Introduction
Luís Inácio Lula da Silva, Brazil’s current president, said that his main opponent in the presidential race, Jair Bolsonaro, looked like a “Court Jester,” and “[did] not wield control over anyone.” At the time, Lula claimed, “Bolsonaro is a hostage of the National Congress; Bolsonaro does not even take care of the budget; the one who takes care of the budget is [House Speaker Arthur] Lira, this has never happened [before].”
This surprising accusation presents us with an apparent oxymoron: how can an undeniably authoritarian ruler who was widely said to threaten democracy can simultaneously seem like a powerless buffoon? Interestingly, the accusations shed light on why Brazil’s democracy survived Bolsonaro. Some of Bolsonaro’s weaknesses are related to Brazil’s constitutional design and some to institutional changes that have taken place since the enactment of the 1988 constitution (Pereira and Melo Reference Pereira and Melo2012). Other reasons for Bolsonaro’s weakness, however, relate to the nature of plebiscitarian leadership, typical of populism.
Brazil’s coalitional presidentialism is characterized by a strong president possessing significant constitutional power, a fragmented party system, and a robust federal structure. Although the president continues to be the agenda setter and still holds several powers, this “strong president equilibrium” has recently been weakened by an empowered legislative branch and a highly assertive and independent Supreme Court. Bolsonaro did not succeed in altering that status quo.
According to the literature on democratic recession and backsliding (Diamond Reference Diamond2015; Mounk Reference Mounk2018; Haggard and Kaufman Reference Haggard and Kaufman2021), which dwells on the dramatic decline of liberal democracy in several parts of the world since the late 1990s, one would swiftly reach the conclusion that Brazil was an ideal candidate for a democratic breakdown.
Gloomy Expectations about Brazilian Democracy
Even before being elected, Bolsonaro elicited gloomy expectations from renowned scholars, intellectuals, and journalists, who argued that he posed the greatest existential threat to Brazilian democracy. Levitsky and Ziblatt (Reference Levitsky and Ziblatt2018), for instance, argued that Bolsonaro scored high in their litmus test of an authoritarian candidate, consisting of four warning signs (rejection of the democratic rules of the game, toleration/encouragement of violence, denial of rivals’ legitimacy, and curtailing civil liberties), and argued that he was more authoritarian than Orbán, Erdogan, Duterte, and even Chávez. Bolsonaro was the most outspoken defender of the previous military dictatorship in his country.
Levitsky and Bizzarro (2018), writing shortly before the elections, claimed that “Bolsonaro could finish off Brazilian democracy… If Bolsonaro wins, Brazil will look more like Venezuela.” Footnote 1 The expectations were indeed bleak: “Under Bolsonaro elections would be less free and fair and the Executive (branch) would constantly abuse power.” The arrival of a far-right populist president in an institutional environment like that of Brazil raised serious concerns that Bolsonaro could abuse his powers and, consequently, undermine checks and balances, harm the rule of law, and, ultimately, threaten its liberal democracy.
In a New York Times op-ed, entitled “Can Brazil’s democracy be saved?” Muggah wrote that “Mr. Bolsonaro represents the greatest existential threat to Brazil’s democracy.” Footnote 2 In a WhatsApp message leaked in May 2020, Celso de Mello, the senior justice of the Supreme Court, wrote: “We must resist the destruction of the democratic order to avoid what happened in the Weimar Republic,” which was overthrown by Hitler. “Brazilian democracy is under serious threat,” agrees Oscar Vilhena Vieira, the dean of the Law School of São Paulo at the Getúlio Vargas Foundation: “The president is not just trying to create an institutional conflict, [but also] trying to stimulate violent groups.” For Matias Specktor, of the Getúlio Vargas Foundation, the problem was not just Bolsonaro, “democracy can die even if you don’t have a strongman.” Footnote 3 If Mr. Bolsonaro is eventually impeached, vice-president, Mourão, a retired army general would succeed him, bringing the army still closer to power. Sérgio Abranches (2022) believes that “Bolsonaro is a real threat to Brazilian democracy. He has been able to neutralize many checks and balance organizations. Bolsonaro has captured the General Public Prosecutor, the federal police, and parts of the judicial system. The only check that has disciplined and stablished limits to the behavior of the President is the Supreme Court.” Footnote 4
In order to make things worse, the COVID-19 pandemic raised additional concerns that Bolsonaro, like other populists around the world, could take advantage of the emergency situation to overreach, erode civil liberties, and put forward an illiberal agenda. It is conventional wisdom that emergency situations require massive delegation of power to the executive, which is the only branch of government with the information, decisiveness, and speed to respond to crises. Checks and balances that ordinarily constrain constitutional governance thus typically cease to exist during times of crisis.
Against the odds and in contrast to pessimistic expectations from many pundits (Diamond and Plattner 2015; Diamond Reference Diamond2022; Levitsky and Ziblatt Reference Levitsky and Ziblatt2018; Graber et al. Reference Graber, Levinson and Tushnet2018; Kyle and Mounk Reference Kyle and Mounk2018), democracy has proved to be resilient. The country has not faced significant institutional erosion, understood in a more minimalist sense; Footnote 5 that is, specifically in terms of the rules of the democratic competitive game and institutions capable of checking on the executive, which remained intact. We agree with Levitsky and Way’s (Reference Levitsky and Way2023) cautious note that “electing a president or prime minister with autocratic tendencies certainly heightens the risk of backsliding, but it should not necessarily be taken as evidence of backsliding.”
Nevertheless, Brazil experienced dramatic policy reversals and massive failures in several policy areas such as the environment, education, and health among others. Here it is important to make a distinction between institutions and policies, which in many cases might not be correlated. Whereas the former represent the rules of the democratic game, the latter are public policies implemented by a particular government. It is not uncommon to observe a fully democratic government implementing poor quality policies and vice-versa. Unlike would be autocrats, Bolsonaro did not try to remove justices or prosecutors by resorting to informal means such as fabricated accusations of corruption. Nor did he propose to enlarge the size of courts or persecute opposition politicians and/or journalists. His bellicose rhetoric, however, is not meaningless. It has negatively affected democratic life, especially the public discourse, in significant ways.
