(Excerpt from a previously published report on using social media to forecast the 2018 Brazilian Presidential Election, Analyzing the 2018 Brazilian Election Social Media Influence Index, July 2018)
Brazil General Election 2018
In July 2017, Social Media Influence (SMI) forecasted that the then-outlier, Jair Bolsonaro, would end up winning Brazil’s presidential election. Assuming a Bolsonaro victory in the October election, SMI will have been the first known quantitative technique to forecast his victory and at a time when pundits did not take his candidacy seriously. SMI continues to forecast Bolsonaro as leading in the first round of voting in October. He is also seen as winning against any candidate in a potential second round of voting.
It is unclear if Lula, the two-term former president who is currently in jail due to corruption, will be eligible to run in the election. Though the probability of Lula officially running is extremely low, he has beaten stacked odds in the past and it would be unwise to completely write-off this scenario. Many claim that he could win if allowed to run, given his relatively high poll figures and his strong brand-name. SMI shows that Lula’s chances are low and have been since early 2017, well before his arrest. According to SMI, Lula has not shown significant amounts of traction and appears to benefit in polls due to positive Social Desirability Bias – which would not translate into actual votes. In other words, according to SMI, Lula would underperform current market expectations and end up losing to Bolsonaro, even if he did run.
This election could still produce some major volatility. Assuming our base case of Lula not running, it is unclear which candidate Lula would support. Additionally, Brazilian elections have multiple important variables that could throw off social media based forecasts such as mandatory voting and free TV time for candidates – both of which, as explained later, would work to the detriment of Bolsonaro.
Currently, SMI forecasts Bolsonaro and Haddad coming in first and second place respectively in the first round of voting (assuming Lula does not run and Haddad is the standard-bearer for the PT). In the follow-on second round of voting, SMI has Bolsonaro winning. Given the multiple bumps in the road between now and election-day, these forecasts can change. However, Bolsonaro’s SMI has been solidly in the lead for the last year and it would be difficult to envision him losing at this stage without a dramatic turn of events.