The prediction market has a clear favorite. The polling evidence does not.
Brazil votes tomorrow in the first round of what has become an exceptionally tight presidential election.
Prediction markets currently give Flávio Bolsonaro roughly a 57% chance of becoming Brazil’s next president, while incumbent Luiz Inácio Lula da Silva trades around 44%.
Our forecast is considerably closer.
Market probability — Flávio Bolsonaro: 57%
ProbGap forecast: 49%
Gap: −8 percentage points
In other words, we do not think Bolsonaro is badly positioned.
We think the market is too confident that he is the favorite.
The disconnect
There is an unusual split in the market.
Lula is expected to finish ahead in tomorrow’s first round, yet Flávio Bolsonaro is favored to ultimately win the presidency.
That scenario is certainly possible.
But the evidence for making Bolsonaro a ~57% favorite in the eventual head-to-head contest looks weaker than the market price suggests.
The final polling picture is extremely close.
A new CNT/MDA poll released on Saturday puts Lula ahead in the first round with 47.8% of valid votes, compared with 42.1% for Bolsonaro.
More importantly for the presidential winner market, the same poll tests the likely second-round matchup.
It finds:
Lula: 47.3%
Flávio Bolsonaro: 43.1%
That is a 4.2-point Lula advantage.
The poll surveyed 2,007 voters between September 30 and October 2 and carries a margin of error of 2.2 percentage points.
But one poll should not determine a forecast.
And the broader polling picture is considerably tighter.
The wider polling picture
Looking across ten recent national polls produces something much closer to a dead heat.
A current polling aggregation has the hypothetical second round at approximately:
Flávio Bolsonaro: 46.0%
Lula: 45.7%
That is effectively a tie.
Individual surveys disagree about which candidate is ahead, and differences of a few percentage points are well within the range that could disappear through ordinary polling error, turnout differences or late movement.
This matters because the prediction market is making a stronger claim.
A 57% market probability does not merely say Bolsonaro can win.
It says he should be a meaningful favorite.
We think the available evidence supports a much narrower race.