and Navia (2009) and Figueiredo and Limongi (2000). Their constitutional strength may be required for good governance in the region, and the resulting governance outcomes are conditional on adequate checks on executive discretion (Alston et al. 2008).

In response to the question "When is a strong president a benefit?", Carey argued "when presidential power can be configured so as to encourage deliberation within legislatures and bargaining between the branches of government, rather than a substitute for or deterrent to legislative-policymaking" (Carey 2009, 173). However, this much more nuanced view of presidential powers is not pursued further in the analysis. More importantly, it is limited to the interaction between legislatures and the executive branch, and does not embed a broader strategic model that includes judicial and other institutions.

Persson and Tabellini (2004), who have developed the most ambitious research program on the impact of institutions on economic development, have aptly identified this missing link in the research agenda. They warn us that "the negative effect of presidentialism is only present among the democracies with lowest scores for the quality of democracy; this suggests that perhaps it is not presidential government per se that is detrimental to economic performance, but rather the combination of a strong and directly elected executive in a weak institutional environment where political abuse of power cannot be easily prevented" (Persson and Tabellini 2004, 96).

Institutional checks on presidents

In the light of Persson and Tabellini's (2004) abovementioned suggestion that the reasons for presidentialism's higher propensity for abuse of power is associated with the broader institutional environment, a key question may be posed: What in Latin America's institutional landscape might explain the variation in the institutional performance of presidentialism in the area? This section and the next explore the role of what I call extended checks and balance institutions — legislatures, the judicial system and the media — in containing presidential abuse.

The bad functioning of democracies in the Latin American region has been associated in public debates with the problem of accountability and of flaws in the mechanisms of political representation. The weakness of checks and balances in the region is associated with a number of features: executives controlling the legislative process, legislatures that are simply rubber stamps of executive decisions, dependent judiciaries, manipulated media and weak control institutions such as ministérios públicos, tribunais de contas and contralorías.

The first two of these features can be found in parliamentary systems without implying violations of democratic accountability. In fact, in these systems, parliaments tend to be weak and single party majoritarian governments behave as "parliamentary dictators" (Powell 2000; Przeworski 2001). A legislature's capacity is usually measured in terms of the number of active standing committees in parliament, the powers held by committees, including the power to approve or shelve a proposal before it reaches the floor, and the quality of expert advice provided by a body of permanent staff, along with several other aspects. The fact that they are weak in Westminster systems cannot be argued to indicate a lack of democracy but an expression of a particular type of constitutional architecture, the majoritarian design.²⁵ In fact, the ability of governments to approve their agenda can be defended on grounds of political efficiency or decisiveness (Cox and McCubbins 2001).

However, in separation of powers systems, legislatures are expected to participate actively in legislative deliberations and to exercise control of governments by reviewing executives' initiatives and overseeing policy implementation. The functioning of legislatures and the interaction of the executive and legislative branches in Latin America has attracted a large body of research in the last decade and a half. The large variation in terms of the role played by legislatures across presidential systems and in Latin America in particular (Morgenstern 2002) has now been mapped out in numerous studies (Cox and Morgenstern 2002; Morgenstern 2002; Saiegh n.d.). At the more proactive end of the continuum, legislatures such as the US Congress have the capability to formulate legislative proposals and thus have a key role in influencing the policy agenda. By contrast, at the other end, legislatures may be rather marginal players, serving as a rubber stamp for the executive's legislative proposals. These legislatures have scant capacity to scrutinize the actions of government. In the middle of this continuum there are cases where the legislatures have been somewhat active but still do not effectively influence the legislative process or oversee the executive.

The traditional view that is associated with the legislatures in the region is that they merely rubber stamp proposals from the executive branch. However, this is an oversimplified view. In fact, the average percentage of recent executive initiatives approved by legislatures is below 50% in countries such as Peru, Costa Rica and Ecuador.²⁶ Only in Mexico and Paraguay this percentage is above 80%. Even when they are approved, there is massive evidence that executive proposals never emerged unscathed from changes introduced by the legislature (Cox and Morgenstern 2002). Weak legislatures come in two types: those that have no institutional capabilities and indeed merely rubber stamp executive proposals and those that can obstruct executive proposals, their institutional weakness notwithstanding. As stressed by Cox and Morgenstern (2002), the latter type — which they call called "recalcitrant legislatures" — is particularly problematic and maybe even more harmful to the democratic process than the former. Legislatures of the former type can be "subservient legislatures" or "parochial-venal" varieties (Cox and Morgenstern 2002). While the subservient ones are entirely dominated by the presidency — such as Mexico's legislature before 1998 —, in the parochial kind, legislators' support can be easily cobbled together by pork-barrel projects.

A number of factors affect the capacity of legislators to influence the policy-making process and to oversee policy implementation: the extent to which the legislature enjoys the confidence of citizens, the number of committees, the average technical quality of law-makers and their expertise and seniority, among others. Other important factors influence legislative capacity: legislators' reelection rates and the importance of the party label. Table 1 shows capacity indicators for the Latin American legislatures: societal confidence in congress, legislators' experience in office, their qualifications, the average number of committees and the effectiveness of law-making bodies.

