Abstract
The books discussed in this review essay all investigate the same basic phenomenon: presidents who do not complete their terms in office, but who are removed without open democratic breakdown. The essay traces the conceptual frameworks used, as they shifted from impeachment to the broader presidential breakdown. It evaluates explanations of the causes of presidential breakdown, and makes suggestions for further research on how the presidential regime type functions. The essay concludes with a call for more empirical investigation of the consequences of presidential breakdown.
Introduction
Democratic regime breakdowns have almost disappeared over time in Latin America, with 1977-1978 as the breakpoint (Álvarez and Marsteintredet in Llanos and Marsteintredet: 42; Pérez-Liñán: 62). 1 With very few exceptions, established democratic regimes have survived after that time. This development was matched, however, with a surge in the number of presidents who did not make it to the end of their designated terms even as democracy itself continued. This pair of basic observations marks the phenomenon of interest in this review article. How should that phenomenon be conceptualized? What explanatory factors might account for it? Do we need to revisit either the basic definition of the presidential regime type or our expectations of the kinds of behaviour that institution generates? On balance, is this a positive development in Latin American politics? 2
Individual chapters in the books under review will be referenced by the last names of the authors only. When the book being cited is not obvious, as here, I include the editors of the book.
I would like to thank David Samuels for helpful comments on an earlier draft of this article.
Conceptualizing Presidential Breakdown
The three books under review illustrate the sequential unfolding of how scholars have characterized the many recent presidents who did not complete their terms even as the political regime around them continued. The earliest volume, edited by Baumgartner and Kada, treats them through the lens of impeachment, the one institutionally recognized mechanism for presidential removal. In most presidential constitutions, presidents can be removed for what the United States constitution calls “high crimes and misdemeanors” through a process that variously uses national congresses (and often their separate houses) and courts to indict and try individual presidents. Kada's conclusion includes a table that usefully summarizes the constitutional removal clauses for global presidential systems. The book's chapters include case studies of impeachments in the United States, the Philippines, Colombia, Madagascar, Russia, Brazil, and Venezuela. Yet not all the cases included for analysis actually follow constitutional removal procedures. Perkins’ chapter on the United States describes numerous partisan skirmishes that took the form of impeachment threats and procedures, and other cases are barely constitutional (e.g., Venezuela) or largely took place in the streets (e.g., the Philippines). It is no wonder that the introduction contends that impeachment “is a fundamentally political process from beginning to end” (5). The Spanish term juicio político (literally political judgement) makes the double meaning clear. In addition, Madagascar and Russia are not usually considered presidential systems (Siaroff 2003). While the volume presents a great deal of unprecedented analysis and information about presidential removals, it also shows how difficult it was to make them fit an existing institutional category like “impeachment in presidentialism”.
Pérez-Liñán's book is transitional in terms of its depiction of the phenomenon of interest. He presents a fine-grained analysis of six impeachment procedures in Latin America's presidential regimes (Brazil, Colombia, Ecuador, Paraguay 1999, Paraguay 2002, Venezuela), juxtaposing the institutional components of impeachment with accompanying non-institutional components including media coverage and popular outrage over presidential scandals. Pérez-Liñán also observes that “recent impeachments constitute the tip of the iceberg of a much broader emerging trend in Latin American politics” (2). The iceberg is a new pattern of instability where presidents are frequently removed from office both with and without formal impeachment procedures. For example, Mustapic (in Llanos and Marsteintredet: 20) counts nine Latin American presidents who resigned after 1980, often without the immediate threat of impeachment. In a final chapter, Pérez-Liñán examines several of these episodes, concluding that they have many of the same causes and actors of the impeachment cases, but reflect alternative mechanisms for resolving crises that involve presidents.