Is Brazil a Near Miss?
Ginsburg and Huq (Reference Ginsburg and Huq2018) argue that while there are many cases in which democracy’s survival might be explained by the absence of a credible threat to the democratic system, we should pay attention to a narrower class of what they call democracy’s “near misses.” These are “the cases in which a democracy is exposed to social, political, or economic forces that could catalyze backsliding, yet somehow overcomes those forces and regains its footing.”
Ginsburg and Huq define a near miss as a case in which a country experiences a “deterioration in the quality of initially well-functioning democratic institutions, without fully sliding into authoritarianism but then partially recovers its high-quality democracy.” True near misses, they claim, are countries in which democracy was under severe threat but survived intact; quick comebacks, such as Colombia under Uribe, are those cases which may have fallen below “a minimum threshold of democratic quality into a nascent competitive authoritarianism, albeit one that nevertheless was quickly reversed.”
The boundaries between these categories are fuzzy and it is difficult to characterize specific cases. Differentiating cases of near misses and cases where democracy was never at risk are also not clear and they are disputable in particular cases. On the face of it, we could classify Brazil as a case of quick recovery. But, considering that Brazil never felt below a threshold of competitive authoritarianism, this would not be entirely accurate either. Pundits exaggerated the threat Bolsonaro posed, but we acknowledge it was not inconsequential. Furthermore, it is also not clear that Brazil is within the subset of cases where the risk of substantial democratic erosion had already become apparent either “1) through institutional deterioration or in the political agenda of an insurgent antisystem faction or leader, but 2) the danger clearly receded with the defeat of the political force behind the threat.”
The country has returned to its previous level of democracy. In fact, it has never been altered in significant ways, as argued before. Checks and balances remained relatively strong in Brazil. The Supreme Court and Congress have systematically ruled against the preferences of the government in key legal disputes. The media is free. Civil liberties have been preserved. The election calendar has been respected (Melo et al. Reference Melo, Negretto and Pereira2023).
The losing party accepted the outcome of the election, despite a Capitol-style incident on January 8, when radical and antidemocratic protesters supporting the former president invaded and ransacked the offices of Congress, the Supreme Court, and the presidential palace. Despite these dramatic and unprecedented events, political crisis and democratic instability were averted. In fact, they generated a bold unified response from almost all political players, from the executive and legislative branches, both at the federal and subnational levels, as well as from the Supreme Court. President Lula, the Speakers of the Chamber of Deputies and the Senate, the Supreme Court’s justices, and state governors rallied to the capital, Brasília, the next day. Parading and holding hands, they provided a robust demonstration of political unity in favor of democracy. The military did not respond to calls by minority radical groups for a coup d’état. The winner of the election is governing with the support of a majority in Congress, including parties and individual politicians associated with the former president. In fact, the presidents of both houses of Congress under Bolsonaro have been reelected with Lula’s active support, who, despite being backed by only a minority in Congress, negotiated a major change in the 2023 budget. Lula has appointed several legislators closely associated with the former government to his cabinet.
Democratic Resilience Despite Numerous Threats
The fact that Brazilian democracy neither died nor experienced a severe process of backsliding despite Bolsonaro’s numerous threats is quite puzzling and begs an explanation. Footnote 6 There was a significant erosion of democratic norms under Bolsonaro, and his confrontational political rhetoric against democratic institutions was unprecedented. However, many of his actions undeniably represented threats, which elicited countervailing responses by institutions.
Bolsonaro, for instance, fired his defense minister for not condoning the politicization of the armed forces, prompting the collective resignation of the army, navy, and air force chiefs. He also called for the impeachment of two Supreme Court justices and went as far filing a petition for the removal of Justice Alexandre de Moraes. The president of the Senate shelved the proposal prompting Bolsonaro not to go ahead with the impeachment of Justice Roberto Barroso. This illustrates the deterrent effect of an independent Congress, which supported Bolsonaro’s policy initiatives but did condone his hegemonic intentions.
Another demonstration of Bolsonaro’s pursuit of hegemonic power is his suspected interference in elections. Investigations by the federal police revealed compelling evidence suggesting manipulation during the presidential election runoff. Silvinei Vasquez, director of the Federal Highway Police, reportedly issued orders for strategic raids with the aim of obstructing voters’ access to polling sites in municipalities primarily in the Northeast region. This region had shown strong support for Bolsonaro’s rival during the first round of elections, where he garnered over 80% of the votes.
The federal police discovered maps on the former minister of justice Anderson Torres’s cell phone, which detailed the distribution of votes in the first round of the election. It was revealed that Torres had personally attended meetings in these Northeastern municipalities just days before the runoff to coordinate these raids. Consequently, both Vasquez and Torres have been arrested.
Following investigations carried out by the federal police, key insights emerged from the testimony provided by Bolsonaro’s direct assistant, Lieutenant Colonel Mauro Cid, who cooperated via a plea bargain agreement approved by the Supreme Court. These revelations shed light on Bolsonaro’s direct involvement in shaping a “legal framework” known as the “minuta do golpe” (coup decree). This framework was intended to provide justification for exceptional measures in the event of a successful popular uprising.
Evidence collected during the investigation suggested that the coup’s planned deadline was December 18, coinciding with the day the Electoral Court was set to confirm Lula’s presidency. The “minuta do golpe,” found both in the residence of the former minister of justice, Anderson Torres, and on his cellphone, called for new presidential elections based on what was deemed illegitimate and unconstitutional decisions by the Supreme Court. It also argued for the necessity of a “state of siege” decree to restore democratic rule in the country.