Confidence in congress varies significantly across countries, being very high in Uruguay and Chile and very low in Ecuador and Guatemala, with Brazil falling somewhat in the middle. Reelection rates also vary widely, mostly because in countries such as Mexico and Costa Rica legislators have term limits and are barred from running for office again. This discourages legislative careers and leads legislators to build alliances with the executive in order to secure post-term appointments. Experience is high in Uruguay and Chile and low in Costa Rica, Mexico and Argentina. Building on these measures and additional indicators such as the strength of committees, legislators' technical expertise and assessments of congress as the place to build a career, the Inter- American Development Bank (IDB) (2006) proposed a classification of Latin American legislatures.

This is reproduced as Table 2.

Table 2 classifies all legislatures in the reactive category and none in the proactive one (which hypothetically corresponds to the US type, where parliaments introduce legislation on their own initiative). The IDB (2006) identifies three types of reactive legislatures: limited, obstructionist and constructive, whose characteristics are somewhat self-explanatory. Limited are those legislatures that are relatively marginal to the political process. Argentina since 1989 is classified in this category. Obstructionist legislatures do not engage in the policy-making process and play the role of a "blunt veto player" — blocking or approving executive proposals. None of the countries are estimated to have high-capacity, proactive constructive legislatures.

Types of legislatures faced by presidents provide distinct incentives for strategies to be pursued in dealing with them, ranging from a bargaining mode of interaction to a more adversarial mode based on presidents' use of their prerogatives (Cox and Morgenstern 2002; Amorim Neto 2006a; 2006b). Until recently, the bargaining mode had remained largely unexamined. As suggested by Cox and Morgenstern (2002), Latin American presidentialism represents a middle ground between the pure separation of powers system à la USA and European parliamentarism. Bargaining involves policy, pork and portfolio appointments. Where the political transaction costs are low as a result of less polarization and an adequate enforcement technology being in place — which is associated with strong institutions — actors can reap rewards from trade-offs; the potential for and scope of bargaining is thus enhanced.²⁷

Notwithstanding the overall weakness of Latin American legislatures, there is great internal variation. More importantly, Brazil, Chile and Uruguay clearly outperform other countries in the region. They are at the top in three categories: average of committees per legislator,²⁸ effectiveness of law-making bodies and average experience of legislators. A more capable legislature indicates ceteris paribus a higher oversight capacity over executive actions (and not only enhanced law-making). Indeed oversight activity requires capacity and political incentives (discussed in the next section). Systematic data on oversight activities is not available and further empirical research is necessary to test the hypothesis that enhanced oversight capacity and presidents with strong constitutional powers are complements. Qualitative evidence and case studies, however, suggest that oversight is stronger in countries such as Brazil and Chile (Leany 2010).

Presidential abuse goes hand in hand with de-institutionalization processes. They are particularly acute in the Andean region and in Venezuela, but there are alarming signals elsewhere. In countries where parties were strongest in Latin America, there occurred a process of disintegration of the party system. In Venezuela, the representation monopoly wielded by Comité de Organización Política Electoral Independiente (COPEI) and Acción Democrática broke down.²⁹ Similar developments can be found in Peru — the vote for Alianza Popular Revolucionaria Americana (APRA), Peru's oldest party and one of the oldest in Latin America, fell to 1.4% of the vote in 2000 (Mainwaring, Bejarano and Leongomez 2006). In Bolivia, traditional parties either almost disappeared, following corruption issues and evidence of links to drug dealing — the case of Movimiento Izquierda Revolucionaria (MIR) — or saw their share of the vote decline abruptly, as happened to the Movimiento Nacional (MRN) (Mainwaring, Bejarano and Leongomez 2006). There has been a collapse of the party system in Argentina as well, expressed, for example, in its denationalization, the proliferation of provincial parties and, most importantly, the virtual breakdown of the Unión Cívica Radical (Leiras 2007; Calvo and Escolar 2005). The flip side of the fragmentation of the party system is the rise of political outsiders, who individually or through their parties have held an increasing share of the vote. The cases of Toledo in Peru, Chavez in Venezuela and Morales in Bolivia are exemplary.

Another sign of de-institutionalization in the region is the increasing importance of social movements and street demonstrations in political developments that oftentimes lead to the ousting of presidents before the end of their terms of office (Pérez-Liñán 2007; Hochstetler 2006). Illustrative cases are Mauad (2000), Sanchez de Losada (2003), Collor (1992), De la Rúa (2001), Perez (1993) and Gutierrez (2002), among others. While these cases suggest a pattern of "stable presidentialism with unstable presidents" (Pérez-Liñán 2007, 204-5), wherein presidents do not finish their terms and there is no democratic breakdown, there are distinct crisis configurations.³⁰ As Pérez-Liñán (2007) argues, expressions of public outrage do not necessarily result in stronger systems of checks and balances, but are suggestive of mechanisms of spasmodic accountability. Legislatures prove capable of punishing wrongdoing without being able to prevent its emergence.

The key to explaining the puzzle posed in the introduction is that it is the quality of the extended system of checks and balances in a country that explains why powerful presidents — such as the Brazilian and the Chilean — wield vast powers yet the system does not degenerate into abuse. Rather, the upshot has been good governance. Presidents exercise powers in a constrained political space. Presidential abuses of power in countries such as Venezuela and Bolivia are clearly associated with attempts by presidents who are constitutionally among the weakest in Latin America to expand presidential powers against recalcitrant assemblies.

