Brazilian Election Betting Guide

Sunday sees round one of the Brazilian presidential election and one of the most intriguing political betting contests ever.

Not much has changed since we previewed this contest a few weeks back. If anything, Lula's advantage in the polls has increased; the Economist's poll-of-polls has him at 51% - if that happens, no second round will be needed. Even if it doesn't, it's hard to see how Bolsonaro makes up any ground in the run-off from the 10% or so of the electorate who might be voting for one of the minor candidates. The Brazilians have an unusually long gap of four weeks until a run-off, so I suppose there is enough time for the dynamics of the race to change. On the other hand, it seems unlikely that many voters will be changing their minds in such a polarised contest. So how come the betting markets still give Bolsonaro around a 30% chance of winning?

Two main worries for Lula backers.

1. The polls are wrong. Perhaps there are a large number of hidden Bolsonaro voters, people who don't want to answer polls or aren't being accurately picked up by the pollsters in some other way. Fellow right wing populist Donald Trump certainly outperformed his polls in the 2020 US election; polling aggregators gave Biden around an 8% national lead whereas the result was under a 5% advantage. Bolsonaro is going to need a lot more of a swing than that to cause an upset. Even an error on the scale of the 2015 UK general election will not be enough. In fact it's quite hard to find any major national election where the polls were out by the scale of the lead that Lula is enjoying.

2. Bolsonaro "steals" the election. President Bolsonaro has certainly been making noises about the integrity of the vote and most people seem prepared for a Trump-like denial of any loss. If the election result is clearly a Lula win, that will be good enough for Smarkets to be settling on, irrespective of whether the incumbent manages to hold on to power via some other method. There are still some other scenarios which could cause a problem. What if Lula were arrested or somehow prevented from participating in a run-off, despite coming out ahead in round one? Depending on the circumstances, something like that clearly could stop him from "winning" the election. From a settlement point of view, it's impossible to issue any categorical assurances to cover all of the possible eventualities.

Despite those concerns, I can't see how Bolsonaro has anything like a 30% chance of winning the election. Still, one leading oddsmaker who works for a company that has customers in Brazil tells me about the amounts of people backing Bolsonaro: "I have never come across a market like this where we just can’t lay something short enough, money doesn’t stop. Can only think that Brazilian customers have possibly and wrongly bought into the whole ‘only God can remove me from office’ stuff and think he just won’t accept the result?".

Image by Alan Santos/PR - https://www.flickr.com/photos/palaciodoplanalto/48787974638/, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=82494253