The 2014 Brazilian Presidential Election: Can We Learn Anything from Prior Elections and Polling? / by Mark Langevin

aaaaaeleicoes-2014 The 2014 Brazilian Presidential Election is coming to a dynamic and still uncertain close. Incumbent President Dilma Rousseff of the Workers Party (PT) scored the highest approval rating for a first year president (2010) since the redemocratization of Brazil in 1985, with 59% rating her government as excellent or good, according to Datafolha. However, President Dilma’s approval rating plummeted before stabilizing at around 41% by the end of 2013 after large public mobilizations against government associated corruption and the poor state of public services spread throughout Brazil in June of 2013. With the economy slowing down and inflation heating up, President Dilma has never recovered and falls far short of her very high approval ratings in the first two years of her term.

The 2014 presidential election became even more dynamic with the sudden death of Eduardo Campos, former Minister of Science and Technology under the Lula government and former Governor of Pernambuco, one of the fastest growing state economies in all of Brazil. Running as the presidential candidate of the Brazilian Socialist Party (PSB) with Marina Silva of the Sustainability Network (the “Rede”) as his vice-presidential running mate, Campos was in third place among voter intentions in mid-August of 2014 when his plane crashed, killing him and the other passengers and crew. Marina Silva who finished with an impressive third place in the first round of the 2010 presidential election, as the head of the Green Party’s ticket, assumed the PSB ticket and rocketed into a dead heat with President Dilma in voter intention polls before her momentum stalled in mid-September of 2014. In a surprise of sorts, Aécio Neves (Brazilian Democratic Socialist Party-PSDB) who trailed behind President Dilma in most of the presidential election intention polls prior to Campos death and Marina’s quick emergence as a contender, fell behind Marina in September only to surpass her in the first round voting on October 5, 2014 with 33.5% of the overall vote, nearly ten percent more than Datafolha’s last public opinion poll measured his voter affinity just days before the balloting(see table below).

Neves surprising vote total gave his candidacy the early momentum in the second round runoff and early polls showed that his campaign had equaled or even surpassed the number of voters intending on casting their ballots for President Dilma. Indeed Datafolha released a poll on October 10, 2014 that reported Neves with 46% of the vote to President Dilma’s 44%; certainly a tie within the margin of error, but a demonstration that Neves had the momentum and could surpass Dilma’s electoral support. However, the most recent Datafolha poll, published on October 22, 2014 shows that while still within the margin of error, President Dilma of the PT appears to have stopped Neves’ momentum and may be leading intentions by a very slight margin.

Who will win on October 26, 2014? Nobody knows for sure, but prior presidential elections and polling may provide some insight into the possibilities as the two candidates prepare to finish their campaigns and count the vote.

Since 1994 there has been a very uneven pattern with respect to the last Datafolha public opinion poll prior to the first or second round presidential elections in Brazil. In 1994 the PSDB’s Fernando Henrique Cardoso increased his support by some 5% per week between the end of July and the first round vote in early October of 1994; winning in the first round with 54.3% of the valid votes, a full 6.3% difference than Datafolha’s last poll before that election. In 1998 he repeated the performance by winning in the first round, again against Lula of the PT, by 53.1%-a difference of 4.1% from the last Datafolha poll before the election. In 1994 and 1998 Datafolha’s measures of voters’ intent to support Fernando Henrique Cardoso fell considerably short, demonstrating this candidate’s tendency to over achieve in these critical elections

The 2002 and 2006 presidential elections witnessed successive victories by the Workers Party’s perennial candidate, Lula, but only in the second round voting. Lula won convincingly in the second rounds of these elections with 61.3% and 60.83% of the valid ballots respectively. However, in both cases Lula underperformed in relation to Datafolha’s latest polls conducted before these two second round electoral victories. In 2002 Lula’s polled 2.7% more in the Datafolha than his vote totals. Similarly, in 2006 Lula beat Geraldo Alckmin of the PSDB by over 20% of the vote, but underperformed by 0.17% compared to the latest Datafolha measures before this second round election. While Lula easily won the 2002 election and his reelection in 2006, his votes fell short of his polling, at least as measured by Datafolha, although the 2006 result was slight and probably not significant.

Changing directions, in 2010 the Workers Party’s presidential candidate, Dilma Rousseff, faced off against the tried and true PSDB candidacy of José Serra, who lost to Lula in 2002. Eight years later Serra also fell short by some 12% of the vote wherein Dilma garnered 56.05% of the overall second round vote, or 1.05% more than the latest Datafolha’s poll report of her popularity among likely voters. In the first round of the 2014 presidential contest incumbent Dilma also won over the votes of some 1.59% more than the Datafolha published poll of October 3, 2014 indicated. Can Dilma count on overachieving again in the second round, as she did in 2010 or earlier in the first round of 2014?

