This bank holiday weekend marked the 60th year of Notting Hill Carnival, and with it came the headline that arrives like clockwork every August: 636 arrests, a fatality, and a fresh round of commentary asking whether Europe’s biggest street party has a crime problem. Cue the usual pearl-clutching. Cue the usual “here we go again.” And cue, right on schedule, the Met’s own admission that the event itself was “overwhelmingly positive and celebratory.”
So which is it? Is Carnival a two-day crime wave, or is it two million people dancing peacefully while a small number of arrests get inflated into a national talking point? I’ll give you a hint: it’s not really a question about Carnival at all. It’s a question about who Britain chooses to watch, how hard it watches them, and a sixty-year relationship between the British state and Black Britain that Carnival has never been allowed to escape.
A carnival born out of racism, not despite it
Let’s start with a fact that gets conveniently dropped from most of this year’s coverage: Notting Hill Carnival exists because of British racism, not in spite of it. It emerged out of the 1958 Notting Hill race riots, in which white mobs hunted Black residents through the streets for several nights while the Metropolitan Police looked on, doing what the Met has historically done best in these situations: not very much.
Claudia Jones, the Trinidadian activist and journalist often called “the mother of Notting Hill Carnival,” organised the first indoor Caribbean Carnival the following year, partly as an act of communal healing. The outdoor street event that grew out of this in the 1960s was never just a party. It was a claim to public space by a community that had just been told, violently, it didn’t belong on that street at all.
Keep that origin story in your back pocket, because it explains almost everything else in this article.
1976 and the birth of a “problem”
The turning point most historians point to is 1976, when confrontations broke out between young Black attendees and police after years of aggressive “SUS” stop-and-search laws that let officers search anyone on nothing more than a hunch; laws disproportionately aimed, funnily enough, at Black youth. From that year on, Carnival became fused in the press and the political imagination with an idea of Black male criminality that Britain has never quite managed to let go of.
Stuart Hall and his colleagues, writing in the landmark Policing the Crisis (1978), argued Black youth in this period were being constructed as the folk devils of the age — a moral panic, to use Stanley Cohen’s term, in which a group is defined as a threat, that threat is amplified by press coverage, and disproportionate control measures follow as a result. Crucially, this had far less to do with actual crime rates than with wider anxieties about race and a changing Britain. Carnival, as the most visible annual gathering of Black Britain, became the perfect lightning rod. Nearly fifty years later, it still is.
Doing the maths on “636 arrests”
Here’s where I want to slow down, because this is the bit the headlines never do.
Around two million people attend Carnival over the two days. This year’s total, once all the arrests had been processed, settled at 636. That’s a figure that sounds enormous in isolation. But run the numbers, and it works out at roughly one arrest for every 3,145 attendees. That’s about 0.03% of the crowd. The Met’s own statement on this year’s policing operation makes exactly this point, framing the arrest total as a small proportion of the roughly 2 million people who attend.
Now compare that to ordinary, everyday policing. In 2023/24, the Metropolitan Police made 96,426 arrests across the whole of London, which works out at roughly 264 arrests on an average day, every day of the year, across the entire city. Carnival, concentrated into a few square miles over 48 hours, produces something in the same order of magnitude as what the Met arrests across the whole of London in a normal couple of days — except this time every single one of those arrests gets a headline, because they all happened in the same postcode over one specific weekend.
Or take it from the other direction. Nationally, England and Wales recorded around 720,000 arrests in 2023/24: roughly 12 arrests per 1,000 people per year. Cleveland, the force with the single highest arrest rate in the country, still only managed 19.8 arrests per 1,000 residents across an entire twelve months. Carnival’s rate — 0.3 arrests per 1,000 attendees, across the single busiest weekend of the year for that patch of London — is lower than what an ordinary English police force manages against its own resident population over a full year, even before you account for the fact that Carnival is deliberately saturated with more officers per square metre than almost anywhere else in the country manages on a normal day.
None of this is a trick with statistics. It’s the opposite: it’s what happens when you actually divide the numerator by the denominator instead of just reading out the numerator in a headline. The “problem” isn’t the rate of offending. It’s that Carnival is one of the only events in Britain where every single arrest, however routine, gets tallied up publicly and read back to the country as evidence of disorder.
And here’s the bit that should really give the headline-writers pause: this year’s Met also described a “welcome reduction” in serious violence for the second consecutive year — the same year the arrest total hit a record high. If arrests were a clean proxy for danger, that sentence shouldn’t be possible. More arrests and less serious violence, in the same twelve months, is only a contradiction if you assume arrests measure crime rather than measuring policing. Once you drop that assumption, it makes perfect sense. A more intensive, better-resourced operation catches more of the low-level stuff (drug possession, minor weapons offences, breaches of existing conditions) while the thing everyone is actually frightened of, serious violence, keeps falling.
