Denialism and Populism: Two Sides of a Coin in Jair Bolsonaro’s Brazil

This article analyses the impacts of the COVID-19 pandemic on Brazil’s populist radical right (PRR), as well as the responses of PRR actors to the pandemic, during the period from March 2020 to October 2021. Despite high death rates and declining popularity in the final months of that period, the Brazilian president consistently maintained a denialist narrative that incorporated key aspects of populist ideology. Based on the analysis of opinion surveys, documents, online messages and secondary sources, we argue that explaining this denialism requires understanding Brazil’s radical-right populism as more than an ideology: it is a social movement. The impacts of the pandemic on Bolsonaro’s PRR government and its responses can only be understood by simultaneously analysing the top-down actions of the leader and the bottom-up role of bolsonarismo – that is, the broad coalition of actors who actively support the radical-right project. The case of bolsonarismo suggests that literature on populism in general would profit from taking right-wing movements more seriously as co-producers of populist rhetoric and practices.

Second, we present Brazil's policy response to the pandemic. Third, we explore the framing strategy Bolsonaro and his supporters have employed to talk about the public health crisis. The fourth section looks at public support for Bolsonaro during this period. Finally, we explore the role of organized social actors, arguing that the president depends on a well-organized grassroots mobilization process that pre-dates his emergence as a populist radical-right leader and which he does not fully control.
The populist radical right in Brazil As noted in the introduction to this special issue (Rovira Kaltwasser and Taggart 2022), Cas Mudde's (2007) analysis identifies three ideas at the centre of European PRR ideology: nativism, authoritarianism and populism. Each of them is present in Brazil's PRR, although with some important differences from the European cases examined by Mudde. Nativismunderstood in Mudde's terms primarily as xenophobic nationalismhas not played an important role in Brazil, in the sense that issues such as closing the borders to immigrants or targeting minority ethnic groups have been largely absent from the public agenda. Since, according to Mudde, nativism is the minimal requirement for defining the populist radical right, some might argue that the Brazilian case should not be included as an example of PRR. However, Brazil's radical right has adopted ideas that are very similar to nativism and play a similar role in the ideology. Unlike the xenophobic characteristic of the European PRR, however, the threat to national identity comes not from outside, but rather from within. National identity is challenged by the perceived corruption of traditional values, which, in turn, is caused by the dissemination of left-wing ideas about social rights and, especially, LGBTQI+ rights and feminism. The functional equivalent of nativism in Brazil is that a traditionalist conception of the nation is under threat.
Brazil's populist radical right also adheres to the hierarchical, 'law and order' ideology that Mudde calls authoritarianism. The defence of traditional order and, especially, of a punitive approach to dealing with criminality are certainly major themes (Avritzer and Rennó 2021). So also are the exaltation of weapons and the imagery of violence, to the extent that during the 2018 presidential campaign followers signalled their support for Bolsonaro by making a pistol-like hand gesture.
But the Brazilian case includes two important adjustments to this authoritarian component. First, if the European PRR has distanced itself from the defence of dictatorship, Bolsonaro's identity is closely connected to nostalgia for Brazil's military regime . Second, as we will explore below, this authoritarian discourse is, ironically, often couched in terms of the need to protect democracy and individual rights. Unlike leaders who took advantage of the pandemic to increase repression, Bolsonaro and the radical right accused subnational governments that imposed COVID restrictions of attacking basic freedoms.
Finally, Bolsonaro's antagonistic worldview is quite similar to the populist ideology of PRR groups described by Mudde. In 2015-16, mobilizations for the impeachment of President Dilma Rousseff gave force to an anti-left narrative that was articulated in terms of 'good' and 'evil' (Dias et al. 2021;Samuels and Zucco 2018). Bolsonaro appeared on the stage soon after, framed as the person who would save Brazil from the left. Indeed, the president is frequently referred to by his supporters as o mito, or 'the legend'.