According to Cid, Bolsonaro was directly involved in the crafting of this decree, offering suggestions and making adjustments. Moreover, Cid revealed that Bolsonaro convened a meeting with the heads of the armed forces to discuss the coup decree. During this meeting, the commander of the Navy expressed readiness to support the presidential call. In contrast, the head of the army explicitly stated his reluctance to participate in any coup plan. The air force commander, also did not endorse the initiative. Consequently, the meeting concluded, and the alleged coup plan was put to rest.
While there were indeed several threats to Brazilian democracy, it is important to highlight that these threats lacked credibility. Bolsonaro neither possessed the institutional means nor garnered the necessary political support to bear the costs associated with implementing democratic setbacks. Brazil’s system of checks and balances was already robust and society remained vigilant. Consequently, these factors not only enabled a reactive response to illiberal initiatives after they were proposed but also served to dissipate and discourage many illiberal initiatives before they could gain momentum.
Is It a Matter of Luck?
Some have argued that this outcome is fortuitous and they have put forward a counterfactual explanation: if Brazilian democracy did not break down during Bolsonaro’s first term of office, it would certainly have been eroded had he been reelected for a second term (Weyland Reference Weyland2020; Smith Reference Smith2020). The essential features of a liberal democratic system, especially the rule of law, would then be at risk during a populist’s second term in power (Mounk Reference Mounk2018). Populist leaders, the argument goes, are usually unrestrained in their second terms. They feel unconstrained by checks and balances such as organizations, by Congress, by independent judges, regulatory agencies, and critics in the media. Populist redesign of the constitution often turns into a continuous process, as power-hungry leaders constantly seek to fortify and extend their domination. Gradually, the argument goes, populists trample democracy to death, as an inexorable and irreversible trend.
Weyland (Reference Weyland2022) is one of the scholars who has also challenged this explanation (see also Przeworski Reference Przeworski2019; Weyland Reference Weyland2020; Treisman Reference Treisman2023). He argues that only 24% of populist leaders who assume office in a democratic country give rise to democratic backsliding, and only one-third of them effect significant democratic deterioration. Instead of democracies dying, it is easier for populist leaders to weaken politically and die as a result of their attempts to asphyxiate democracy.
Brownlee and Miao (Reference Brownlee and Miao2022) come to similar conclusions: democratic backsliding is much less likely to lead to autocracy than is currently assumed. By examining all cases of backsliding and breakdown from the 1920s to 2020, they argue that “the tale of backsliding preceding breakdown has been much less common than the story of backsliding—and recovery—within electoral democracies.” In other words, a century worth of evidence (1920–2019) suggests that democracies are sturdier than they look.
Drawing on a variety of sources, Little and Meng (Reference Little and Meng2023) find no evidence of global democratic decline over the last decade either. Using objective indicators of democracy, such as incumbent performance in elections rather than subjective indicators which rely on expert coder judgement, they claim that recent declines in average democracy scores can only be driven by changes in coder bias. Admittedly, these contrasting findings largely stem from the diverse ways scholars conceptualize and measure both democracy and backsliding (Knutsen et al. Reference Knutsen, Kyle Marquardt, Michael Coppedge, Juraj Medzihorsky, Teorell, Gerring and Lidberg2023).
Coder subjectivity could account for part of the observed 53% decline in Brazil in the V-Dem’s Liberal Democracy Index from 2015 to 2022 (from .78 to .51), which seems to be triggered by the impeachment of the former president Dilma Rousseff in 2016. The decline precedes Bolsonaro (the index only fell from .60 to .51 under Bolsonaro). This new sensibility to democratic erosion reflected public debates prompted by Diamond (Reference Diamond2014) and the increasing availability of information on the subject. But other factors might explain resilience to backsliding and reversal to democracy. Brownlee and Miao (Reference Brownlee and Miao2022) mention the stabilizing moment of multipartyism and political competition. They also stress the importance of economic development. As an aside, Brazil is at the income threshold at which wealth is protective of the onset of backsliding according to their analysis. Weyland (Reference Weyland2022) considers the ineptitudes of populist leaders in dealing with emergencies, but also shows how political competition, the party system, and the effects of systems of government affect the ability of populists to undermine democracy. Boese et al. (Reference Boese, Edgell, Hellmeier, Maerz and Lindbergh2021) argue that restrictions imposed by the legislature prevent the executive from engaging in an undemocratic escalation of its powers, but it is the independence of the judiciary that provides resilience against collapse. These claims resonate well in the Brazilian case, as our analysis will seek to demonstrate.
Gamboa’s (Reference Gamboa2022) meticulous case studies of Colombia and Venezuela shed light on the factors contributing to the containment of democratic backsliding processes. In this work, she provides in-depth insights into the intricate dynamics at play within each country’s political landscape by focusing on the opposition strategies. Where the opposition resorts to unconstitutional means of resistance, the autocrats have an excuse for enhancing hegemonic control.
There are similarities in the way the judicial system and political parties worked in containing Uribe in Colombia, as discussed in Ginsburg and Huq (Reference Ginsburg and Huq2018) and Gamboa (Reference Gamboa2022), and Bolsonaro in Brazil. However, in the case of Brazil a key actor was the so-called parties of the Centrão, the largest legislative group in Congress. While nominally supporting Bolsonaro, its leaders strategically maneuvered to block his more illiberal initiatives. Our analytical narrative focuses on party fragmentation in a context of relatively autonomous legislative and judicial institutions as well as on federalism.
We do not observe the counterfactual of a second term under Bolsonaro given that he lost the reelection. Therefore, it is critical to examine in detail the political and institutional mechanisms explaining why Brazilian democracy survived a far-right populist, but also why it demonstrated the strength and capacity to deal with a president uncommitted to democratic principles. We are inspired by Ginsburg and Huq (Reference Ginsburg and Huq2018), Gamboa (Reference Gamboa2022) and Bermeo’s (Reference Bermeo2022) argument that the questions raised by this new scholarship, which challenges the received wisdom on threats to democracy, invites more intensive country-based research.
