Polymarket prices Luiz Inácio Lula da Silva at 64.5¢ to win the 2026 Brazilian presidential election — a 64.5% market-implied probability — while a poll-average model corrected for Brazil’s documented right-wing polling error puts the true number near 68.5%, a modest 4-point edge on the incumbent.
The case rests on a single adjustment: Brazilian polls have systematically understated the Bolsonaro-aligned vote, and any honest estimate has to discount Lula’s current lead before converting it to a probability. Do that, and 64.5¢ looks slightly cheap rather than obviously rich. It breaks if the discount should be larger than 2022’s runoff error suggests, or if Lula — who turns 81 on October 27, 2026 — is not the effective candidate on polling day. This Deep Dive walks the math, the case against, and the levels that would flip it.
The Bet at a Glance:
• Market: “Will Luiz Inácio Lula da Silva win the 2026 Brazilian presidential election?” — Polymarket, August 6, 2026
• Price: 64.5¢ = 64.5% market-implied probability; $8.08m volume, $363,366 liquidity — Polymarket API snapshot, August 6, 2026
• Our estimate: 68.5% — poll average corrected for 2022 runoff polling error, then discounted for candidate risk
• Edge: +4.0pp on the YES side — within model noise, not a conviction call
• Catalyst / date: first round October 4, 2026; runoff late October if no candidate clears 50% — AS/COA
• Main rival: Flávio Bolsonaro at 26.6¢; Renan Santos 6.7¢; all others 2.1¢ or below — Polymarket, August 6, 2026
• Disconfirmation: a poll-error haircut of 4.5pp rather than 2.2pp drops the estimate to 56.2%, below the market price
Methodology: how the 68.5% was built
The estimate starts with the two most recent published runoff polls. Quaest, fielded July 10 to 13, 2026 with 2,004 respondents and a margin of error of ±2 points, had Lula 45% to Flávio Bolsonaro’s 37%. Datafolha had the same pairing at 47% to 43%. The unweighted mean is a Lula lead of 6.0 points.
That lead is then discounted using the 2022 runoff as the calibration case: final polls showed Lula 52% to 48%, a four-point lead, and the actual result was 50.9% to 49.1%, a 1.8-point win — a 2.2-point overstatement of Lula’s margin. Applying that same 2.2-point haircut gives an adjusted lead of 3.8 points.
Converting a lead to a probability requires an uncertainty term. Two months out, with Quaest’s own series swinging 10 points between April and July 2026, a standard deviation of 6.5 points is defensible. That yields a 72.1% raw win probability, multiplied by 0.95 for candidate and health risk, giving 68.5%.
The caveats are real. Two polls is a thin average, one historical election is a thin calibration, and the 0.95 health factor is a judgement rather than a measurement. The sensitivity table below shows how much the answer moves when those inputs change.
The market and the price
Polymarket’s Brazilian presidential market is one of the deeper political contracts currently trading. The Lula contract alone shows $8.08m in volume against $363,366 in liquidity as of August 6, 2026, with the full candidate set carrying well over $100m in combined volume. That matters: thin markets produce prices that are noise rather than signal, and this one is not thin.
The field prices as a two-horse race with a long tail. Lula sits at 64.5¢, Flávio Bolsonaro at 26.6¢, the activist Renan Santos at 6.7¢, and Goiás governor Ronaldo Caiado at 2.1¢. Every other named candidate — including São Paulo governor Tarcísio de Freitas, Michelle Bolsonaro, Fernando Haddad and Jair Bolsonaro himself — trades at 0.4¢ or below.
| Candidate | Polymarket price | Implied probability | Our estimate | Edge (pp) |
|---|---|---|---|---|
| Lula (PT) | 64.5¢ | 64.5% | 68.5% | +4.0 |
| Flávio Bolsonaro (PL) | 26.6¢ | 26.6% | 25.0% | −1.6 |
| Renan Santos | 6.7¢ | 6.7% | 4.5% | −2.2 |
| Ronaldo Caiado (PSD) | 2.1¢ | 2.1% | 2.0% | −0.1 |
Sources: Polymarket API snapshot, August 6, 2026; our estimates from the poll-average model described above. Prediction-market cents read directly as probabilities before fees and spread; the four listed prices sum to 99.9¢, so overround is negligible.