The role of the extended system of checks and balances — which includes the high courts, the contralorías (or tribunais de contas), the ministério público and the media — may provide the key to understanding the success of "exaggerated presidentialism". Indeed the countries that possess the most independent and effective judicial and audit institutions, as well as autonomous and pluralistic media, are the ones that have been performing systematically better. Admittedly, existing measures of judicial independence are based on different conceptions of independence and do not provide consistent values across countries (Ríos-Figueroa and Staton 2008). However, the existing measures suggest that Brazil and Chile are among the top performers. Chile has the top score in five out of seven existing indicators compiled by Ríos-Figueroa and Staton (2008), whereas Brazil ranks first or second in six of them.³¹ In the widely cited Feld and Voigt de Jure indicator, Uruguay, Chile, Brazil and Costa Rica are the top performers in terms of judicial independence (IDB 2006) (See Figure 2).³² At the other extremity of the spectrum, Venezuela, Bolivia, Argentina, Honduras and Ecuador are the worst performers.

The effectiveness of high courts depends on mechanisms that are path dependent (Stone-Sweet 2000, 139-50), including judicial activation and litigation, caseload and the associated legitimization of Supreme Court interpretation. Court stability and independence in the past evolve into institutional practices of the past. Thus, court independence in the new democratic context in Chile, Brazil and Uruguay may have been facilitated by the long-term stability they inherited. Brazil, Uruguay and Chile are also the top performers in the long-term investigation of judicial independence in Latin American countries presented in Perez-Liñan and Castiglione (2009). In this study, judicial independence is proxied by the turnover rates of Supreme Court justices in the period 1904-2006. Brazil displayed the lowest score overall (0.08) — slightly higher than the United States (0.06) — in the three periods discussed in the text (1904-1944, 1945-1977 and 1978-2006), while El Salvador, Argentina and Colombia were the countries with the most unstable high courts, particularly in the second period (1945-1977), when their scores were respectively 0.31, 0.24 and 0.27.

In terms of the effectiveness of national audit institutions, the top performers are also the countries that have the best judicial institutions in the region: Chile's Contraloría General de la República, Brazil's Tribunal de Contas da União and Costa Rica's Contraloría General de República (Figure 3).³³ Colombia's Contraloría General de la República also fares well and has the best score for the Andean region, regarding both types of institutions, which is consistent with the country's recent overall performance in terms of democratic governance. Indeed, despite the instability caused by the guerrilla movement, the country has experienced none of the problems of neighbouring Ecuador, Peru, Bolivia or Venezuela. By contrast, Argentina's Auditoría General de la Nación and Peru's Contraloría General de la República are the weakest audit institutions in the region, in terms of criteria such as autonomy, credibility, enforcement and timeliness of audit work.

In addition to judicial and audit institutions, an independent media is also key to explaining good governance. Chile, Brazil and Uruguay boast the most diversified and independent media in the region. Chile and Uruguay are the only countries classified as having freedom of the press for the period covered by Freedom House (2002-2009) (www.freedomhouse.org). Likewise, Chile, Uruguay, and Brazil (along with Argentina) have been consistently among the top 4 countries in the Reporters Without Borders ranking (http://en.rsf.org).

The Determinants of the Independence of Autonomous Institutions

The preliminary and admittedly scant comparative evidence reviewed thus far points to the plausibility of the alleged link between the strength of institutions such as audit and judicial bodies and democratic stability and presidential restraint. The key issue then hinges on the determinants of their independence and effectiveness. This is a topic that has attracted considerable attention in the current research agenda. While independence is a necessary condition for the establishment of the rule of law, the former may exist in the absence of the latter (Helmke and Rosenbluth 2009). Independent autonomous institutions emerge as a result of the strategic interaction among political actors in competitive contexts (Whittington 2003; Vanberg 2008). The extent to which these institutions enjoy autonomy is associated with constitutional choices made in critical junctures in the past (Knight 2001). Constitutional choices made in competitive settings tend to generate independent institutions. Key elements in their resulting institutional design facilitate autonomy. At the level of their functioning (as opposed to the political process creating them), political competition likewise explain their activism and independence level. In other words, autonomous institutions are endogenous. The effects of political competition come in two forms, power alternation and power fragmentation.³⁴ Their effects are intertwined but analytically differentiable (Stephenson 2003).

The argument about the effects of power alternation is forcefully made in Chavez, Ferejohn and Weingast (2003) and Ferejohn (1999). Political competition and fragmentation increase coordination costs in the elected branches and make it harder for them to directly attack the judiciary and control institutions after decisions that affect their interests are made. Thus, the more competitive a political system is, the more autonomous judicial institutions and autonomous institutions such as audit bodies will become.

Power alternation in turn generates incentives for politicians to delegate independence to courts and similar institutions in contexts of intense competition because they fear that while in opposition they would be better off under an independent institution than under courts that can be manipulated by political rivals. Thus, political uncertainty provides incentives for delegation to these bodies. Politicians buy political insurance against the risk of being out of office (Ginsburgh 2003, 18-25). Furthermore, because judicial institutions are usually reactive — they respond to the actions of authors that demand their intervention —, more political competition to lead different parties to bring their disputes to the judiciary, thereby leading to more independence.

For the Latin American case, there is a growing body of literature exploring the effects of elite competition and power fragmentation on the emergence of autonomous institutions (Finkel 2008; Chavez 2004; Helmke 2005; Ríos-Figueroa 2007; Pozas-Loyo and Ríos-Figueroa 2010; Helmke and Ríos-Figueroa n.d.). Thus, case studies in Peru, Argentina and Mexico have all converged on the conclusion that power alternation is key to the emergence of autonomous institutions. These hypotheses have been tested at the subnational level in Brazil (Melo, Pereira and Figueiredo 2009; Melo, Pereira and Werneck 2010), Argentina (Leiras, Giraldi and Tuñon 2009), and Mexico (Beer 2010).

