The program Bolsonaro expanded, Lula created—and voters remembered.
No Brasil de maio de 2022, uma pesquisa Datafolha revelou uma distância de 21 pontos percentuais entre Lula e Bolsonaro na corrida presidencial, sinalizando que o país caminhava, uma vez mais, para um confronto entre dois projetos irreconciliáveis de nação. A vantagem do ex-presidente — sustentada por mulheres, jovens, trabalhadores de baixa renda e até pelos beneficiários do programa social expandido pelo próprio incumbente — sugeria não apenas uma preferência eleitoral, mas um julgamento coletivo em formação. O centro político, sem candidatos capazes de romper a polarização, assistia ao seu próprio esvaziamento.
- A margem de 21 pontos é a maior já registrada pelo Datafolha desde o início do acompanhamento da corrida, em maio de 2021, indicando uma consolidação acelerada do eleitorado em torno de Lula.
- Bolsonaro intensificou viagens de campanha, expandiu o Auxílio Brasil e atacou o sistema eleitoral — mas nenhuma dessas estratégias conseguiu fechar o abismo nas pesquisas.
- A ironia mais reveladora da pesquisa: entre os próprios beneficiários do Auxílio Brasil, programa que o governo reconfigurou como trunfo eleitoral, Lula vence por 59% a 20%.
- O campo do centro desmorona — Ciro Gomes estagna em 7%, Doria retirou sua candidatura, e os indecisos chegaram ao menor patamar histórico da série, sinalizando que o eleitor já escolheu seu lado.
- As declarações de Bolsonaro sobre a segurança das urnas eletrônicas, sem evidências, geraram críticas de rivais, instituições e até sinais de preocupação dos Estados Unidos, ampliando o desgaste do incumbente.
Em maio de 2022, o Datafolha registrou o maior intervalo já medido na corrida presidencial brasileira: Lula com 48% das intenções de voto no primeiro turno, contra 27% de Bolsonaro. A pesquisa, realizada com 2.556 eleitores em 181 municípios, mostrava um campo político que se fechava rapidamente em torno de dois polos, deixando pouco espaço para terceiras vias.
A força de Lula atravessava demografias de forma consistente. Entre mulheres, liderava por 49% a 23%. Entre jovens de 16 a 24 anos, por 58% a 21%. No Nordeste, chegava a 62%. Entre os que ganham até dois salários mínimos, alcançava 56%. O dado mais simbólico, porém, vinha dos beneficiários do Auxílio Brasil — o programa social que Bolsonaro havia expandido como estratégia eleitoral e que nascera como Bolsa Família nos governos Lula: entre eles, o ex-presidente vencia por 59% a 20%.
Bolsonaro mantinha vantagem apenas em nichos específicos: entre os mais ricos, entre empresários e, por margem estreita dentro do erro amostral, entre evangélicos. Sua campanha havia apostado em viagens intensivas, alianças com o centro e, de forma crescente, em ataques ao sistema eleitoral — sugerindo, sem provas, que a eleição poderia ser comprometida. As declarações geraram críticas amplas, inclusive de observadores internacionais.
Desde março, Lula havia crescido de 43% para 48%, enquanto Bolsonaro permanecia praticamente estagnado. Na pesquisa espontânea, o ex-presidente subiu de 30% para 38%; o incumbente recuou levemente. O número de indecisos caiu ao menor patamar da série. Com a saída de João Doria e de Eduardo Leite da disputa, o centro se esvaziava antes mesmo do início oficial da campanha. O que restava era, essencialmente, um duelo entre dois projetos de Brasil — e o eleitorado, em sua maioria, já havia feito sua escolha.
In late May, as Brazil's presidential race entered its final stretch, a new Datafolha poll delivered stark numbers: Luiz Inácio Lula da Silva held 48 percent of first-round voting intentions, while incumbent Jair Bolsonaro trailed at 27 percent. The 21-point gap represented the widest margin the polling institute had recorded since it began tracking the race in May 2021, and it signaled a political landscape hardening into a two-person contest.
The survey, conducted over two days among 2,556 voters across 181 municipalities nationwide, showed the race collapsing toward polarization. Ciro Gomes, the third-place finisher from the Democratic Labour Party, registered just 7 percent. The remaining nine candidates—a mix of centrist and fringe contenders—barely registered, each pulling 2 percent or less. Seven percent of voters cast blank or null ballots, while 4 percent remained undecided. The margin of error was two percentage points.