Is Lula at 64.5¢ value? Marginally, on our numbers — but the margin is thin enough that it should not be mistaken for conviction. The market has already done most of the work an outsider might think it has missed. A naive reading of the July polling, which shows Lula ahead by eight points in the Quaest runoff scenario, would imply a win probability somewhere north of 80%. The market is at 64.5%, which means traders are collectively applying a discount of roughly 15 to 20 points to the headline polling. That discount is rational and well-founded in Brazilian electoral history. Our model applies a smaller, more precisely calibrated discount drawn from the 2022 runoff specifically, and lands 4 points above the market. The disagreement is about the size of a correction both sides agree is necessary.
Why Brazilian polls understate the right
The 2022 general election is the reference case, and the miss was not marginal. Going into the first round, polls showed Lula leading by as many as 14 points. The actual result was Lula 48.4% to Bolsonaro’s 43.2% — a 5.2-point gap. Datafolha and Ipec, the two most closely watched pollsters, had Bolsonaro at 36% and 37% respectively once null and blank votes were removed. He finished on 43.2%, an understatement of six to seven points.
The explanations offered at the time were structural rather than technical, which is precisely why they should be expected to persist into 2026.
“Far-right parties often draw support from people who don’t have a lot of trust in others,” and those supporters “are more likely to decline requests to be interviewed.”
— Christopher Hanretty, Professor of Politics, Royal Holloway, University of London (Al Jazeera)
Gustavo Flores-Macias, professor in the Department of Government at Cornell University, put it more bluntly in the same analysis, saying pollsters “clearly haven’t figured out” how deep support for Brazil’s right-wing runs. A further complication is that Brazil has not run a census since 2010, leaving pollsters to estimate the demographic composition of the electorate — religion and household income in particular — from ageing data.
None of those conditions has been resolved in the four years since. If anything, the imprisonment of Jair Bolsonaro raises the plausibility of response bias among his supporters rather than lowering it.
The case for Lula at 64.5¢
Three things support the YES side beyond the raw polling.
The first is direction of travel. Quaest’s own series has moved consistently in Lula’s favour through 2026: in April the poll had Flávio Bolsonaro ahead 42% to 40%; by June Lula led 44% to 38%; by the July field dates it was 45% to 37%. A candidate trending upward two months out is in a materially better position than the same lead arrived at from above.
The second is the opposition’s coalition problem. Flávio Bolsonaro’s candidacy was formalised on July 25, 2026 in São Paulo, and reporting on the convention noted that few Brazilian political heavyweights attended — a signal of the difficulty he has had assembling a broad anti-Lula front. The right’s vote is also split across Renan Santos at 6.7¢ and Ronaldo Caiado at 2.1¢, which complicates first-round consolidation even if a runoff would eventually unify it.
The third is that the fragmentation is priced asymmetrically. Flávio at 26.6¢ against a first-round polling position of 33% to 37% implies the market already doubts his ability to convert. Our estimate of 25% for Flávio is close to the market’s, which means the 4-point disagreement really is concentrated on the Lula line rather than spread across the board.
What is the prediction for the 2026 Brazilian election? Our model makes Lula a 68.5% favourite to win, most likely via a runoff in late October rather than an outright first-round victory. The first-round polling — Lula at 39% against Flávio Bolsonaro’s 33% in the July Quaest survey, or 36% to 23% in the Genial/Quaest series with Tarcísio de Freitas included — sits well short of the 50% threshold needed to avoid a second round. That runoff is where the polling-error adjustment does its work, because a head-to-head consolidates the right’s fragmented first-round vote behind a single candidate. The projected runoff margin after correction is roughly 3 to 4 points in Lula’s favour, which is a win but not a comfortable one, and comfortably inside the range where a repeat of 2022’s polling miss would change the result.