The enhanced role of high courts in Chile, Brazil, Uruguay and Colombia is consistent with more competitive political systems and constitutional choices under competitive conditions (Pozas-Loyo and Ríos-Figueroa 2010; Taylor 2008; Lemos 2010). Colombia's very fragmented constitutional assembly of 1991 produced a constitutional chart where judicial powers were ample and strong. Similarly, in Brazil's constituent assembly of 1987-1988, the Ministério Público and the tribunais de contas were significantly expanded. More importantly, their functioning has been effective because of the increasingly competitive political system in post-1988 Brazil. Lastly, Mexico also illustrates nicely the effects of power de-concentration on the emerging autonomy of institutions such as the Suprema Corte de Justicia, the Auditoría Superior de la Federación and the Instituto Federal Electoral (Ackerman 2007; Merino et al 2010).

Power concentration, on the other hand, largely explains the weakness of autonomous institutions following the collapse of the two-party system in Argentina in the new democratic setting. The calamitous debacle of Radical governments in Argentina's new democracy paved the way for the deterioration in the strength of checks to executive discretion (Levitsky and Murillo 2005). Indeed, enhanced participation by the courts in the region reflects both the judicialization of politics (Sieder, Angel and Schilolden 2005) and the expansion of judicial power resulting from the democratization process. The recent reassertion of judicial power reflects power de-concentration under presidentialism in several countries.

In sum, a configuration of factors that include political competition and power alternation in a context of free and independent media seem to be conducive to the rise of autonomous institutions that are necessary to check executive power. The key point is that an open and professional media, reducing the information asymmetry between power-holders and citizens, and the breaking of political monopolies, is essential for the emergence of an independent judiciary, which in turn will have a decisive role in checking governments.

As occurs with the executive and legislative branches across Latin America, there is also great institutional diversity among autonomous institutions. Further research is needed to map out the institutional landscape (see Ríos-Figueroa, 2009 and Navia and Ríos-Figueroa 2005, for example) and, more importantly, to uncover the precise causal mechanisms linking the varying institutional arrangements with good governance. The interaction between political concentration, media independence and autonomous institutions is key to understanding this link. Autonomous institutions are endogenous to the political system. Indeed, a major threat to the judicial system and to autonomous institutions is non-compliance with their decisions. The executive branch controls much of the means of enforcement and can therefore undermine the authority of these institutions. Power fragmentation may erode the conditions for successful direct attacks on these institutions, but media support for the courts is also crucial to counter non-compliance (Staton 2010). Undoubtedly, institutional design matters — and constitutional choices made at critical junctures (and the power balance underlying them) have long-lasting effects. However, the costs of reversing judicial decisions are a function of the effects of such interventions on public opinion, and ultimately on belief systems. Autonomous institutions have long lasting effects but at the same time they are endogenous to political competition. As political competition declines and power concentration increases, governments will attempt to affect de facto independence of the institutions. Conversely, as completion increases, there emerges incentives for the autonomous institutions to become more autonomous. The methodological challenge is to disentangle the effect of institutional design, which is the effect of the power balance in key critical junctures, from the effects of changes in political competition within them. Testable hypotheses of the claims made in this paper require the obtention of reliable measures of the independent variable (presidential abuse of power), as well as adequate proxies for the autonomy of check institutions/rule of law, presidential powers and political competition, which are only partially available at this stage.

Conclusions

Latin America's presidencialismo is again at centre stage. Nonetheless, after almost three decades of academic debates on its alleged intrinsic flaws, hardly any observers now blame presidentialism for the governance problems in the region. The view that ungovernability looms large in presidential systems has been discredited. For one thing, presidents have been removed from office without any democratic breakdowns. More importantly, the institutional landscape varies markedly across the countries, and while several countries have experienced crises that seem unrelated to institutional design, others seem to be on the road to good governance. While it has been recognized that parlamentarism and presidentialism differ along many relevant dimensions — ranging from the frequency of coalition governments to their contrasting influence on economic and social development or potential for making chief executives accountable —, the argument that systems of government affect governability has been largely discredited. Similarly, the view that coalition governments are crisis-laden has also been superseded by a much more complex view. Some specificities of Latin American presidentialism have also come to the fore — they represent a halfway house between a pure separation of powers system à la USA and European parliamentarism. Coalition governments seem to be the rule, and this entails hammering out agreements over policy, pork and portfolio appointments.

Paradoxically, strong presidents and strong legislatures have produced good governance, as happened in Chile and, to a lesser extent, in Brazil. This is only possible where checks and balances are firmly established. An extended notion of checks and balances has been proposed which includes the media and institutions of horizontal accountability — ministérios públicos, tribunais de contas, Ombudsman's offices — to highlight their importance for democratic governance. A contrasting example comes from Argentina, where polarization and weak checks on the executive have produced abuse of power. Venezuela and Bolivia, in turn, provide examples of countries where presidents enjoyed few proactive and reactive powers and the legislature was not fragmented, creating a situation of crisis which prompted presidents to pursue a strategy of unilaterally expanding presidential powers. Thus, the combination of weak presidents and recalcitrant legislators engendered explosive inter-branch conflicts. A major claim of the analytical overview in this paper is that for a strong system of checks and balances to take root, political competition and power fragmentation are essential both at the level of constitutional choice and at the level of its effective functioning. A much more complex view of presidential powers is needed — one that embeds them in models of strategic interaction between the branches of government and autonomous institutions.