The Rise of a Weak Strongman
Bolsonaro’s rise to the presidency was the product of a perfect storm: the combination of a gargantuan corruption scandal (Petrolão and its anticorruption operation, Lava Jato) and an unprecedented crisis resulting from economic mismanagement and the demise of a commodity boom (Melo Reference Melo2016). The fallout was a window of opportunity for an outsider of sorts—Bolsonaro was in his seventh term of office as a lackluster congressman, who had changed parties seven times. Bolsonaro chose one of the Brazil’s micro “parties for hire”—the PSL—to run for the presidency in 2018 and consistently counted on the support of a core group of firebrand online militants, amounting to 15% of the electorate. This was enough to secure him a place in the runoff, where he benefited from the PT’s high rejection rates among voters and won by a 10% margin. Bolsonaro is therefore best characterized as a “minority populist president.”
Bolsonaro’s victory reflected more of a problem of supply in the political market rather than demand: the PT’s arch rival party in previous races—the PSDB—was also hurt by corruption scandals involving its leader Aécio Neves. Although Bolsonaro repeatedly lashed out at the political horse trading of coalitional presidentialism, referring to it as velha política, demand for a radical right politician such as himself was marginal. His support emanated from a largely insignificant constituency. Bolsonaro certainly tapped into the “abashed right” (Power Reference Power2000), which became politically voiceless beginning at the end of the military regime, and also into the rapidly growing neopentecostal evangelical’s constituency. But the core authoritarian bolsonaristas are a relatively small—though noisy and unwieldy—group. His victory was the byproduct of the PT’s high levels of rejection among swing voters (Hunter and Power Reference Hunter and Power2019).
A large group of legislators rode Bolsonaro’s coattails to victory; but he and his loyal followers left the party, which merged with the “Democrats” (a center-right party) to create a new party, called “União Brasil.” Bolsonaro remained partyless for slightly more than half of his term of office, counting on a cacophonic group of supporters which he accommodated in the federal bureaucracy, only to fire them when they became dysfunctional. Later, he sought the support of the Centrão, a heterogenous group of over 200 center-right and some radical right legislators. The Centrão core consists of three parties—the PL, the Progressistas, and the Republicanos—among other small parties whose main characteristic is a relentless progovernment stance and non-programmatic, rent-seeking behavior, which is a product of its size and pivotal position in the political spectrum. As legislators from the Centrão are not ideologically driven, they did not face high costs or feel constrained by voters to be part of Bolsonaro’s coalition. Bolsonaro’s dependence on the Centrão reflected the fact that Brazil’s party system fragmentation—its effective number of political parties 16.4, in 2018—was the world’s highest, anywhere, anytime.
Bolsonaro’s mismanagement of the pandemic and his son being accused of participation in illicit activities made the rapprochement with the Centrão a matter of political survival, a move that culminated in the dismissal of one of the bulwarks of his presidency, former Lava Jato judge Sérgio Moro. In his second year in office, Bolsonaro dropped his rejection of parties and antisystem rhetoric and morphed into an entirely new political animal: a firebrand president in alliance with ancien régime political bosses of the Centrão, whom he had lambasted during his campaign. In doing so, Bolsonaro became the court jester Lula ridiculed.
Bolsonaro alienated an important electoral constituency—the Lava Jato’s supporters—while building a key legislative alliance with the Centrão. This was a very successful move because he managed to secure the speakership of the House in early 2021, which was previously held by a centrist rival. During this process, he acquiesced to the Centrão’s demands and fired a number of his firebrand allies from the Olavista faction, Footnote 7 such as his boisterous education minister, who resisted the rapprochement with the Centrão. Although the military remained in the government—controlling key positions of the federal bureaucracy (Hunter and Vega Reference Hunter and Vega2022)—they were increasingly overshadowed by Centrão politicians.
Bolsonaro famously stated, “I was obliged to form a coalition.” The architect of this new strategy, Minister Luís Eduardo Ramos, candidly described the metamorphosis affecting Bolsonaro and overstated its legislative impact in these terms: “we met with party whips and said: do you want to participate in the government? And that way we started building support in Congress. From that moment onwards (April 2021), we were able to pass our agenda.” Footnote 8
Although Bolsonaro was able to build a coalition with the Centrão he did so in a vulnerable situation: the bargaining power was no longer in the hands of the executive but rather on the hands of the Centrão. Thus, the leaders of the Centrão were able to set up the price of the negotiations and, therefore, became the government guarantor. As long as Bolsonaro was able to reward them accordingly, legislators had no incentives to get rid of him.
A Domesticated Populist
Generally speaking, the backsliding argument starts from the assumption that institutions and society are defenseless and, above all, helpless victims of elected populists (Kyle and Mounk Reference Kyle and Mounk2018). So, once in power, populists would almost always unilaterally weaken the liberal components of a democracy. The boiling frog argument (Luo and Przeworski Reference Luo and Przeworski2023) used as a metaphor for the opposition’s inability to coordinate against backsliding downplays the role of forces that countervail backsliding.
Brazil possesses a complex set of consensual style institutions that allow for the representation of diverse interests within the political system. The country uses open-list proportional representation (PR) with high magnitude districts, which leads to high party fragmentation. Arguably, party fragmentation, and weak parties would facilitate a strongman’s dominance because it could split an autocrat’s democratic opposition (Bermeo Reference Bermeo2022). But another scenario emerged in the Brazilian case: Bolsonaro was unable to dominate conservative parties and was forced to bargain with them in order to build a legislative shield against an impeachment. Congress did not capitulate, having managed to forestall executive overreach. The protagonists in this process were not opposition parties (as emphasized in Gamboa Reference Gamboa2022), but parties who provided conditional support to the government, while simultaneously constraining his actions.