Several article-length pieces in the mid-2000s came to a similar conclusion that the barely-constitutional impeachment processes were similar enough to presidential resignations (following pressures from protesters and partisans) that they could be treated as one phenomenon. Close study of particular instances showed that the difference in mechanisms of presidential removal was often a matter of timing and expediency – would the president resign before congress put together some kind of removal process, which might or might not be formal impeachment depending on circumstances? The most common label chosen for the phenomenon was the process-neutral “interrupted presidency” (Kim and Bahry 2008; Marsteintredet and Berntzen 2008; Negretto 2006; Valenzuela 2004), but it is indistinguishable from “presidential fall” (Hochstetler 2006), “presidential failure” (Hochstetler and Edwards 2009), or Pérez-Liñán's (2007) “presidential removal”. All of these works attempted to make sense of the patterns among multiple cases, most often using quantitative analyses to trace the impact of a similar set of causal variables.
The volume edited by Llanos and Marsteintredet is a logical successor to this work. It introduces yet one more term, “presidential breakdown”, 3 for these cases of early presidential exit and continues to work with the causal variables identified in the earlier works. Several chapters explore comparative patterns, including Mustapic's analysis of the roles of congresses and a chapter by Álvarez and Marsteintredet that compares causal processes in democratic breakdown and presidential breakdown. Most of the book returns to national case studies, but this time does so with the analytical weight of the cross-national studies behind them. The chapters revisit the comparative conclusions, often arguing that explanations of particular cases require emphasizing some of the shared causal variables over others and/or introducing additional contextual factors.
Since it is the term used by one of the works under review, I will use it in the rest of the article to indicate the full set of presidents who did not complete their full terms, regardless of the mechanism of their removal.
At the same time, the volume's organization shows a second agenda, which is to stress that while there have been a striking number of presidential breakdowns with some shared characteristics and causal processes, there are also important subcategories of breakdown (see also Marsteintredet and Berntzen 2008). Argentina, Bolivia, and Ecuador are covered in a section on cases of repeated breakdown; Brazil, Paraguay, and Venezuela are instances of impeachment; the fates of presidents in the Dominican Republic, Guatemala, and Peru appear as presidential breakdowns restoring democratic order. The last set of countries is sometimes excluded from the cross-national studies of presidential breakdown on the grounds that they are not democratic enough to present the phenomenon of interest. Including these “gray” cases here, along with a short epilogue on the breakdown of Zelaya's presidency in Honduras in 2009, provides considerable material for probing both the conceptual limits and the normative assessment of the phenomenon.
The volume's inclusion of the gray cases is a useful corrective to the tendency to finesse the sometimes-undemocratic trappings of presidential breakdowns by including a control variable of one of the indices of democracy in the comparative studies. It is not always easy to draw a firm line between regime and presidential breakdown, as Honduras clearly shows. In an epilogue, Llanos and Marsteintredet say the final outcome will make the category clear, but the hemisphere is still divided over the issue several years later, with a new president well into his term. The threat to President Correa of Ecuador in 2010 is another example where people in the thick of events cannot agree whether there was a coup attempt or not. The military quite clearly participated in the Ecuadoran removal of Mahuad in 2000, but that case is commonly included among the instances of presidential breakdown and the military's involvement usually earns at best a footnote in discussions of the “new” instability. It might be helpful to introduce an intermediary empirical category for these cases that fall somewhere between total regime breakdown that results in a long-term military government and true presidential breakdown where the democratic regime is largely unscathed. In any event, the Llanos and Marsteintredet book is the most complete treatment of recent presidential breakdowns in Latin America that exists, and is a valuable summation of current knowledge of the phenomenon.