Rather than articulating populism as antagonism against a particular social group or sector of the elite, however, Bolsonaro's anti-establishment rhetoric is targeted at all politicians. Although prior to 2018 he had been in Congress for nearly three decades, he successfully played the role of an outsider who came to the presidency to clean up corruption and fight against the forces that threatened to destroy the traditional values upon which the nation was founded. In contrast to the US under Donald Trump and to cases of PRR in Europe, in Brazil, populist right-wing parties played no important mediating role between Bolsonaro and his supporters. As observers such as Juan Pablo Luna (2021) have noted, the role of party systems as legitimate political intermediaries is in crisis not only in Brazil, but throughout Latin America.
In Bolsonaro's anti-establishment rhetoric, his lack of strong party ties is portrayed as a positive sign of his distance from corrupt institutions. As the president himself declared in a speech delivered shortly after his election, 'Popular power is no longer dependent on intermediation. The new technologies have enabled a direct relationship between the elected and those represented.' 8 Indeed, communication between Bolsonaro and his supporters largely occurs online.
The literature on online communication strategies has shown that populist rhetoric fits well with the affordances of social media platforms (e.g. Engesser et al. 2017). The relevance of new digital technologies remains, however, a contentious issue in academic debates on populism. In a recent article, for example, Mudde (2021: 589) argues against overstating the importance of social media in explaining the rise of populism: 'without adoption by traditional media, the political effects of social media remain fairly limited'. The Brazilian case provides dramatic evidence against this formulation in three ways. First, an increasing proportion of the population relies on social media channels to inform themselves about politics. Brazilian public opinion survey data show that, in the last decade, social media platforms increasingly replace television as a source of information about politics (Stabile and von Bülow 2021). Second, Mudde's argument focuses on Facebook and on Twitter, failing to account for the role of messaging apps such as WhatsApp and Telegram. Nonetheless, in many countries PRR actors have used the particular affordances of these platforms intensively and strategically to further their goals. 9 Third, and most importantly given our purposes in this article, digital technologies not only allow followers to receive communication directly from populist leaders (Weyland 2017: 58), but also enable them to communicate their approval or disapproval and to launch their own campaigns.
Populist communication is thus a two-way street, rather than simply a top-down process of indoctrination. Notwithstanding the official presidential rhetoric, what has emerged in Brazil is not the absence of intermediation, but rather the creation of new forms of intermediation that are organized largely outside of the party system. Supporters of the president are thus not a 'largely disorganized mass' (Weyland 2017: 50). On the contrary: we argue that the populist project requires mass mobilization that mayas in the Brazilian casebe the result of social movement organization that precedes the rise of the populist leader. A similar argument has been proposed by Kenneth Roberts (2015), who argues that mass social protest often precedes and sets the stage for populist outsiders.
However, we disagree with Roberts's analysis both in respect of the definition of social movements and in respect of his understanding of populism. Roberts understands social movements as bottom-up 'autonomous forms of collective action' that may influence populism but that fundamentally contradict the typical 'top-down' mobilization of mass constituencies by populists (Roberts 2015: 682-683). The concept of populism, he argues, can only be applied in cases 'where socio-political mobilization is controlled from above and dependent on a dominant authority figure to weld together diverse popular interests and articulate a shared political project for "the people"' (Roberts 2015: 685).
Following Mario Diani (1992), we define social movements differently, as networks of actors that engage in collective action on the bases of both shared collective identities and shared orientations towards a conflict. Movements are thus characterized by the contentious ideas that bring their members together to question existing social, political and cultural conditions (or others' efforts to change them) and can be either progressive or conservative. They are not 'autonomous' in the sense of being separated from the political system. Furthermore, although scholars have often presumed that social movements engage primarily in 'noninstitutionalized politics', much recent work has demonstrated that movements often combine public, contentious forms of collective action (such as street protests) with behind-the-scenes negotiations with elected officials and party leaders (Abers and von Bülow 2011;Abers et al. 2014;Goldstone 2003;Rossi 2017). In some cases, movements engage in 'institutional activism' by occupying government positions in legislatures or the bureaucracy Banaszak 2010;Pettinicchio 2012). The collaborative networks connecting actors in civil society to others in the political system have been mostly studied in the context of leftist governments (populist or otherwise), but we argue that they can be a key aspect of the PRR as well.
