Brazil Election Odds 2026: The Favourite Just Flipped

Published 13 September 2026 · Prices frozen 13 September 2026, 16:12 UTC

For eleven months the Brazil election odds said the same thing: Lula wins a fourth term. On the evening of 10 September that stopped being true, and $149.3m of open interest — the largest world-politics market on Polymarket — now makes Flávio Bolsonaro a narrow favourite to run a country his father is serving 27 years in prison for trying to seize. Most coverage of this race is still quoting the old number. We pulled every open Brazilian contract on Polymarket, diffed a month of CLOB price history to the hour, checked Kalshi’s competing ladder and read the two markets that flatly disagree with each other — and the interesting finding is not who is ahead. It is that the exchange thinks Lula wins on 4 October and loses on 25 October.

Where the Brazil election odds stand today

Five separate Polymarket contracts price this election and they are almost never shown together. One asks who becomes president. One asks which two candidates reach the runoff. One asks whether there is a runoff at all. One asks who finishes second on 4 October. One prices the first-round margin. Put on a single board, they describe a market that has stopped believing in an incumbent and has not yet stopped believing he leads the first count.

Two features of that board matter more than the headline price. The first is the liquidity gradient: the presidency carries $149.3m and the outright-first-round contract carries $230,000. That is a six-hundred-fold difference in stake size, and it should be a six-hundred-fold difference in how seriously you take a two-point move. The second is how closed the field is. Renan Santos, the Free Brazil Movement co-founder running for the Missão party, is third at 1.85¢. Augusto Cury — a psychiatrist and self-help author polling between 6% and 10% — trades at 0.45¢. Everyone else is at 0.15¢ or below, including Jair Bolsonaro himself.

The week the favourite changed

Diff the CLOB price history from 13 August to now and the month has a clean shape. Lula bled steadily for four weeks, Flávio climbed steadily, and the two lines crossed at roughly 18:00 UTC on 10 September — the same day two pollsters published surveys putting Flávio ahead in a runoff for the first time.

The spread between the two candidates went from Lula +35.55 to Flávio +2.15 in a month, a 37.7-point swing. But look at the rows around it. Both men got more likely to reach the runoff over the same period, and the chance of anyone avoiding a runoff fell. The market did not decide the race was more volatile. It decided the race was narrowing to exactly two men and that the second of them had stopped being the underdog.

Why Lula can lead round one and still lose

Here is the gap worth understanding. The second-place market prices Flávio Bolsonaro at 77¢ to finish second on 4 October. If he finishes second, somebody finishes first, and on that ladder the only realistic candidate is Lula. So the same exchange that gives Lula 47.5¢ to be president gives him something close to 77% to top the first round. That is a thirty-point wedge, and it is not an error.

Brazil elects a president with more than 50% of valid votes or not at all. A fragmented first round — Cury at 6-10%, Renan Santos at 3-6.5%, two governors splitting a few points more — favours the candidate with the most durable floor, and that is Lula’s north-eastern base. A runoff three weeks later is a straight binary, and that is where the votes currently parked with Cury, Caiado and Zema have somewhere to go. The market is pricing the consolidation, not the count.

Ten points of disagreement between two markets on one exchange about one event is the kind of thing that exists only in a thin book. The margin ladder has traded $470,000 across eleven outcomes; the second-place market has traded $7.0m across five. When they disagree, our methodology says to weight by depth and say so out loud rather than average them into a number neither market supports.

What 4 October actually decides

Less than the date suggests. The first round elects a president only if someone clears 50% of valid votes, and the market gives that 6.6¢. The overwhelmingly likely outcome is that 4 October resolves three derivative markets — second place, third place, margin of victory — and the headline question rolls on to 25 October.

This matters for anyone reading a first-round result as a verdict. In 2022 Lula led the first round by 5.2 points and won the runoff by 1.8. A first-round lead in Brazil is a starting position, and the Polymarket board is currently pricing Lula to hold one and lose from it.

The Supreme Court crisis the odds are really pricing

The repricing did not come from the campaign trail. On 1 September Justice André Mendonça — a Bolsonaro appointee — unsealed federal police documents indicating that the banker Daniel Vorcaro had repeatedly sought legal advice from Justice Alexandre de Moraes. Vorcaro’s Banco Master collapsed amid fraud allegations running to billions of dollars, and de Moraes is the judge who sentenced Jair Bolsonaro to 27 years. The Supreme Federal Court then cancelled two consecutive sessions, on 9 and 10 September, and Chief Justice Edson Fachin called an extraordinary sitting of all ten justices for 15 September.