Causes of Presidential Breakdown
Taking the instances of presidential breakdown as a set, what do these studies tell us about why it is that so many Latin American presidents have been unable to complete their terms in office? The three books play similar roles on this subject too. The volume on impeachment edited by Baumgartner and Kada presents an initial effort to try to make sense of what was then a brand new phenomenon. Pérez-Liñán works with a set of institutional and non-institutional causal factors that became the dominant framework for thinking about what forces pushed presidents out, as the comparative articles used similar variables. The book edited by Llanos and Marsteintredet further investigates that dominant framework. Many of the individual chapters, with their case study focus, insist on the importance of causal variables unique to the country of study. Chapters at the beginning and end of the book address the causal arguments more systematically, however, underlining some of the comparative conclusions while undermining others, and suggesting some new variables that might be relevant for future comparative study. Here I briefly present the Baumgartner and Kada book's arguments and then discuss the other two together, sorting causal factors along the institutional versus non-institutional dimension before a final discussion of additional factors that deserve study.
As Kada notes in her conclusion (137), there had been no impeachments outside the United States until the 1990s, so the book tracked uncharted territory empirically. The descriptive and legalistic studies of US impeachments offered few theoretical signposts either (Baumgartner: 3-5). With impeachment the defining construct and global presidentialism the empirical scope, the book focused on the balance of power between the legislative and executive branches and factors likely to directly affect it: legal provisions for impeachment, the structure of party politics, and presidential popularity. An “other factors” clause left room for widening the analysis (Baumgartner: 8). Kada (144-145) finds that all the cases covered in the book show the importance of the voting threshold for impeachment, partisan composition, presidential patronage, and public opinion of the president. While the Baumgartner and Kada book is referenced in the subsequent debates (e.g., Hochstetler 2006; Pérez-Liñán 2007), this particular set of causal factors does not carry forward. In fact, only the question of partisanship and, to some degree, public opinion, remains on the theoretical agenda, with the others disappearing.
The main body of studies of presidential breakdown departs from different starting points that presumably account for their turn to other causal variables. As already noted, the Baumgartner and Kada volume focused on impeachment, while the later literature conceived of the target concept more broadly. Second, almost all the subsequent empirical work focused on the Latin American region (exceptions include Hochstetler and Edwards 2009; Kim and Bahry 2008). Third, the theoretical starting point was the comparative literature on the relative merits of presidentialism versus parliamentarism, with Juan Linz's work forming a constant backdrop (Linz 1978, 1990, 1994) – even though the focus on Latin America has meant that the universe of cases considered comes only from the functionally-presidential side. Fourth, the second and third points tended to place presidential breakdown in a complex of concepts where its opposite was democratic regime breakdown in presidentialism, not normally functioning presidentialism. Explanations of the former were used to generate hypotheses about presidential breakdowns in democratic regimes. Following a convention in some of the literature that also orients the Llanos and Marsteintredet volume, I divide those hypotheses into institutional and non-institutional variables here. 4
The debates briefly introduced here are discussed in more detail in Hochstetler and Edwards (2009).
The institutional variables originate directly in the studies of Linz and others about the regime instability problems of presidential regimes. In a first generation of studies, they had focused on the ways that the basic rules of the presidential regime type itself presented serious problems for democratic consolidation (Elgie 2005). Executives and legislatures were selected separately and did not depend on each other for survival. This gave each a claim to democratic legitimacy and no straightforward way to adjudicate between the claims when the two branches deadlocked. Linz and others thought they probably would deadlock – conflict seemed inevitable given the hubris of directly elected presidents, rigid term lengths, and the possibility that the executive's party would not control one or more branches of the legislature. Linz expected regime breakdown as a result, but later scholars thought the same factors might account for presidential breakdown as well. In fact, Álvarez and Marsteintredet (in Llanos and Marsteintredet) perform exactly this long-awaited statistical analysis, and find that the same basiclevel variables do help account for both breakdown outcomes, with some variations in specifics.
In terms of particular variables, essentially every analysis includes consideration of whether the president's party or party coalition controls the legislature; it is the only major causal variable in all three of the books under review. Minority executives are more common in presidentialism, and they are more likely to face conflicts with congresses since neither formally relies on the other for survival in office. This is seen to reduce their incentives to form coalitions across parties, to cooperate in more general terms, and to sustain party discipline. In their book-length study of the relationship between political parties and separation of powers, Samuels and Shugart (2010: 120-122) find that impeachment is almost never (1 per cent of the time) carried out as an intra-party accountability mechanism in presidentialism, unlike the way that parliamentary parties regularly challenge their ruling executives (30 per cent of changes in prime ministers). Thus majority support for the president in the legislature is a powerful protection against efforts by the legislature to shorten presidential terms, while minority status is hazardous.
