Brazilian political and media systems

Despite enabling the rise of two right-wing populists with substantial similarities, Brazil differs from the United States in significant ways in terms of its political and media systems. First, contrary to the US model, in which elections are decided by an electoral college (Bromwich, 2016), the Brazilian electoral system is based on popular vote and it occurs in two rounds, with the second being when the two most voted presidential candidates participate on a runoff (Toffoli, 2014). Voting in Brazil is mandatory for most of the electorate (other than those aged 16–18 and 70+, and illiterates), currently around 86% (Velasco et al., 2016). At 79.3%, voter turnout in the 2018 Brazilian national election (G1, 2018) was higher than the 60.2% for the 2016 US Presidential election (Wilson, 2017). Citizens also have the option to vote null or blank, which Brazilian scholars consider an expression of ‘electoral indifference’ (Silva et al., 2017). In 2018, these votes represented about 9.5% of all the ballots cast (TSE, 2018).

Second, whereas the American political system is dominated by two major political parties, Republican and Democratic, in Brazil there are 35 political parties, 28 of them with elected officials in Congress (Almeida et al., 2016). Third, Brazil has the ‘horário eleitoral gratuito’, an allocated free timeslot established by law, usually 25 minutes in length twice daily during elections, in which parties can present their political programs simultaneously on all TV and radio stations.

Fourth and finally: the Brazilian media system is economically dependent on the state through official advertising, and newspapers are elite-oriented and low circulation (Hallin & Papathanassopoulos, 2002). Contemporary journalism in Brazil has increasingly adopted norms and routines from US journalism, through a hybrid approach that selects which rules to adapt and which ones to ignore (Mourão, 2016). For instance, reporters often view the ideal of objectivity with cynicism, but it is frequently evoked in the newsroom to preserve autonomy, as a mechanism of defense against accusations of bias, and for journalists to distance themselves from activists (Albuquerque & Silva, 2009; Mourão, 2016; Waisbord, 2000).

Nonetheless, Brazilian journalists perceive more political, economic, and organizational influences on their work than US journalists (Hanitzsch & Mellado, 2011). Brazilian journalism also makes more use of personalization frames in their election coverage, focusing on personal characteristics of the candidates rather than on the parties’ stances and the political process (Carvalho & Cervi, 2018), which may contribute to weaken political parties in the country (Lima, 2009).

Framing presidential campaigns

This chapter adopts Barr’s (2009) definition of populism as ‘a mass movement led by an outsider or maverick seeking to gain or maintain power by using anti-establishment appeals and plebiscitarian linkages’ (p. 38). He argues that the concept of populism needs to be understood as a combination of three factors: appeals (an anti-establishment rhetoric claiming the elite’s failure in representing ordinary citizens and asserting the populist’s independence from the status quo); location (whether a political independent, a representative of a marginal party, or a person who radically reshapes their own party); and linkages (citizens provide passive support for a politician and confer legitimacy to their leader via public opinion polls and mass demonstrations).

Bolsonaro fits Barr’s (2009) definition of a populist because during the presidential campaign he (1) focused on corruption scandals and constantly attacked political elites, particularly the Workers’ Party which had been in power for the previous 15 years, (2) was a representative of a marginal party, although he has been in politics for almost 30 years, first as a city councilman and then as a congressman, and (3) emphasized a direct connection with voters via social media and during his rallies, constantly promoting endorsements from his supporters. But Bolsonaro’s far-right populism is a clear departure from other populist presidents in Brazil’s history, who had been traditionally associated with leftist ideas, such as Getúlio Vargas (1930–1945; 1951–1954), Juscelino Kubitschek (1956–1960) and João Goulart (1961–1964) (de Castro & Ronci, 1991).

Although populism as a political movement is nothing new, it is only more recently that scholars have begun to examine the role of media in amplifying populist messages and promoting populist actors – whether intentionally or not (Aalberg et al., 2017; Blassnig et al., 2019). Esser and his colleagues (2017) examined this through two lenses: one looks at populism by the media, that is, news organizations or journalists who deliberately propagate populist messages in their editorials or media coverage; the other looks at populism through the media, where news outlets serve as platforms for populists, amplifying the reach of their messages. This chapter focuses on this second perspective, examining how mainstream newspapers in Brazil reported on the candidacy of a populist, that is, how media framed Bolsonaro’s presidential bid during the 2018 national election.