Lula's strength cut across demographic lines. He commanded 49 percent among women compared to Bolsonaro's 23 percent. Among voters aged 16 to 24, he led 58 to 21. In the Northeast, traditionally a stronghold for the left, he reached 62 percent against Bolsonaro's 17. Among those earning up to two minimum wages monthly, Lula held 56 percent to Bolsonaro's 20. The gap was equally pronounced among Black voters (57 to 23), Catholics (54 to 23), and the unemployed (57 to 16).
One metric proved particularly telling: among recipients of Auxílio Brasil—the rebranded version of the Bolsa Família welfare program that Lula's own government had created—the incumbent managed only 20 percent support while Lula captured 59 percent. The program, which Bolsonaro had expanded as an electoral strategy, had failed to translate into political gain. Bolsonaro retained advantages only in narrow slices: among those earning more than ten minimum wages monthly (42 to 31), among business owners (56 to 23), and among evangelical voters, where he held a slim numerical edge within the margin of error (39 to 36).
The race had shifted since Datafolha's previous survey in late March. Then, Lula had stood at 43 percent and Bolsonaro at 26, a 17-point spread. The new gap of 21 points reflected both Lula's growth and Bolsonaro's inability to gain ground despite intensified campaigning. In the spontaneous portion of the survey—where voters named candidates without prompting—Lula reached 38 percent, up from 30 percent in March, while Bolsonaro actually declined slightly to 22 percent from 23 percent. The number of undecided voters had compressed to 29 percent, the lowest point in the polling series and a sign that voters were consolidating around the two frontrunners.
The intervening months had seen both campaigns shift strategy. Lula had announced his running mate, former São Paulo governor Geraldo Alckmin of the Socialist Party of Brazil, and accelerated his campaign schedule, though not without stumbles—he had been forced to walk back statements on abortion rights that drew criticism. Bolsonaro, meanwhile, had ramped up campaign travel and enlisted support from centrist parties while simultaneously escalating attacks on the judiciary and raising unfounded questions about the security of Brazil's electronic voting systems. He had begun suggesting, without evidence, that the election itself might be disrupted. These claims of potential coup-like activity served to energize his most committed supporters but drew criticism from rival candidates, institutional leaders, and international observers, including signals from the United States that Brazil's electoral system was secure.
The poll had been conducted before former São Paulo governor João Doria withdrew from the race on May 23, but researchers had tested his candidacy in a separate scenario where he registered 3 percent—insufficient to alter the overall picture. Doria's exit followed the earlier withdrawal of Rio Grande do Sul's former governor Eduardo Leite, who had been positioned as a backup option for the centrist PSDB party. The race, in other words, had already begun its winnowing before the official campaign period fully commenced. What remained was a contest between two antagonists, with the middle ground collapsing.
Notable Quotes
Bolsonaro's escalation of coup-like rhetoric serves to keep his radical base unified but draws criticism from other candidates, institutional leaders, and international pressure.— Datafolha analysis
The Hearth Conversation Another angle on the story
Why does a 21-point lead matter so much at this stage? Couldn't things shift dramatically?
The margin itself is large, but what's more significant is the composition. Lula is winning across nearly every demographic—young and old, rich and poor, every region except the wealthy enclaves. For Bolsonaro to close a gap like that, he'd need to flip entire voter blocs, not just persuade a few undecideds.
The Auxílio Brasil numbers are striking—Bolsonaro expanded the program but Lula still dominates among recipients. Why didn't that work?
Because people distinguish between a policy and the person offering it. The program was created under Lula. Bolsonaro expanded it, yes, but voters seem to credit the original architect. You can't simply buy back a legacy that isn't yours.
What about Bolsonaro's rhetoric on election integrity? Is that a closing strategy or something else?
It's a consolidation play. He's not trying to win over the middle—he's trying to keep his base locked in and energized. But it's a risky bet. It alienates other candidates, it draws international scrutiny, and it doesn't seem to be moving his numbers.
The undecided voters dropped to 29 percent. Does that mean the race is essentially over?
Not over, but set. The undecideds are at historic lows, which means most people have already chosen. What remains is turnout and whether anything unexpected happens. But the structural advantage is Lula's to lose.
Why did Ciro Gomes and the centrists fail to gain traction?
They were squeezed. Lula moved left, Bolsonaro moved right, and the space between them collapsed. Voters faced a choice between two clear visions and largely stopped looking for a middle ground. Ciro at 7 percent is what's left of that appetite.