Notes

1 Numerous episodes are illustrative of these developments. For example, in Ecuador, Gutiérrez unconstitutionally fired members of the Supreme Court, while the government in Venezuela disclosed the identities of voters in the plebiscite, and in Bolivia, Morales was reported to have encouraged mobs to block access to the Assembly buildings during key legislative sessions. For recent accounts of these developments in the Andes, see Mainwaring, Bejarano and Leóngomez (2007), Lehoucq (2008) and Corrales (2009).

2 Numerous episodes are illustrative of these developments. For example, in Ecuador, Gutiérrez unconstitutionally fired members of the Supreme Court, while the government in Venezuela disclosed the identities of voters in the plebiscite, and in Bolivia, Morales was reported to have encouraged mobs to block access to the Assembly buildings during key legislative sessions. For recent accounts of these developments in the Andes, see Mainwaring, Bejarano and Leóngomez (2007), Lehoucq (2008) and Corrales (2009).

3 The expression comes from Siavelis (1997; 2002). Uruguay — also a case of exaggerated presidentialism — would also be in this group of countries, whereas Argentina has seen great institutional deterioration. Although Uruguay does not share some of the features mentioned such as open list proportional representation or large number of effective parties, party support is fragmented due to the high level of intra-party factionalism. The predictions for Uruguay were that it would also be doomed to fail.

4 Admittedly, Chile's constitutional reforms of 1989 and 1991 eliminated some blatantly undemocratic features of the 1980 constitution, such as the president's ability to dissolve the chamber of Deputies at least once during his or her term.

5 Figueiredo, Jacobi and Weingast (2006) have termed such an integrated approach the "new-separation-of-powers approach".

6 Useful reviews of the debate are Carey (2005) and Cheibub (2007).

7 Presidents in Costa Rica more than compensate their weak powers by their enormous influence over the legislature as a result of the crucial role they play in legislators' appointments to the bureaucracy after they have served their one term of office (after which they cannot be re-elected).

8 These are more than compensated by their vast partisan powers. This is postulated as a general pattern by Shugart and Mainwaring (1997); where presidents have little in the way of constitutional prerogatives, they also tend to have more partisan powers. The normative model behind this proposition is obviously the United States.

9 Thus it was argued that where party leaders have no control over candidate selection, the ballot structure allows for preference vote (open list) and elections are non-concurrent, the result would be a fragmented party system and minority presidents. This is so because the lack of control over candidate selection was believed to undermine party leaders' ability to secure the party line. In turn, open lists and proportional systems where votes are not pooled according to party affiliation would weaken the influence of parties during the electoral race, fostering legislators' individualism because they would become less dependent on the party for electoral success. Moreover, when elections for legislators and presidents are held at different moments in time, there is a decoupling of their electoral fate. If a president becomes less popular during midterm elections, the votes for the coalition members will probably be affected and the president will lose his/her support base.

10 Shugart and Mainwaring (1997, 397) argued that "whereas party coalitions in parliamentary systems generally take place after the election and are binding, in presidential systems they often take place before the election and are not binding past election day. Executive power is not formed through postelection agreement among parties and is not divided among several parties that are co-responsible for governing, even though members of several parties often participate in cabinets".

11 For a dissenting view, see Valenzuela (2008), where he essentially reasserts his earlier view.

12 Simply due to the fact that while all types of coalitions are possible in the former, in the latter they should include the president's party.

13 By definition, parliamentary governments can approve all of their legislative proposals, whereas presidents can be defeated by coalitions of parties.

14 A surprising and counterintuitive finding is that, in fact, minority governments are more successful at approving their agenda than majority governments (Cheibub, Przeworski and Saiegh 2004; Cheibub 2007).

15 A key empirical finding is that the higher the fragmentation, the higher the probability of coalition rate.

16 Cheibub, Przeworski and Saiegh (2004) test these claims using a dataset containing data on all democracies between 1946 and 1999.

17 These claims are based on extensions of formal models of coalition formation under parliamentarism, developed by political economists. Coalition formation is modelled as a non-cooperative game (Austen-Smith and Banks 1988).

18 The alleged influence of presidentialism on democratic survival has also been disputed. However, as the likelihood of coups has decreased significantly over the last twenty years or so, this issue has lost its appeal in the intellectual debate. Parliamentarism as an alternative has also disappeared from the reform agenda in the region, most notably in countries where it was seriously considered in the recent past, such as Brazil, Argentina and Chile.

19 This association is due to the fact that military rule was more frequent under presidential regimes. See Cheibub and Limongi (2002) and Cheibub (2007).

20 In fact, the interaction between assemblies and presidents has been modelled as a distinctive bilateral veto game, where the president defines his or her strategy depending on the type of legislature he or she anticipates. Depending on which type of legislature (recalcitrant, workable, subservient or parochial-venal), the president will adopt an optimal strategy to deal with it (Cox and Morgenstern 2002); more on types of legislature in section 2.

21 I based the empirical analysis on the econometric analysis of 106 cabinets appointed in 13 countries of the Americas.

22 The empirical analysis is based on the performance of minority presidential governments in Latin America (74 cases).

23 This was the case, for instance, of the last constitutional reform in Uruguay, when the president's agenda powers were extended.