In emphasizing the role played by social movements as active PRR actors, we build on Robert Jansen's (2011: 82) definition of populism as a political project that combines popular mobilization with populist rhetoric. We argue that this project can be promoted by incumbents, by opposition leaders, by social movements, or by alliances that bring together all or some of these actors. In Brazil, new right-wing actors and organizations began to emerge at least two decades before Bolsonaro's election (Rocha et al. 2021: esp. ch. 2). In 2015 and 2016, these organizations used populist rhetoric to organize mass protests in favour of President Rousseff's impeachment. The populist framing mechanisms of antagonismwhich divided the country between the 'good people' and its enemiesand reductionismthe focus on corruption as the source of all of the country's problemsenabled these actors to build unified diagnoses and targets. Those mechanisms also allowed them to present multiple motivating messages, all of which appealed to a wide audience (Dias et al. 2021). Bolsonaro's successful rise to power cannot be understood without considering his ability to act as the representative of this new right-wing field and of the rising antisystem sentiment. Likewise, his surprising response to the pandemic cannot be understood without considering the interactions with the social movement that emerged out of this earlier process of organization and that is constitutive of the PRR political project.
During the pandemic, denialism has interacted with the three (adapted) ideological components of Mudde's model in interesting ways. Denialist discourse emphasized a populist reading of the pandemic as a conspiracy fabricated by the opposition to cause economic collapse, with the express purpose of undermining Bolsonaro's government. The functional equivalent of nativism was expressed through the constant use of the Brazilian flag and the propagation of anti-Chinese sentiment. Most notably, the odd combination of authoritarianism and a discourse of freedom appeared in the frequent mobilizations in favour of Bolsonaro during the pandemic. Protests organized by government supporters called for the military to intervene in support of the president and against the legislature and the judiciary. 10 The justification was that, by supporting state and local governments that imposed lockdowns, those institutions undermined the efforts of a democratically elected president to defend economic freedom and individual rights.
Radical-right populism and denialism can thus be understood as two sides of the same coin. As argued by Rebecca , Bolsonaro's responses to the pandemic added a new cleavage to an already deeply divided society: denialists versus those who saw the pandemic as a serious threat. This new cleavage did not fit neatly along right-left divisions. It did not even unify those who had voted for Bolsonaro in 2018, and in the first months of the pandemic it created tensions with two ministers of health around issues such as social distancing and the use of chloroquine . In fact, centrist and right-wing governors and other political authorities, including many who had supported Bolsonaro's election, criticized his denialist approach , as did many mainstream conservative media outlets, one of which described COVID-19 as 'Bolsonaro's personal Chernobyl'. 11 Right-wing groups once united in support of Bolsonaro's campaign were now divided, with some even calling for the president's impeachment. 12 The response of the Bolsonaro government to the pandemic Since Bolsonaro took office in January of 2019, his government has promoted a state of permanent conflict with the legislature and the judiciary. When the pandemic took hold, the leader of the Chamber of Deputies, Rodrigo Maia, intensified this conflict by declaring that only pandemic-related initiatives would be put to a vote. This measure effectively blocked Bolsonaro's broader political agenda and put pressure on the government to face the public health crisis.
