Hostname: page-component-cd4964975-g4d8c Total loading time: 0 Render date: 2023-03-27T19:06:33.499Z Has data issue: true Feature Flags: { "useRatesEcommerce": false } hasContentIssue true

Tenuous Pacts and Multiparty Coalitions: The Politics of Presidential Impeachment in Latin America

Published online by Cambridge University Press:  09 March 2022

David De Micheli*
Department of Political Science, Division of Ethnic Studies, University of Utah
Jose T. Sanchez-Gomez
Department of Government, Cornell University and Instituto Desarrollo, Paraguay
Kenneth M. Roberts
Department of Government, Cornell University
*Corresponding author. Email:
Rights & Permissions[Opens in a new window]


When and why do legislatures impeach presidents? We analyse six cases of attempted impeachment in Paraguay, Brazil and Peru to argue that intra-coalitional politics is central to impeachment outcomes. Presidents in Latin America often govern with multiparty, ideologically heterogeneous coalitions sustained by tenuous pacts. Coalitions are tested when crises, scandals or mass protests emerge, but presidents can withstand these threats if they tend to allies’ interests and maintain coalitions intact. Conversely, in the absence of major threats, presidents can be impeached if they fail to serve partners’ interests, inducing allies to support impeachment as acts of opportunism or self-preservation.

Spanish abstract

Spanish abstract

¿Cuándo y por qué los poderes legislativos proceden con la destitución a presidentes? Analizamos seis casos de intentos de destitución presidencial en Paraguay, Brasil y Perú para sostener que la política entre las coaliciones es central en los resultados de dichas deposiciones. Los presidentes en Latinoamérica con frecuencia gobiernan con coaliciones multipartidarias e ideológicamente heterogéneas sostenidas por pactos débiles. Estas coaliciones son puestas a prueba cuando emergen crisis, escándalos o protestas masivas, aunque los presidentes pueden sostenerse antes estas amenazas si atienden los intereses de sus aliados y mantienen las coaliciones intactas. Por el contrario, en la ausencia de amenazas mayores, los presidentes pueden sufrir destituciones si no logran servir los intereses de sus aliados, induciendo a los aliados a apoyar la deposición como actos de oportunismo o autopreservación.

Portuguese abstract

Portuguese abstract

Quando e por que os poderes legislativos destituem presidentes? Analisamos seis casos de tentativas de impeachment no Paraguai, Brasil e Peru para argumentar que a política de coalizão é fundamental para os resultados do processo. Os presidentes da América Latina geralmente governam com coalizões multipartidárias e ideologicamente heterogêneas sustentadas por pactos tênues. As coalizões são testadas quando surgem crises, escândalos ou protestos em massa, e os presidentes podem resistir a essas ameaças se atenderem aos interesses dos aliados e mantiverem as coalizões intactas. Por outro lado, na ausência de grandes ameaças, os presidentes podem sofrer impeachment se não atenderem aos interesses dos parceiros, induzindo aliados a apoiar o impeachment como atos de oportunismo ou autopreservação.

Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (, which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright © The Author(s), 2022. Published by Cambridge University Press


In November 2020, mass protests forced Peru's Interim President Manuel Merino to resign from office, less than a week after he led a congressional impeachment of the popular president Martín Vizcarra and claimed the office for himself. Although the social backlash triggered by this impeachment was rare – Peruvians widely interpreted it as self-serving political cover for corrupt legislators – the political machinations that produced it were not. The Peruvian case was the latest in a series of presidential impeachments in Latin America that have laid bare the constitutional measure's political uses and, potentially, its abuses. In Paraguay, for example, President Fernando Lugo was removed from office in 2012 after impeachment proceedings that lasted less than 24 hours. In 2016, Brazilian President Dilma Rousseff was impeached amidst a massive corruption scandal, despite the absence of any evidence to implicate her personally.

Striking in these latter two cases is that the impeachments were orchestrated not by opposition-controlled legislatures, nor simply in response to popular demands or severe crises. Instead, presidents seemingly shielded by legislative coalitions were abandoned by their allies, who joined opposition forces in supporting articles of impeachment. On paper, legislatures are endowed with impeachment powers as constitutional mechanisms of horizontal accountability, an institutional corrective for executive misconduct. But in practice, of course, impeachments are not straightforward applications of the law; they can rarely, if ever, be decoupled from the political self-interests and strategic calculations of parties and politicians.

That impeachments are inherently political is hardly a novel assertion. It is an aphorism found in scholarship and ‘folk wisdom’ in popular interpretations of impeachment proceedings. But what, exactly, does it mean to say that impeachments are inherently ‘political’? Conventional explanations attribute presidential failures to crisis, scandal, mass protest and the absence of a legislative majority held by the president's party or coalition. We contend, however, that under minority presidents who govern with multiparty coalitions, the impact of factors like crises, scandals and protest is filtered through intra-coalitional politics, which better accounts for whether and when legislatures remove presidents from office.

While a number of factors – including coalitional politics – have been shown to shape impeachments, we complement existing scholarship by centring intra-coalitional dynamics as a proximate cause of divergent outcomes, separating this from facilitating background factors.Footnote 1 Additionally, we empirically demonstrate the causal processes at work and improve upon qualitative approaches by employing a case selection strategy that allows for testing propositions, making causal inferences and weighing alternative explanations. We show that impeachments ultimately hinge on contingent political alignments and self-interested calculations tied to the unravelling of tactical alliances among actors who are otherwise rivals and competitors. In Latin America's fragmented party systems, presidents rarely command single-party legislative majorities and often rely on ideologically heterogeneous, multiparty coalitions built on and sustained by tenuous pacts. Rather than a function of legal wrongdoing, poor presidential performance or the partisan composition of legislatures, we see impeachments as a more proximate function of coalition management and conflict. Shifting political conditions such as crises or scandals can alter strategic calculations, incentives and alignments in ways that unsettle coalitions constructed among strange political bedfellows. Presidents who effectively manage their coalitions during crises, however, can prevent legislative allies from removing them in acts of political opportunism, self-preservation or retaliation. Conversely, presidents who mismanage their coalitions may be impeached even in the absence of legal wrongdoing, severe crisis or scandal. Effective coalition management does not follow a ready-made formula and is not determined by ideological or programmatic compatibility; it belongs more to the art of politics than the science of it, and it places a premium on a leader's political adeptness, and not simply her political resources. What matters, above all, is how leaders deploy the resources they command to tend to the political interests of coalition partners.

The cases referenced above are examples of how coalition allies resorted to impeachment when they felt presidents had reneged on tacit commitments to protect allies’ interests. To evaluate our arguments, we leverage divergent impeachment outcomes across four embattled presidents – Lugo in Paraguay, Rousseff and her successor Michel Temer in Brazil, and Pedro Pablo Kuczynski in Peru. Kuczynski and Temer both managed to evade impeachment threats, though this was only temporary for Kuczynski, who resigned when facing a second impeachment vote. Temer's survival of multiple threats illustrates that outcomes can hinge on the presence of a legislative shield rather than presidential malfeasance; Kuczynski's fate demonstrates that impeachment is not simply an ideological weapon deployed by conservative legislators against leftist presidents. In short, our case selection strategy leverages divergent outcomes across cases and successive, intra-case time periods. Combining longitudinal and cross-sectional comparisons allows us to draw valid inferences from the analysis of causal processes in settings that isolate and control for different explanatory variables.

This combination also allows us to measure our key explanatory variable (coalition management) independent of the outcome of interest (presidential impeachment), thus avoiding the inferential pitfalls associated with circular or tautological reasoning. When looking behind impeachment proceedings and the public rationales offered by legislators, it becomes clear how the conventional factors identified in previous scholarship do not account sufficiently for divergent outcomes of presidential survival or impeachment. Instead, our analysis casts impeachments as a symptom of ‘coalitional presidencies’,Footnote 2 in which the constitutional lever of impeachment can be wielded as a routinised instrument of partisan, factional or personal advantage.

This article contributes most directly to theorising on presidential failure and instability in Latin America by offering a more systematic account of highly contingent coalitional politics and the causal mechanisms by which they produce presidential impeachments. More specifically, our argument distinguishes between background and proximate causes of impeachment and identifies the political dynamics and shifting calculations within the president's coalition – rather than static factors like the partisan composition of the legislature – as the key factor driving divergent impeachment outcomes. Additionally, this study contributes to the growing scholarship on coalitional presidentialism in Latin America, which has identified coalition formation and management as sources of democratic endurance. By contrast, our analysis draws attention to the mechanisms of coalitional unravelling and its consequences for political stability, horizontal accountability and the functioning of democratic institutions. Ultimately, our analysis sheds light on the causes of presidential impeachment and raises questions about whether legislative behaviour surrounding presidential impeachment indeed serves as a mechanism of accountability, or as a partisan weapon for opportunistic and self-interested political elites.

In what follows, we first review current explanations in the impeachment literature before detailing our argument that intra-coalitional political dynamics are central to divergent impeachment outcomes. We then discuss our case selection and empirical strategy and present our process-tracing analysis. The final section discusses the implications of our findings for the impeachment literature and the quality of democratic institutions in Latin America.

Crisis, Outrage and Minority Presidents

Although cloaked in the language of constitutional legality, presidential impeachments are necessarily political acts. Impeachment is central in the literature on interrupted presidencies and political instability, which attributes presidential downfalls to a confluence of factors: crisis or scandal provoked by presidential wrongdoing or poor performance; the mobilisation of public opinion and protest; or the vulnerability of a president whose party holds a minority of seats in the legislature. Kathryn Hochstetler and Margaret Edwards, for example, find support for these factors in their quantitative analysis.Footnote 3 Aníbal Pérez-Liñán similarly emphasises the importance of the president's ‘legislative shield’, and the role of the media, in mobilising public outrage over scandals.Footnote 4

Previous literature makes clear that these factors increase the risk of impeachment. What remains unclear is whether these are direct and proximate causes of impeachment or facilitating conditions that offer a pretext for opportunistic rivals to strike against incumbent presidents. This is particularly true regarding economic crises, political scandals and mass protest. Given the common stipulation that impeachment be reserved for ‘high crimes and misdemeanours’, legislatures are compelled to provide a legal rationale for impeachment proceedings. Crisis and scandal are easy culprits, but their emergence does not preordain serious threats of impeachment. Presidents may be caught red-handed in acts of corruption or other misdeeds, yet this is no guarantee that legislatures will impeach them. Indeed, Hochstetler and Edwards find that presidents mired in corruption scandals are more likely to face challenges to their presidencies but are less likely to fall.Footnote 5 Crises or scandals can provide political cover for what might otherwise be a risky legislative strategy – impeaching a president – but scandal alone does not compel legislatures to impeach.