Party fragmentation has moderated illiberal governments even when has not been associated with ethnic politics. This claim is consistent with most cross-national quantitative studies which find no robust relationship between party fragmentation and democratic breakdown (Hicken et al. Reference Hicken, Baltz, Vasselai, Coppedge, Edgell, Knutsen and Lidberg2022).
Critically Brazilian parties are programmatically weak but highly centralized. Paradoxically, parties are very weak in the electoral arena, but strong inside Congress. Individual legislators tend to present relatively high party discipline (Pereira and Mueller Reference Pereira and Mueller2003) and the internal organization of Congress is also strictly partisan: the rules governing nomination to committees, committee work, and parliamentary rights are based on the parties’ seat share. Parties also benefit from an unparalleled level of public funding. When the Supreme Court banned corporate donations to political parties in 2015, parties reacted by creating a multibillion public electoral campaign fund in 2017, the FEFC. It distributed R$4,9 bi (US$ 950 million) to Brazilian parties in 2022. The FEFC funds came in addition to public funding, the FEAFPP, which distributed a total of R$930 bi (US$ 193 million) in 2022 to 24 parties. União Brasil, Brazil’s largest party in 2022, was awarded R$758 million and the PT, R$500 million, from the FEFC, in addition to R$168 million R$104 from the FEAFPP, respectively. Footnote 9
Over the last decade, Congress gradually approved measures that strengthened the legislative branch. Presidents were banned from impounding budgetary amendments (earmarks), which weakened their strategic advantage in executive-legislative relations. Constitutional amendments in 2015 (EC 2015) and 2019 (EC 100) made appropriating a certain amount of earmarks mandatory, thereby guaranteeing an annual flow of funds for pork barrel projects. Presidents’ ability to issue provisional measures with the force of law were also curtailed.
These changes in public funding and legislative prerogatives altered the “strong president” equilibrium that prevailed before Bolsonaro, and which he was unable to revert. Access to multimillion public resources from earmarks and electoral and party financing funds meant that parties became much more autonomous vis-à-vis the executive branch. As the next section will illustrate, this autonomy resulted in significant defeats in Congress for Bolsonaro.
Brazil also possesses a federal system with multiple veto points, or points at which legislation can bog down. Powerful state governors countervail presidential power. The governor of São Paulo, João Doria, provides a compelling case, acting as a formidable rival to Bolsonaro. Controlling a vast administrative machinery employing 710,000 employees (outnumbering the federal machinery with 624,000) Footnote 10 and vast resources, Doria successfully engaged in a highly publicized horse race against the federal government for the production of a COVID vaccine. He managed to initiate production in a state government institution.
Furthermore, checks and balances worked, especially in the Supreme Court, public prosecutor, federal police, Audit Courts, autonomous regulatory agencies, and a free and competitive media. All have imposed losses on incumbent presidents (Melo and Pereira Reference Melo and Pereira2013). The legislative and judicial branches have been instrumental, even on issues as “specific” as administrative transparency. For example, Bolsonaro issued three decrees to significantly weaken transparency and freedom of information (Michener Reference Michener2023), and three times the measures were defeated: first in Congress in 2019 (the president’s first legislative defeat, decree 9.690) and two others were struck down by the Supreme Court (decrees 9.759 and 928). In short, Bolsonaro’s illiberal initiatives, which were many, have met with vigorous institutional responses, a sign of Brazil’s democratic resilience. Multiple institutional and partisan vetoes have worked to safeguard democracy from illiberal initiatives.
Bolsonaro did not enjoy presidential hegemony, which Pérez-Liñán et al. (Reference Pérez-Linãn, Schmidt and Vairo2019) describe as the president’s ability to exercise political control over other institutions, particularly the legislature and the judiciary—a major driver of democratic instability in Latin America. Rather, the combination of a politically weak president (with declining constitutional powers, i.e., Bolsonaro), and an environment with strong checks and balances, facilitates democratic survival even under populist administrations. In this institutional environment, “minority populist presidents” lack the political capacity—both in terms of political support and formal powers—for an incumbent takeover.
Bolsonaro’s confrontation with the Supreme Court started with a number of highly publicized episodes during the pandemic, involving his failed attempt at meddling with the federal police. Bolsonaro appointed Alexandre Ramagem as the head of the federal police. Ramagem, who had previously served as Bolsonaro’s personal security chief following an assassination attempt during the 2018 presidential campaign, had grown close to the Bolsonaro family. However, this decision triggered a chain of events. Justice Minister Sérgio Moro announced his resignation, accusing Bolsonaro of political interference in the federal police’s operations due to Ramagem’s appointment. Moro claimed, “the president wanted someone in the federal police that he could contact to gather information and intelligence reports.” Footnote 11 In response, the Supreme Court swiftly nullified Bolsonaro’s appointment of Ramagem as the new head of the federal police, citing a breach of constitutional principles, including impersonality, morality, and public interest.
He suffered several setbacks during the pandemic (Ginsburg and Versteeg Reference Ginsburg and Versteeg2021; Llanos and Weber Reference Llanos, Weber, Llanos and Marsteintredet2023), despite the fact that unlike other populists he tried to shift the blame to governors and mayors rather than concentrate actions on the federal government. Consequently, unlike other leaders, he never benefited from the “rally around the flag” effect. Quite the opposite: his ineptitude in dealing with the pandemic is consistent with populist behavior in general (Weyland Reference Weyland2022; Michener Reference Michener2023).
Nonetheless, we must reject explanations solely based on the (in)competence of populist leaders. Levitsky and Way (Reference Levitsky and Way2023) argue that leaders like Viktor Orbán in Hungary and Nicolás Maduro in Venezuela succeeded because they had a clear plan for their authoritarian strategies. In contrast, figures like Donald Trump in the United States and Jair Bolsonaro in Brazil faltered due to their improvisational approach.