This chapter adopts the definition of framing as structures that are socially shared and symbolically organize our social world, and media frames as patterns of selection, interpretation, and presentation of social discourse (Gitlin, 1980; Reese, 2001). When examining frame-building, Hänggli (2012) emphasized that there is a symbiotic relationship of mutual dependence between politicians and journalists because, on one hand, political actors tend to rely on traditional media channels to reach the public, and on the other hand, journalists rely on these political actors as sources, seeking their input for their news production. Araújo and Prior (2020) examined editorials published by both Brazilian and international newspapers during the 2018 national election, and found that Brazilian opinion pieces normalized Bolsonaro’s candidacy and promoted false equivalencies between Bolsonaro’s far-right authoritarianism and the Workers’ Party center-left government, often incorporating populist tropes such as the division of the society (us vs. them), moral degradation, and anti-establishment rhetoric in their editorial pieces. The analysis in this chapter presents the opportunity for testing the phenomenon of frame building by examining the extent to which Bolsonaro’s social media discourse, including his frames, issues and emotional appeals, have resonated with the media and have been incorporated in news coverage of the presidential campaign.

News coverage of elections in the United States tends to frame politics as a strategic game (Cappella & Jamieson, 1997), focusing on tactics of campaigns and on the strategies used by candidates to stay ahead in the polls instead of on public policy. In part this is because journalists usually look for new developments and changing scenarios that provide new storylines on a regular basis, whereas the position of a candidate on a public issue is more static (Hayes, 2013). Patterson (2000) argues that it is a direct result of commercialism: the game can be reported without taking sides in the political debate to avoid alienating either. The game also embodies conflict, which is considered newsworthy. The problem that citizens end up learning little about what the candidates stand for, and base their choice on electability rather than whether the candidate serves or supports their interests.

Interestingly, however, horse-race coverage is not a traditional feature of political news in Brazil, even though candidates often focus on matters of electability in their campaign discourse: the rhetoric of many contenders during electoral cycles in the country is historically based on the ‘voto útil’ (tactical ‘useful voting’), trying to convince citizens to vote on the candidate who would be more likely to defeat the Workers’ Party rather than on the candidate of their personal choice (Barth, 2006; Cruz, 2018).

Previous research shows that countries with multiparty political system and regulated media markets – both characteristics of Brazil’s political and media systems – tend to cover policy issues more often (Aalberg et al., 2012; Skorkjær Binderkrantz, & Green-Pedersen, 2009). However, in contrast, personalization is a historical feature of Brazilian news coverage, and news articles tend to focus on candidates’ image rather than their or their parties’ stances on issues or government proposals (Carvalho & Cervi, 2018). This practice may also therefore disproportionately provide a spotlight for populist politicians, as they tend to deliver the raw material that sparks outrage and controversy, attracting news audiences: ‘the often disruptive, and frequently politically incorrect, discourse of populist leaders fits the media narrative machine’s criteria for newsworthiness perfectly’ (Araújo & Prior, 2020: 5).

Digital campaigns in Brazil

Digital technologies, particularly social media, have been playing a fundamental role in presidential campaigns during the most recent electoral cycles around the world. During the 2010 Brazilian presidential election, Twitter was already embedded in the political campaign as a tool for candidates, journalists, and the public. Research at the time showed that the runner-up, non-incumbent Jose Serra, was the candidate who most effectively used Twitter (Steffen, 2011), but in terms of voter engagement, deliberation and mobilization on social media, it was voters for Dilma Rousseff, from the incumbent’s Workers’ Party, who were most active (Nunomura, 2013). In contrast, Facebook has been used primarily for broadcasting, with little interaction with supporters (Rossini et al., 2016).

Social media offers an opportunity for candidates from ‘partidos nanicos’ (tiny political parties) to achieve popularity online, sometimes even appearing on the trending topics in the country (Marques, Sampaio & Aggio, 2013). Whereas previous Brazilian studies focused on the extent to which social media was incorporated in candidates’ routines rather than examining its content, this chapter examines what was communicated by Bolsonaro and how journalists covered his social media discourse in the larger context of presidential campaigns. While digital campaigns in the United States have been receiving considerable attention by both news media and academia (Berger & Milkman, 2012), the social media political strategies and their outcomes in other areas of the globe such as Latin America remain underexplored.