The case against: why the market may be right
The strongest argument against our number is that the 2.2-point haircut is too small. We chose it because it comes from the 2022 runoff, the closest structural analogue to the scenario being priced. But the same election produced a six-to-seven-point understatement in the first round. If the correct adjustment is nearer 4.5 points, the model outputs 56.2% — comfortably below the market price, and the edge inverts from a modest buy to a clear sell.
The second argument is candidate risk, which our 0.95 multiplier may understate. Lula turns 81 on October 27, 2026, three weeks after the first round and around the likely runoff date. He would be the oldest person ever inaugurated as Brazilian president. The market prices his most plausible PT replacement, Fernando Haddad, at 0.1¢ — meaning traders see essentially no path for the party without Lula at the top of the ticket. If the true probability of Lula not being the effective candidate is 8% rather than 5%, the estimate falls by roughly two further points.
The third is that opposition consolidation may be more achievable than the convention turnout suggested. The Liberal Party has publicly framed Flávio as a unifying figure rather than a continuity candidate.
Flávio is “a balanced politician, experienced and prepared for dialogue.”
— Luciano Zucco, head of the Liberal Party in Brazil’s lower house (HNGN)
If that framing lands with centre-right voters currently parked with Caiado and Santos, the runoff arithmetic tightens quickly.
Sensitivity: how much the assumptions matter
Because the edge is only 4 points, the honest thing to publish is the range rather than the point estimate. The table below varies the two soft inputs — the polling-error haircut and the uncertainty term — holding the 0.95 candidate-risk factor constant.
| Uncertainty (σ) | Haircut 0.0pp | Haircut 2.2pp | Haircut 4.5pp |
|---|---|---|---|
| 5.0 points | 84.1% | 73.8% | 58.7% |
| 6.5 points | 78.1% | 68.5% | 56.2% |
| 8.0 points | 73.5% | 64.8% | 54.6% |
Sources: our model, built on Quaest (July 10–13, 2026) and Datafolha runoff polling, calibrated against the 2022 runoff result of 50.9% to 49.1%. All cells include the 0.95 candidate-risk multiplier.
The market price of 64.5¢ sits inside this grid at almost dead centre — it corresponds closely to a σ of 8.0 with a 2.2-point haircut. In other words, Polymarket is pricing a slightly more uncertain race than we are, not a fundamentally different one. Of the nine cells, five sit above the market price and four below. That is not the signature of a mispriced market; it is the signature of a fairly priced one where reasonable models disagree at the margin.
Where this bet breaks
The lean on Lula at 64.5¢ rests on assumptions that could fail. It breaks if any of these fire:
- A poll-error haircut above roughly 3.5 points becomes justified. If August and September polling shows Flávio Bolsonaro outperforming his surveyed numbers in any state or municipal contest, the first-round-style correction of six to seven points becomes the better analogue, and the model output drops below the market price.
- Lula’s health becomes a live campaign issue. He turns 81 on October 27, 2026. Any hospitalisation, withdrawal from scheduled debates, or formal discussion of a replacement candidate invalidates the 0.95 multiplier immediately — and with Haddad priced at 0.1¢, the market would reprice violently rather than gradually.
- The centre-right consolidates before the first round. If Ronaldo Caiado at 2.1¢ or Renan Santos at 6.7¢ withdraws and endorses Flávio Bolsonaro, the runoff arithmetic changes before the runoff exists. Watch for a single-candidate agreement in September.
- The price moves through 72¢. Above that level the market has closed the gap to our central estimate, and there is no remaining edge to capture even if the thesis is correct.
