Who will win Brazil's 2026 presidential election (including any runoff)?

rationale:

State of the Race and First-Round Polling With exactly two weeks until the first round on October 4, the contest is structurally locked into a near-certain runoff between incumbent President Luiz Inácio Lula da Silva and Flávio Bolsonaro. While Lula consistently leads first-round polling, his margin is narrow and shrinking. Mid-September surveys place Lula around 38–39% and Flávio at 34–36%; for example, Datafolha (15–17 Sept) has Lula at 39% and Flávio at 36% 2 sources. The residual "third-lane" candidates are stuck in the single digits, with Augusto Cury around 6%, Ronaldo Caiado near 4%, Renan Santos at 3%, and Romeu Zema trailing at 2% g1.globo.com. Aggregators such as BBC/PollingData similarly reflect this tight first-round plurality of roughly 39% to 34% bbc.com.

Runoff Simulations The decisive evidence for the final outcome lies in the second-round runoff simulations, which point to a statistical dead heat. Because the non-polarized electorate—especially supporters of Cury, Caiado, and Zema—is predominantly right-leaning, Lula's first-round lead evaporates in head-to-head matchups 2 sources. Some institutes still give Lula a slight edge, such as Datafolha at 46–44 g1.globo.com and Nexus/BTG at 47–46 en.wikipedia.org. However, an increasing number of polls show Flávio numerically ahead, including Quaest at 42–40 g1.globo.com, PoderData at 46–44 poder360.com.br, and AtlasIntel at 47.2–46.8 2 sources. Consequently, runoff aggregates now show a genuinely tied race or a marginal 45–44 advantage for the challenger bbc.com. Factoring in the anticipated runoff polling margins reinforces the conclusion that this ostensible tie implies a substantive advantage for Flávio because of historical right-wing polling underestimates.

Momentum and Campaign Trends Over the past month, momentum has trended unambiguously in Flávio's favor. Quaest's tracking captures this erosion vividly: Lula's runoff lead was eight points in July, fell to five in early August, then three, then one, and reversed to a two-point deficit by mid-September g1.globo.com. This shift is partly driven by the STF/Banco Master crisis, which has energized the right and damaged Lula's standing 2 sources. Lula is also underperforming his 2022 metrics, running roughly 5–6 points worse than at this exact stage of the previous cycle poder360.com.br. While the incumbent commands a formidable political machine, free TV time, and late social spending boosts like a recent Bolsa Família increase www1.folha.uol.com.br, his approval rating remains stagnant at roughly 48% approve to 50% disapprove 2 sources, capping his upside.

Historical Base Rates and Polling Bias Two conflicting historical base rates complicate the outlook. On one hand, Brazilian incumbents generally win reelection, and the first-round polling leader has won every presidential runoff since redemocratization (a 7-for-7 record) en.wikipedia.org. On the other hand, the structural mechanism behind that perfect first-round leader record—that the leader also led the runoff polls—is completely absent here. Furthermore, Brazilian polls systematically underestimated the Bolsonaro-aligned right in both 2018 and 2022 veja.abril.com.br. In 2022, Datafolha underestimated Jair Bolsonaro's first-round performance by roughly seven points. If even a fraction of this rightward polling miss persists, a nominal tie in current surveys implies a substantive advantage for Flávio.

Final Assessment Balancing these dynamics makes Flávio the narrow favorite, with a 50% chance of victory against 47% for Lula. The combination of adverse momentum for the incumbent, a right-leaning pool of minor-candidate transfers, anticipated runoff polling margins, and a documented history of polls undercounting the conservative base outweighs the structural advantages of Lula's incumbency and the first-round-leader heuristic. Caiado and Zema are functionally eliminated from contention and lose hypothetical runoff matchups by wide margins 2 sources, leaving Caiado with a 1% chance and Zema with effectively 0%. The remaining 2% is reserved for the "Other candidate" option, absorbing tail risks such as a late surge by Cury, TSE cassation rulings, or unforeseen candidate incapacitation en.wikipedia.org.