The Dilma campaign’s capacity to overachieve in the final days and the sum of her votes may depend on Aécio Neves’ momentum and where it lies down the final stretch. Datafolha’s recent published polls show a technical tie, but one punctuated by an early Neves lead only to be overtaken by Dilma during the last week. Yet, the margin of 4 points, as indicated in Datafolha’s October 22, 2014 poll, indicates that while Neves’ early momentum has flattened, he is still capable of winning. The question is whether Neves’ surprising overachievement in the first round of the 2014 election, with a margin of 9.55% over the latest Datafolha measures (October 3, 2014), signals a trend toward overachievement, reflects poll error, or whether the first round was such a dynamic process, given Campos’ death and Marina Silva’s meteoric rise and subsequent fall, that it cannot shed much light on the probable outcome of the second round?

The number of voters refusing to cast a valid vote may decide this election (this includes abstentions, null and blank ballots), if they surpass the average of some 25% taken together. However, given the close polling it would seem that voter interest may be peaked by this electoral duel-unless a sizable contingent of Marina Silva’s first round voters decide to stay home on Sunday. The “negative” campaigning, especially by Dilma’s campaign, may reflect an effort to steer Marina Silva’s voters away from the polarized choice of casting a ballot for Neves, but such campaigning might also encourage such voters to “hold their noses” to cast a ballot for Neves and against Dilma. The abstention factor may produce the outcome, but if it does not, then the fact that Neves’ campaign has flattened in terms of voters’ intentions may indicate that the PSDB’s candidate surprising over performance in the first round will not be repeated on October 26, 2014.

If abstentions are near the average and Neves cannot overachieve, based on the October 22, 2014 Datafolha poll, then Dilma is best positioned to win with a very close margin, possibly produced by overachievement-but this is unlikely. Indeed, President Dilma’s modest over performance in the second round of the 2010 elections was preceded by her underachievement in the 2010 first round balloting wherein Datafolha’s latest poll measured her support at 54%, but she only garnered the votes of 46.91%. This case casts a more sober shadow upon any optimism that she can repeat her 2010 second round modest over performance.

It is more likely that the winner squeaks out a victory with a margin of less than 2% of the valid votes-given only two days left for campaigning and short of any last minute campaign surprises. This election may come down to history; whomever wins in Minas Gerais (both candidates grew up in the state and Neves was its two-term governor), and possibly Pernambuco (Eduardo Campos’s home state as well as Lula’s) will likely win at the national ballot box. Neves should be the favorite to win Minas Gerais given his direct leadership as both its governor and senator-not to mention his storied grandfather’s (Tancredo Neves) pervasive esteem among Mineiros; but a similar historical context did not help presidential aspirant Al Gore in the United States in 2000-who lost his home state of Tennessee.

Rather than forecast the winning candidacy through such uncertainty and dynamic conditions, it might be time to toast Brazilian democracy and count the votes.

For a full report of the successive differentials between Datafolha voter intention measures and electoral outcomes for Brazilian presidential elections since 1994, see the table below.

Year Candidates* *winner Last Datafolha Poll Before Election Electoral Result Poll/Vote Difference
1994-First Round Fernando Henrique Cardoso (PSDB)* 48% 54.3% +6.3%
1994-First Round Lula (PT) 22% 27% +5%
1998 –First Round Fernando Henrique Cardoso (PSDB)* 49% 53.1% +4.1%
1998 –First Round Lula (PT) 26% 31.7% +5.7
2002-Second Round Lula (PT)* 64% 61.3% -2.7%
2002-Second Round José Serra (PSDB) 36% 38.7% -
2006-Second Round Lula (PT)* 61% 60.83% -0.17%
2006-Second Round Geraldo Alckmin (PSDB) 39% 39.17% -
2010-Second Round Dilma Rousseff (PT)* 55% 56.05% +1.05%
2010-Second Round José Serra (PSDB) 45% 43.95% -
2014-First Round Dilma Rousseff (PT) 40% 41.59% +1.59%
2014-First Round Aécio Neves (PSDB) 24% 33.55% +9.55%
2014-First Round Marina Silva (PSB) 22% 21.32% -
October 22, 2014-DataFolha
2014-Second Round Dilma Rousseff (PT) 47% ? -
2014-Second Round Aécio Neves (PSDB) 43% ? -