Rio's homeless census shows 8,000 people, but federal data suggests triple that

Over 8,000 people living on Rio's streets face food insecurity, with 30% reporting days without eating and 2% resorting to garbage for food; 83% identify as Black or mixed-race.
The gap between those two numbers is not a rounding error
Rio's municipal census counts 8,195 homeless people, but federal data suggests 23,431—a discrepancy that may reflect how the city measures its crisis.

Em uma das maiores cidades do Brasil, dois números tentam descrever a mesma realidade — e a distância entre eles revela algo mais profundo do que uma divergência estatística. O censo municipal do Rio de Janeiro registrou 8.195 pessoas em situação de rua em 2024, enquanto pesquisadores federais, usando dados do próprio governo, contaram 23.431 — quase o triplo. Publicado em silêncio, sem anúncio público, e ignorando as solicitações do Ministério Público Federal, o levantamento da prefeitura levanta uma questão que transcende a metodologia: o que uma cidade escolhe enxergar diz muito sobre o que ela escolhe resolver.

  • A prefeitura do Rio publicou seu censo de pessoas em situação de rua sem qualquer comunicado oficial — os dados simplesmente apareceram em um portal de dados, sem campanha, sem reconhecimento público.
  • Um estudo da UFMG usando o CadÚnico federal aponta 23.431 pessoas em situação de rua na cidade, quase três vezes o número oficial de 8.195 — uma diferença que sugere subcontagem sistemática, não imprecisão acidental.
  • O orçamento da Secretaria de Assistência Social caiu 7 milhões de reais e opera hoje no nível mais baixo em uma década, enquanto 30% dos entrevistados relatam ter passado um dia inteiro sem comer.
  • O Ministério Público Federal cobra o censo desde 2024, não obteve resposta e recorreu à Justiça Federal pedindo intervenção urgente diante da omissão da gestão municipal.
  • A prefeitura, questionada sobre os atrasos, o silêncio e a indiferença às solicitações dos promotores, respondeu apenas com notas sobre metodologia — sem responder às perguntas feitas.

No início de janeiro, a prefeitura do Rio de Janeiro divulgou um censo contando 8.195 pessoas em situação de rua na cidade — um aumento de 4,2% em relação a 2022. A publicação aconteceu sem anúncio na página inicial do portal Data Rio e sem nenhuma comunicação da Secretaria de Assistência Social, cujo orçamento encolheu 7 milhões de reais e hoje está no patamar mais baixo em dez anos.

O problema é que outro levantamento, conduzido por pesquisadores da Universidade Federal de Minas Gerais com dados do CadÚnico de 2025, chegou a um número muito diferente: 23.431 pessoas — quase o triplo do total oficial. A divergência não é pequena o suficiente para ser atribuída a diferenças metodológicas menores. Ela sugere que o censo municipal pode estar capturando apenas uma fração da população que realmente vive nas ruas do Rio.

O próprio censo municipal, ainda que incompleto, descreve uma realidade de privação severa. O Centro concentrou o maior número de pessoas contadas — 1.300 —, e 83% dos entrevistados se identificaram como negros ou pardos. A maioria não completou o ensino fundamental; 10% não conseguiam ler um bilhete simples. Quando perguntados sobre o que precisariam para sair das ruas, a resposta mais comum foi emprego. Apenas 1% disse estar na rua para ficar mais perto do trabalho.

A insegurança alimentar atravessa esse retrato com força: 45% dependem de doações para comer, mais de 30% passaram um dia inteiro sem se alimentar na semana anterior ao censo, e 2% relataram ter comido do lixo. São números que descrevem não uma escolha, mas um abandono estrutural.

O Ministério Público Federal solicita esses dados desde 2024 e nunca recebeu resposta. Promotores entraram com ação judicial pedindo intervenção federal, acusando a prefeitura de negligência grave. Questionada sobre os atrasos, o silêncio na divulgação e a omissão diante das solicitações dos promotores, a gestão municipal respondeu com uma nota sobre metodologia e a afirmação de que os serviços foram mantidos — sem responder ao que foi perguntado.