Unable to completely ignore the pandemic while at the same time minimizing its impacts, the federal government's response was erratic and at times contradictory. Bolsonaro's minister of the economy initially declared that the best way to reduce the economic impacts of the pandemic would be to move forward with austerity reforms (Oliveira 2020). But soon after, he launched a relief package that included the early payment of the thirteenth salary (usually paid at the end of the year), new rules allowing companies to reduce working hours and salaries with support from the federal government to pay part of the wages, and various measures to help small businesses and the airline industry. 13 The government also proposed an emergency relief cash transfer programme (see below). Another key move came in June 2020, when the national health research agency, FIOCRUZ, signed a deal with AstraZeneca to test its vaccine in Brazil. 14 Yet such actions were fragmented and uncoordinated. By July 2020, the federal government had invested only one third of the emergency funding approved by Congress to fight the pandemic (Lotta et al. 2020). In a federative country such as Brazil, national government coordination of subnational decision-making would be crucial. 15 By all accounts, however, it did not occur (Ortega and Orsini 2020). For example, the government created a 'COVID-19 Crisis Committee' without subnational representation. Emergency funding from the national health system failed to reach subnational institutions, while states began to compete in the struggle to acquire health equipment and supplies (Abrucio et al. 2020;Fernandez and Pinto 2020). There was no national policy promoting mass testing or contact tracing (Ferigato et al. 2020;Lucas et al. 2021). The government also lagged behind on policies to deal with the consequences of the pandemic on non-health-related activities such as school closures (Lucas et al. 2021).
Added to the lack of a coordinated pandemic strategy was the Bolsonaro administration's active opposition to efforts by subnational governments to implement restrictions on mobility. In March 2020, many Brazilian states and cities, some of which were governed by right-wing politicians, issued decrees closing schools and shops and prohibiting public events and gatherings (Agência Brasil 2020a). The federal government's first reaction was to issue a decree that gave the presidency the power to define 'essential' activities that could not be halted during the pandemic. It then proceeded to characterize religious activities, lotteries, civil construction, beauty parlours, gyms and industrial activities as essential (Máximo 2020). It also launched the 'Brazil Cannot Stop' campaign to disseminate the idea that the disease was not dangerous enough to justify shutdowns (Asano et al. 2020: 8). In mid-April, Bolsonaro fired his minister of health for openly supporting social isolation policies and World Health Organization recommendations more generally. In July, the president vetoed legislation requiring the use of masks (Asano et al. 2020: 12). Legislative and judicial authorities reversed many of these initiatives. For example, four days after the publication of the 'essential activities' decree a Supreme Court judge issued an injunction declaring it to be unconstitutional, and vetoes were overturned by Congress (Agência Brasil 2020b).
While denying the gravity of the crisis, the government invested in its own strategy for dealing with COVID-19: the use of chloroquine and other unproven medications. On 23 March 2020, the government agency responsible for regulating medications, ANVISA, published a protocol that allowed chloroquine and hydroxychloroquine to be distributed without a prescription through government programmes. 16 Bolsonaro's second minister of health, Nelson Teich, had been in office for less than a month when, in mid-May, he resigned in protest after being pressured to defend the medication's effectiveness. He was replaced by an army general, Eduardo Pazuello, who had no medical training. Just a few days later, on 22 May, a ministry protocol recommended the use of chloroquine in all COVID-19 cases, including minor ones (Asano et al. 2020: 10).
The legislature, the courts, governors, mayors, civil society organizations and individuals launched their own initiatives to fill the void. 17 The most important national programme to combat the economic effects of the pandemic was, no doubt, the Emergency Aid cash transfer programme. This programme was originally proposed by a broad-based national coalition of NGOs, trade unions, social movements and progressive politicians who called for an emergency income supplement for the poor. Under pressure, the government agreed to a payment of R$200 a month (approximately US$35) for a three-month period. However, after an intense lobbying effort by the above-mentioned coalition (Orofino et al. 2021), Congress approved three times that amount on 31 March. Law 13982/2020 provided cash transfers of R$600 a month for a three-month period to low-income individuals and informal workers. The payment was doubled for single mothers. It was a major win for those who defended a strong, 'Keynesian' response to the economic crisis.
Civil society, opposition political parties and subnational actors also appealed to the Supreme Court in their efforts to influence public policies. The court issued decisions requiring the government to adopt measures to protect indigenous groups, instruct federal employees to observe technical and scientific criteria in health-related decision-making, and allow states to purchase vaccines in the case of federal non-compliance to a national vaccination plan, among other things (Asano et al. 2020). Although Brazil's Constitution states that healthcare is the joint responsibility of federal, state and municipal authorities, such Supreme Court decisions fuelled pro-government arguments that the president was being usurped of his powers by the political establishment.