Among the works under review, Pérez-Liñán presents the most detailed and multi-faceted analysis of what he calls the “legislative shield,” showing that it is only partially based on partisanship, a conclusion with which Mustapic's chapter (in Llanos and Marsteintredet) on the role of congress agrees. Notwithstanding the logical plausibility of this set of institutional arguments, the quantitative studies have had some trouble in isolating the explanatory weight of minority government – perhaps because minority governments are so common in Latin America, including in countries that have never had a presidential breakdown. Even so, all the case studies in the Llanos and Marsteintredet volume, with the exception of Buitrago's chapter on Bolivia, come down on the side of seeing executive-congress interactions as most central to the presidential breakdown outcome. The editors’ conclusion flags this the most important causal conclusion of the book (213).
On the non-institutional side, one major focus has been mass protests demanding the president leave office early (Hochstetler 2006; Hochstetler and Edwards 2009; Kim and Bahry 2008; Pérez-Liñán 2007). All of the South American cases have such protests, as do most other instances of recent presidential breakdowns except for the Dominican Republic. As such, the existence of anti-president protest and related variables (like the president's use of violence against protest) has tended to be strongly associated with presidential breakdowns. Successful mass protests are broader-based than unsuccessful ones (Hochstetler 2006; Pérez-Liñán 2007). All of these authors also consider the incidence of scandals involving the president as a root cause of protest and/or presidential breakdown. Pérez-Liñán's book again presents a particularly well-developed version of these arguments, showing how popular outrage grows out of presidents’ scandalous behaviour, media coverage of it, and citizens’ economic dissatisfaction with presidential performance. As outrage grows, presidents are impeached, unless they can put together a set of loyal legislators who will protect them. This is partially a matter of partisanship, but also presidential leadership (133), a factor underlined in Llanos's chapter (in Llanos and Marsteintredet).
As the brief summary of Pérez-Liñán's book suggests, economic variables are also considered in much of this literature as an additional set of non-institutional variables. These range from neoliberal policy orientations to concrete economic performance results, usually for inflation or unemployment rates. These are sometimes lagged and sometimes not. I do not present the arguments in much detail since the results have been so inconsistent, as reported by particular authors and in the literature as a whole. While economic problems do form a backdrop for many presidential breakdowns, there are once again economic problems in many locations where presidents survive through their terms. Llanos and Marsteintredet make the plausible argument that economic problems create challenges for presidents to manage, but that the management of them matters more than the problems themselves (217). In any event, economic choices and outcomes seem to be more an aggravating factor than one that brings down presidents who lack other problems.
Final causal conclusions are difficult to assess. As Llanos and Marsteintredet note in their concluding chapter, presidents typically face a “perfect storm” of causal factors, making it difficult to sort out which are most critical. Perhaps more fundamentally, presidential breakdowns are the result of long causal chains and causal conclusions depend both on how far back in the chain one goes and what assumptions are made about causal sequencing and historical counter-factuals (213). For example, both the Pérez-Liñán and Llanos and Marsteintredet volumes consider many of the same variables, but assume a different causal sequence. Pérez-Liñán presumes that scandals and protests are the precipitating factors, and legislative factors come in afterwards, as a shield that either protects the president from public outrage or not (132-133). In Llanos and Marsteintredet's volume, most chapters begin with a detailed accounting of the failing relationships of presidents with their congresses over time, with protest appearing as “only the last factor in a chain of events leading to presidential breakdowns” (216). Logically, this situation could still include protest as a “tipping” or final cause, although one would either need non-breakdowns (the book includes only positive instances of breakdown) or counter-factual analysis to see if that is true. Similarly, it is possible that the troubles of Pérez-Liñán's presidents begin in the legislature before the scandals that eventually bring them down.
