Brazil's runoff was a referendum on the frontrunner; prediction markets at 87% were entirely wrong
Voters despised the incumbent, so the first-round winner drew all the fire onto himself. This let the trailing Lula complete the reversal in the runoff.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Bolsonaro lost first, not to Lula, but to himself
Bolsonaro faced money laundering charges from his time as a congressman five years earlier; this year he was exposed on WhatsApp asking Banco Master's owner for millions of dollars; before the vote, a magazine revealed him on a boat with a pharmaceutical lobbyist and a prostitute; the plane he used to flee to America was tied to an Argentina money-laundering case and a Texas drug-trafficking case. But all these scandals existed before the first round. What actually changed the outcome: the first round was supposed to be a confidence vote on Lula's 45% support, but once Bolsonaro unexpectedly led, all the fire in the runoff trained on him. Winning the first round became his defeat.
— James BosworthWhatsApp groups proved more effective than traditional campaigning
Lula, who had never used a smartphone before, got his first phone during the campaign. His team set up the ‘Lula Army’WhatsApp group with over 10 million members—more than 5% of Brazilian voters. When a mayor blocked public transit on election day, the group immediately coordinated voters to the polls. Several 30-second clips from his TV debate were perfectly formatted for social media shorts. Whether by design or accident, the effect was excellent.
— James BosworthPrediction markets betting 87% on Bolsonaro were entirely wrong
Before election night, Bosworth himself assigned Bolsonaro a 64% chance of winning; prediction markets at the time were broadly priced at 80%, peaking at 87%. Lula won. When asked if it was a ‘20% tail-risk event,’Bosworth was direct: no, they really got it wrong. They were simply overconfident. Even professional forecasters admitted this wasn't about being right—it was just ‘who was wrong by less.’
— James BosworthJust 1 to 2 percent of voters decided this election
Between the two rounds, turnout changed less than 2%, and the share of invalid or blank ballots barely budged. On Brazil's scale, 1 to 2% meant millions of votes. But Lula's margin was just 4 million. A razor-thin fraction of swing voters decided the entire race.
— James BosworthTrump claims 9-for-9 in Latin America; that number is fabricated
Trump keeps saying he's gone 9 for 9 in recent Latin American elections. Bosworth was direct: Trump made that number up entirely; it just happens to roughly match his actual record. Bosworth reads this wave of right-wing wins as a global anti-incumbent cycle rather than ideological realignment. But regardless of the cause, every Latin American leader—left, center, or right—will have to directly confront Trump over the next two years, until January 2029.
— James BosworthRubio wants Venezuela's elections in six months; the regime prefers 2028 or 2029
Under the plan Rubio (not Trump) is pushing, Delcy Rodríguez would declare the presidency vacant, establish a new electoral commission, then hold elections six months out—optimally July next year, at latest September or November. But Rodríguez herself wants to delay elections to 2028 or 2029. Bosworth identifies this as the precise conflict between Rubio and other Trump officials, and between Rubio and Venezuelan authorities. Rubio wants the political credit: ‘Venezuelan democracy, on my watch.’
— James BosworthLula and Machado hate each other but might secretly align
Lula has been too friendly with Chávez and Maduro—leftist dictators—so Venezuela's opposition resents him, holding him partly responsible for the failure to complete a power transition. Now Trump treats Venezuela as a quasi-colony, and that troubles Lula: South America should be Brazil's sphere. Bosworth doesn't expect them to embrace openly, but thinks Lula might support new elections because it aligns with his goal of Venezuela escaping U.S. control.
— James BosworthMachado could win 90 percent but might need to concede the next election
If free elections were held tomorrow—Machado versus Delcy Rodríguez—Machado would win by over 90%, excluding overseas votes. (Roughly a quarter of Venezuelans have already left.) But winning is only the start. She needs to seize actual military and police control—part of why the Trump administration was hesitant about her—then fix ‘one of the worst economic collapses in history’and rebuild multiparty democracy with real checks and balances. That probably means accepting that she or her party loses the next election.
— James BosworthIn their own words · checked verbatim
So there are two answers to that. One is that Flavio was a terrible candidate, and the other was that Lula was a great candidate.
James Bosworth3:07
In some ways, coming in first place in the first round was the worst thing that could happen to Flavio Bolsonaro because all of a sudden, all of the attention got put on him and it let Lula catch up from behind.
James Bosworth5:14
I will tell you, they missed. They were overconfident on this election.
James Bosworth8:14
But in Brazil, that's a couple million votes. And in an election won by 4 million votes, you know, a shift of 1% or 2% across those rounds really did make the difference.
James Bosworth9:16
Trump was going around saying he was nine for nine in elections recently. He just pulled that number out.
James Bosworth14:25
This year, there's not going to be an election. There's really going to be, there's about a six to nine month run up for any election that has to happen because there are steps that must occur.
James Bosworth17:27
Better than 90 percent, assuming everything was free and fair, Maria Karina Machado wins.
James Bosworth24:34
And she needs to somehow restore the checks and balances, which means she probably loses the next election, or her party does.
James Bosworth25:36
Figures
| Lula Army WhatsApp group members | Over 10 million; more than 5% of Brazilian voters | 6:14 |
| Bosworth's odds for Bolsonaro night before the vote | 64% | 8:14 |
| Prediction markets' peak odds for Bolsonaro | 87% | 8:14 |
| Turnout change between the two rounds | Less than 2% | 9:16 |
| Lula's final margin of victory | 4 million votes | 9:16 |
| Machado's odds versus Delcy Rodríguez | Over 90% | 24:34 |
| Venezuelan population that has left the country | Roughly one-quarter | 24:34 |
| Bosworth's confidence in Democrats winning the House | 96% | 26:36 |
Glossary
- status of forces agreement
- An agreement defining the legal status and liability of foreign military personnel; distinct from establishing a permanent military base.
- chavistas
- Supporters of Maduro and his ruling party coalition; named after late president Hugo Chávez.
How to listen
Investors and analysts tracking Latin American political trends, prediction market accuracy, Venezuela's transition situation, and oil investment opportunities.
You can skip the closing chat about Ecuador's Quito mayoral election and U.S. House race; it's low-information.