Widespread protest and public support for impeachment are also likely to increase its odds. But protests are sometimes a final straw, emerging after impeachment proceedings begin.Footnote 6 Moreover, protests can cut both ways, creating public demand for impeachment or providing embattled presidents with ‘popular shields’.Footnote 7 And though protests have brought down presidents (in Argentina, Ecuador and Bolivia, for example), protesters lack authority to remove presidents and thus count on legislatures (or resignations) to enact their will. As with scandal, protest and public demand might provide political cover for impeachment, but they do not guarantee it. Friendly or collusive legislatures can protect even the most embattled presidents from removal.Footnote 8

Finally, current literature emphasises the partisan composition of the legislature in explaining presidential failures. Explicitly or implicitly, scholars argue that legislatures inherently pose threats to presidents when their parties hold a minority of seats, depriving presidents of a ‘legislative shield’.Footnote 9 But the view that partisan or ideological make-up preordains ‘the opposition’ to challenge presidential authority underappreciates the prevalence and function of multiparty coalitions, which are built to bridge this gap.Footnote 10 Impeachment threats against minority presidents do not follow inevitably from scandal or protest, nor mechanically from the partisan composition of the legislature; their emergence must be accounted for. This requires attention to the central role the president's coalition can play in disarming – or escalating – potential threats of impeachment.

Coalitional Politics and Presidential Impeachments

We argue that whether, when and why legislatures impeach presidents is conditioned by intra-coalitional political dynamics. We draw on scholarship that centres presidential systems around the president's governing coalition. Whereas earlier scholarship expected presidentialism to produce inter-branch conflict,Footnote 11 more recent scholarship has drawn attention to ‘coalitional presidentialism’,Footnote 12 which enables minority presidents to compensate for their legislative minority status by building multiparty coalitions. Coalition-building has been common in Latin America, where proportional representation is associated with fragmented multiparty systems and typically prevents a president's party from capturing a legislative majority.Footnote 13 To govern without partisan majorities in the legislature, presidents must assemble coalitions among actors whose cooperation is contingent across a range of factors, such as cabinet seats, influence over public policy, the selection of a vice-president and/or legislative chambers’ leadership. Less formally, coalitions may also be built around tacit agreements or expectations of shared control over pork-barrel spending and government appointments, alternation in office, or mutual protection against judicial scrutiny.

Previous scholarship on impeachment has largely cast executive−legislative political dynamics as friendly or antagonistic based on the partisan composition of the legislature or a posture adopted early in the president's term. We, however, see this relationship as dynamic − contingent on, not foreclosed by, partisan or ideological misalignment, and responsive to broader and evolving political conditions. We emphasise minority presidents’ need to rely on coalitions, even of strange bedfellows, to enact their agendas and remain in office. Coalition partners, in turn, agree to cooperate in exchange for their own benefits, often with little regard for party or ideology. In Latin America, minority presidents have been more common than not,Footnote 14 yet cooperating with the putative opposition has enabled many to govern effectively.Footnote 15

In elaborating this argument, we rely on Paul Chaisty and colleagues’ concept of the ‘presidential toolkit’, which allows minority presidents to build and sustain coalitions by granting concessions and benefits to coalition partners – promises of political influence or future power, finite cabinet and highly-sought posts, pork-barrel resources, policy concessions, the exchange of favours, etc.Footnote 16 Patronage and spoils come with the office, but presidents are also responsible for the leadership skills and style they bring to bear on coalition allies. Leadership is notoriously elusive to theorise, but scholars have nonetheless emphasised its importance in shaping impeachment outcomes. Mariana Llanos and Ana Margheritis, for example, attribute Fernando de la Rúa's downfall in Argentina to poor leadership.Footnote 17 Similarly, Christopher Martínez emphasises the conditions that facilitate the election of political neophytes, who lack the skills and experience to maintain political relationships.Footnote 18 In essence, these scholars are identifying what Chaisty and colleagues describe as coalition management, the ‘process of continuous political cultivation’ of their often heterogeneous coalitions.Footnote 19 As we will see, this continuous cultivation of coalition partners, or its breakdown, was vital to the divergent outcomes of impeachment processes, as none of the presidents in question adopted the strategies that Pérez-Liñán labels ‘isolation’ or ‘confrontation’ in dealing with Congress; all opted for the alternative strategy of ‘negotiation’ that he outlines.Footnote 20 Impeachments occurred when negotiations ceased to protect the narrowly personal or partisan self-interests – not the programmatic or ideological goals – of coalition partners. Notably, the presidents who were impeached in Brazil, Paraguay and Peru had little or no legislative experience prior to assuming the presidency, and only Temer – the one politician we analyse with a long career as a congressional deal-maker – successfully evaded all impeachment efforts. What matters, we argue, is not simply which or how many resources are available, but how adeptly the president deploys these resources to tend to allies’ interests.

To successfully manage their coalitions, minority presidents must first act as formateurs, establishing pacts that sustain the coalition. Explicit accords may be buttressed by more tacit ‘understandings’ or mutual expectations and, once established, terms of a pact can be renegotiated over time as conditions change. At a maximum, ‘successful’ coalition management enables the president's legislative policy agenda. Remaining in office is clearly a prerequisite for this; but at a minimum, coalition partners must shield the president from legislative impeachment, even if they decline to help the president legislate and make policy. Under these latter conditions, minority presidents might become lame ducks, unable to advance their agendas, but they would not be impeached. Impeachment only occurs when erstwhile allies decide that lame-duck status, or executive−legislative gridlock, is not enough to neutralise presidents they no longer support; their interests require that they join the opposition and vote for impeachment, transforming the legislative shield into a sword.

Dramatic decisions to defect and support impeachment, we argue, are not ‘bolts from the blue’, even if legal proceedings occur quite suddenly (as in Paraguay). Instead, impeachments are preceded by, and are observably distinct from, serious strains within multiparty coalitions that call into question the adherence of one side or another to the terms of the political pact. How publicly these grievances are aired may vary, but their existence is critical to our argument about the centrality of coalition management to impeachment outcomes; in their absence, we might risk using the impeachment itself as evidence for breakdown or mismanagement of multiparty coalitions, a clear form of circular or tautological reasoning. Valid causal inference requires that we conceptualise and identify our independent variable, the effectiveness of coalition management, prior to and separate from our dependent variable of impeachment outcomes.

We accomplish this in our analysis because the strategic interaction that leads to impeachment, in each case, is an iterative process that unfolds over time across a series of sequential stages. In early iterations of the process, opposition forces may submit impeachment petitions but are blocked by the president's multiparty legislative shield. Subsequently, serious conflicts between the president and coalition partners provide evidence that their alliance is fraying, and that key partners are questioning whether the coalition continues to serve their personal or partisan interests. Coalition partners may, for example, vote against the president's legislative agenda, criticise the president in the press, resign from cabinet positions, fraternise with opposition forces or grumble behind closed doors. These are all signs that the president is struggling to manage her coalition, but the political consequences for the president fall short of impeachment. Only in subsequent iterations when the president's continuation in office no longer serves, and may in fact threaten, their interests do coalition members seize opportunities to impeach presidents.

So conceived, coalition (mis)management falls along a spectrum with varying levels of cohesion or conflict. Although there is no clear threshold of mismanagement or conflict that determines precisely whether a coalition partner will support impeachment, our comparative analysis suggests that a critical stage is reached when coalition parties, factions or leaders come to perceive the president's continuation in office as a threat, rather than an asset, to their own ambitions. Once coalition partners perceive that a president no longer opens doors to, but instead inhibits, their access to power or benefits, they are likely to seize opportunities to employ the sword of impeachment. Coalitions unravel, then, in ways that pit the president's interests against those of her coalition partners, inducing allies to support impeachment in acts of self-interest, self-preservation or opportunism.

In this framework, factors commonly identified as correlates of failed presidencies – executive malfeasance, scandal, crisis and mass protest – are important, but should be understood as exogenous shocks that thrust impeachment onto the political agenda and test the coalition. Likewise, conflicts over policies or ideology are a test of coalitional cohesion, but often a secondary consideration in coalition break-ups. These all qualify as risk factors for any president, part of the background or ‘permissive’ conditions that make impeachment possible. They are not, however, the direct or proximate cause of impeachment, as they alone cannot account for divergent outcomes once a credible impeachment threat has emerged, justified or not. As we will see, legislatures (and coalition partners) are equally capable of manufacturing justifications for impeachment as they are of shielding malfeasant presidents or removing presidents even without public support.

It is important to recognise that intra-coalitional politics and the legislative–executive relationship are dynamic, not static. The relationship is continuously updated throughout a president's term, not fixed early on, according to how broader political conditions shape the strategic incentives of coalition partners. Under conditions of crisis or scandal, even presidents who initially established functional working relationships with Congress and coalition partners but who fail to tend to the interests of coalition partners can see these relationships unravel, leaving the president exposed to the legislative sword.

Elements of our argument are supported by existing scholarship. Gabriel Negretto's quantitative study of minority presidents and coalition formation finds that minority presidents are no more likely to fall, but that majority governments are more likely to survive;Footnote 21 studies by Leiv Marsteintredet and colleagues find that presidential failure correlates with the resignation of cabinet ministers and with the inclusion of a vice-president from outside the president's party;Footnote 22 and Andrés Mejía Acosta and John Polga-Hecimovich link presidential instability in Ecuador to the inclusion of legislators exercising discretion over impeachment in governing coalitions.Footnote 23 Similarly, one analysis of Lugo's impeachment in Paraguay suggests the legislature used impeachment as a ‘political weapon’ against a weakened president ‘when the opportunity arose’ – that is, when Lugo's mishandling of a crisis ‘finally shattered his troubled coalition’.Footnote 24 In various cases, scholars suggest that coalitional management and cohesion are critical to presidents’ survival of impeachment threats, but the contingent dynamics of intra-coalitional politics are not always centred or specified as a proximate cause of divergent outcomes. Nor do extant studies empirically demonstrate these causal processes at work, or employ small-N case selection strategies that allow for testing theoretical propositions. Our purpose, then, is to specify and elaborate this argument theoretically and show that divergent impeachment outcomes cannot be explained by the facilitating conditions that inspire impeachment threats alone. We show that factors like crisis, scandal and mass protest are filtered through coalitional politics, and that evasion of impeachment depends on the internal cohesion of multiparty coalitions formed around minority presidents. These coalitions are contingent on short-term, tactical considerations of partisan, factional and personal advantage, especially in contexts where they bring together parties with divergent ideological profiles and programmatic objectives.