Method

This study employed a content analysis to examine predominant frames, issues, and sources in news coverage of Jair Bolsonaro’s presidential campaign in Brazil. It also examined parallel frames used by Bolsonaro on Twitter to compare the candidate’s tweets with media coverage. The time frame used for sampling was August 16 to October 28, 2018, which is the official timeline for when acts of campaigns are allowed during presidential elections by Brazilian law (Toffoli, 2014).

Stories mentioning Bolsonaro and containing the keywords candidato, eleições or eleição in the headline and lead paragraph were retrieved from Factiva database. Two Brazilian newspapers were analyzed, both elite-oriented publications chosen for their circulation and national importance (Hanitzsch & Berganza, 2012; Kucinski, 1998), and articles were retrieved using a proportionate and systematic random sampling (Mourão, 2016), with a final corpus corresponding to 15% of the total of news articles: Folha de S. Paulo (N = 129) and O Globo (N = 63). During the mentioned time frame, there were 304 tweets from candidate Jair Bolsonaro. Using the same sampling method, 50% of all his tweets were retrieved for analysis.

Two coders analyzed the corpus of news articles and tweets, with Krippendorff’s alpha being used to assess reliability. Newspaper stories were coded for two binary metaframes: ‘strategic’ framing (campaign strategies, motivations of political actors, voter demographics and analysis of campaign conduct) versus issue-driven framing (analysis and evaluations of policy issues and themes of public interest), (α = 0.72); and descriptive (fact-based, event-centered and straightforward account, with journalist acting as an observer) versus interpretative (evaluation or explanation of a situation signaling a more subjective approach, with journalist acting as an analyst) (α = 0.70). They were also coded for the presence of five metaframes that are not mutually exclusive: ‘horse race’, (focusing on who is ahead and who is behind on the race, comparing fundraising amounts, or discussing polls about the candidates, α = 0.73), personalization (α = 0.75), sensationalism (α = 0.70), attacks made by Bolsonaro against opponents (α = 0.70) and Bolsonaro as the target of physical or verbal attacks (α = 0.96). Finally, the articles were coded for the presence of top five electoral issues according to national polls (Datafolha, 2018; Ibope-CNI, 2018): economy (α = 0.83), health care (α = 0.74), public safety (α = 0.85), corruption (α = 0.74) and education (α = 1.00).

The candidate’s tweets were coded for the presence/absence of the following dimensions, expanding on framing categories used on previous research (Parmelee & Bichard, 2011): campaign trail (candidate’s daily activities and events, α = 0.70), personal (α = 0.71), ideology (information on policies and issues, α = 0.84), attacks on the opposition (α = 0.79), attacks on the media (α = 0.82), call to action (calling for supportive action such as voting, liking, volunteering, donating) (α = 0.74), endorsement (testimonials from supporters or official endorsements from party, other candidates, or celebrities, α = 0.90), and sensationalism (α = 0.76). Coders also analyzed the issues mentioned above, as well as four overarching emotional appeals in Bolsonaro’s tweets: anger (α = 0.71), fear (α = 0.73), happiness (α = 0.72) and sadness (α = 0.70).

Results

The findings comprise two lines of inquiry. Firstly, it compares the predominant frames and issues in news coverage of Bolsonaro’s presidential bid and in his own tweets. Secondly, it examines correlations between the Twitter activity and the news coverage. The news articles were primarily descriptive, adopting a fact-based and event-centered approach, with the journalist functioning as an observer. Newspaper Folha de S. Paulo was significantly less likely to produce interpretative pieces [χ2 (1) = 22.25, p < .001]. However, there were significant exceptions that will be explored below.

Table 19.1 compares the frames in the two newspapers and Bolsonaro’s tweets. The most frequent frame adopted by the candidate in his tweets was attacking the opposition (46%), however, this was the least frequently used frame in the newspapers. Bolsonaro would often frame his tweets in ideological terms (44%), such as his opposition to Communism and praise for the free market economy, and he would frequently position the political left as enemies of freedom. The media overwhelmingly relied on the strategic frame as opposed to the issue-oriented frame, largely ignoring Bolsonaro’s ideological appeals, but as expected, the horse race frame was largely absent, particularly from Folha de S. Paulo.