The prevalence of scandals, economic mismanagement, and presidential defiance of democratic rules in these cases does support Llanos and Marsteintredet's conclusion that “presidents appear to be at least partially responsible for their own demise” (217), and some brief empirical modelling supports that. While this is an important observation – and one that presidents in the region might do well to note – thoroughly incompetent and venial presidents do not fail on their own. They are usually more than happy to go on governing. Thus we should not over-focus on presidents to the exclusion of legislative action, court decisions, and street protests when explaining presidential breakdown.
Clearly, these works have not closed the causal analysis of presidential breakdown. On the contrary, they continue to point to more avenues for future research. Many of the explanatory variables in the Llanos and Marsteintredet case studies stress unique causes of breakdown – such as Bolivia's history of social movement contestation (Buitrago), Ecuador's unique coalitional politics (Mejía Acosta and Polga-Hecimovich), or the internecine fighting in Paraguay's dominant Colorado Party (Nolte). Taken as a group, however, the chapters do suggest a number of variables that might usefully be investigated more fully across the region. The concluding chapter discusses a number of these, from more systematic consideration of the maturity of democracy to the role of cabinets and cabinet formation. For example, international factors have been almost wholly missing from the comparative studies (216), incorporated primarily to explain when the hemispheric context supported regime versus presidential breakdown. Yet foreign actors appear repeatedly in individual cases, including in Bolivia, Guatemala, and Paraguay, usually encouraging presidents to go peacefully, but sometimes preventing illegal removal. Thus they might be another important final arbitrating variable that should be better understood. Another area for further research would be to follow Llanos and Marsteintredet's suggestion to make the maturity of a democracy part of the analysis; they note that presidential scandals are considerably more egregious in less mature democracies, and other variables also presumably function differently in the gray cases.
Do We Need to Change the Definition of Presidentialism?
Presidentialism is one of the basic regime types of representative democracy. A classic formulation appears in Sartori's magisterial study of comparative constitutions, which says that a regime is presidential
if and only if the head of state i) results from popular election, ii) during his or her pre-established tenure cannot be discharged by a parliamentary vote, and iii) heads or otherwise directs the governments that he or she appoints (Sartori 1994: 84).
Linz details that this means presidents are elected for a rigidly fixed term and that both the president and legislature enjoy their own individual democratic legitimacy (Linz 1994: 6). Later influential work has streamlined this definition somewhat, defining presidentialism as requiring the separation of origin and survival of both president and congress, as well as the fixed terms of each (Shugart and Carey 1992: 19; see also Samuels and Shugart 2010: 27-28). 5 Does the advent of presidential breakdown imply that this definition needs to be changed?
Their definition also includes that the executive “names and directs the composition of the government” and that s/he have “some constitutionally granted lawmaking authority” (Shugart and Carey 1992: 19).
As a first principle, it seems most useful analytically to keep the classic definition of presidentialism at least as an ideal type, which is more or less approximated by actual political practices. Beyond that, most instances of presidential breakdown do not violate the ideal type. Classic presidentialism includes the possibility of impeachment for individuals who do not meet the law of the land as they execute their positions. Presidential resignations were not expected (Linz 1994: 10), but they do not obviously deviate from the defining characteristics. The most problematic cases for the definition were when congresses used long-standing institutional clauses meant to facilitate presidential removal in extraordinary situations of illness or other personal presidential incapacity to govern. These clauses have historically been considered fully compatible with the idea that a parliamentary vote cannot discharge a head of state during his or her pre-established tenure, but their status might need to be refined if congresses continue to use them to remove presidents with simple voting majorities for reasons that do not stand up to impeachment standards.
