(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)
Summary
Social media based forecasting techniques are very similar to traditional polls and surveys in many ways, except that they leverage much larger sample sizes in real time and with more limited data distortion. Using social media and alternative on-line datasets to forecast elections worked well in the 2016 US and 2017 French Elections as ZettaCap’s Social Media Influence (SMI) forecasts beat polls, betting markets, and traditional analysts in predicting election outcomes.
For the 2018 Brazilian Election, SMI forecasts a Bolsonaro victory in the second round of voting, which is not in-line with polls or betting markets. Bolsonaro’s SMI has been extremely strong since around this time last year which qualifies him as the de-facto front-runner. Bolsonaro’s extremely negative Social Desirability Bias has produced lower non-SMI consensus expectations than what is clear in social media. SMI points to the PT candidate, either Lula or Haddad, as qualifying to run against Bolsonaro in the second round of voting. SMI shows other principal candidates as being relatively weak contenders.
In contrast to SMI, polls currently show Bolsonaro suffering a loss in the second round of voting, assuming he makes it that far, due to an apparent ceiling of support, reminiscent of a Le Pen scenario in France. Betting markets show Alckmin winning, even though he has not been able to show significant traction in polls or in social media. Due to extremely divergent forecasts among SMI, polls, and betting markets, Brazil appears to be the most difficult election to forecast of the last few years.
Though the first round of voting is less than 3 months away, a lot can still impact forecasts. Upcoming free TV and radio time for candidates should influence voting intentions, but as they greatly benefit establishment party candidates, Bolsonaro will be at a disadvantage. Also, upcoming debates can have a considerable impact.
Brazilian assets have significantly underperformed in relative terms over the last decade. The upcoming election could serve as a trigger to invert this trend. A non-trivial trend reversal could indicate a multi-year bull run for Brazilian assets in general and especially equities. The potential return over the coming 12 months is considerable.