The personalization frame appeared in almost half of the media coverage and about a quarter of Bolsonaro’s tweets, and news framing also positioned the candidate as a target of attacks in almost half of the stories. This is in no small part because Bolsonaro was stabbed during the campaign trail, and this incident played a central role in Brazilian coverage. In the following days, newspapers covered the investigations and discussed Bolsonaro as a physical target, comparing the incident with previous cases of violence against politicians in the country, including murder and assassination attempts between political rivals in the 1960s and 1990s (O Globo, 2018b).

Table 19.2 compares the most mentioned issues by the media and Bolsonaro. More than half of the analyzed stories did not address any policy issue during the electoral cycle and in many others, the issues were less prominent than strategic framing. Whereas news coverage focused on the economy, public safety, and corruption, Bolsonaro mostly tweeted about the latter. News coverage on issues included criticisms of as well as information on Bolsonaro’s policies. For example, on the economy, reports focused on his proposal of creating a tax on financial transactions and privatizing public companies, his lack of concrete plans to address economic recovery, and speculated on future economic instability in the country depending on the election results. Pearson’s correlations between issues and frames in media coverage showed that, when news media reported on corruption, news articles were more likely to adopt the attack frame (r = .204, p < .01). Likewise, news articles reporting on public safety, particularly those addressing gun policy, were more likely to adopt sensationalism (r = .183, p < .05).

As for emotional appeals, most of Bolsonaro’s themes of fear revolved either around public safety or comparing the leftist party with a totalitarian regime. Anger was a common rhetoric, and it was significantly correlated with both sensationalism (r(152) = .58, p < .001) and attacks (r (152) = .49, p < .001). The sensationalist rhetoric was evident in 32% of Bolsonaro’s tweets, with focus on messages around violence and crime in the country (Table 19.3).

To examine the extent to which Bolsonaro’s tweets influenced news coverage, both datasets were merged and compiled into a single file, with each row representing one day in the presidential campaign. Then, a series of linear regression analyses was conducted to examine the effects of the independent variables (candidate’s tweets) from each N day on the dependent variables (news coverage) from each N+1 day. Results shown on Table 19.4 indicate that when Bolsonaro addressed voting issues in his tweets, news coverage would be more likely to present an interpretative metaframe (β = .400, p < .01) and would be more likely to quote him in the stories (β = .454, p < .01). For instance, a news article in Folha de S. Paulo reproduced a tweet by Bolsonaro (@jairbolsonaro, 9/27/18) responding negatively to his vice-presidential running mate’s criticism of the Christmas bonus, which is guaranteed by longstanding labor law, but interpreted this as part of an internal crisis within the party, adding historical information on instances where the government discussed abolishing the benefit, and experts weighing in as to whether it was a constitutional right (Folha, 2018).

Interestingly, when Bolsonaro appealed to anger, media would be significantly more likely to adopt the interpretative frame (β = .515, p < .01), include the candidate as a source in the news articles (β = .480, p < .01), and mention voting issues (β = .659, p < .001). This significance, however, must be interpreted in light of the electoral cycle. Bolsonaro’s use of anger intensified towards the end of the presidential campaign: from less than 5% in the first three weeks of the campaign to 71% during the last week of the electoral cycle, a trend reflected in news reporting. However, a qualitative analysis a posteriori indicated that the interpretative frame was adopted in relation to the issues mentioned by the candidate rather than to analyze and interpret his emotional appeals.

Discussion

Brazilian media overwhelmingly prioritized strategy over policy, contradicting previous assumption that a country with a multiparty political system and a more regulated media market would present issue frames more often (Aalberg et al., 2012; Skorkjær Binderkrantz & Green-Pedersen, 2009). A possible explanation for this is two-fold. First, there was an ongoing discussion about the banned candidacy of former president Luis Inacio Lula da Silva, and the possible ramifications of that decision. Second, the novelty brought by Bolsonaro’s candidacy was heightened after his stabbing, which brought more focus to the unusual elements of the presidential race rather than the discussion of candidates’ stances on issues and public policies of national concern. Framing politics as a strategic game is detrimental to democracy, because it leads to political cynicism and erosion of trust in political institutions (Cappella & Jamieson, 1997), and in Brazil it could contribute to tactical voting (Barth, 2006), benefiting populist politicians, as well as blank and null votes signaling electoral indifference (Silva et al., 2017).