The lack of federal government coordination of the pandemic response, combined with the efforts to reduce the capacity of subnational governments to implement effective responses and with the promotion of unproven miracle cures, has left a trail of increasing inequality and inefficiency (Fernandez and Dantas 2020). The impacts of the pandemic were felt disproportionately across gender, race and class divides. One study found that the increase in non-violent deaths (a proxy for COVID mortality) in the state of São Paulo between 2019 and 2020 was 11.5% for white Brazilians and 25.1% for black Brazilians, a powerful indicator considering that the black population is younger on average and hence would be expected to have lower mortality rates (Marinho et al. 2021).
The economic effects of the pandemic were also distributed unequally. A national survey in May 2021 confirmed that food shortages were greater among black Brazilians, those living in the country's poorest region (the north-east), women and those with lower educational levels (Amâncio 2021). This is in part explained by variation in the ability to comply with stay-at-home measures. In November 2020, 27.1% of those with university degrees were working from home. 18 This percentage fell to 4.4% for those with a high-school degree or an incomplete university degree, and to less than 1% for all others. 19 Framing the pandemic: populist denialism and informational chaos The term 'informational chaos' summarizes how Bolsonaro and his supporters talked about the pandemic. Most of the literature on the circulation of information about the pandemic in Brazil has focused on the diffusion of disinformation (see, for example, Machado et al. 2020;Recuero et al. 2022), understood as the intentional act of producing and disseminating false information (Humprecht 2018: 3). We build on this literature by arguing that the informational chaos produced was closely aligned with the populist project.
A mix of denialism, disinformation and antagonism sustained the variety of interpretations of the nature of the crisis offered by President Bolsonaro and his supporters: that COVID-19's impacts on health were not that serious; that only older people and those with co-morbidities were at risk, and thus that those who died would have done so anyway; that a major recession in the Brazilian economy could cost more lives and suffering than the pandemic; that simple cures were being suppressed by nefarious forces; that the virus itself had been created by the Chinese as part of a cunning scheme to gain world dominance. The fact that some of these ideassuch as the massive global conspiracy theoryseemed to contradict others such as that the disease was not really that deadlywas not an issue. More importantly, for bolsonaristas, the denialist interpretation of COVID-19 could be easily adapted and readapted to fit the populist framework of antagonism: the disease was a hoax invented to harm the government; or it could be easily cured if establishment institutions would allow for the use of alternative medications; or it was real, but part of a global communist plot. Taken together, the circulation of these frames kept Bolsonaro's populist radical-right constituency in a state of constant mobilization.
Data manipulation contributed to this chaotic informational context, as a study of the Ministry of Health's communication policies on Instagram shows ). The authors identified three data manipulation strategies: fragmentation (presenting data in parts, to make it harder to visualize the number of deaths), word play (substituting the word 'death' for 'life' and emphasizing the number of people that had 'recovered' from the disease) and, finally, simply suppressing information .