However, limiting our analysis to the competence of populists while disregarding the role of congressional and judicial institutions along with the vigilance of society is overly simplistic. Such an approach would suggest that leaders like Orbán and Maduro could effortlessly undermine American or Brazilian democracy, or conversely, that Trump and Bolsonaro, being deemed incompetent, would be incapable of disrupting democracy in Hungary or Venezuela. This line of reasoning lacks nuance.
Bolsonaro’s administration was systematically defeated in the Supreme Court, which resorted to its criminal jurisdiction against key presidential allies (Arguelhes Reference Arguelhes2022). The more the Supreme Court felt threatened by Bolsonaro’s illiberal initiatives, especially during the COVID pandemic and during his nasty campaign against the electronic ballot, the greater the number of times in which the justices of the Supreme Court behaved unanimously against Bolsonaro. In effect, shortly after Bolsonaro took office in 2019, the Supreme Court started working as a “unified continent” rather than as “eleven isolated islands,” reverting a widely used metaphor to characterize the court’s lack of collegiality and its individualistic behavior. According to Justice Luís Roberto Barroso, “what united us is the defense of democracy.” Footnote 12
Although the 1988 Brazilian Constitution established a strong and independent judicial system, the Brazilian Supreme Court has become a political protagonist since the Mensalão corruption scandal trial in 2012. In addition to being a constitutional and an appellate court, it is also a criminal court in cases involving highly ranked officials.
Even during Bolsonaro’s electoral campaign in 2018 he stirred tensions with the Supreme Court by supporting the idea of increasing its size. During his term in office, Bolsonaro falsely claimed to have uncovered evidence of fraud in the electronic ballot. Following baseless allegations about electronic voting, the Superior Electoral Court imposed the largest fine (R$20 million, US$ 3.9 million) it has ever imposed, for bad faith litigation, on the PL, Bolsonaro’s party. As his demands for a printed ballot were rejected both in Congress and in the Supreme Court, the president decided to attack specific justices, threatening two of them with impeachment. However, the president of the Senate, Senator Rodrigo Pacheco, whose election was supported by Bolsonaro, never put the request to a vote. Bolsonaro’s strategic alliance with the Centrão, of which Pacheco was part, failed in its confrontation with the Supreme Court. The Senate shelved the president’s motions to impeach justices Luís Roberto Barroso and Alexandre de Moraes in August 2021. Bolsonaro’s threats of contempt for court decisions experienced the same fate. Footnote 13 The president accumulated judicial defeats early in his administration, including an appointment to the powerful federal police. Similarly, the minister of the environment, one of the most vocal bolsonaristas, resigned amid allegations that he had obstructed a police investigation into illegal logging.
Bolsonaro was unable to curb, pack, or even reform the judiciary. All his attempts ended in virtual failure (Da Ros and Taylor Reference Da Ros and Taylor2022). No other Brazilian president has suffered so many judicial defeats. Arguelhes (Reference Arguelhes2022) claims that the Supreme Court was able to constrain Bolsonaro as a result of three interconnected factors: first, the president’s attitude towards the COVID pandemic, which paved the way for the Supreme Court to unanimously side against Bolsonaro’s preferences, allowing state and local governments to adopt their own restrictive measures regardless of the national policy recommendation. Second, the court expanded its criminal jurisdiction throughout the Bolsonaro presidency, directly initiating criminal investigations in cases such as online crimes against the Supreme Court ministers and the honor of the court. Third, a kind of informal alliance emerged between the court and Congress in which the latter consistently blocked all Bolsonaro’s initiatives and attacks against the former, including attempts to create a congressional committee to investigate the judiciary and formal requests to impeach justices of the court.
The public prosecutor in the Federal Tribunal of Accounts also acted as an important check on Bolsonaro’s overreach. Between 2019 and 2022, it filed 618 requests for investigation of wrongdoing, almost ten times the number of requests filed during the previous government (2015–18). These requests ranged from denunciations of irregularities in budget amendments to Bolsonaro’s misuse of public bank loans. Footnote 14
The Supreme Court’s reaction to executive overreach culminated in the so-called fake news inquiry, in which the court rather than the public prosecutor initiated a criminal inquiry. It led to the arrest of a bolsonarista federal deputy for threats against the court’s justices, as well as indictments and social media bans on numerous influencers and YouTubers. A widely publicized national debate ensued on whether the courts had violated democratic procedures (Arguelhes and Recondo Reference Arguelhes and Recondo2023). Whatever the merit of the court’s actions, they nonetheless demonstrate Bolsonaro’s powerlessness vis-à-vis the judiciary.
Civic actors and the independent media constrained Bolsonaro. The major TV networks, including the powerful Rede Globo, were highly critical of the president and his government. All major newspapers—O Globo, Folha de São Paulo, and Brazil’s premier conservative outlet, O Estado de São Paulo—joined forces in opposition (Michener Reference Michener2023, 16). These news outlets unified to question the administration’s propagation of disinformation, best exemplified by their formation of a ‘Press Consortium’, which compiled subnational data on the pandemic to compensate for the Bolsonaro administration’s dubious commitments to transparency. In short, the news media served as a counterpoint, particularly given the president’s strong presence on social media and his propagation of fake news.
Bolsonaro vs Congress: A Tale of Numerous Defeats
By directly associating the specific style of coalition presidentialism practiced by the previous Workers’ Party (PT) administrations with corruption, Bolsonaro fed the electorate a kind of aversion to politics itself, thus filling an open space for the emergence of his populist profile and agenda.
Therefore, Bolsonaro refused to build a coalition and sought to establish direct connections with voters. In Congress, he worked by building ad hoc and cyclical majorities in favor of executive preferences in specific bills. He sought to motivate legislators by mobilizing public opinion against them. Bolsonaro governed in a kind of permanent campaign of polarization; a conspiratorial tone formed a fundamental part of his government’s crusade against faceless enemies, who emerged anew every day and everywhere.