24 However, the institutional debate is not exhausted by the debate on the influence of regime type on governability. Other normative values are crucial for democracy, such as accountability, representativeness, economic development and human development (Gerring, Thacker, and Moreno 2009). Research on these issues has been vibrant recently, but does not form a body of cumulative research similar to that found in the area of systems of government reviewed in this section. This caveat notwithstanding, it has been argued in large N research that parliamentarism is superior to presidentialism in avoiding the two canonical problems arising in political delegation: adverse selection and moral hazard. Parliamentarism is better at controlling adverse selection problems in legislative recruitment (due to the enhanced role of parties in the process), while being less effective than presidentialism in controlling moral hazard problems (because the latter encourages more interbranch checks). Under parliamentarism, parties internalize the costs of the executive's misbehaviour and therefore have incentives not to oversee the executive (Strom 2003). Presidentialism, in turn, is generally held to offer more identifiability of governments and, consequently, more potential for accountability (Samuels, and Shugart n.d; Powell 2000).

25 Other relevant features of such designs include unitary states and absence of constitutional review, for example.

26 Data refer to different years from the 1980s to the 2000s (IDB 2006).

27 See Spiller and Tommasi (2007) for an extended analysis.

28 A small value indicates that legislators are assigned to one or to a small number of committees, thus with higher chances of specialization.

29 COPEI and AD supported an independent candidate in the 1998 election that elevated Chavez to the presidency.

30 In some cases, presidents are impeached by the legislature following procedures specified in the constitution; in others, they leave their posts amidst massive public outrage, usually after media scandals involving corruption and/or proved connections to drug cartels. More often than not, these two aspects are intertwined (Perez-Liñán 2007; Hochtetler 2006). The role played by the legislature varies, ranging from a marginal role, where street protests dominate, to a proactive role following denouncements by legislators. Presidents in these cases invariably count on minority support and have declining popular approval, or were attempting to implement unpopular economic reforms. The former case raises concern and suggests that the legislature is being bypassed in its constitutional role of overseeing the executive's actions.

31 I have omitted Henisz's indicator, which is in fact a dummy independent/non-independent. In this account, Chile has the only independent Judiciary in the region.

32 The figure refers to de jure judicial independence, that is, provisions aimed at insulating these institutions from the influence of the executive branch.

33 The data is for an index of effectiveness computed by Santiso (2007) on the basis of scores for independence, credibility, enforcement and timeliness and are available for a smaller set of Latin American countries.

34 Andrews and Montinola (2004) explore empirically a related theme: the effect of veto players on rule of law.

Bibliographical References

Ackerman, John. 2007. Organismos autónomos y democracia: el caso de Mexico: Mexico City: Siglo Veintiuno.

Alston, L., Marcus Andre Melo, Bernardo Mueller, and Carlos Pereira. 2008. On the road to good governance: Recovering from economic and political shocks in Brazil. In "Policy-making in Latin America": How politics shapes policies, ed. E. Stein, M. Tommasi and P. Spiller. Cambridge: Harvard University Press.

Amorim Neto, Octavio. 2006a. Presidencialismo e governabilidade nas américas. Rio de Janeiro: Editora FGV.

___. 2006b. The presidential calculus: Executive policy making and cabinet formation in the Americas. Comparative Political Studies 39 (4): 415-40.

Andrews, Josephine. T., and Gabriela. R. Montinola. 2004. Veto players and the rule of law in emerging democracies. Comparative Political Studies 37 (1): 55-87.

Austen-Smith, David, and Jefrey Banks. 1988. Elections, coalitions and legislative outputs. American Journal of Political Science 82 (2): 405-22.

Beer, Caroline. 2010. Electoral competition and institutional change in Mexico. Notre Dame: University of Notre Dame Press.

Calvo, Ernesto, and Marcelo Escolar. 2005. La nueva política de partidos en Argentina. Crisis política, realineamiento partidarios y reforma electoral. Buenos Aires: Prometeo Libros.

Carey, John. 2009. "What sort of strong presidents?". In Making presidentialism work, ed. Andrew Ellis, J. J. Oroszco Henríquez and D. Zovatto. Goteborg: IDEA.

___. 2005. "Presidential versus parliamentary government". In Handbook of new institutional economics, ed. Claude Ménard and Mary Shirley. Dordrecht: Springer.

Chavez, R. B. 2004. The Rule of law in nascent democracies: Judicial politics in Argentina. Stanford: Stanford University Press.

Chavez, R., J. Ferejohn, and B. Weingast. 2003. A theory of the politically independent Judiciary. Paper presented at the Annual Meeting of the American Political Science Association, in Philadelphia, USA.

Cheibub, José Antônio. 2007. Presidentialism, parliamentarism and democracy. New York: Cambridge University Press.

Cheibub, Jose A., and Fernando Limongi. 2002. Democratic institutions and regime survival: Parliamentary and presidential democracies reconsidered. Annual Review of Political Science 5: 151-79.

Cheibub, José A., Adam Przeworski, and Sebastian Saiegh. 2004. Government coalitions and legislative success under presidentialism and parliamentarism. British Journal of Political Science 34: 565-58.

Colomer, Josep, and Gabriel Negretto. 2005. Can presidentialism work like parliamentarism?" Government and Opposition 40 (1): 60-89.

Corrales, Javier. 2009. For Chavez, still more discontent. Current History 108 (715): 77-82.