Beyond official governmental channels, the online radical right-wing networks that had grown exponentially since the beginning of the 2000s became sounding boards for informational chaos during the pandemic. 20 Based on an analysis of 802 messages sent through WhatsApp groups, Felipe Bonow Soares et al. (2021) found that after two televised speeches by the president (on 24 March and 31 March 2020), in which he criticized social distancing measures and urged Brazilians to go back to their 'normal' lives, there was a 'huge spike' in online disinformation messages. According to the authors, most disinformation (42%) was connected to how social distancing measures would hurt the economy. Other topics included the governors and mayors and their 'hidden interests' against Bolsonaro (36%); how China had engineered the virus (27%); how the media (21%), the Congress (13%), the leftists (16%), the Brazilian Supreme Court (9%), the soon to be fired Health Minister, Mandetta (6%) and others were all conspiring to defeat Bolsonaro. (Soares et al. 2021: 5) The ability of the president and his supporters to disseminate informational chaos through online networks, alongside the general uncertainty surrounding the spread of the virus and its consequences, produced a cacophony of voices. This made it more difficult for governors and mayors to institute social isolation policies. For example, a few weeks before the city of Manaus (in the Amazon region) was struck by its second, mammoth wave of COVID-19 in January 2021, the state governor cancelled a lockdown decree. In one tweet, on 27 December 2020, a bolsonarista federal deputy commemorated the decision: 'The pressure of the people also worked in Manaus. The Governor of Amazonas @wilsonlimaAM went back on his lockdown decree. Congratulations, Amazonense people, you made your power felt!' 21 The effect of pandemic politics on public opinion Bolsonaro's popularity rose and fell in waves that moved roughly in the inverse direction of the waves of COVID-19 contagion, at least until March 2021. As can be seen in Figure 1, support for the president dropped with the onset of the pandemic in March 2020, even before the number of deaths began to rise, but, as Cesar Zucco and Daniela Campello (2021) note, after the economic effects of shutdown policies could already be felt. In the following months, however, approval ratings rose to the highest levels since Bolsonaro took office.
Key to explaining the rise of Bolsonaro's popularity in mid-2020 was the implementation of the 'Emergency Aid' programme mentioned above. 22 The policy, approved in late March 2020, heated up the economy and actually reduced poverty in comparison to the pre-pandemic period. By June, almost half of the country's population lived in homes where at least one person received payments (IBGE 2020). One study found that, without the programme, average income loss to workers due to the pandemic would have equalled 18% in July 2020. Instead, average income increased by 24% compared to pre-pandemic levels (Gonzalez and Barreira 2020).
Although the president had originally opposed the policy, he successfully proceeded to take credit for it and eventually extended it for three more months, and then again for another four months (at half the original value). The effect on Bolsonaro's popularity seems to have been an indirect one. Although it increased soon after the payments kicked in, an October 2020 survey found that low-income respondents who received the payments approved of the government at the same level as those who had not received them. For Fabiano Santos and Tiago Ventura (2020), the injection of so much money into the economy and the sense that the government was responding to the crisis may have produced broad-based optimism that went far beyond the direct beneficiariesa feeling that the situation was under control.
Throughout 2020, the president was partially successful at distancing himself from the impacts of the pandemic. One December poll found that 52% of respondents agreed that he had no responsibility for COVID-related deaths, while another 38% argued that he was one of the culprits, but not the main one (Amâncio 2020). But, in the beginning of 2021, disapproval of Bolsonaro's personal role in managing the crisis began to increase. By then, the president was taking more of the blame than any other actor in the political system: 42% of respondents considered him to be the main guilty party for the crisis, followed by 20% who blamed governors and 17% mayors (Gielow 2021a). A survey conducted by the Institute of Democracy and Democratization of Communications (IDDC) in April 2021 found that 67.7% of respondents agreed that 'the president gave little importance to the impact of the new coronavirus, undermining the fight against the pandemic in Brazil'. 23 The same poll showed that less than 30% of respondents evaluated the federal government's role in the vaccination programme favourably.
With negative public opinion on the rise, the opposition in Congress was able to launch a Senate inquiry to investigate the government's handling of the pandemic. Meeting from May to October 2021, the inquiry broadly publicized accusations of corruption and inefficiency in the acquisition and distribution of vaccines. The mounting evidence against the federal government may help explain increasingly negative public opinion results, despite the diminishing deaths and the return of the emergency aid programmeat lower ratesbetween April and October. Apparently, these improving conditions were no longer sufficient to buttress popularity. In mid-September, one poll found that only 22% of respondents thought that the government was 'good' or 'very good', the worst result since the beginning of Bolsonaro's presidency (Gielow 2021b).