What to watch
The first round is October 4, 2026, with roughly 156 million eligible voters. If no candidate clears 50% of valid votes, a runoff follows later in October — and on current polling that runoff is the overwhelmingly likely path, since neither Lula’s 39% nor Flávio’s 33% in the July Quaest first-round series is close to the threshold.
Between now and then, three windows matter. The August and September Quaest and Datafolha releases will show whether Lula’s April-to-July trend continues or stalls; a stall at a four-point lead is meaningfully different from a continued climb. Any formal withdrawal-and-endorsement among the minor right-wing candidates is the single fastest way the runoff maths tightens. And the campaign debate schedule is the most likely venue for the candidate-risk question to become concrete.
Readers tracking how prediction markets have handled comparable political contracts can compare this with our analysis of the Polymarket 2026 Senate market, and for a different Brazilian market entirely, our look at Brazil at 7¢ in the World Cup outright. The domestic policy backdrop is covered in our reporting on the Brazil casino legalisation bill.
TL;DR
Polymarket has Lula at 64.5¢ (64.5% implied) to win Brazil’s October 4, 2026 election. Correcting the current six-point average runoff lead by the 2.2-point error observed in the 2022 runoff, then applying a 0.95 candidate-risk factor, gives a true estimate of 68.5% — a 4-point edge on the YES side. That edge is real but small, and it inverts to a sell if the correct polling haircut is 4.5 points rather than 2.2. The single trigger to watch is Lula’s health: he turns 81 on October 27, 2026, and the market prices his replacement at 0.1¢.
FAQ
What are the odds for Lula to win the 2026 Brazilian election?
Polymarket priced Lula at 64.5¢ on August 6, 2026, which reads directly as a 64.5% implied probability. Flávio Bolsonaro was the next shortest at 26.6¢, with Renan Santos at 6.7¢ and Ronaldo Caiado at 2.1¢. The Lula contract carried $8.08m in volume and $363,366 in liquidity.
Is Lula at 64.5¢ value?
On our model, marginally — the estimate is 68.5%, a 4-point edge. But that figure is highly sensitive to the polling-error correction applied. Using a larger haircut drawn from the 2022 first round rather than the runoff produces 56.2%, which would make the price expensive rather than cheap.
Why do Brazilian polls understate right-wing candidates?
In 2022, polls put Bolsonaro at 36% to 37% in the first round; he received 43.2%. Analysts attributed the gap to lower institutional trust among right-wing voters making them likelier to decline interviews, and to Brazil’s lack of a census since 2010, which forces pollsters to estimate electorate composition from outdated demographic data.
When is the 2026 Brazilian presidential election?
The first round is October 4, 2026, with about 156 million eligible voters. If no candidate wins more than 50% of valid votes, a runoff between the top two follows later in October. Current polling makes a runoff the overwhelmingly likely outcome.
Why is Jair Bolsonaro not the candidate?
Jair Bolsonaro was declared ineligible on June 30, 2023 and barred from office through 2030. He was subsequently arrested and sentenced over an attempted coup, and has been serving that sentence under humanitarian house arrest. His son Flávio Bolsonaro was formalised as the Liberal Party’s candidate on July 25, 2026.
What would change this analysis?
Four triggers: evidence that the polling error is larger than 2022’s runoff gap, any development around Lula’s health, a withdrawal-and-endorsement consolidating the centre-right behind Flávio Bolsonaro, or the price moving through 72¢, at which point the modelled edge disappears entirely.
This article is informational analysis only and is not betting or financial advice. Odds and prediction-market prices move constantly; every price quoted is a timestamped snapshot, not a live line. There is no such thing as a guaranteed bet — past results and model estimates do not guarantee outcomes. Do your own research.
Gambling carries financial risk and can be addictive. 18+ / 21+ depending on jurisdiction. Never bet more than you can afford to lose. If you or someone you know needs help, visit GamCare (UK), call 1-800-GAMBLER (US), or see our Responsible Gambling page.
