The race is too close to call with confidence.
In the vast and politically consequential state of Minas Gerais, Brazil's 2026 presidential contest reveals itself as a story still being written. A fresh survey places President Lula and Senator Flávio Bolsonaro in a statistical embrace — separated by four points, yet united by a margin of error that renders the gap meaningless. The numbers remind us that in democracy's unfolding, certainty is often the first casualty of honest measurement.
- A four-point Lula lead over Flávio Bolsonaro sounds like an advantage — until the two-point margin of error turns it into a dead heat.
- With 8% of voters either undecided or prepared to cast blank ballots, a decisive bloc remains unanchored and capable of tipping the outcome.
- A second scenario pitting Lula against ex-governor Romeu Zema produces an exact 42–42 tie, signaling that the field of contenders itself is still unsettled.
- Minas Gerais — one of Brazil's most populous and electorally powerful states — is refusing to be claimed by any candidate heading into 2026.
A poll released Thursday by Real Time Big Data captures Minas Gerais as genuinely contested ground in Brazil's approaching presidential race. Surveying 1,600 voters on May 19 and 20, the institute found President Lula at 44 percent against Senator Flávio Bolsonaro's 40 percent in a hypothetical runoff. The four-point gap, however, dissolves under scrutiny: the survey's two-point margin of error means the race could just as plausibly be tied at 42 apiece. Eight percent of respondents said they would vote blank, null, or hadn't yet decided.
The pollsters also modeled a different second-round contest — Lula versus Romeu Zema, the former Minas Gerais governor now aligned with the Novo party. That simulation produced a perfect tie at 42 percent each, with 7 percent indicating blank or null votes and 9 percent still undecided. The existence of both scenarios underscores how much depends on the first round: which opponent Lula faces in a runoff could reshape the entire calculus.
Conducted at a 95 percent confidence level and registered with Brazil's Superior Electoral Court under protocol BR-07114/2026, the survey carries institutional weight even as its findings resist clean interpretation. What emerges most clearly is a basic truth about this moment in Brazilian politics: no candidate has yet earned the right to treat Minas Gerais — or perhaps anywhere — as settled territory.
A new poll released Thursday paints a picture of Minas Gerais as genuinely unsettled terrain in Brazil's 2026 presidential race. The Real Time Big Data survey, which contacted 1,600 voters over May 19 and 20, shows President Lula holding 44 percent support against Senator Flávio Bolsonaro's 40 percent in a hypothetical runoff matchup. That four-point gap sounds decisive until you account for the margin of error: two percentage points in either direction. The numbers, in other words, collapse into a technical tie.
What this means in practical terms is that the two candidates could end up anywhere from Lula at 42 and Flávio at 42, to Lula at 46 and Flávio at 42. The pollsters are essentially saying the race is too close to call with confidence. Eight percent of respondents said they would cast blank or null votes, or remained genuinely undecided about how they would vote in such a scenario.
The institute also modeled an alternative second-round contest, this one between Lula and Romeu Zema, the former governor of Minas Gerais who now represents the Novo party. In that simulation, the two men landed at exactly 42 percent each—a perfect tie. Seven percent indicated they would vote blank or null, while nine percent remained undecided. The existence of this alternative scenario suggests that Minas Gerais, one of Brazil's most populous and politically consequential states, remains genuinely fluid. The outcome of the first round will matter enormously in determining which of these runoff matchups actually occurs.
The survey was conducted with a confidence interval of 95 percent, meaning the pollsters are 95 percent certain their results fall within the stated margin of error. The methodology and findings have been registered with Brazil's Superior Electoral Court under protocol BR-07114/2026. As the 2026 race takes shape, these numbers underscore a basic reality: no candidate has yet secured the kind of commanding position that would allow them to treat any state, much less a state as important as Minas Gerais, as settled territory.
A Conversa do Hearth Outra perspectiva sobre a história
Why does a four-point lead collapse into a tie?
The margin of error is two points either way. So Lula could actually be at 42, and Flávio could be at 42. The poll is telling us it doesn't know which one is really ahead.
But Lula is ahead in the actual numbers shown.
He is, in this snapshot. But the snapshot has built-in uncertainty. It's like saying someone is taller, but the measuring tape itself is imprecise enough that they could actually be the same height.
What does the Zema scenario tell us?
That Minas Gerais is genuinely open. If Flávio doesn't make the runoff, Lula faces a different opponent entirely—and they tie at 42-42. The state isn't locked in for anyone.
Who are the undecided voters?
Eight percent in the Flávio matchup, nine percent against Zema. Not huge numbers, but in a tied race, they're the whole ballgame.
Does this poll settle anything?
It settles that nothing is settled. It's a snapshot of genuine uncertainty in a state that will likely decide the presidency.