While the politics of informational chaos was not effective in guaranteeing Bolsonaro's popularity throughout the whole period, denialist ideas have been accepted by a significant minority of voters, Bolsonaro's core constituency. The 2021 IDDC survey mentioned above found that 40% of those who voted for Bolsonaro in the second round of the 2018 election agreed that chloroquine was effective against COVID-19 compared with 18% of those who voted for the opposition candidate, Fernando Haddad. Other studies confirm that perceptions vary according to political positions. Based on data from a national probabilistic online panel of 2,400 respondents, completed on 3 May 2020, Ernesto Calvo and Tiago Ventura (2021: 8) show that, at least among online users, the perception of risk associated with the pandemic was correlated with electoral identification. Thus, 23% of respondents who supported the opposition candidate in the second round of the 2018 election considered it very likely that they would lose their jobs or become infected with COVID-19, while only 12% of Bolsonaro supporters had the same perception (see also Gramacho et al. 2021 for similar conclusions).
Fighting such disinformation has proven very difficult, in part because, as data from the 2018 and 2019 IDDC surveys show, the president's supporters believe that most false news originates from the mainstream media (and especially from the dominant television network, Globo), and not from politicians or social media (Stabile and von Bülow 2021). In this environment, information diffusion on social media tends to create 'echo chambers' with little access to fact-checking initiatives (Soares et al. 2021).
The politics of bolsonarismo: a PRR movement Understanding how the pandemic has affected the Bolsonaro government requires an analysis of not only how public policies and framing strategies were received by the general population, but also how organized social groups actively participated in the creation of mobilization campaigns that supported these policies and narratives. As mentioned, this grassroots mobilization builds on right-wing organizing that played a key role in Brazilian politics years before Bolsonaro's election (Dias et al. 2021;Rocha et al. 2021). These actors unified around Bolsonaro's candidacy, breeding the phenomenon of bolsonarismo. Understood as a social movement in support of the president, bolsonarismo is part of a process of right-wing mobilization that is larger than the leader himself, and precedes his rise to power.
During the pandemic, one of the most notable organized groups to defend pandemic denialism included members of the medical establishment. In August 2020, a group of physicians participated in an event with Bolsonaro in the Presidential Palace, during which they delivered a letter that argued in favour of the so-called 'early treatment' of COVID-19, with the 'off-label' use of chloroquine, ivermectin and other medications. 24 The doctors belonged to the 'Brazil Beating the COVID-19 [pandemic] Movement', created in mid-April. 25 In March 2021, the movement's Instagram page had approximately 40,000 followers. The organization posted publications not only in support of the use of the medications mentioned above, but also against lockdown policies. 26 Another group, 'Doctors for Life Covid-19', published three statements between May 2020 and February 2021, signed by thousands of physicians, also defending 'early treatment' and demanding that the Supreme Court re-establish the full competence of the federal government for pandemic policies. 27 Its website offered visitors a list of physicians who agreed with the early treatment protocol, most of whom advertised 'telemedicine' services. 28 The efforts of these medical professionals created a veneer of expert authority (van Leeuwen 2007: 49) to the disinformation campaign and helped produce informational chaos around the pandemic. Indeed, rather than couching denialism in an anti-scientific discourse, they sought to legitimate their claims on the basis of what looked like scientific research. They relied on the support not only of the president, but also of medical associations such as the Federal Council of Medicine. They turned the table on critics of the government (including other medical organizations, such as the Association of Infectious Disease Specialists) by arguing that those who positioned themselves against 'early treatment' protocols were politicizing the pandemic and were thus responsible for the high number of deaths, which could have been prevented by simply using affordable existing medications. They cited numerous studies that supposedly confirmed the efficacy of the treatment, but that were based on unpublished and/or on preliminary or partial research results. 29 After investigating these activities, the Senate inquiry mentioned above included the names of 13 physicians and one biologist in the list of 78 individuals recommended for prosecution (Senado Federal 2021: 1112-1123. The content produced by these groups was widely distributed by other bolsonarista actors through online channels. One study, based on a bolsonarista Telegram group, analysed 867 messages published between 19 March and 24 September, in which 429 unique users mentioned treatment with chloroquine (Nascimento et al. 2020). It found that at least 20% of the messages mentioned some kind of scientific source. As the authors explain, many of these messages offered testimonial evidence from physicians who identified their own use of chloroquine (and other drugs) with success in the fight against COVID-19 (Nascimento et al. 2020). They thus presented their vision of how to fight the pandemic not as a challenge to science, but as a matter of patriotism and honesty. The depiction of courageous doctors who fought the politicized World Health Organization, the powerful drug industry and corrupt politicians, thus connected denialist discourse to nationalist ideas and to a populist ideology of struggle against dangerous elites.