Polymarket has a market on exactly this. It is small — $60,850 of lifetime volume — but it moved before the election market did, and it moved further.

The sequence is the argument. De Moraes’s exit odds doubled in the first three days of September, before either of the polls that made headlines. Both Lula and Flávio Bolsonaro have demanded investigations into justices, along predictable party lines — but the asymmetry is obvious: the judge under pressure is the one whose rulings put the opposition’s founder in prison and kept him off the ballot. A market that thinks his position is weakening is a market that thinks the legal scaffolding around this election is weakening too.

The polls behind the Brazil election odds

Three pollsters published in the four days before the market crossed over, and all three agree on the shape: Lula ahead in the first round, the runoff a coin flip.

Every runoff number there is inside its own margin of error, which is precisely why the presidency market sits at 49.65¢ to 47.50¢ rather than anywhere decisive. The first-round columns are the more interesting read: a 4 to 5.6 point Lula lead over Flávio, with Cury taking 6-10% and Renan Santos 3-6.5%. That is the fragmentation the second-place market is pricing, and it is why the two questions separate.

Note what the market does not do with Cury. He has polled as high as 11%, his name saw a search spike after the first televised debate, and he trades at 0.45¢ to be president. The margin ladder gives him no victory band at all. Polling third in a two-round system is, in market terms, worth almost nothing.

Kalshi lists Brazil and nobody trades it

Kalshi carries four Brazilian presidential events: KXBRPRESADVANCE-26OCT04 on runoff qualification, KXBRAZILPRES1R-26OCT04 on an outright first-round win, KXBRPRESIDENT2-BRPRES26-2 and KXBRPRESIDENT3-BRPRES26-3 on second and third place, plus KXBRPRES1MOV-BRPRES26 on the margin. On 13 September 2026 every market in all of them returned no bid, no ask, no last price, no volume and no open interest.

More telling than the empty books is what is missing from the list. Kalshi has no market on who becomes president of Brazil — the single question carrying $149.3m on Polymarket. Its ladders price the derivatives and skip the headline. They also still list Tarcísio de Freitas across every event, months after the São Paulo governor chose re-election over a national run and endorsed Flávio; Polymarket prices him at 0.05¢.

What could move these odds before 4 October

Three dates and one number. The first date is 15 September, when the full Supreme Federal Court sits in extraordinary session to decide whether de Moraes should face investigation. That is two days from publication and it is the only scheduled event between now and polling day capable of moving both the institutional market and the electoral one at once.

The second is 3 October, when the campaign period formally ends and the last polls land. The third is 4 October itself, which resolves the runoff-qualification market and the second-place market while leaving the presidency open for another three weeks.

The number is the runoff transfer. Cury, Caiado, Zema and Renan Santos will collect somewhere between 10% and 20% between them on 4 October, and on 25 October those votes go somewhere or stay home. No published poll has yet tested the second preferences of Cury voters specifically, which is the largest genuinely unknown quantity in this race. When one does, expect the presidency market to move more on it than it moved on either poll last week.

Related on GeoOdds

• Brazil presidential election — live candidate probabilities and chart
• Which candidates advance to the runoff — live odds
• First round second place — live odds
• First round margin of victory — the full ladder
• Brazil Chamber of Deputies winner — live party odds
• Polymarket vs Kalshi: Venezuela leader odds — the same cross-exchange method
• Israel election odds 2026 — the other election the markets split in two

Sources: Polymarket Gamma and CLOB public APIs (events 45915, 45924, 45925, 204972, 205029, 257477); Kalshi public trading API (KXBRPRESADVANCE-26OCT04, KXBRAZILPRES1R-26OCT04, KXBRPRESIDENT2-BRPRES26-2, KXBRPRESIDENT3-BRPRES26-3, KXBRPRES1MOV-BRPRES26); Wikipedia on the electoral calendar and Jair Bolsonaro’s ineligibility and sentence; ABC News / Associated Press on the Supreme Court crisis; AS/COA poll tracker; Rio Times on the 10 September PoderData and AtlasIntel releases. Prices captured 13 September 2026, 16:12 UTC and not updated since. Nothing here is investment advice — see our disclaimer.