Rio de Janeiro Crime Statistics 2026: What the Data Says

The two trends everyone reports as one

Rio's crime data in 2026 is not a single story and anyone telling you “Rio is getting safer” or “Rio is getting worse” is quoting half of it. Violence against people is falling hard: the Instituto de Segurança Pública recorded 1,778 violent deaths in Rio de Janeiro state in the first half of 2026, down 9.9% on the same period of 2025, and the January-to-April count of 1,242 was the lowest for those months since 1991. Street robbery fell 20% over the same half-year, to 24,697 recorded occurrences.

Theft of property, meanwhile, went up. Vehicle theft rose 14.5% across the half-year to 13,436 cases and cargo theft rose 10.1% to 1,733, even though both fell sharply in June alone. And the single indicator that describes most visitor experience — phones — has been rising for years: the Brazilian Public Security Yearbook recorded 72,187 mobile phones robbed or stolen in Rio state during 2025, up 22.7% from 58,813 in 2024, at a time when the national figure fell 7.7%. Rio was one of only two states in Brazil where that number grew.

If you want one sentence: the chance of being hurt in Rio is falling, and the chance of losing your phone is not. That is a very different risk profile from the one the headlines imply, and it changes what you should actually do differently on the ground.

Violent death: the falling number that barely touches your trip

“Letalidade violenta” is ISP's umbrella indicator: intentional homicide, robbery followed by death, bodily harm followed by death, and deaths resulting from police intervention. It is the number international coverage of Rio is built on, and it is genuinely improving — the first four months of 2026 produced the lowest count in 35 years of the series.

It is also the number least relevant to a visitor, because of where it happens. Violent death in Rio is concentrated in specific territories contested by armed groups, almost all of them far from the Zona Sul, Centro's daytime core, Barra or the airport corridor. A tourist in Ipanema is not statistically exposed to the thing that drives this indicator, in the same way that a tourist in central Chicago is not exposed to that city's homicide geography.

The honest counterweight: a six-month trend is not a promise about next month. In July 2026, violent lethality jumped 25% after six consecutive months of decline. Police operations, disputes between armed groups and single incidents move this number abruptly, which is exactly why we treat our same-day safety page as a planning-context page and not a forecast.

Street robbery: down 20%, and still the number that matters

“Roubo de rua” is ISP's composite of robbery against pedestrians, robbery on public transport and robbery of mobile phones. It is the closest thing in the official data to “what might happen to a visitor”, and it fell 20% across the first half of 2026 to 24,697 recorded occurrences — 3,820 in June alone, down 21.9% on June 2025.

That is real improvement and it is worth saying plainly, because the number has been falling steadily enough that advice written in 2022 now overstates the risk. The municipal patrol programme publishes its own perimeter-level figures showing the steepest falls in parts of Botafogo, Centro, Tijuca and Campo Grande — the stretches with the densest targeted patrolling.

Where it has not improved is composition. Phone robbery now accounts for close to 38% of all street robbery in recent ISP reporting, and Centro has seen the sharpest growth of any part of the city — in one January-to-August comparison, over 1,300 phone robberies against roughly 610 the year before. So street robbery is less likely than it was, and more likely than ever to be specifically about the thing in your hand.

Four words you have to separate: roubo, furto, letalidade, registro

Brazilian crime data distinguishes roubo from furto, and English-language coverage collapses them constantly. Roubo is taking by force or threat — a snatch, a mugging, a weapon. Furto is taking without violence — a phone lifted off a café table, a bag out of an unattended car. They are separate legal categories with separate statistics, and the two move independently: Rio's roubo numbers are falling while furto numbers rise. If a figure does not say which one it is, you cannot use it.

Letalidade violenta, as above, is a bundle of four causes of death and not a homicide rate. Comparing it to another country's homicide rate overstates Rio by including police-intervention deaths and robbery-related deaths.

And every figure on this page is a registro count — a tally of police reports actually filed. Stolen phones are massively under-reported by visitors, who leave the country rather than spend a morning at a delegacia, so the true number is higher than 72,187 and the true improvement in street robbery is less certain than the percentage suggests. A falling count can mean falling crime, falling reporting, or both.

Five rules for reading any Rio crime number

First, check whether it is the state or the city. ISP publishes Rio de Janeiro state, which includes Baixada Fluminense municipalities with very different risk profiles from the city, and journalists switch between the two mid-article. A state figure tells you almost nothing about Leblon.

Second, check the denominator. An absolute count rising in a metropolitan region of twelve million is not the same as a rate rising, and most of the figures quoted in travel coverage are raw counts with no population base at all.

Third, check the comparison period. Month-on-month in Rio is noise: a single operation, a holiday weekend or a heatwave moves it. Year-on-year for the same months is the minimum honest comparison, and a rolling twelve months is better.

Fourth, check which indicator. “Crime is up in Rio” is usually true of some indicator in some month — there are dozens. Ask which one, then ask whether it has any mechanism that could reach a visitor. Cargo theft rising 10% is a logistics story, not a holiday story.

Fifth, separate data from geography. Rio's risk is more spatial than it is temporal: which street, which hour, which direction you walk out of the metro matters more than which quarter the data covers. That is what our neighbourhood-level safety map is for, and no citywide statistic substitutes for it.

What the data actually implies for your trip

Treat the phone as the asset at risk and plan around it specifically — that is what the numbers say, and it is a much narrower, more solvable problem than “Rio is dangerous”. Our same-day safety check covers the official sources to read before you go out, and the neighbourhood safety map covers 45 areas by day and night, which is the resolution these statistics cannot give you.

Second, buy travel insurance that actually covers electronics theft, and read the clause on unattended items and police reports — most policies require a boletim de ocorrência filed in Brazil, which means you need to know where the tourist police station is before you need it. We cover what to do in the first hour in robbed in Rio: what to do, and the policy comparison in our Rio travel insurance guide.

Third, don't let the improving numbers talk you out of basic habits. A 20% fall in street robbery still leaves nearly 25,000 recorded robberies in six months. The point of reading the data is not reassurance — it is working out which of the many things you have been told to worry about are the two that are actually load-bearing.