The second round of Brazil’s presidential elections, taking place Sunday, could be close according to polls showing President Dilma Rousseff of the Workers Party (PT) favored by 52 percent over challenger Aécio Neves of the Social Democratic Party (PSDB) with 48 percent. As we noted with Mexico’s 2012 presidential elections, media coverage can have a strong impact on election turnout and voter preferences, and there is compelling research [doc] that Mexico’s TV duopoly was decisive in determining the outcome of Mexico’s 2006 election. While Brazil’s media is mostly opposed to the PT, it hasn’t been able to swing recent presidential elections – but it appears it’s not for a lack of trying.

Manchetômetro (literal translation: “headline meter”), an independent project affiliated with the Instituto de Estudos Sociais e Políticos at the Universidade do Estado do Rio de Janeiro, has analyzed major media coverage of the main candidates and parties in Brazil. Focusing on TV coverage on Brazil’s largest audience TV news program “Jornal Nacional,” and front-page coverage in Brazil’s three largest newspapers (Folha de S. Paulo, O Globo and O Estado de São Paulo), Manchetômetro has documented a pattern of lopsided coverage that has disproportionately put Dilma Rousseff and the PT in a negative light, while opposition candidate Aécio Neves of the PSDB (and first round challenger Marina Silva) have received much more positive treatment and much less negative coverage, proportionately. Manchetômetro’s analysis also reveals that this bias against the PT is not new; coverage favored Fernando Henrique Cardoso (of the PSDB) over Rousseff’s predecessor, Lula da Silva during the 1998 election, and there was also more favorable treatment of the economic situation under Cardoso than the current economic downturn under Dilma, though the former was significantly worse than the latter.

TV Coverage: “Jornal Nacional”

Globo’s decades-old “Jornal Nacional” is considered to be a leading news program on Brazilian television; with around 18 million viewers (in 2013) it has a total audience of about twice the combined viewers of the four leading U.S. Sunday morning news talk shows (“This Week,” “Face the Nation,” “Meet the Press” and “Fox News Sunday”). The program’s influence may not be surprising considering Globo’s dominant position in Brazil’s media landscape. In a 2009 book, Giancarlo Summa (who did communications for Lula’s 2006 campaign) noted that Globo controlled 61.5 percent of UHF television stations, 31.8 percent of VHF TV, 40.7 percent of newspaper distribution, 30.1 percent of AM radio, 28 percent of FM radio and - through an association with Rupert Murdoch - 77 percent of the cable TV market.

Manchetômetro found a strong bias against Dilma Rousseff in “Jornal Nacional” coverage. During the campaign period (beginning July 6, and up to October 22), 21 percent of the coverage of Rousseff was negative, the most of any candidate. The percentage of positive coverage that Rousseff received was just 1 percent, much less than any other candidate. (Rousseff also had 78 percent “neutral” treatment.)

By comparison, coverage of Aécio Neves was 5 percent favorable, 4 percent negative and 91 percent neutral. Marina Silva had 6 percent positive coverage and no negative treatment at all. Her predecessor who had been her running mate, the late Eduardo Campos, received 18 percent favorable coverage and 15 percent negative treatment. Each of the opposition candidates received more than twice as much positive treatment on the show as Rousseff.

Manchetômetro also found disproportionately negative coverage of the PT during the campaign period, with 28 percent negative treatment on “Jornal Nacional” and no favorable coverage. While the PSDB also received more negative coverage than positive, it did receive some favorable coverage (1 percent) and much less negative treatment proportionately (5 percent) than the PT. The PT received far more negative coverage on the program than any of the other parties examined (the Democratic Movement Party, PMDB; the Socialist Party, PSB; and the PSDB). Manchetômetro found a similar pattern in 2014 coverage prior to the campaign period, but even worse, with just 0.4 percent positive coverage and 85 percent negative coverage of the PT -- more than twice as much negative total air time than for any of the other parties.

Newspapers: Front Page Coverage

Analysis by Manchetômetro shows that media bias against the PT and Dilma Rousseff is not restricted to TV. The media monitoring project examined front page coverage in the three leading newspapers Folha de S. Paulo, O Globo and O Estado de São Paulo, and explained its decision to focus on A1 treatment this way (our translation):

Newspaper covers have much greater communicative power than do mere fragments of news publications.

The headlines and cover photos are the most communicative elements in the publication, seen either by subscribers and their families, the people who buy newspapers at newsstands or even by the people who pass the newsstands every day, where the newspaper covers are exposed to public scrutiny.