Cox, Gary, and Scott Morgenstern. 2002. "Latin America's reactive assemblies and proactive presidents". In Legislative politics in Latin America, ed. Scott Morgenstern and Benito Nacif. . Cambridge: Cambridge University Press.

Cox, G. W., and M. D. McCubbins. 2001. "The institutional determinants of economic policy outcomes". In Presidents, parliaments, and policy, ed. S. Haggard and M. D. McCubbins. New York: Cambridge University Press.

Crisp, Brian. 2007. "The nature of representation in Andean legislatures and attempts at institutional reengineering". In The crisis of democratic representation in the Andes, ed. Scott Mainwaring, Ana Maria Bejarano and Eduardo Leongómez. Stanford: Stanford University Press.

Ferejohn, John. 1999. Independent justices, dependent judiciary: Explaining judicial independence. Southern California Law Review 72: 353-84.

Figueiredo, Argelina, and Fernando Limongi. 2000. Presidential power, legislative organization and party behavior in Brazil. Comparative Politics 32 (2): 151-70.

Figueiredo, Rui P. de, T. Jacobi, and B. Weingast. 2006. "The new separation-of-powers approach to American politics". In The handbook of political economy, ed. Barry Weingast and Donald A. Wittman. Cambridge: Cambridge University Press.

Finkel, Jodi. 2008. Judicial reform as political insurance: Argentina, Peru, and Mexico in the 1990s. Notre Dame: University of Notre Dame Press.

Gerring, John, Strom Thacker , and Carola Moreno. 2009. Are parliamentary systems better?. Comparative Political Studies 42 (3) 327-59.

Ginsburg, Tom. 2003. Judicial review in new democracies: Constitutional courts in Asian cases. Cambridge: Cambridge University Press.

___. 2005. Courts under constraints: Judges, generals and presidents in Argentina. Cambridge: Cambridge University Press.

Helmke, Gretchen, and J. Ríos-Figueroa, eds. n.d. Courts in Latin America, Cambridge University Press.

Helmke, Gretchen, and Frances Rosenbluth. 2009. Regimes and the rule of law: Judicial independence in comparative perspective. Annual Review of Political Science 12: 345-66.

Hochstetler, Kathryn. 2006. Rethinking presidentialism: Challenges and presidential falls in Latin America. Comparative Politics 38 (4): 401-18.

Inter-American Development Bank (IDB). 2006. The politics of policies: Economic and social progress in Latin America, 2006 Report. David Rockfeller Center for Latin American Studies.

Jones, Mark. 1997. "Evaluating Argentina's presidential democracy": 1983-1995. In Presidentialism and democracy in Latin America, ed. Scott Mainwaring and Matthew Shugart. New York: Cambridge University Press.

Knight, Jack. 2001. "Institutionalizing constitutional interpretation". In Constitutional culture and democratic rule, ed. John Ferejohn, Jack Rakove and Jonathan Riley. New York: Cambridge University Press.

Lehoucq, Fabrice. 2008. Bolivia's constitutional breakdown. The Journal of Democracy 19 (4): 110-24.

Leiras, Marcelo. 2007. Todos los caballos del rey. La integración de los partidos políticos y el gobierno democrático de la Argentina, 1995-2003. Buenos Aires: Prometeo Libros.

Leiras, Marcelo, A. Giraldi, and G. Tuñon. 2009. Who wants an independent court? Political competition and Supreme Court instability in the Argentine provinces, 1984-2008. Paper presented at the Comparative Politics Workshop, Yale University, in New Haven, USA.

Lemos, Leany. 2010. Legislative oversight of the executive branch in six democracies in Latin America. Paper presented at the Oxford-Princeton Global Leaders Workshop. Princeton, USA.

Levitsky, Steven, and Maria V. Murillo. 2005. Argentine democracy: The politics of institutional weakness. University Park, PA: Penn State Press.

Linz, Juan. 1990. The perils of presidentialism. The Journal of Democracy 1 (1): 51-69.

___, and Arturo Valenzuela, eds. 1994. The failure of presidential democracy: The case of Latin America. Baltimore: Johns Hopkins University Press.

Mainwaring, Scott, and Matthew Shugart, eds. 1997a. Presidentialism and democracy in Latin America. New York: Cambridge University Press.

___. 1997b. "Presidentialism and the party system". In Presidentialism and democracy in Latin America, eds. Scott Mainwaring and Matthew Shugart. New York: Cambridge University Press.

Mainwaring, Scott, Ana Maria Bejarano, and E. Leongómez, eds. 2006. The crisis of democratic representation in the Andes. Stanford: Stanford University Press.

Melo, Marcus André, Carlos Pereira, and Carlos Figueiredo. 2009. Political and institutional checks on corruption: Explaining the performance of Brazilian audit institutions. Comparative Political Studies 42 (9): 1217-42.

Melo, Marcus André, Carlos Pereira, and Heitor Werneck. 2010. Delegation dilemmas: Coalition size, electoral risk and regulatory governance in new democracies. Legislative Studies Quarterly 35 (1): 31-56.

Merino, Mauricio, Sergio Ayllón, and Guillermo Cejudo. 2010. La estrucutra de rendición de cuentas em México. México: Edición UNAM-CIDE.

Morgenstern, Scott 2002. "Explaining legislative politics in Latin America". In Legislative politics in Latin America, eds. Scott Morgenstern and Benito Nacif. Cambridge: Cambridge University Press.

Navia, Patricio, and Julio Ríos-Figueroa. 2005. The constitutional adjudication mosaic of Latin America. Comparative Political Studies 38 (2): 189-217.