Mustapic (in Llanos and Marsteintredet) makes the only suggestion of a change in the actual definition in these works, arguing that the frequency of presidential breakdown means that the presidential regime type should be defined by fixed terms for congresses, but not for presidents (28). The dissolution of a congress would signal the rupture of democratic presidentialism, while early termination of a president would be accepted as well within the limits of the regime type. Mustapic supports her argument with the observation that several acknowledged authoritarian regimes, such as those of Stroessner in Paraguay and Pinochet in Chile, established ways for presidents to dissolve congress (but not vice versa), and those clauses were part of their authoritarian quality (28). As noted above, I do not consider early terminations of most presidents to require changing the current definition. On the other hand, the addition of more-specific discussion of fixed congressional terms to the defining understanding of presidentialism is consistent with current understandings of the concept. Linz’ discussion is almost wholly focused on presidents, but the rigidity of congressional terms is quite central to the definitions associated with Shugart and his co-authors.
If the ideal type of presidentialism can be left largely unchanged, there is a separate question of how well Latin American constitutions fit these definitions. The fit has always been approximate despite the general inclination to refer to all of the Latin American countries as presidential. To cite several well-known examples, until recent constitutional changes, the Bolivian Congress selected the president if no candidate won an absolute majority, violating the clause about separate origins, although the president was not thereafter accountable to the legislature. From 1979-1992, Peru formally had a prime minister accountable to the legislature, in addition to a popularly elected head of state. These kinds of provisions already made the empirical distinction between presidential, parliamentary, and semi-presidential systems “inadequate” (Siaroff 2003), although it is unlikely that these particular clauses made presidentialism work very differently in these countries.
The question arises again in the context of changes made in recent constitutions, including changes in two of the countries most prone to presidential breakdown. The changes were at least ostensibly designed to decrease the instances of breakdown. Ecuador's 2008 constitution
created a citizens’ council to exercise ‘social control’ over all branches of government, and gave its president the right to disband the congress once for ‘reiterated and unjustified obstruction’ of national development plans, political crisis, or ‘internal commotion’ (in which case both branches then face elections) (Hochstetler and Samuels 2011: 135).
Bolivia joins Venezuela in having a constitutional provision for citizens to initiate referenda to remove presidents (and other elected officials) (Hochstetler and Samuels 2011: 135). Honduras has edited its constitution to leave it quite unclear who among the congress, supreme court, and military can remove the president, an ambiguity that contributed to the presidential crisis there (Llanos and Marsteintredet, Epilogue: 232). There may well be other examples of which I am unaware; just these suggest some reason to be wary of assuming that all Latin American regimes are straightforwardly presidential in a Linz/Shugart sense. The provisions in Ecuador and Honduras, in particular, have the potential to rebalance power between the executive and the legislature in ways the standard presidential regime type does not anticipate, probably moving power to the executive from the legislature, although in different ways.
Do We Need to Change the Expectations of Behaviour in Presidentialism?
Rather than changing the definition of presidentialism, the better approach is to observe that there have been violations of its rules, both in the institutions established and in the behaviours displayed. More fundamentally, the rules as laid out by Sartori and others do not appear to condition behaviour in the ways that Linz expected. This is the case not just in the phenomenon of presidential breakdown, but also in many of the pathways thought to lead to problems of presidentialism. One direction for future research is to go back to the defining characteristics of the presidential regime type, and rethink the behavioural expectations that flow from it. This section considers some starting points for that kind of analysis.