Given space limitations, we refer readers to the online Appendix for extended discussion of case selection, methodological decisions and alternative explanations, which we review here before presenting case analyses.Footnote 25 We present process-tracing analysis of six attempted impeachments in three countries, employing a most similar/different outcome (MSDO) research design to explain divergent outcomes across similar cases of embattled minority presidents dependent on multiparty legislative shields.Footnote 26 To account for divergent outcomes and assess our propositions, we compare presidents who faced impeachment attempts, but not all of whom ultimately fell. Process-tracing methods are most appropriate given our emphasis on the dynamic interactions and tactical considerations of coalition partners, which unfold successively within cases. Process tracing allows us to closely examine the precise factors in complex causal configurations leading to impeachment in each case and to adjudicate between background and proximate causes. We draw on high-quality sources of domestic journalism; in the Paraguayan case we supplement with elite interviews.

We selected cases with a crisis or scandal that challenged the president and thus raised a credible impeachment threat. Under these conditions, removal from office is possible, but not guaranteed, as it hinges on the complex configuration of political crises mediated by coalition management. Table 1 displays the selected cases, their outcomes and values on explanatory factors. Temer and Kuczynski each faced two serious impeachment attempts, providing additional within-country variation that highlights the dynamic and contingent nature of coalitional politics behind impeachment. Of these four presidents, three ultimately fell. Temer's survival and Kuczynski's initial evasion provide crucial variation for understanding divergent impeachment outcomes. All four are also minority presidents, which reflects the reality of Latin American contexts and the conditions for the intra-coalitional dynamics we theorise. In other institutional contexts with majority rule by an internally fragmented or undisciplined party, the management of intra-party factions may well provide analytical parallels to multiparty coalition management, but it lies beyond the scope of this analysis to explore such impeachment dynamics.

Table 1. Outcomes and Explanatory Factors

Note: a Kuczynski lacked a governing coalition but was ideologically aligned with the legislative opposition.

Source: Authors’ elaboration.

Importantly, this case selection provides variation on our main independent variable, coalition management. Additionally, it allows us to control for various factors thought to shape impeachment outcomes (e.g. president's ideology and ideological alignment with allies). By comparing cases cross-sectionally and longitudinally, we are able to demonstrate how the dynamics of coalition management shape the politics of impeachment. Our analyses below pay close attention to our theory-building cases of Lugo and Rousseff, which are similar on many variables and combine multiparty coalitions with coalitional disintegration. We then present streamlined analyses of Temer and Kuczynski to leverage variation on our key independent variable and assess alternative explanations. We elaborate on these and other methodological considerations in the online Appendix.

The Rise and Fall of Lugo

The 2008 presidential victory of Lugo, a leftist former bishop of the Catholic Church, took most observers by surprise, given the historical underdevelopment of leftist parties and movements in Paraguay, and the longstanding dominance of two conservative parties with roots in the nineteenth-century oligarchic order.Footnote 27 A proponent of peasant movements and liberation theology, Lugo presented a significant challenge to Paraguay's political establishment. It was evident from the start that instability would plague Lugo's administration. Talk of Lugo's removal began immediately and became standard fare in the discourse of Congress, the media and the economic elite. Nevertheless, protected by the legislative shield of a multiparty coalition, Lugo governed for nearly four years without serious impeachment attempts – until June 2012, when pre-existing strains in his coalition came to a head, igniting proceedings that removed Lugo in under 24 hours.Footnote 28

The broad coalition that elected Lugo was built to challenge the long-dominant Colorado Party in power since 1948. The Colorados’ electoral dominance began to erode following democratisation in 1989, but opposition forces remained fragmented and thus could not defeat the Colorados in presidential elections. A political crisis created an opportunity in 2006, however, when Colorado President Nicanor Duarte sought to gain full control over his party and amend the national Constitution to allow his re-election. These moves prompted opposition forces to coalesce, a broad movement to develop in protest, and Lugo to emerge as an independent political leader above party politics, given his background in the church. Lugo appealed to a diverse coalition of opposition partisans and independents, generating momentum for a presidential bid.Footnote 29

Although his campaign generated enthusiasm, Lugo's initial coalition lacked the nationwide political organisation needed to defeat the Colorados. To address this organisational deficit, Lugo formed an alliance with the centre-right Liberals, the Colorados’ conservative rivals who had not held power since 1940. The Liberals supported his candidacy and gained the vice-presidential spot on Lugo's ticket. The alliance was a marriage of convenience between politicians with distinct ideologies, but who shared an interest in ending Colorado dominance. In Lugo's words, this alliance was forged ‘for practical reasons’, because ‘we did not have electoral experience, and the Colorado Party is an election-winning machine … The Liberals are the only [opposition] party that had national reach.’Footnote 30 Lugo's Liberal running-mate, Federico Franco, echoed this view: ‘We couldn't get there with a Liberal candidate in the same way that Lugo couldn't get there without Liberals’ support. He needed the muscle, the strength of the Liberals’ structure, people knowledgeable about the electoral process.’Footnote 31

Lugo thus entered the 2008 campaign leading a multiparty, ideologically heterogeneous coalition, the Alianza Patriótica por el Cambio (Patriotic Alliance for Change, APC). This electoral front included the Liberals and several small leftist and centre-left parties, as well as civic alliances formed by labour and peasant unions. Adding to this heterogeneity, the Liberals were not a disciplined organisation, and had three primary internal factions led by Vice-President Franco and Senators Efraín Alegre and Blas Llano.

In 2008, Lugo's APC won the presidency with 41 per cent of the vote. This electoral victory awarded the APC sizeable seat shares in both houses of Congress, though Liberal politicians held the overwhelming share of those seats. Although the coalition did not have enough seats to pass legislation on its own, it had enough to form a ‘legislative shield’ against impeachment – so long as the Liberals remained loyal to Lugo.Footnote 32

Control over the executive branch became an enormous point of tension, however, as different parties – and Liberal factions – demanded their own quota of power. To stabilise his coalition, Lugo resorted to the distribution of cabinet positions to allied partners.Footnote 33 Yet the allocation of cabinet positions itself created unrest within the coalition: Liberal politicians felt they were not adequately compensated for their contributions to Lugo's victory, and they wanted greater influence over appointments.Footnote 34 Leftist allies were also unsatisfied, feeling that Lugo should have appointed cabinet officials favouring a more radical leftist agenda.

Lugo struggled to appease these politically and ideologically fractious blocs within his governing coalition. His government expanded social programmes that buttressed Lugo's own popularity but did little to ease tensions within the governing coalition. Indeed, these policies created unease among conservative elites inside and outside the coalition, who feared leftist influence in the administration and likened Lugo's government to the radical ‘Bolivarian’ Left in Venezuela, Bolivia and Ecuador.

Not surprisingly, then, the combination of conservative opposition and intra-coalitional dissent meant that political opponents (including some Liberal members) routinely raised the possibility of impeachment. Three days after Lugo took office, the senate's legal advisor suggested Lugo undergo trial for the way he appointed ministers elected as legislators. Few took the accusations seriously, but it was a harbinger of things to come. Six months later, in March 2009, a US embassy cable reported rumours that political leaders from the Colorado and Unión Nacional de Ciudadanos Éticos (National Union of Ethical Citizens, UNACE) parties were ‘working together to assume power via (mostly) legal means should President Lugo stumble in coming months. Their goal: capitalise on any of Lugo's missteps to break the political deadlock in Congress, impeach Lugo, and assure their own political supremacy.’Footnote 35 In May 2009, another cable reported that some politicians believed the Liberal vice-president was part of a plot to overthrow Lugo.Footnote 36 In February 2010, tensions around impeachment gained international attention when Vice-President Franco and other politicians were alleged to have discussed the possibility of impeachment with US military officials.Footnote 37 Politicians also continued offering justifications for Lugo's impeachment to the public, claiming that he was a threat to the ‘democratic system’Footnote 38 and pursued a ‘leftist totalitarian project’.Footnote 39 This opposition discourse was encouraged by mainstream media outlets, including the influential ABC Color. Some charges were formally presented and advanced partially through Congress,Footnote 40 but none garnered significant support before 2012 because the Liberals protected Lugo despite their evident disagreements.

In 2012, however, a series of factors coalesced to erode Lugo's multiparty coalition, setting the stage for a Liberal defection and Lugo's eventual impeachment. First, Lugo had gradually distanced himself from key bases of support, especially among the Liberals. Lugo's relationship with his vice-president began to erode almost immediately following his election, as he excluded Franco from the process of selecting cabinet ministers.Footnote 41 This left Franco feeling snubbed, and it violated the tacit understanding that the Liberals would have significant influence in Lugo's coalition. Second, in 2011 Lugo's allies proposed constitutional amendments to permit presidential re-elections. Lugo's minister of public infrastructure, Alegre (Liberals), publicly declared his opposition to such reforms and shortly thereafter was removed from his cabinet post.Footnote 42 Matters only worsened as Lugo increased the presence of Colorado leaders in his government, further straining his relationship with the Liberals.Footnote 43 Including some Colorado politicians (although not major leaders) in his cabinet was a strategy to gain influence in minor factions within the Colorado Party and moderate the opposition of a few Colorado legislators.Footnote 44

Third, significant peasant mobilisations and land conflicts emerged in rural areas, increasing political tensions between the government and conservative elites. Although Lugo's administration had not implemented any agrarian reform and was unsuccessful in recovering public lands unduly allocated to oligarchs under previous administrations, the conservative opposition accused Lugo of encouraging class struggle. Major land occupations by peasant organisations in 2011 posed a serious threat to powerful landowners. And while the government successfully prevented armed confrontations, Lugo was blamed for encouraging conflict.Footnote 45 Fear of Lugo's leftist leanings was transparent among conservatives inside and outside the ruling coalition.

As the 2013 election approached, tensions swelled around the selection of a successor to Lugo, who could not legally be re-elected. Lugo's unwillingness to explicitly back a Liberal candidate angered those in the party who felt the Liberals should lead the ticket as a form of political reciprocity. As Franco reports, ‘We believed [Lugo] would produce rotation, that he would govern the country trying to correct course, and then would enable the possibility for a Liberal to continue this process in the next period.’Footnote 46 Instead, Lugo began associating himself with another outsider politician, a progressive journalist widely viewed as his preferred successor, confirming his disregard for the Liberals in their eyes. In short, Lugo did not honour Liberals’ expectations of intra-coalitional reciprocity and leadership rotation, and the coalition seriously frayed as Liberal leaders became convinced that Lugo was an obstacle to their own presidential ambitions.