Negretto, Gabriel 2009. Political parties and institutional design: Explaining constitutional choice in Latin America. British Journal of Political Science 39: 117-39.

___. 2006. "Minority presidents and democratic performance in Latin America". Latin American Politics and Society 48 (3): 63-92.

Pérez-Liñán, Aníbal. 2007. Presidential impeachment and the new political instability in Latin America. New York: Cambridge University Press.

___, and Andrea Castiglione. 2009. Presidential control of high courts in Latin America: A long-term view (1904-2006). Journal of Politics in Latin America 1 (2): 87-114.

Persson, Torsten, and Guido Tabellini. 2004. Constitutions and economic policy. Journal of Economic Perspectives 18 (1): 75-98.

Powell, G. Bingham 2000. Elections as instruments of democracy: Majoritarian and proportional visions. New Haven: Yale University Press.

Pozas-Loyo, Andrea, and J. Ríos-Figueroa. 2010. Enacting constitutionalism: The origins of independent judicial institutions in Latin America. Comparative Politics 42 (3): 293-311.

Przeworski, Adam 2001. "Accountability social en América Latina y Más Allá". In Controlando la política: Ciudadanos y medios en las nuevas democracias latino americanas, eds. E. Peruzzotti and C. Smulovitz. Buenos Aires: Temas Grupo Editorial.

Ríos-Figueroa, J. 2009. The institutional setting for constitutional justice in Latin America. Paper presented at the Conference Judicial Politics in Latin America, CIDE, March, in Mexico City, Mexico.

___. 2007. The fragmentation of power and the emergence of an effective judiciary in Mexico, 1994-2002. Latin American Politics and Society 49 (1): 31-57.

Ríos-Figueroa, J., and J. Staton. 2008. Unpacking the rule of law. A review of judicial independence measures. Working Paper 21, Committee on Concepts and Methods, IPSA.

Saiegh, Sebastian. n.d. Active players or rubber-stamps? An evaluation of the policy-making role of Latin American legislatures. San Diego: University of California. http://dss.ucsd.edu/~ssaiegh/legislatures.pdf.

Samuels, David, and Matthew Shugart. n.d. Presidents, parties, prime ministers. New York, Cambridge University Press.

Santiso, Carlos. 2007. Auditing for accountability: The political economy of government auditing and budget oversight in emerging markets. PhD diss., Washington, DC, Johns Hopkins University.

Shugart, Matthew, and John Carey. 1992. Presidents and assemblies: Constitutional design and electoral dynamics. New York: Cambridge University Press.

Siavelis, Peter. 2002. "Exaggerated presidentialism and moderate presidents": Executive-Legislative relations in Chile. In Legislative politics in Latin America, ed. Scott Morgenstern and Benito Nacif. Cambridge: Cambridge University Press.

___. 1997. "Executive-Legislative relations in post-Pinochet Chile": A preliminary assessment. In Presidentialism and democracy in Latin America, ed. Scott Mainwaring and Matthew Shugart. New York: Cambridge University Press.

Spiller Pablo and Mariano Tommasi. 2007. The institutional foundations of public policy in Argentina. Cambridge: Harvard University Press.

Staton, Jeffrey. 2010. Judicial power and strategic communication in Mexico. New York: Cambridge University Press.

Stein, Ernesto, and Mariano Tommasi eds. 2008. Policy-making in Latin America: How politics shapes policy. Cambridge: Harvard University Press.

Stephenson, Matthew C. 2003. "When the devil turns...": The political foundations of independent judicial review. Journal of Legal Studies 32 (1): 59-89.

Stone-Sweet, Alec. 2000. Governing with judges: Constitutional politics in Europe. New York: Oxford University Press.

Strom, Kaare. 2003. "Parliamentary democracy and delegation". In Delegation and accountability in parliamentary democracies, ed. K. Strom, W. Miller and T. Bergman. New York: Oxford University Press.

Taylor, Matthew. 2008. Judging policy: Courts and policy reform in democratic Brazil. Stanford: Stanford University Press.

Valenzuela, Samuel. 2008. "Los partidos políticos y los desafíos de la gobernabilidad democrática en América Latina." In Reforma de los partidos políticos en Chile, ed. Arturo Fontaine, Cristián Larroulet, Jorge Navarrete and Ignacio Walker. Santiago: UNDP, Consorcio CIEPLAN, CEP, Libertad y Desarrollo and Proyectamérica.

Vanberg, Georg. 2008. "Establishing and maintaining judicial independence". In The Oxford handbook of Law and Politics, ed. Keith E. Whittington and Daniel Kelemen, 99-118. Oxford: Oxford University Press.

Whittington, Keith E. 2003. Legislative sanctions and the strategic environment of judicial review. International Journal of Constitutional Law 1 (3): 446-74.

Submitted in July, 2008
Accepted in December, 2009

*This article builds on sections of an earlier draft prepared for the Plataforma Democrática Project (CIEPLAN-IFHC-Centro Edelstein) and presented at the seminar Desequilibrios de Poder in América Latina, Universidad del Salvador, Bogotá, 2010. I would like to thank Julio Ríos-Figueroa and Octavio Amorim Neto for detailed criticisms and numerous suggestions for the paper. I would also like to thank Carlos Pereira, Bernardo Mueller, Luis Schiumerini and Marcelo Leiras for enlightening discussions about the paper's central argument.