Flávio Bolsonaro beat the polls and won Brazil's first round. The market responded by making him an overwhelming favorite. We think it has gone too far.
Brazil has given us the information we were waiting for.
When we published PG-0011 before the first round, prediction markets gave Flávio Bolsonaro a 57% chance of becoming Brazil's next president.
Our forecast was 49%.
The polls suggested something close to a coin flip.
Then Brazil voted.
Flávio Bolsonaro finished first with 47.03% of valid votes, ahead of Luiz Inácio Lula da Silva on 45.16%. Neither reached 50%, sending the election to a runoff on October 25. TSE
The prediction market reacted violently.
Flávio is now trading around 84% to win the presidency.
That is a 27-point increase from our original market snapshot. polymarket.com
Some of that repricing is justified.
We are changing our forecast too.
But not nearly as much as the market.
What changed?
The most important fact is straightforward:
the polls underestimated Flávio Bolsonaro.
Immediately before the election, the UOL polling aggregate had Lula at 40.5% and Flávio at 37.2%.
Its runoff average was essentially tied:
Lula 45.1%
Flávio 44.8%
Instead, Flávio won the first round by almost two percentage points. UOL Notícias
That is real information.
Actual votes should carry substantially more weight than pre-election polling.
Our original 49% forecast therefore no longer represents our best estimate.
But the market has made a much larger adjustment.
At 84%, it is effectively saying that Flávio is now an overwhelming favorite.
We think that conclusion goes too far.