Matters came to a head in June 2012 when a land conflict left 11 peasants and six police officers dead in a rural province. While the causes of this massacre remain unclear, what is certain is that ‘the days following were a mess’, according to Lugo's former advisor, Gustavo Codas, ‘because sectors of the Left wanted to protest a government they saw repressing peasants, and elites wanted to overthrow Lugo because they saw a government supporting peasant rebellions’.Footnote 47 This outbreak of violence inflamed the coalition's internal conflicts, especially after Lugo responded by appointing another Colorado politician, Rubén Candia Amarilla, to his cabinet to oversee the National Police. Amarilla was close to Colorados who were relatively moderate in their opposition, and as a prior attorney general Amarilla adopted repressive measures against peasants. Lugo believed appointing him could appease Colorado factions and conservative elites in the opposition, but the gambit backfired. The appointment angered the Left and social movements, and Lugo overestimated the support he would gain from conservative elites, who were already supporting his ouster. Moreover, Lugo failed to coordinate a crisis response with the Liberals, further diminishing what support remained for him in the party.

Lugo's former minister and chief of cabinet, Miguel López-Perito, confirms that the final straw for Lugo's coalition was the appointment of Amarilla. Had Lugo named a Liberal instead, López-Perito reports, the Liberals might not have joined the opposition and supported Lugo's impeachment.Footnote 48 Three days after the massacre, Senator Llano ‘and the heads of the Liberal Party in Congress called upon representatives from the Left and ministers close to Lugo to take a decisive approach toward the Liberals and the removal of Amarilla’, recalls former senator Alberto Grillón.Footnote 49 Lugo, however, ignored their demands, sealing his own fate. With Amarilla remaining in his post, the final important faction of the Liberals, led by Llano, abandoned Lugo's administration.

Though Lugo faced no mass protests, enjoyed broad public supportFootnote 50 and presided over a stable economy, this sequence of events ultimately broke his ‘legislative’ shield and left him vulnerable to impeachment.Footnote 51 On 21 June 2012, impeachment charges were brought for ‘improper performance of duties’, passing the lower house by a vote of 76−1, with three abstentions. The following day, the upper house also voted overwhelmingly in favour of impeachment, clearly demonstrating Lugo's weakness in Congress once the ‘legislative shield’ of the Liberals was removed. Lugo stated that he would accept the outcome of the vote to ‘avoid bloodshed’,Footnote 52 but insisted that the process amounted to a parliamentary coup that violated his right to due process.

Vice-President Franco, who assumed the presidency with support from the main Liberal factions, defended Lugo's ouster, claiming that ‘it's the same as the vote of no confidence’.Footnote 53 Since Paraguay is not a parliamentary system, such a rationalisation clearly shows that relevant political actors see the impeachment process as a political act rather than a legal one. Ultimately, Lugo was able to paper over ideological conflicts within his coalition and survive impeachment threats so long as he could appease disparate Liberal factions and allow Liberal leaders to anticipate accessing power themselves. The coalition, however, was predicated on a series of tacit agreements and expectations regarding power-sharing arrangements and leadership rotation; once Lugo's compliance with these expectations was called into question, congressional allies broke with the president and eventually joined his opponents in impeachment proceedings under the cover of political scandal. The impeachment of Rousseff in Brazil followed a similar political logic.

The Demise of Rousseff

In Brazil's 2014 election, 28 parties won seats in the lower house and 18 parties won seats in the upper house of Congress. Such fragmentation exacerbates the challenges of maintaining a functional coalition, even without aggressive investigations into political corruption. Yet this was the case in Brazil in 2016 after revelations of a massive bribery scheme implicating all major political parties threatened coalition allies, undermined the tacit understandings that sustained the president's coalition, and increased risk of presidential impeachment.

Key members of Partido dos Trabalhadores (Workers’ Party, PT) President Rousseff's second-term coalition belonged to the Partido do Movimento Democrático Brasileiro (Brazilian Democratic Movement Party, PMDB), which became implicated in the bribery scheme along with the PT. Generally seen as right-of-centre, the PMDB is characterised as a clientelist party known more for its pursuit of cabinet positions and spoils than for its ideology. Rousseff inherited many of her party alliances from her predecessor, Luiz Inácio ‘Lula’ da Silva, a former labour union leader who founded the PT and served two terms as a minority president (with PMDB support) from 2003 to 2010. Lula hand-picked Rousseff to succeed him and brokered the inclusion of Temer (PMDB) on Rousseff's ticket as the vice-presidential candidate in the 2010 elections.Footnote 54 Initially this alliance paid off politically: the PMDB and PT held the most seats in Congress, and PMDB politicians were elected to leadership positions in the legislature for much of Rousseff's tenure. Renan Calheiros (PMDB) was elected leader of the senate in 2013 and Eduardo Cunha (PMDB) leader of the lower house in 2015. Thus Rousseff, leader of the centre-left PT, relied heavily on the centre-right PMDB to govern Brazil's fragmented political system.

Although this coalition was relatively functional through Rousseff's re-election in 2014, the relationship soured when revelations of a massive bribery scheme implicated an impressive proportion of the political class across all major parties. The scheme centred on the state-controlled energy company Petrobras and consisted of a cartel of engineering firms competing for government contracts, from which politicians would extract bribes.

Investigations were first revealed in March 2014 – well before Rousseff's re-election in October 2014 – but it eventually became clear that what initially appeared as the isolated improprieties of select individuals was instead a massive scheme in which firms, Petrobras executives, politicians and party officials all participated for personal and partisan gain. Once charged or found guilty, participants began offering testimony in exchange for plea deals. Authorities conservatively estimated that upwards of Reais$42 billion were lost to this scheme, though this was not clear at the time of the 2014 election.Footnote 55 Informants alleged that funds were used in Rousseff's 2010 campaign,Footnote 56 but the electoral court initially approved her campaign's finances following the 2014 election, subduing these accusations.Footnote 57

Yet Brazil's three main political parties, the PT, PMDB and Partido da Social Democracia Brasileira (Brazilian Social Democratic Party, PSDB), continued to face allegations from informants, who claimed kickbacks were hidden, laundered or funnelled into party coffers. In February 2015, a former Petrobras executive testified that the PT benefitted handsomely from this scheme, triggering a major backlash against Rousseff's government in public opinion.Footnote 58 Polls conducted shortly before and after this news broke showed that Rousseff's disapproval ratings increased from 44 to 62 per cent.Footnote 59 With commodity prices falling and the economy entering into recession after a period of robust growth, anti-PT sentiment quickly translated into mass protests: in March 2015, an estimated 2.5 million Brazilians took to the streets in over 200 cities calling for Rousseff's impeachment.Footnote 60 The political crisis deepened in August 2015 when the electoral court reopened its auditing of the PT's 2014 campaign finances, suggesting there was evidence to the swirl of allegations.Footnote 61

As anti-PT sentiments crystallised in public opinion, calls for impeachment remained by and large the rallying cry of protesters.Footnote 62 Few politicians echoed these calls initially, despite some submitting impeachment petitions. Indeed, before eventually accepting a petition, lower-house leader Cunha rejected more than ten requests for Rousseff's impeachment submitted after her 2014 re-election – a clear demonstration of a functional legislative shield. In retrospect, many politicians, aware of their own ties to the bribery scheme and unsure of how far investigations might go, were likely opting for prudence by not vocally supporting impeachment. With time, however, it became clear that prosecutors were determined to investigate and charge as many officeholders as possible, heightening uncertainty and risk for politicians and placing new strains on Rousseff's coalition.

As investigations proceeded and prosecutors took aim at Cunha, Rousseff's coalition began to unravel largely because Rousseff was unwilling to offer the protection from prosecutorial reach that Cunha and other congressional leaders expected of their ally. In July 2015, news broke of a testimony accusing Cunha of receiving a US$5 million bribe in the Petrobras scheme.Footnote 63 That day, Cunha publicly accused Rousseff of coordinating with prosecutors to coax an allegation from informants. The following day, Cunha publicly broke with Rousseff and declared himself in opposition to the government, establishing a pattern whereby Cunha would retaliate against Rousseff as investigations into his activities crept closer.

In August 2015, prosecutors moved to indict Cunha for diverting US$40 million of public funds.Footnote 64 Though his future was increasingly insecure, Cunha had still not come out publicly in favour of impeachment. No politician had yet been arrested due to the scandal, and Cunha likely believed he could survive by relying on the judicial protections of his office or powerful allies to subvert investigations. Similarly, senate leader Calheiros – who would not be indicted for corruption until late 2016 – publicly stated his opposition to impeachment, saying that impeaching Rousseff would ‘set fire to Brazil’.Footnote 65 The situation grew more tense, however, when the lower house's ethics commission began considering a request to investigate Cunha for failing to disclose Swiss bank accounts allegedly used to accept bribes – a charge that could remove Cunha from Congress and deprive him of judicial privileges that could delay his trial considerably.Footnote 66 Adding to the political intrigue, the ethics commission's decision would hinge on the votes of three PT representatives.

With Cunha's situation growing desperate, threatening Rousseff with impeachment behind closed doors offered him much-needed political leverage, despite the absence of evidence to implicate Rousseff personally in the scandal. Mindful of the damage that Cunha's retaliation could inflict on Rousseff and the PT, Lula met with PT deputies to try to organise a ‘truce’ between the party and Cunha.Footnote 67 To no avail, the ethics committee voted to investigate Cunha.Footnote 68 Hours later, Cunha announced he would accept a petition for Rousseff's impeachment.Footnote 69 Notably, Rousseff's alleged offence in this petition was not corruption – the basis for protesters’ calls for impeachment and Cunha's public rationale – but a contorted charge of violating federal budget laws by temporarily transferring funds from state banks to government social programmes. This accounting practice had also been employed by Rousseff's democratic predecessors,Footnote 70 and Cunha himself had publicly declared it to be insufficient grounds for impeachment just eight months prior.Footnote 71

Though Cunha's role was critical in initiating impeachment proceedings, other key coalition members, Calheiros and Temer, similarly advanced the impeachment process, albeit in less conspicuous ways. Less than one week after Cunha agreed to a vote on impeachment, Temer sent an open letter to Rousseff in which he claimed she had no confidence in him or his party and that he felt like a ‘decorative’ vice-president – a strong signal that any loyalty to Rousseff was fading.Footnote 72 Meanwhile, Calheiros never publicly stated his position on impeachment – even as he became the target of a growing number of corruption investigations himself – but his coordination with Temer and Cunha in advance of impeachment votes suggests a concerted effort. In March 2016, the PMDB formally withdrew from Rousseff's government, saying no politician could hold office in the federal government in the name of the party.Footnote 73 In April 2016, with Rousseff's coalition in tatters, and with Cunha under a looming threat of being suspended from office by the Supreme Court, the lower-house leader scheduled an impeachment vote.Footnote 74 The following day, anticipating the outcome of the lower house's vote, Calheiros scheduled a vote in the senate.Footnote 75

Rousseff's impeachment seemed likely, but it remained unclear if the lower house could reach the necessary two-thirds majority. Rousseff made a last-ditch effort to persuade legislators to vote against impeachment. In an editorial in the widely read Folha de S.Paulo newspaper, Rousseff accused her coalition partners outright of using impeachment to thwart corruption investigations.Footnote 76 But with a significant portion of Congress implicated (or soon to be) in the scandal, Rousseff's appeal likely only reminded politicians that it would be less risky to install Temer as president – given his own implication in corruption allegations – than to leave Rousseff in power. The following day, the lower house voted in favour of impeachment, easily surpassing the two-thirds threshold.Footnote 77 Sixty-one per cent of legislators who had joined Rousseff's coalition after her re-election voted in favour of impeachment.Footnote 78 Within a month, the senate also voted to open impeachment proceedings, suspending Rousseff and installing Temer as president.