Less than a year into his administration, Bolsonaro decided to leave the political party that he had used as an electoral vehicle, the Social Liberal Party (PSL). In effect, Bolsonaro never committed to partisan stability, switching parties nine times since he began his political career in 1989 as a city of Rio de Janeiro councilor.
Bolsonaro’s strategy of governing without a coalition failed. Two shocks, which drastically reduced his popularity, led the government to surrender to the coalition-building imperative.
The first was the COVID-19 outbreak. Bolsonaro’s disapproval rating correlated to the number of deaths in the country. Bolsonaro minimized the virulence of the pandemic and its health consequences. “It is at most a little flu,” said Bolsonaro, downplaying the threat of the pandemic in Brazil. Under the huge distress of the pandemic, a significant portion of voters, including those self-identified as right-wing and center-right—supposedly, the core of Bolsonaro’s voters—refused to follow the president’s recommendation of relaxing social distancing and rejected his policy performance during the pandemic (Pereira et al. Reference Pereira, Medeiros and Bertholini2020).
The second shock was a set of corruption scandals. First came a scandal involving his oldest son and senator, Flávio Bolsonaro, who was accused of money laundering. Then Bolsonaro and his administration became ensnared by corruption allegations. These included the invoicing of COVID-19 purchases of the vaccine Covaxin, the use of off-budget resources known as “the secret budget” to buy legislative support, and kickback schemes involving evangelical ministers in the ministry of education. Finally, suspicions surfaced surrounding more than fifty properties the Bolsonaro family had purchased, in cash, over the last two decades or so.
In less than two years of his administration, the opposition filed over sixty formal impeachment requests, which prompted Bolsonaro to build a coalition with the Centrão. This was a “large minority,” not a minimal winning coalition. The number of seats occupied by the ten parties of the Centrão totaled only 204 out the 513 existing seats in the Chamber of Deputies. In other words, it was fundamentally a “negative” coalition, possessing only the capacity to veto undesirable legislative initiatives such as an impeachment. It was not a proactive or reform-promoting coalition that would be able to put forward an illiberal agenda. It was nevertheless a sign that Bolsonaro had been brought into line; he would have to play by the rules of Brazilian coalition presidentialism.
Bolsonaro himself justified his decision of building a coalition with the Centrão by stating:
I was obliged to form a coalition […] it is necessary for governability. If someone has a problem with a congressman from the Centrão, it was you who put him here. With only 150 legislators, I would not go anywhere. The word Centrão is pejorative… I belonged to one of the Centrão parties (Popular Party-PP) for about 20 years …in fact, I am a politician from Centrão.
Bolsonaro’s political subordination to the institutionalized game of multi-party presidentialism, forcing him to betray his voters by assembling a coalition with the Centrão parties, was a clear institutional pathway to survival for a politically vulnerable government.
Presidents in such circumstances, including populist ones, need pivotal parties with a non-ideological profile to govern. Such parties prevent extreme and radical institutional changes from being approved. They constrain coalition governments from negotiating extreme solutions and, consequently, avoid illiberal actions. They provide democratic stability. In other words, multipartyism can foster moderation and containment.
Notwithstanding the ability of Bolsonaro’s minority coalition (with the Centrão) to block opposition motions for impeachment, it was insufficient for the government’s legislative success in Congress. Figure 1 shows Bolsonaro’s legislative success rate, as measured by the percentage of executive bills approved by Congress. It stands as the lowest among all presidents since re-democratization. These numbers suggest that Congress worked as a true veto point to Bolsonaro, constraining his legislative initiatives. Some key defeats are exemplary: the constitutional amendment on the electoral ballot was not approved. Also, two of Bolsonaro’s key constitutional amendments proposals—one on regulation of online content—were also shelved by the Senate speaker, in an unprecedented move.
Another outlier is the proportion of provisional decrees published by Bolsonaro. As Figure 2 shows, Bolsonaro has the lowest number of provisional decrees converted into law (115), signifying the worst conversion rate of provisional measures into law among former presidents (Bolsonaro: 45%, Rousseff: 69%, Lula 2: 79%, Lula 1: 84%, FHC 2: 64%, and FHC 1: 88%). Provisional decrees have immediate force of law soon after they are issued. However, they need to be approved within 120 days by Congress in order to become a full law. This failure rate contrasts with the fact that Bolsonaro issued the highest absolute number of provisional decrees (254) of any government. These decrees represented the largest share of legislative activity, at 76.5%. The conclusion is obvious—Bolsonaro had extreme legislative difficulties in Congress.
As Figure 3 shows, Bolsonaro was by far the president with the highest number of presidential vetoes overridden by Congress (thirty compared to four under Temer, three under Rousseff and two under Lula) (Hunter and Power Reference Hunter and Power2023). In other words, Congress was ready to check and contradict Bolsonaro even when the president tried to veto legislative initiatives or to put forward his conservative and illiberal agenda.
Despite poor legislative success, Bolsonaro was perhaps unsurprisingly the president who published the highest number of executive orders (1462, as compared to FHC: 1255; Lula: 1230; Rousseff: 839). Unlike regular bills, executive orders are not immediately submitted to the direct scrutiny of the legislature; instead, they take immediate effect after publication. This strategy is typical of presidents aiming to bypass an adversarial legislature, but they tend to be limited to administrative issues. Their large number should not be interpreted as executive overreach: the analysis of their content shows that they were mainly due to the emergency conditions of the pandemic.
Bolsonaro’s capitulation to the Centrão materialized in a sort of hyper-delegation of budgetary powers because the power of the purse was partly transferred from the president to Congress. This transfer amounted to an unprecedented arrangement, the so-called orçamento secreto (secret budget). This explains both the Centrão’s and Bolsonaro’s electoral competitive performance in the 2022 elections.