Disinformation campaigns were coupled with traditional social movement routines, such as street protests. Pro-government marches and protests occurred regularly throughout the country during the early months of the pandemic. During much of 2020 and parts of 2021, weekly demonstrations were held in front of Congress in support of the president. Protests were marked by a combination of anti-lockdown and authoritarian symbols. Some held signs saying, 'We don't want the vaccine, we want chloroquine' (Longo 2020), and 'The real virus is corruption'. Others chanted, 'Brazil will never be red', associating lockdowns with communism. Calls for the military to stop the courts, the legislatures and subnational authorities from blocking Bolsonaro's agenda were also common at these protests. Authoritarian references to military intervention and populist identification of the left as the enemy thus joined up with denialism on the streets.
As Bolsonaro's popularity plummeted in the first months of 2021, his supporters became even more active. They also became more extreme. In May 2021, a new campaign emerged on social media and in street protests throughout the country with the catch-phrase 'I authorize'. The words refer to authorizing the president to carry out a military coup to force subnational governments to end shutdown policies and to stop the Supreme Court and Congress from vetoing Bolsonaro's policies. This mobilization reached its height in the commemoration of the country's Independence Day, on 7 September 2021, when bolsonarista groups organized massive protests across the country. The rhetoric that dominated those mobilizations demonstrates how at least some populist radical-right actors react when they perceive that their political project is under threat. That reaction points towards the strengthening of extreme right-wing sectors and ideas, a trend that originates both in bottom-up mobilization and from the leader's actions.

Conclusion
Over the course of 20 monthsfrom March 2020 to October 2021the effects of the COVID-19 pandemic on Bolsonaro's popularity have proven volatile. After an initial loss in approval in the early months, Brazil's president was able to regain and even increase his popularity in the second half of 2020, while avoiding personal blame for the rising death rate. In 2021, however, the situation changed dramatically with the arrival of a new and deadlier virus wave which coincided with a lethargic vaccination process and the launching of an investigation by legislators into the government's handling of the pandemic. Yet, despite plummeting approval ratings, the president's discourse remained consistently denialist, disinformative and antagonistic.
In this article, we have proposed that, to understand Bolsonaro's pandemic responses, we must consider the role of grassroots right-wing actors whose organizing efforts date from well before Bolsonaro's election. Their use of populist frames enabled the creation of a broad coalition and set the stage for Bolsonaro's ascent to power. The case of Bolsonaro suggests that the literature on populism in general, and on the radical right in particular, would benefit from taking the role of right-wing social movements more seriously as co-producers of populist rhetoric and actions. Social movement mobilization should not be understood merely as an important precedent that makes way for populist leaders, but as constitutive of populism as a political project.
It is too soon to evaluate the full implications of the pandemic either for Brazil's democracy or for its current PRR government. Given the extreme volatility of events, the analysis presented here can thus only be preliminary. What we can be confident of, however, is that Bolsonaro's responses to the pandemic made him more prone to criticism not only from the opposition, but also from members of his own government and from former allies in civil society. Never before had an elected president of Brazil been accused of genocide.
At the same time, both the president and bolsonaristas radicalized their rhetoric by openly calling for authoritarian solutions. In this way, they blurred the boundaries between what Mudde (2007) called the radical right and the extreme right. Aggressive disputes between the president's supporters and detractors gained new levels of ferocity during the pandemic, as both groups accused each other of connivance with the deadly effects of COVID-19. The pandemic has thus pushed an already polarized Brazilian society towards even greater disunion.