By ushering the PT out of power, Rousseff's impeachment began a rightward political shift in Brazil, and it intensified ideological polarisation. There are no signs, however, that ideological conflict was the proximate cause of coalitional disintegration and the collapse of Rousseff's legislative shield. After all, the PMDB governed in alliance with the PT from Lula's first term in office – when the PT was still shedding its far-left profile – and it continued to shield Rousseff from impeachment threats throughout 2015 despite massive right-wing protests and right-wing opposition legislators. Were ideological motivations paramount, the PMDB clearly could have moved to impeach Rousseff in 2015 and claimed a popular mandate to do so. Only when PMDB leaders themselves were confronted with serious corruption charges, and Rousseff declined to employ any tools at her disposal to shield them from prosecution, did they abandon Rousseff and deploy the sword of impeachment.

Post-impeachment developments further support the argument that impeachment served the interests of malfeasant allies. Secret recordings of Temer and allies discussing impeachment as a cover-up quickly emerged, and Cunha, Calheiros and politicians from the other mainstream conservative party were removed from the legislature on corruption charges. And Cunha himself was sentenced to 15 years in prison for corruption. With Brazil's major parties seriously discredited and imprisoned front-runner Lula prohibited from running for president in 2018, the path was cleared for Brazil to elect a far-right, anti-establishment, law-and-order candidate, Jair Bolsonaro, as president. Rousseff's impeachment, therefore, proved to be a critical first strike in a larger right-wing political offensive that weaponised the law to incapacitate flawed party institutions and fill the political void with a new brand of militarised populism.

But although the unravelling of Rousseff's governing coalition opened the door for this larger right-wing offensive, the timing and sequencing of events, as chronicled above, strongly suggest that the defection of PMDB congressional leaders was driven by personal and political considerations rather than ideological convictions. Brazil's corruption scandal – in particular, the efforts of politicians to shield themselves from its prosecutorial reach – shattered the multiparty coalition that offered a measure of governability in a context of hyper-fragmentation. It undermined the logic of coalitional presidentialism, and left Brazil exposed to the institutional gridlock, paralysis and polarisation famously attributed to multiparty presidentialism.

Ironically, the scandal culminated in the impeachment of one of the few leading politicians who was not directly implicated in allegations of corruption. Rousseff's impeachment had to do with her unwillingness to provide political and legal cover for corrupt members of her coalition, rather than her own malfeasance. The outcome of this case, therefore, hinged on the tacit understandings and expectations that coalition members have of their partners. These expectations condition legislative support for minority presidents, and they leave such presidents vulnerable to abandonment or betrayal when left unfulfilled.

The Survival of Temer

If Rousseff's impeachment demonstrates the fateful consequences of coalitional disintegration, then Temer's survival illustrates the benefits of coalitional cohesion. The contrast between these two presidents offers explanatory leverage since Temer presided over the same political and economic crises that roiled Rousseff's second term yet was credibly accused of corruption. The public also overwhelmingly supported his impeachment: 81 per cent of Brazilians supported his removal, greater than the support for Rousseff's impeachment.Footnote 79 If malfeasance or popular demands explained impeachment, Temer would have faced similar, if not greater, risks. Yet Temer did not meet Rousseff's fate, despite two attempts by the judiciary to compel legislative impeachment. Twice Temer was shielded by congressional allies who shared his goal of staving off prosecution.

Although threats to Temer's presidency did not originate in the legislature, his judicial indictments offer comparative insight. First, as with legislative impeachment, indictments suspend the sitting president, and conviction can lead to complete removal from office. Second, prosecution of a president can proceed only after a lower-house vote. One can certainly imagine that in the absence of the judiciary's aggressive prosecution, discussions of impeaching Temer would not have occurred in Congress. Indeed, after Temer assumed the presidency, more than 25 petitions for his impeachment were submitted, yet Temer's allies rejected all such requests.Footnote 80 Thus although prosecutors forced Congress’ hand, Congress was not obliged to comply, and Temer's allies remained veto players exercising discretion over his fate.

This became clear as Temer's legal exposure became untenable. His misdeeds and attempt to buy Cunha's silence led to two separate judicial indictments in 2017, first for ‘passive corruption’, then for obstruction of justice and organised crime.Footnote 81 Prosecutors asked Congress to remove Temer from office and pay R$10 million in fines. Were investigations to proceed and find Temer guilty, he would be tried by the Supreme Court, which would investigate him in an expedited trial and make a final decision, leaving Temer with no further legal recourse. Were Congress not to acquiesce, investigations would be postponed until Temer left office, when he could be tried at a snail's pace, with ample opportunity for lengthy appeals processes.

Thus for Temer – as for his congressional allies in similar jeopardy – deploying all strategies to avoid legal investigations was paramount. Fortunately for Temer, he could count on allied leaders in Congress to rally opposition to these indictments, in particular lower-house leader Rodrigo Maia of the right-wing Democratas Party and senate leader Eunício Oliveira of Temer's PMDB. Both allies were also implicated in the bribery scheme, giving them personal reasons to keep the president on their side.Footnote 82 Unlike Rousseff, Temer and his key coalition partners were aligned in their opposition to investigations, and they could use their institutional leverage and political resources to maintain a legislative shield.

Twice Temer's allies shielded him from prosecutorial reach, but despite maintaining a right-wing coalition built around his PMDB, this was not a foregone conclusion. The Comissão de Constituição e Justiça (Congressional Justice Commission, CCJ), tasked with reviewing indictments and making recommendations before the chamber vote, was led by Temer's co-partisan Sérgio Zveiter, who announced publicly that he would vote to recommend that investigations proceed.Footnote 83 Fearing that other committee members would do the same, Temer's allies (limited by rules setting quotas for parties in congressional committees) had no choice but to play ‘political chess’, swapping out dissident PMDB deputies and other right-wing politicians who, Temer feared, would vote against him.Footnote 84 The CCJ ultimately defied Zveiter,Footnote 85 but Temer, still fearing the chamber vote, began offering pork to his allies and worked to accelerate the chamber's vote to minimise opportunities for critics to voice dissent.Footnote 86 Ultimately, Temer's coalition remained intact and his first indictment was defeated, but Temer's anxiety surrounding the vote suggests that the outcome was not predetermined. Temer's survival placed a premium on his ability to co-opt and manage an undisciplined coalition.

A similar process played out with Temer's second indictment. Since Temer knew that his own party was not unified to protect him, he and his allies switched from carrots to sticks in an effort to enforce party discipline. Dissident PMDB deputies were stripped of influential cabinet posts following the first vote,Footnote 87 and Zveiter and five other PMDB deputies were suspended from party functions, prompting Zveiter to leave the PMDB altogether.Footnote 88 PMDB leaders also intimidated other potential defectors by expelling ‘traitors’ from the party.Footnote 89 This party resolution failed, but the overall strategy was successful in limiting defections; in October 2017 the lower house again voted in Temer's favour, postponing his investigations, keeping him in office and prolonging future prosecution.

Temer's tumultuous presidency illuminates the centrality of coalitional dynamics in determining whether presidents are ousted. Crises or malfeasance alone are insufficient to explain impeachment, as Temer was personally and credibly implicated in allegations of corruption and obstruction of justice. Popular will, moreover, played little role in Congress’ actions, as the public overwhelmingly supported Temer's removal throughout both proceedings. Also clear in this case is the uncertainty around chamber votes, despite the ideological and partisan alignments between Temer and his coalition; the main threat to Temer was the lack of party discipline, not ideological opposition. Ultimately, Temer drew heavily upon his ‘presidential toolkit’ to ensure votes went in his favour: he distributed pork, reallocated cabinet posts and committee assignments, disciplined party dissidents and manipulated the timing of votes. Like Rousseff, Temer was vulnerable to impeachment; but unlike her, he limited defections and maintained a relatively cohesive coalition built, in part, around shared interests in evading prosecution.

The Survival and Demise of Kuczynski

Finally, we turn to Kuczynski, the centre-right president of Peru, who evaded one impeachment attempt only to succumb to a second, when he read the writing on the wall and submitted his resignation. Like Temer, Kuczynski faced dismal public approval ratings, but he was a conservative leader – and prominent technocrat with a background in international finance – who was ideologically compatible with conservative veto players in Congress. The Peruvian case demonstrates that impeachments are neither determined by, nor contingent on, ideological conflict between the president and Congress. Neither are they restricted to leftist presidents opportunistically removed by conservative legislators, as in Paraguay and Brazil under Rousseff. Impeachment is a potent and malleable political weapon deployable for a wide range of political purposes, not simply an instrument to dispose of presidents with whom legislators disagree.

Kuczynski was twice threatened with impeachment as evidence of his allegedly corrupt misdeeds emerged.Footnote 90 Unlike Temer in Brazil, however, Kuczynski lacked a solid governing coalition to form a legislative shield. He narrowly survived the first impeachment threat by capitalising on factional rivalries within the right-wing Fuerza Popular (Popular Force, FP), the main congressional opposition party of Keiko Fujimori, Kuczynski's opponent in the 2016 presidential election. But with revelations that Kuczynski's survival was the product of a corrupt bargain between the president and congressional politicians, his opportunistic alliance unravelled, sealing his fate.