Most of Linz's expectations about presidentialism and its defects were grounded in more prosaic relationships of executives and legislatures. He proposed causal pathways between presidentialism and democratic breakdown that travelled through divided government, legislative deadlock, and the like. The study of presidential breakdown has travelled some of those same routes. As such, it has been shaped by a view of presidentialism in crisis even though detailed analysis shows that Linz's causal pathways were either not particularly common in presidentialism, or not more common in presidentialism than in parliamentarism, and thus not especially good indicators of when presidentialism was likely to break down (Cheibub 2007). What Cheibub's book does not provide is a clear vision of how presidentialism does in fact work, if not as Linz expected it to. Here, a valuable set of propositions about the ordinary workings of presidentialism is in a recent volume by David Samuels and Matthew Shugart (2010). The book offers new arguments about how basic regime type deeply conditions the nature of partisanship, quality of representation, and other behaviours and outcomes likely to be key to when presidents will be unable to complete their terms in office. In their view, intra- and inter-party relationships are the basis of executive-legislative relationships; those partisan relationships, in turn, look very different in presidential and parliamentary systems. Presidents have controls over their own parties that were previously overlooked, but also are even more vulnerable in situations of divided government (Samuels and Shugart 2010). While they are mostly focused on impeachment as a form of removal of presidents, related studies of partisan and inter-branch relationships in presidentialism should illuminate the underlying situations that make presidents more vulnerable to presidential breakdown of any kind.
Linz also had more specific behavioural expectations directly relevant to presidential breakdown. He thought that
impeachment would be difficult to achieve, that presidents would be unlikely to resign voluntarily; and that publics would not allow presidents they had elected to leave office early (Hochstetler and Samuels 2011: 129, referencing Linz 1994: 10).
This essay has already discussed the considerable work done on impeachment, but the other two developments also deserve additional research. Why and when do presidents resign voluntarily? That this is a puzzle becomes clearer when Latin American and South American presidents are compared to their counterparts in other global presidential regimes. In the latter, presidents often simply wait out challenges to completing their turns in office (Hochstetler and Edwards 2009). Their comparative analysis offers some initial hypotheses to explain the difference, but the subject has barely been raised. What combination of external pressures, regime dynamics, and internal psychology bring presidents, especially elected ones, to the point of resignation? Are there commonalities among presidential calculations, or does the topic require a turn to the examination of particular presidents as several of the chapters in Llanos and Marsteintredet would suggest?
What makes publics turn on their presidents? One of the least-commented aspects of the mass protests against presidents is how few of the presidents have been able to mobilize supporters for their continuing in office. Chávez in Venezuela does – although his term is threatened by a military coup and a recall referendum rather than the usual mechanisms; Cubás in Paraguay does; but the vast majority of presidents were removed undefended. Protests demanding presidents leave office often become visible only in the final stages of a presidency, but where exactly do they come from? To what extent can they be understood as the working out of the presidential regime type, versus the external factors (scandal, economic performance) with which they are more often associated? How do voters withdraw their mandate? Given that even the largest protests are much smaller than the numbers of voters who put presidents in office, should analysts speak of mandate withdrawal (Perez-Liñán 2007: 211)?
The study of presidential breakdown is still in an expansive stage. Each new work opens as many questions as it answers. We now know more about how presidentialism does not shape behaviour than how it does. This section provides just a few examples of some directions for further work that would help account for observed patterns.
Are Presidential Breakdowns a Salutary Development in Latin American Politics?
Presidents fail, democracy continues. What implications does this development carry for Latin American democracy? Are Latin American governments more accountable, more stable, or better able to govern now that presidential breakdown is fairly common – or are all of those dimensions now worse? Most academic works on presidential breakdown leave such questions to the final pages or paragraphs of their studies, focusing on causes rather than consequences.
From such final musings, Hochstetler and Samuels (2011: 130-134) gleaned a “pessimist vs. optimist” framework for characterizations of the probable consequences of presidential breakdown. Pessimists point out that presidential systems were not designed for such events and require disruptive and delegitimizing processes to remove presidents; in the meantime, important governing goes undone; and vice presidents and other replacements are often not well prepared to govern. Overall, pessimists suspect that “challenges and failures leave deep wounds that potentially deform subsequent presidencies”, a new “peril” of presidentialism (Hochstetler and Samuels 2011: 130). Optimists do not deny that presidential breakdowns represent crises for their political systems, but consider any likely negative effects to be short lived. In fact, the ability to remove ineffective or scandal-ridden presidents may be considerably less traumatic than living with those leaders for full terms. Replacements come into office with vivid reminders of the need to govern inclusively and well, and often have broad support coalitions.