Rio de Janeiro's municipal government released a census in early January counting 8,195 homeless people across the city—a modest 4.2% increase from the previous count in 2022. The study was published quietly on Data Rio, the city's official data portal, with no announcement on the homepage and no press campaign. The Social Assistance Secretary's office, which has seen its budget shrink by 7 million reais and now operates at its lowest funding level in a decade, made no mention of the findings either. Meanwhile, a separate study conducted by researchers at the Federal University of Minas Gerais, using 2025 data from CadÚnico—the federal government's registry for social welfare programs—counted 23,431 homeless people in Rio. That figure is nearly triple the city's official count, and it suggests the municipal census may be capturing only a fraction of the actual population living on the streets.

The discrepancy raises hard questions about what the city knows and what it chooses to measure. The Federal Public Ministry has been requesting this census data since 2024 and has received nothing. Prosecutors have filed a lawsuit asking the federal courts to intervene, accusing Rio's municipal government of gross negligence in its policies for the homeless population. The city has not responded to those inquiries. When asked why the 2024 census took so long to publish, why it received no public announcement, and why the city ignored the federal prosecutors' requests, the municipal government offered only a statement about methodology and a claim that all services have been maintained or expanded despite the budget cuts. It did not answer the questions.

The census itself paints a portrait of severe deprivation. The Center district had the highest concentration of homeless people counted—1,300 individuals—followed by Botafogo with 495. Most of those surveyed were men, and more than 70 percent were between 18 and 49 years old. Eighty-three percent identified as Black or mixed-race. The majority had not completed primary school; 10 percent could not read a simple note. When asked why they were sleeping on the street, the most common answer was family conflict, followed by alcoholism or drug use. Half said they still maintained contact with family members.

When asked what they needed to leave the streets, the overwhelming majority said employment. Seventeen percent said housing, and 10 percent said money. Only 1 percent reported being on the street to stay closer to work. The picture that emerges is not one of transience or choice but of structural abandonment: people cut off from family stability, lacking education and steady income, with nowhere else to go.

Food insecurity runs through the population like a current. Forty-five percent said they depend on donations to eat. Two percent reported having eaten from garbage. More than 30 percent said they had gone an entire day without food in the previous week. These are not abstract statistics—they describe the lived reality of thousands of people in one of Brazil's largest cities, people the municipal government counted at 8,195 but whom federal data suggests number closer to 24,000. The gap between those two numbers is not a rounding error. It is the difference between a problem the city believes it understands and a crisis it may not be measuring at all.

The Federal Public Ministry has been requesting census data since 2024 and has received nothing, filing a lawsuit accusing the city of gross negligence in homeless policies.
— Federal Public Ministry
La Conversación del Hearth Otra perspectiva de la historia
Inventor

Why would a city publish a census and then essentially hide it?

Model

Because the numbers are uncomfortable. A 4% increase in homelessness, even if modest, is still an increase. And if you're already cutting the social assistance budget, you don't want to draw attention to a growing problem.

Inventor

But the federal data suggests the real number is three times higher. How does that happen?

Model

The municipal census is counting people they can find and interview on the street or in shelters. The federal registry is broader—it's based on administrative data from welfare programs, which catches people who've registered with the government even if they're not actively on the streets that day. Different methods, vastly different results.

Inventor

So the city might genuinely not know how many homeless people there are?

Model

Or it might not want to know. The Federal Public Ministry has been asking for this data for two years and getting silence. That's not incompetence—that's a choice.

Inventor

What does the census reveal about who these people are?

Model

They're mostly men in their 20s through 40s, most are Black or mixed-race, most didn't finish school. Half still talk to their families. They want jobs, not charity. But they're also going hungry—a third of them didn't eat for entire days.

Inventor

And the city cut the budget for helping them?

Model

By 7 million reais. It's now at a 10-year low. The government says services were maintained, but you can't maintain services with less money. Something has to give.

Inventor

What happens next?

Model

The courts will decide whether to intervene. But the real question is whether Rio will ever count these people accurately, or whether it will keep the numbers low enough to avoid action.

Quieres la nota completa? Lee el original en G1 ↗
Contáctanos FAQ