With the FP holding an outright majority in Congress, Kuczynski struggled to govern effectively from the start of his term. He assumed office after narrowly defeating Keiko Fujimori, a daughter of the former populist authoritarian leader Alberto Fujimori. Despite winning the presidency, Kuczynski's personal party vehicle, Peruanos por el Kambio (Peruvians for Change, PPK), held only 14 per cent of seats in the legislature, and opposition from Fujimori's FP was fierce. Without a majority coalition, Kuczynski struggled to enact his policy agenda, and repeated congressional investigations forced several cabinet members to resign.Footnote 91

Kuczynski thus seemed to be particularly vulnerable to impeachment should the opportunity arise: he lacked a legislative shield, and low approval ratings would suggest public support for impeachment and provide the opposition with political cover. Yet removing a sitting president was difficult even under these conditions. The FP held a majority of seats in Congress, but a simple majority suffices only to begin impeachment proceedings. To convict and remove the president, Congress needs a two-thirds majority – a threshold the FP could only reach in concert with other parties, while remaining cohesive itself.

Thus when news broke that Kuczynski was involved in a bribery scheme with Brazilian engineering firm Odebrecht – allegations he unequivocally denied – he sought to evade impeachment by forging an alliance with a dissident faction within the FP, capitalising on the political rivalry between Keiko Fujimori and her brother, FP congressman Kenji. Even before the 2016 election, the rivalry between these siblings had attracted considerable public and media attention.Footnote 92 Keiko and Kenji disagreed on whether to seek a pardon for their father, Alberto Fujimori, himself imprisoned on charges of corruption and human-rights violations committed during his authoritarian presidency. Kenji, leading a small but rebellious faction within the FP, favoured a pardon for his father. Keiko, on the other hand, opposed a pardon, fearing that her father's return to public life would undermine her political standing as the party leader.Footnote 93

Eyeing this schism, Kuczynski was well aware of the political leverage his pardon authority offered.Footnote 94 As new allegations and testimonies implicated Kuczynski in corruption,Footnote 95 the president sought an alliance with Kenji, who was facing punishment from his party for non-compliance.Footnote 96 A pact with the factional leader could provide Kuczynski with a legislative shield, in exchange for a pardon of Alberto Fujimori. Subsequently, in December 2017, congressional investigators revealed testimony of Odebrecht executives tying Kuczynski to the bribery scheme.Footnote 97 The opposition in Congress moved swiftly to introduce impeachment proceedings.Footnote 98 Kenji, meanwhile, implored his fellow legislators to respect due process and allow Kuczynski an opportunity for self-defence.Footnote 99

As Congress prepared for an impeachment vote on 21 December 2017, Kuczynski openly declared the proceedings a coup attempt.Footnote 100 The opposition was determined to impeach, but fragmentation on the Left and Right cast doubt on the vote: the Left was reluctant lest the FP capitalise on Kuczynski's removal to consolidate power, and factional rivalries threatened the opposition's numerical advantage. When votes were cast, the opposition fell short of the necessary two-thirds majority by nine votes.Footnote 101 Leftist parties abstained, as did Kenji and nine other FP politicians who joined him in breaking the party line.Footnote 102 Kuczynski's alliance with Kenji's rebellious faction of ten votes thus proved decisive to his survival.

Three days later, Kuczynski upheld his end of the bargain and pardoned Alberto Fujimori for ‘humanitarian reasons’,Footnote 103 prompting further decline in his approval ratings and large-scale protests in the streets.Footnote 104 Kuczynski's political woes worsened as the congressional investigative committee announced new revelations of corruption along with new impeachment proceedings.Footnote 105 The opposition rebuffed accusations of acting without cause by releasing videos in which Kuczynski's lawyer was caught offering pork to Kenji and other legislators in exchange for votes in opposition to impeachment.Footnote 106 The videos strongly suggested that Kuczynski's initial survival was the product of bribery and back-room politicking, and the revelations deprived Kuczynski of the same tools to build an alliance and evade impeachment a second time. With his impeachment vote set and no coalition in place to secure a legislative shield, Kuczynski recognised that his demise was inevitable and resigned from office on 21 March 2018.

Kuczynski's tumultuous and short-lived presidency ultimately hinged on his ability to deploy presidential powers to exploit political factionalism in the opposition and assemble opportunistic alliances to guard against impeachment votes. Unlike Rousseff and Lugo, Kuczynski was a conservative president ousted by a majority-conservative Congress, indicating the limits of ideological explanations of impeachment outcomes. Instead, this impeachment highlights how political rivalries and competition for power between and within parties can lead to shifting political alliances held together by contingent and self-interested political bargains. Such tenuous pacts may just as easily shield malfeasant presidents as they do expel them.


Presidential impeachments are not readily explained by crises, scandals, mass protest or minority presidential status. Neither are they simply a function of ideological conflict between presidents and legislatures. While all these factors were present in Rousseff's case, her impeachment is only indirectly attributable to them. Lugo was impeached despite the absence of mass mobilisation, public demand, major crisis or legal wrongdoing on the president's part. The cases of Temer and Kuczynski also highlight that presidents cannot safely assume that partisan or ideological alignments will protect them from serious challenges. Only Temer, who shared with his coalition partners a strong interest in remaining out of prosecutorial reach, evaded impeachment. In the cases of Lugo and Rousseff, the legislative shields seemingly provided by multiparty coalitions were quickly transformed into legislative swords once key partners determined that the president's continuation in office had become a threat to their interests.

Recent presidential impeachments in Latin America thus demonstrate the tenuous character of multiparty governing coalitions forged by minority presidents, and the centrality of presidential leadership in the management of these coalitions. The political pacts of coalitional presidentialism were highly tactical in nature, and they rested on tacit assumptions that presidents would protect the interests of coalition partners who shared few, if any, of the president's programmatic or ideological commitments. The internal cohesion of these coalitions was reasonably robust – even in contexts of economic crisis, corruption scandals, social protest and ideological conflict – so long as presidents tended to the interests of their partners. Coalitions disintegrated, however, once allies lost confidence in the president's willingness to safeguard their interests. Sudden defections left presidents vulnerable to impeachment proceedings predicated on the loss of political support rather than legal transgressions by the executive branch.

This does not mean that crises, scandals and protest are unimportant. They matter, in part, because they create permissive conditions or pretexts for coalition partners to escape from alliances that no longer serve them. This may be especially true in ideologically misaligned coalitions that are tactical and contingent in character; under such conditions, partners have little affinity for the president or her party and do not want to serve the latter's long-term political interests. Partners may even be looking for a ‘way out’ of a coalition that has outlived its original purpose. Crises, scandals and protests may also factor into the equation because they force presidents to make difficult choices between risky or unpalatable alternatives that are threatening to their partners and/or their own core constituencies. This was evident in Rousseff's refusal to protect her allies implicated in the corruption scandal, and Lugo's appointment of a Colorado minister in response to rural violence. These crisis responses infuriated allies and contributed to their defections. Crises, therefore, may alter the political context in ways that strain governing coalitions; they can foreground new, unforeseen issues or bring old ones to a head in ways that were not ‘covered’ by the original terms of a political pact.

Though our analysis is limited in temporal scope, prior research indicates that the coalitional dynamics we identify are not unique to more recent cases of impeachment and have indeed been present in earlier cases – though coalitional dynamics are not highlighted as proximate causes in these analyses.Footnote 107 The contribution of this study is the centring of coalitional politics as a proximate cause of impeachment against presidents of diverse ideological persuasions. Our study also contributes a detailed empirical demonstration of these causal processes at work. Nonetheless, the limited temporal scope of our study means the field would benefit from more systematic and fine-grained analyses of the coalitional dynamics of impeachment attempts. A focus on coalitional politics might also help differentiate impeachments from other types of presidential ‘failures’ that have distinct causal mechanisms and scope conditions.

Finally, this analysis carries implications for understanding democratic stability and accountability, and raises questions about how impeachment and coalitional presidentialism shape democratic governance. Latin Americanists generally agree that impeachment safeguards democracy by enabling presidential removal by legal means.Footnote 108 The increase in presidential impeachments since the 1990s in Latin America may thus reflect the institutionalisation of constitutional mechanisms to remove incumbents who might have been overthrown by military coups in the not-so-distant past. Outcomes short of regime breakdown, however, do not necessarily signal the proper functioning of democratic institutions. Our analysis offers a less sanguine view of how coalitional politics condition the ability or willingness of legislatures to exercise horizontal accountability in good faith.

In these cases, impeachment was not deployed to remove malfeasant presidents from office. Self-interested politicians retaliated against presidents who failed to defend their interests. Publics demanding removal saw their will carried out only when this was aligned with elite interests. Even in the absence of public demand, elites manufactured contorted charges to justify impeaching democratically elected presidents. So too did they employ legislative and investigative machinery to shield malfeasant presidents from prosecutorial reach – as any observer of US politics under Donald Trump should also recognise.

In these cases, then, coalitional presidentialism did not necessarily enhance the rule of law or the quality of representation in multiparty systems, and it certainly did not enhance democratic stability.Footnote 109 To the contrary, it often contributed to instability and institutional dysfunction by allowing legislatures to undermine popular will and neglect their duty of holding malfeasant presidents accountable. Impeachment can be a mechanism of horizontal accountability, but it can also be weaponised as a highly partisan and opportunistic tool for protecting the interests of political elites. Impeachment diminished to an institutional surrogate for a parliamentary vote of no confidence may well enhance flexibility at the cost of institutional legitimacy; the routine and self-interested weaponisation of a constitutional lever designed to safeguard against manifest wrongdoing or abuses of power too easily reveals the sordid side of democratic competition. It also reinforces the contingency and fragility of governing coalitions, making it difficult to build the kinds of durable ruling blocs that are needed to implement and sustain major reforms, or address serious structural challenges like social and economic inequality. Such a bias toward the status quo can spread cynicism among the public and discredit parties and political elites from across the political spectrum – a ready-made formula for the type of anti-establishment populist demagoguery that Bolsonaro rode to power in Brazil. Given its high political stakes, weighty legal implications and intrinsic moral opprobrium, impeachment is a powerful political tool – one that is best held in reserve, and sparingly deployed.


The authors would like to thank François Gélineau and Aníbal Pérez-Liñán for helpful suggestions on an earlier draft of this article.

Supplementary material

To view supplementary material for this article, please visit


1 In qualitative comparative analysis (QCA) parlance, we distinguish between necessary and sufficient conditions that produce outcomes and ‘background characteristics’ that constitute the ‘domain of investigation’. Rihoux, Benoît and Ragin, Charles, Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques (Thousand Oaks, CA: Sage, 2009), p. 20CrossRefGoogle Scholar.

2 Chaisty, Paul, Cheeseman, Nic and Power, Timothy, Coalitional Presidentialism in Comparative Perspective: Minority Presidents in Multiparty Systems (New York and Oxford: Oxford University Press, 2018)CrossRefGoogle Scholar.