The interpretive meta-frame, a recurring feature of American news stories, was largely absent from the articles examined in this study, especially within newspaper Folha de S. Paulo, which prioritized descriptive and event-centered stories, nine out of ten times. Traditionally, US journalists adopt interpretive strategies that are shared within the profession and entail a collective interpretation of events, ultimately building authority to support journalistic empowerment (Zelizer, 2009). However, in Brazil, the interpretive frame was reserved for Bolsonaro’s issue-related messages, rather than the strategic meaning of events. When Bolsonaro’s social media discourse referred to policy issues, media departed from the traditional descriptive framework, and was significantly more likely to examine and evaluate the public policies mentioned by him.

Another interesting finding is that Brazilian media would often position Bolsonaro as a target of attacks from others while occasionally reproducing his attacks towards other candidates and political parties. This is a problem, as previous research has shown that when news stories report on an attack campaign, whether these stories are policy-focused or personality-focused, the audience is significantly more likely to perceive the news stories as more negative and more biased, as well as to perceive articles reporting attacks as less informative and less useful (Min, 2004). In addition, personal attacks in particular have a demobilizing effect on voters, with reduced intention to vote (Min, 2004), though this is less likely to apply to Brazil, where turnout tends to be high because voting is mandatory.

Personalization frames appeared in about half of the news articles, which reflects Bolsonaro’s social media messaging – although it only appeared in one in four tweets, its use was textbook. For instance, he would refer to himself as an ‘honest president, patriot, and with God in his heart’ (@jairbolsonaro, 08/16/2018), referencing his military service and religion (@jairbolsonaro, 08/28/2018). He was also very personal when broadcasting his recovery after being stabbed during the campaign trail. Previous research shows that news media favor this kind of framing (Carvalho & Cervi, 2018).

In contrast, when attacking his opponents, Bolsonaro focused on the Workers’ Party (PT), only occasionally mentioning its candidate, Fernando Haddad by name, and briefly making indirect remarks towards former president Luis Inácio Lula da Silva, one of the founders of PT. This strategy may be in part due to Lula’s popularity among the electorate. Indeed, before being banned from candidacy by the Electoral Supreme Court, due to accusations of corruption while in office, Lula was the candidate leading the polls by a wide margin (Gielow, 2018). Bolsonaro was also able to capitalize on the issue of corruption in light of the wave of protests calling for the impeachment of then-president Dilma Rousseff in the years of 2015 and 2016, and the fact that former president Luis Inácio Lula da Silva had been detained in April of 2018 and remained in prison for 580 days, accused of money laundering and receiving bribes (Londoño & Casado, 2019). His anti-corruption and anti-establishment rhetoric were a key populist appeal picked up by the newspapers, which is an example of populism through the media: ‘charismatic leaders, harsh rhetoric, and stirring issues hit all the right keys of newsworthiness’ (Esser et al., 2017, p. 5).

Whilst studies of US election reporting found that commercial news values were prioritized above journalistic standards of balance and accountability, which helped to normalize and legitimize Trump’s candidacy (Cushion & Thomas, 2018), this concern could also be applied to news coverage in Brazil, to some extent. Bolsonaro’s rhetoric on corruption had certainly resonated with the media, but newspapers mostly focused on the economy and criticized his policies. However, the overwhelming reliance on strategic frames and the reproduction of attacks uttered by candidates in news reports is indeed cause for concern, as it may contribute to the cynicism and apathy of voters during elections, ultimately choosing to withdraw from the democratic process.

It should be noted that these findings cannot be generalized to the entire electoral process nor to all candidates in the political spectrum. However, the study presents a framework that could be applied to other right-wing populist candidates on social media. Another limitation is that, despite the array of social media platforms, this study focused on Twitter. There is also a limitation concerning the media outlets analyzed, but it would be expected that these traditional news outlets would have an impact on news content transfers between media as well as reporting decisions from journalists at other news outlets beyond the ones chosen here (Harder, Sevenans & Aelst, 2017).