Conclusions: The Frog Jumped Out of the Boiling Water
Most contributors to the investigations on democratic backsliding observe the likelihood of democratic survival from the perspective of the aggressor. As in Luo and Przeworski (Reference Luo and Przeworski2023) mentioned, “the boiling frog syndrome” metaphor is leveraged to explain why social and political actors fail to react in a coordinated and timely manner. Most cases of backsliding, however, are followed by the return to the status quo ex ante, or a condition not far from it (Treisman Reference Treisman2023; Brownlee and Miao Reference Brownlee and Miao2022; Ginsburg and Huq Reference Ginsburg and Huq2018; Gamboa Reference Gamboa2022). The outcome is contingent on a number of factors, including the influence of constitutional structures as well as the political dynamics that constrain the behavior of populist leaders. These factors, to return to the metaphor, helped the frog leaping out of the boiling water thus avoiding the degradation of Brazilian democracy.
Pundits were extremely pessimistic about the prospects for Brazil’s democracy and underestimated the factors that mitigated Bolsonaro’s attempts to erode this country’s democratic rule of law. We argue that the election of a populist autocrat, like Bolsonaro, might be a necessary, but definitively not a sufficient condition for democratic erosion. We agree with Levitsky and Way (Reference Levitsky and Way2023) that “it is relatively easy for an authoritarian-leaning politician to win power in a democracy (Bolsonaro in Brazil or Trump in the United States, for example), but consolidating an authoritarian regime is much harder.”
We argue that the institutional elements of Brazil’s political system, especially its consensual features—multipartyism, a strong legislature and judicial system, federalism, and a historical legacy of resilience—provide antidotes against populism’s illiberal and antidemocratic initiatives. We borrow Weyland’s (Reference Weyland, Kaltwasser, Taggart, Espejo and Ostiguy2017) definition of populism which understands it “as a political strategy through which a personalistic leader seeks or exercises government power based on direct, unmediated, uninstitutionalized support from large numbers of heterogeneous followers.”
Our view is consistent with the argument of protective belts and long-term equilibria, which sets in after short-term fluctuations disappear (Coppedge et al. Reference Coppedge, Edgell, Knutsen, Lidberg, Coppedge, Edgell, Knutsen and Lidberg2022). Brazil’s system of coalitional presidentialism is characterized by numerous veto points. Coordination among the various institutional actors is difficult, and coalitions become a necessity. However, Brazilian coalitions are not rooted in policy agreements; instead, they revolve around the distribution of tangible benefits and pork barrel politics. Parties and politicians with hegemonic preferences grapple with the demands of compromise and consensus, the proliferation of veto points, slow decision-making processes, and the imperative to provide compensatory incentives to secure cooperation.
Two counterfactual questions help clarify the argument. If Bolsonaro had been a more effective leader would the outcome have been different? Did these institutional factors matter considering Bolsonaro’s ineptitude, as exemplified by his mishandling of COVID? Our answer is yes. Institutions forced Bolsonaro to compromise, play the coalitional presidentialism game, and completely abandon his antisystem rhetoric and practice. We characterize Brazil as a case of quick recovery within the broader cases of near misses’ democracy (Ginsburg and Huq Reference Ginsburg and Huq2018) with two important caveats: 1) Brazil would never fall below a threshold of competitive authoritarianism, and 2) the threat posed by Bolsonaro was real but not entirely credible in the light of Brazil’s political institutions and political dynamics.
Bolsonaro was opposed by a vibrant and vigilant society; virtually all major newspapers and media conglomerates actively opposed his government, providing a countervailing power to his strong presence in social media. As he was consistently defeated both by Congress and the judiciary, Bolsonaro had no alternative but to make adjustments to his initial antipolitical discourse, which had been effective in electing him in 2018. His narrative neither provided him with the capacity to govern nor did it undermine political institutions. He was, therefore, obliged to activate a kind of “survival mode.”
He never completely abandoned though his confrontational rhetoric with core institutions. Bolsonaro realized that, as much as he attempted, he did not have the political conditions to succeed. He could not count on a majority support of voters for a radical agenda, and was elected because of a strong rejection of his opponent. Society, Congress, and control organizations have demonstrated their strength and ability to impose restrictions and successive defeats on the president.
Faced with the very strong rejection of how he managed the pandemic, including among his core voters, it became increasingly clear that he had lost electoral competitiveness. The strategy of Bolsonaro was, therefore, to walk on the razor’s edge. If he had moderated his speech and confrontational attitude too much, signaling that he had totally surrendered to coalition presidentialism, he would have reduced the chances of having his term in office shortened, but he would have run the risk of seeing his electoral base of support lose cohesion or even disintegrate. On the other hand, if he crossed the red light in his confrontation, he could become even more isolated, losing electoral competitiveness and, more importantly, jeopardizing the political viability of completing his term of office, and going to jail. Therefore, although Bolsonaro had to calibrate his populist antisystem narrative, he could not entirely give up his bellicose and authoritarian speech in order to survive politically.
That is the reason why Bolsonaro, like other populists in stable democracies, continued to confront democratic institutions despite the fact that his chances of succeeding were fairly slim. It became his modus operandi of survival no matter what. Playing the democratic game, sooner or later, would make him uncompetitive. To keep surviving, populists need to keep pushing … needing to go to the very limit in order to build, at least, an exit strategy.
Nevertheless, Bolsonaro ended up being convicted by the Electoral Court for abuse of power and misuse of the media at a meeting at the presidential palace in which he inveighed against the electronic voting system invoking false claims and directly attacked justices from the Electoral Court and the Supreme Court. In view of this conviction, the former president is barred from office for eight years. Although this punishment took place when Bolsonaro was no longer president, it is worth mentioning that it was the first time in Brazilian history that the Electoral Court imposed such a drastic decision on a former president. This Electoral Court’s decision can be interpreted as a clear reaction to the institutional clashes and various personal attacks that the Supreme Court and the Electoral Court ministers suffered from Bolsonaro himself, from his family members and from members of his administration.
Appendix