3 Hochstetler, Kathryn, ‘Rethinking Presidentialism: Challenges and Presidential Falls in South America’, Comparative Politics, 38: 4 (2006), pp. 401–18CrossRefGoogle Scholar; Edwards, Margaret, ‘Understanding Presidential Failure in South America’, Latin American Politics and Society, 57: 2 (2015), pp. 111–31CrossRefGoogle Scholar.

4 Pérez-Liñán, Aníbal, Presidential Impeachment and the New Political Instability in Latin America (New York: Cambridge University Press, 2007)CrossRefGoogle Scholar.

5 Hochstetler, Kathryn and Edwards, Margaret, ‘Failed Presidencies: Identifying and Explaining a South American Anomaly’, Journal of Politics in Latin America, 1: 2 (2009), pp. 3157CrossRefGoogle Scholar.

6 Llanos, Mariana and Marsteintredet, Leiv (eds.), Presidential Breakdowns in Latin America: Causes and Outcomes of Executive Instability in Developing Democracies (New York: Palgrave Macmillan, 2010)CrossRefGoogle Scholar.

7 Pérez-Liñán, Aníbal, ‘A Two-Level Theory of Presidential Instability’, Latin American Politics and Society, 56: 1 (2014), pp. 3454CrossRefGoogle Scholar.

8 Helmke, Gretchen, Institutions on the Edge: The Origins and Consequences of Inter-Branch Crises in Latin America (Cambridge and New York: Cambridge University Press, 2017)CrossRefGoogle Scholar; Pérez-Liñán, Presidential Impeachment.

9 Pérez-Liñán, Presidential Impeachment.

10 de Abranches, Sérgio Henrique Hudson, ‘Presidencialismo de coalizão: O dilema institucional brasileiro’, Dados, 31: 1 (1988), pp. 538Google Scholar; Negretto, Gabriel, ‘Minority Presidents and Democratic Performance in Latin America’, Latin American Politics and Society, 48: 3 (2006), pp. 6392CrossRefGoogle Scholar.

11 Linz, Juan, ‘The Perils of Presidentialism’, Journal of Democracy, 1: 1 (1990), pp. 5169Google Scholar; Mainwaring, Scott, ‘Presidentialism, Multipartism, and Democracy: The Difficult Combination’, Comparative Political Studies, 26: 2 (1993), pp. 198228CrossRefGoogle Scholar.

12 Abranches, ‘Presidencialismo de coalizão’; Chaisty, Cheeseman and Power, Coalitional Presidentialism.

13 Mainwaring, ‘Presidentialism, Multipartism, and Democracy’.

14 Arturo Valenzuela, ‘Latin American Presidencies Interrupted’, Journal of Democracy, 15: 4 (2004), pp. 5–19.

15 Abranches, ‘Presidencialismo de coalizão’; Negretto, ‘Minority Presidents’.

16 See Chaisty, Cheeseman and Power, Coalitional Presidentialism, for a full treatment of this framework.

17 Llanos, Mariana and Margheritis, Ana, ‘Why Do Presidents Fail? Political Leadership and the Argentine Crisis, 1999–2001’, Studies in Comparative International Development, 40: 4 (2006), pp. 77103CrossRefGoogle Scholar.

18 Martínez, Christopher, ‘Presidential Instability in Latin America: Why Institutionalized Parties Matter’, Government and Opposition, 56: 4 (2021), pp. 683−704CrossRefGoogle Scholar.

19 Chaisty, Cheeseman and Power, Coalitional Presidentialism, p. 77.

20 Pérez-Liñán, Presidential Impeachment, p. 148.

21 Negretto, ‘Minority Presidents’.

22 Llanos and Marsteintredet (eds.), Presidential Breakdowns; Leiv Marsteintredet and Fredrik Uggla, ‘Allies and Traitors: Vice-Presidents in Latin America’, Journal of Latin American Studies, 51: 3 (2019) pp. 665–88.

23 Andrés Mejía Acosta and John Polga-Hecimovich, ‘Coalition Erosion and Presidential Instability in Ecuador’, Latin American Politics and Society, 53: 2 (2011), pp. 87–111.

24 Marsteintredet, Leiv, Llanos, Mariana and Nolte, Detlef, ‘Paraguay and the Politics of Impeachment’, Journal of Democracy, 24: 4 (2013), pp. 111–13CrossRefGoogle Scholar.

25 The online Appendix is available at under the ‘Supplementary materials’ tab.

26 Rihoux and Ragin, Configurational Comparative Methods.

27 Roberts, Kenneth M., Changing Course in Latin America: Party Systems in the Neoliberal Era (New York: Cambridge University Press, 2014)Google Scholar.

28 Pérez-Liñán, ‘A Two-Level Theory of Presidential Instability’; Marsteintredet, Llanos and Nolte, ‘Paraguay and the Politics of Impeachment’.

29 Nickson, Andrew, ‘The General Election in Paraguay, April 2008’, Electoral Studies, 28: 1 (2009), pp. 145–9CrossRefGoogle Scholar.

30 Fernando Lugo, interviewed by Jose T. Sanchez-Gomez, 1 July 2014.

31 Federico Franco, interviewed by Jose T. Sanchez-Gomez, 23 June 2014.

32 Pérez-Liñán, Presidential Impeachment.

33 Chaisty, Cheeseman and Power, Coalitional Presidentialsim.

34 Blas Llano, interviewed by Jose T. Sanchez-Gomez, July 2014.

35 ‘Paraguayan Pols Plot Parliamentary Putsch’, Wikileaks Public Library of US Diplomacy (Paraguay Asunción, 28 March 2009), available at, last access 10 Jan. 2022.

36 ‘Circling Sharks in Landlocked Paraguay’, Wikileaks Public Library of US Diplomacy (Paraguay Asunción, 6 May 2009), available at, last access 10 Jan. 2022.

37 ‘Grosera carta de ministro a embajadora de EE.UU.’, ABC Color, 5 March 2010, available at, last access 10 Jan. 2022.

38 ‘Antes de seis meses, hay que sacar a Lugo, advierte Jaeggli’, ABC Color, 13 Nov. 2009, available at, last access 10 Jan. 2022.

39 ‘Calé trata de cobardes a los liberales que no quieren hacer juicio político’, ABC Color, 31 Aug. 2009, available at, last access 10 Jan. 2022.

40 Katia Gorostiaga, ‘Del acuerdo inicial a la ambigüedad: Incidencia de las estrategias presidenciales en las destituciones anticipadas’, Master's Thesis, Facultad Latinoamericana de Ciencias Sociales (FLACSO), Mexico City, 2016.

41 Federico Franco, interviewed by Jose T. Sanchez-Gomez, 23 June 2014.

42 Efraín Alegre, interviewed by Jose T. Sanchez-Gomez, 3 July 2014.

43 Gorostiaga, ‘Del acuerdo inicial a la ambigüedad’.

44 Gorostiaga shows that Lugo gradually included Colorado affiliates in his cabinet, moving from two to four Colorados as heads of a total of 12 government departments.

45 Juan Gayoso, ‘Cartes alerta sobre olor a sangre en Ñacunday’, Ultima Hora, 10 March 2012, available at, last access 10 Jan. 2022; Antonella Brignardello, ‘Advierten sobre posible violencia en el campo’, ABC Color, 25 Jan. 2012, available at, last access 10 Jan. 2022.

46 Federico Franco, interviewed by Jose T. Sanchez-Gomez, 23 June 2014.

47 Gustavo Codas, interviewed by Jose T. Sanchez-Gomez, 15 July 2017.

48 Miguel López-Perito, interviewed by Jose T. Sanchez-Gomez, 15 July 2017.

49 Alberto Grillón, interviewed by Jose T. Sanchez-Gomez, 16 July 2017.

50 As noted in Pérez-Liñán, ‘A Two-Level Theory of Presidential Instability’, Paraguayan public opinion data is scarce. Latin American Public Opinion Project (LAPOP) data indicate Lugo's net approval was slightly negative in early 2012. But polls conducted following impeachment revealed 58 per cent rated Lugo's administration as excellent or good. ‘Lugo mantiene mejor imagen que Federico Franco, según encuesta’, Ultima Hora, 16 July 2012, available at, last access 10 Jan. 2022.

51 Pérez-Liñán, ‘A Two-Level Theory of Presidential Instability’.

52 ‘Entrevista con Fernando Lugo: “Empezamos la resistencia pacífica”’, interview by Martín Granovsky, Cuba Sí, 25 June 2012, available at, last access 10 Jan. 2022.

53 ‘Entrevista al presidente de Paraguay, Federico Franco, en Los Desayunos’, Los Desayunos (RTVE, 2 April 2013), available at, last access 10 Jan. 2022.

54 Jeferson Ribeiro, ‘PT e PMDB fecham acordo para candidatura de Dilma em 2010’, Globo, 20 Oct. 2009, available at,,MUL1348659-5601,00-pt+e+pmdb+fecham+acordo+para+candidatura+de+dilma+em.html, last access 10 Jan. 2022.

55 Bibiana Dionísio, ‘PF estima que prejuízo da Petrobras com corrupção pode ser de R$42 bi’, Globo, 12 Nov. 2015, available at, last access 10 Jan. 2022.

56 Rodrigo Rangel, ‘Campanha de Dilma em 2010 pediu dinheiro ao esquema do “petrolão”’, Veja, 27 Sept. 2014, available at, last access 10 Jan. 2022.

57 ‘TSE aprova com ressalvas contas de Dilma e de Comitê Financeiro para presidente da República’, Tribunal Superior Eleitoral, 11 Dec. 2014, available at, last access 10 Jan. 2022.

58 Fabio Serapião and Wanderley Preite Sobrinho, ‘PT recebeu até 200 milhões de dólares em propina, estima delator’, CartaCapital, 5 Feb. 2015, available at, last access 10 Jan. 2022.

59 ‘Avaliação da presidente Dilma Rousseff’ (São Paulo: Datafolha, 27 Nov. 2015), available at, last access 10 Jan. 2022.

60 ‘Mapa das manifestações pelo Brasil’, Globo, accessed 26 May 2017, available at, last access 10 Jan. 2022.

61 Mariana Oliveira, ‘Ministro do TSE pede investigação das contas de campanha de Dilma’, Globo, 21 Aug. 2015, available at, last access 10 Jan 2022.

62 Nunes, Felipe and Melo, Carlos Ranulfo, ‘Impeachment, Political Crisis and Democracy in Brazil’, Revista de Ciencia Política, 37: 2 (2017), pp. 281304CrossRefGoogle Scholar.

63 Graciliano Rocha and Bela Megale, ‘Delator na Lava Jato diz que pagou propina de US$5 mi a Eduardo Cunha’, Folha de S.Paulo, 16 July 2015, available at, last access 10 Jan. 2022.