The headlines and the cover pages are considered most relevant by the editors of the newspaper because they best summarize the contents of the entire publication, and supposedly attract more readers.

Manchetômetro coordinator João Feres Júnior also explained in an email: “A previous study done by us on the 2010 presidential election shows that the proportions of positive, negative and neutral valences on the cover are similar to the ones on the newspaper taken as a whole.”

Manchetômetro found that during the campaign period in 2014, the total number of positive, negative, and neutral headlines for each candidate, respectively, broke down like this (as of October 20):

  • Coverage for Neves was 14 percent positive (38 headlines), 14 percent negative (37 headlines) and 73 percent neutral.

  • For Rousseff, it was 1 percent positive (just five headlines), 43 percent negative (188 headlines), and 56 percent neutral.

  • For Campos (the PSB candidate who died in mid-August), it was 14 percent positive (14 headlines), 16 percent negative (17 headlines), and 71 percent neutral.

  • For Marina Silva, coverage was 7 percent positive (24 headlines), 19 percent negative (59 headlines), and 74 percent neutral.

Proportionately, each of the opposition candidates had a minimum of seven times as much positive coverage as Rousseff, and usually 14 times as much. Rousseff meanwhile saw more than twice as much negative treatment as any of the opposition candidates both in terms of a percent of total coverage and by number of headlines.

Front page treatment of political parties during the campaign period showed a similarly disproportionately negative tone toward the PT:

  • PT: 2 percent positive (four headlines), 69 percent negative (121 headlines), and 30 percent neutral.

  • PSDB: 11 percent positive (seven headlines), 40 percent negative (26 headlines), and 48 percent neutral.

  • PMDB: 2 percent positive (one headline), 47 percent negative (26 headlines) and 51 percent neutral.

  • PSB:  0 percent positive (zero headlines), 28 negative (19 headlines) and 72 percent neutral.

Among all the leading parties, the PT had the highest percentage of negative treatment by far.

A History of Bias in the Brazilian Media: Cardoso vs. Lula and Dilma

According to Manchetômetro, disproportionately negative media treatment of the PT and PT presidential candidates extends to ex-president Lula da Silva. Looking at ongoing coverage, the project has found that former president Fernando Henrique Cardoso of the PSDB has received proportionately more favorable coverage than Lula. As of October 22, coverage for Cardoso was 38 percent positive (10 mentions), 23 percent negative (six mentions) and 38 percent neutral (10 mentions); while treatment of Lula was 4 percent positive (6 mentions), 38 percent negative (53 mentions) and 58 percent neutral (81 mentions). This is consistent with media coverage of the 1998 elections, in which Manchetômetro noted that 23 percent of Cardoso’s coverage in the three leading newspapers was favorable and 12 percent negative, versus 3 percent favorable and 36 percent negative for Lula.

It is also interesting to see how Brazil’s major media has covered the economic issues surrounding presidential elections. As we have noted in a recent paper, Brazil’s economy has slowed down over the last three years, but since it took power in 2003, the Workers Party administrations have made significant gains in economic growth and social progress. From 2003 to 2014, GDP per capita (adjusted for inflation) grew at a rate of 2.5 percent, compared to 0.8 percent annual per capita growth during the Cardoso government (1995-2002). Nevertheless, even though many key economic indicators in 1998 were far less favorable than in 2014, there has been much more negative media coverage of the economy during the 2014 campaign period than in 1998.

Manchetômetro looked at coverage of the economy in O Estado de S. Paulo in both the 1998 and 2014 campaign periods and found that in 1998 there were 50 positive news stories about the economy, 76 negative stories, and 17 neutral ones. Since the campaign period began this year, there have been 9 positive stories on the economy, 127 negative stories, and 23 neutral stories. This means that in 1998 about 52 percent of cover stories on the economy were negative and 35 percent were positive, while this year 80 percent of cover stories on the economy have been negative and only about 6 percent positive. Manchetômetro finds that despite the fact that most economic indicators are better this year than they were in 1998, O Estado de S. Paulo coverage makes the Brazilian economy appear to be in worse shape than it was in 1998, when a significant part of the presidential campaign took place amidst a financial crisis.

Professor of Economics at the Federal University of Rio de Janeiro Carlos Pinkusfeld comments in a Manchetômetro post:

The old adage says that truth is the first casualty of war. In “electoral battle,” discussion does not escape this maxim. Not so much by presenting misleading data, but by deliberately confusing interpretation and doomsday predictions.