All of the books under review take somewhat mixed positions, but the impeachment books lean to some pessimism if impeachment – and the presidential and/or legislative behaviour that has driven it – continues to be common. Kada worries about the deterrent value of impeachment when presidents in Brazil and Venezuela subsequently win court cases on their behaviour. She thinks impeachment is probably better than outright coups, but notes
…[o]ne could even argue that impeachment is worse than coups in that while coups are clearly counter to democratic principles, impeachment is not, but lends a legal cloak to the protagonists who attempt to topple a president (152).
Similarly, Pérez-Liñán concludes that the “politicized and spasmodic” accountability embodied in Latin American impeachments leads him to a position “somewhere between a pessimistic image of Latin America as a vast institutional wasteland and an optimistic image of societies governed by well-established ‘rules of the game’” (209). Llanos and Marsteintredet keep to the spirit of their edited volume in their conclusion that presidential breakdowns have different implications for different countries (224). Brazil (Cheibub Figeiredo) was, unusually, greatly strengthened by impeaching President Collor de Mello in 1992 while El Salvador (Bjune and Petersen) is actually even worse off after removing President Serrano in this gray case than people thought at the time. Breakdowns are especially disruptive when repeated, as in Argentina, Bolivia, and Ecuador, but especially good when presidents were autocratic.
Notwithstanding Kada's fear, most observers agree that presidential breakdown is a better alternative to political stalemate and conflict than a military coup. Even if president-topplers gain a legal cloak, it should be easier to work on improving the functioning of imperfect democracies than to find legitimacy and accountability in outright authoritarian regimes. What all of this pessimism reminds us is that the substitution of presidential for regime breakdowns is far from enough to be reassured about democratic quality in Latin America. The gray cases in the Llanos and Marsteintredet volume illustrate this well, but even less ambiguous cases show numerous signs that the political rules of the game have not been fully internalized by presidents, congresses, publics, and every actor involved.
On the other hand, these are interpretive conclusions, with little direct empirical analysis of the consequences of presidential breakdown. Hochstetler and Samuels (2011) have the only systematic comparative analysis of the successors of presidential breakdown that I know of. They use country risk data (itself a combination of reputational data and concrete data on topics such as economic outcomes) and some very partial public opinion data about support for democracy to see how presidential breakdowns affect these measures. They compare outcomes for replacement presidents and presidents who are challenged but manage to complete their terms to “normal” Latin American presidents who govern their full terms and never face efforts to remove them. They find that countries where presidential breakdown was threatened or achieved show very little difference in these measures compared to countries that had normal presidents. As they note, “[t]his is not to say that any of these presidents is performing well, only that there are few clear differences between these populations” (141). In the end they conclude with the optimists that presidential breakdown is, on average, less disruptive and negative than pessimists fear. This does not exclude the possibility of disequilibration in some instances, and they also single out the frequent-removers of Argentina, Bolivia, and Ecuador for special concern (142). Their analysis does not test for longer-term effects and many other sources of data could be brought to the analysis. What this article shows is that the question of the impact of presidential breakdown on democracy, legitimacy, and governability is susceptible to direct empirical analysis rather than only interpretive speculation, and deserves more work.
As I have argued several times in this essay, study of the topic of presidential breakdown is still quite young. The books under review manifest considerable achievement in understanding the phenomenon, but also have opened up new areas of research rather than giving final answers to old ones. Presidential breakdown itself may eventually prove to be a phenomenon of particular times and places (indeed Venezuela and perhaps others show that continuismo never went too far from the scene), but it has clearly redefined the expectations of actors on the ground in the post-transition period. Overall, this area of study shows that presidents and presidentialism in Latin America continue to be important orienting actors and institutions in political outcomes there, even as they share political space with legislators, parties, publics, and other actors who have shown some surprising developments themselves.