64 Vladimir Netto and Mariana Oliveira, ‘Janot apresenta ao STF denúncia por corrupção contra Cunha e Collor’, Globo, 20 Aug. 2015, available at, last access 10 Jan. 2022.

65 Agência Brasil, ‘Impeachment de Dilma seria “botar fogo no Brasil”, diz Renan’, CartaCapital, 11 Aug. 2015, available at, last access 10 Jan. 2022.

66 Ranier Bragon, ‘Com apoio de metade do PT, PSOL e Rede pedem cassação de Cunha’, Folha de S.Paulo, 13 Oct. 2015, available at, last access 10 Jan. 2022.

67 Vera Rosa and Adriano Ceolin, ‘Lula pede a deputados do PT que deem “trégua” para Eduardo Cunha’, Estadão, 16 Oct. 2015, available at,lula-pede-ao-pt-tregua-para-cunha--imp-,1780636, last access 10 Jan. 2022.

68 ‘Bancada do PT decide votar contra Cunha no Conselho de Ética’, UOL, 2 Dec. 2015, available at, last access 10 Jan. 2022.

69 Felipe Amorim and Marina Motomura, ‘Eduardo Cunha aceita pedido de impeachment da oposição contra Dilma’, UOL, 2 Dec. 2015, available at, last access 10 Jan 2022.

70 Nunes and Melo, ‘Impeachment, Political Crisis and Democracy in Brazil’, p. 288.

71 Pedro Venceslau and Elizabeth Lopes, ‘Cunha rejeita tese de impeachment de Dilma por “pedaladas fiscais”’, Estadão, 19 April 2015, available at,cunha-rejeita-tese-de-impeachment-de-dilma-por-pedaladas-fiscais,1672754, last access 10 Jan. 2022.

72 Valdo Cruz, Daniela Lima and Marina Dias, ‘Em carta-desabafo a Dilma, Temer diz que foi desprezado’, Folha de S.Paulo, 8 Dec. 2015, available at, last access 10 Jan. 2022.

73 Nathalia Passarinho and Fernando Calgaro, ‘Por aclamação, PMDB oficializa rompimento com governo Dilma’, Globo, 29 March 2016, available at, last access 10 Jan. 2022.

74 Ranier Bragon and Debora Alvares, ‘Cunha decide começar votação do impeachment por deputados do Sul’, Folha de S.Paulo, 12 April 2016, available at, last access 10 Jan. 2022.

75 Ricardo Brito, Isabela Bonfim and Julia Lindner, ‘Renan articula votar afastamento de Dilma pelo Senado no dia 11 de maio’, Estadão, 13 April 2016, available at,renan-articula-votar-afastamento-de-dilma-pelo-senado-no-dia-11-de-maio,1855750, last access 10 Jan. 2022.

76 Dilma Rousseff, ‘Democracia: O lado certo da história’, Folha de S.Paulo, 16 April 2016, available at, last access 10 Jan. 2022.

77 Carol Siqueira, ‘Câmara autoriza instauração de processo de impeachment de Dilma com 367 votos a favor e 137 contra’, Agência Câmara Notícias, 17 April 2016, available at, last access 10 Jan. 2022.

78 Nunes and Melo, ‘Impeachment, Political Crisis and Democracy in Brazil’, p. 287.

79 Gustavo Simon, ‘Maioria dos brasileiros pede saída de Michel Temer, afirma Datafolha’, Folha de S.Paulo, 24 June 2017, available at, last access 10 Jan. 2022.

80 Igor Gadelha, Daiene Cardoso and Isadora Peron, ‘Maia afirma que vai rejeitar os pedidos de impeachment’, Estadão, 28 Oct. 2017, available at,maia-afirma-que-vai-rejeitar-os-pedidos-de-impeachment,70002064158, last access 10 Jan. 2022. Llanos and Pérez-Liñán find that politicians submit impeachment for position-taking and not because they expect petitions to advance, suggesting petitions do not indicate the seriousness of impeachment threats. Still, allies’ outright rejection of impeachment suggests a functional legislative shield against legislative opponents and the judiciary, which played a critical role here by forcing consideration of impeachment. Llanos, Mariana and Pérez-Liñán, Aníbal, ‘Oversight or Representation? Public Opinion and Impeachment Resolutions in Argentina and Brazil’, Legislative Studies Quarterly, 46: 2 (2021), pp. 357−89CrossRefGoogle Scholar.

81 Afonso Benites, ‘Rodrigo Janot apresenta segunda denúncia contra Temer’, El País, 15 Sept. 2017, available at, last access 10 Jan. 2022; Renan Ramalho and Vitor Matos, ‘Janot apresenta ao Supremo denúncia contra Temer por corrupção passiva’, Globo, 26 June 2017, available at, last access 10 Jan. 2022.

82 Debora Alvares, ‘Citado em delação, Eunício Oliveira faz acordo para ser eleito com folga’, Folha de S.Paulo, 23 Jan. 2017, available at, last access 10 Jan. 2022; ‘Rodrigo Maia é acusado de corrupção e lavagem de dinheiro pela PF’, Globo, 9 Feb. 2017, available at, last access 10 Jan. 2022.

83 João Fellet, ‘Relator vota pela aceitação de denúncia contra Temer; entenda peso da decisão’, BBC News Brasil, 10 July 2017, available at, last access 10 Jan. 2022.

84 Bernardo Caram and Alessandra Modzeleski, ‘Partidos da base de Temer fazem novas trocas na CCJ’, Globo, 12 July 2017, available at, last access 10 Jan. 2022.

85 Bernardo Caram, Alessandra Modzeleski and Fernanda Calgaro, ‘CCJ rejeita parecer que recomendava continuidade da denúncia contra Temer’, Globo, 13 July 2017, available at, last access 10 Jan. 2022.

86 ‘Veja estratégias do governo para derrubar inquérito contra Temer na Câmara’, Globo, 12 July 2017, available at, last access 10 Jan. 2022.

87 Thiago Faria, ‘PMDB suspende deputados que votaram a favor de denúncia contra Temer’, UOL, 10 Aug. 2017, available at, last access 10 Jan. 2022.

88 Gustavo Maia, ‘Relator de denúncia contra Temer deixa PMDB após punição: “Perseguição política”’, UOL, 11 Aug. 2017, available at, last access 10 Jan. 2022.

89 Juliana Braga, ‘Divergentes do PMDB podem ficar sem punição na segunda denúncia’, Globo, 11 Oct. 2017, available at, last access 10 Jan. 2022.

90 ‘Encuesta de Opinión Pública a Nivel Nacional, Marzo 2018’ (Datum Internacional, March 2018), available at, last access 18 Jan. 2022.

91 ‘Los ministros que dejaron el Gabinete en el gobierno de PPK’, El Comercio, 5 Jan. 2018, available at, last access 10 Jan. 2022.

92 Jacqueline Fowks, ‘Kenji Fujimori não foi votar na sua irmã’, El País, 7 June 2016, available at, last access 10 Jan. 2022.

93 Jacqueline Fowks and Carlos Cué, ‘Os Fujimori, um drama familiar que domina o Peru’, El País, 31 Dec. 2017, available at, last access 10 Jan. 2022.

94 ‘PPK y Kenji Fujimori se encontraron de forma casual: ¿Qué se dijeron?’, La República, 7 July 2017, available at, last access 10 Jan. 2022.

95 ‘La Lista Negada’, Caretas, 30 Nov. 2017, available at, last access 10 Jan. 2022; ‘Comisión Lava Jato pide a PPK que se pronuncie sobre vínculos con Odebrecht’, 30 Nov. 2017, available at, last access 10 Jan. 2022.

96 ‘Kenji Fujimori fue nuevamente suspendido 120 días por Fuerza Popular’, La República, 28 Nov. 2017, available at, last access 10 Jan. 2022.

97 ‘Odebrecht afirma que pagó US$782,000 a PPK’, Caretas, 7 Dec. 2017, available at, last access 10 Jan. 2022.

98 ‘Presentan moción de vacancia presidencial’, Caretas, 15 Dec. 2017, available at, last access 10 Jan. 2022.

99 ‘Kenji Fujimori exhorta al Congreso “respetar presunción de inocencia” de PPK’, La República, 15 Dec. 2017, available at, last access 10 Jan. 2022.

100 ‘PPK: “Estamos ante un golpe disfrazado de interpretaciones”’, La República, 20 Dec. 2017, available at, last access 10 Jan. 2022.

101 ‘PPK se queda: Congreso no consigue los votos y el fujimorismo fracasa’, La República, 22 Dec. 2017, available at, last access 10 Jan. 2022.

102 Alexander Villarroel Zurita, ‘Los fujimoristas que se abstuvieron de votar por la vacancia de PPK’, El Comercio, 22 Dec. 2017, available at, last access 10 Jan. 2022.

103 ‘Alberto Fujimori recibe indulto y Kenji muestra el preciso instante’, La República, 25 Dec. 2017, available at, last access 10 Jan. 2022.

104 ‘Encuesta de Opinión Pública a Nivel Nacional, Enero 2018’ (Datum Internacional, Jan. 2018), available at, last access 18 Jan. 2022; ‘Alberto Fujimori: Miles marcharon contra indulto y en rechazo a PPK’, La República, 28 Dec. 2017, available at, last access 10 Jan. 2022.

105 ‘Vacancia presidencial: Presentaron moción multipartidaria contra PPK’, La República, 8 March 2018, available at, last access 10 Jan. 2022; ‘Congreso admitió nueva moción de vacancia presidencial contra PPK’, La República, 15 March 2018, available at, last access 10 Jan. 2022.

106 ‘Videos demostrarían supuesta compra de votos para evitar vacancia de PPK’, La República, 20 March 2018, available at, last access 10 Jan. 2022.

107 Valenzuela, ‘Latin American Presidencies Interrupted’; Llanos and Marsteintredet (eds.), Presidential Breakdowns; Acosta and Polga-Hecimovich, ‘Coalition Erosion and Presidential Instability in Ecuador’.

108 Marsteintredet, Leiv and Berntzen, Einar, ‘Reducing the Perils of Presidentialism in Latin America through Presidential Interruptions’, Comparative Politics, 41: 1 (2008), pp. 83101CrossRefGoogle Scholar; Pérez-Liñán, Presidential Impeachment.

109 Chaisty, Cheeseman and Power, Coalitional Presidentialism.

Figure 0

Table 1. Outcomes and Explanatory Factors

Supplementary material: PDF

De Micheli et al. supplementary material

De Micheli et al. supplementary material

Download De Micheli et al. supplementary material(PDF)
PDF 447 KB