What an arrest figure actually measures
An arrest figure is not a neutral measurement of how much crime happened. It’s a measurement of how much policing happened. This year the Met deployed more than 7,000 officers into those few square miles of west London and, for the first time, layered live facial recognition on top. Put that much surveillance into any space on Earth, and you will generate hundreds of arrests, almost regardless of the underlying behaviour of the two million people walking through it.
Compare that to the exclusion zone the Met threw up around central London during the Euro 2024 final, purely on the possibility that football fans might misbehave. Nobody ran headlines about “hundreds of arrests at the football” that summer, despite dozens of officers being injured at genuinely disorderly events elsewhere in the same period. The differential isn’t really about behaviour. It’s about which crowds get treated, in advance, as a policing problem to be solved rather than an event to be enjoyed.
None of this is to wave away the real harm — 68 weapons were seized this year, including a handgun, and a man tragically died after fleeing officers into a canal near Meanwhile Gardens, a death the IOPC is now, rightly, investigating. But it is to say that “636 arrests” does an enormous amount of rhetorical work in a headline, and that work only functions if nobody stops to ask how the figure was produced, or what it looks like once you put it next to literally any other British policing statistic.
The facial recognition question
This year’s Carnival also saw a major expansion of live facial recognition, credited with well over 120 arrests and used to check attendees against the sex offenders register. What gets left out of the celebratory police statements is that LFR technology has repeatedly been shown, in independent research, to misidentify Black faces at higher rates than white ones — Big Brother Watch has documented this for years, and has specifically flagged the Met’s repeated deployment of the technology at Notting Hill Carnival. And yet where does the Met choose to pilot and intensify this technology most aggressively? Not the Chelsea Flower Show. Not Wimbledon. Carnival.
I don’t think that’s a coincidence. I think it’s a fairly precise illustration of who this country has decided is worth watching most closely. The same expanding-surveillance-under-a-security-pretext logic I wrote about in the War on Terror’s enormous effect on racism, just aimed at a different “threat” this time.
It means a meaningful chunk of this year’s “record” arrest figure is itself an artefact of a new surveillance tool being switched on for the first time, not evidence that Carnival got any more dangerous than it was last year. Of the 636 arrests eventually recorded, 121 came directly from facial recognition hits. Strip those out, and the “record-breaking” framing looks a lot less like a story about rising disorder and a lot more like a story about a new camera being switched on.
The stop-and-search picture
Facial recognition isn’t the only place this shows up. London Assembly data on the 2023 Carnival — the most detailed breakdown publicly available — found 812 people stopped and searched, of whom 283 were recorded as Black and just 60 as White. Almost half the total, 396 people, had their ethnicity logged as “unknown,” which is its own quiet scandal: nearly one in two stop-and-search records at the country’s biggest Black cultural event apparently weren’t filled in properly.
This isn’t a Carnival-specific quirk. A landmark 2026 study by King’s College London and MOPAC, analysing 150,000 stop-and-search interactions across the whole of London, found Black Londoners were stopped and searched 3.4 times more often than white Londoners overall. In addition, in the 24 worst-affected wards, that gap widened to a Black Londoner being up to 48 times more likely to be stopped than a white one.
The same research found that when officers wrote down their grounds for stopping someone, those grounds were reliably vaguer for Black Londoners than for white ones — “wearing concealing clothing,” “being in a high crime area,” phrases that could describe almost anyone doing almost anything, but which somehow attach themselves to Black bodies far more often than white ones.
Put Carnival’s own stop-and-search numbers next to that citywide pattern and the picture stops looking like a coincidence. It starts looking like Carnival is simply where London’s ordinary policing habits show up at their most concentrated and most visible, which is exactly the argument I made at greater length in is the UK institutionally racist?: the disparity isn’t confined to one bad event or one bad year, it’s baked into the everyday routines of British policing, and Carnival just happens to be the one weekend a year those routines get photographed.
“But what about the stabbings?”
I can already hear the objection. What about the knife crime? What about the sexual offences? Surely a few hundred arrests reflect something real?
Sure, some of it does. Carnival is not immune from the same social problems as the rest of the country, and pretending otherwise helps nobody. But notice how selectively this argument gets deployed. Nobody demands a “reckoning” about knife crime after Carnival’s much smaller, predominantly white-organised festival counterparts.
The demand for Carnival specifically to justify its own existence, year after year, in a way almost no other British mass event is required to, is itself the phenomenon worth studying. That’s not a coincidence either. It’s what six decades of “Black event equals security risk” framing looks like when it’s been fully internalised. Right down to the maths nobody bothers to do before printing the headline.
Finally…
Carnival’s organisers, and the Met’s own Commander Claire Smart, have pointed to a genuine and welcome fall in serious violence over the last two years, credited to closer cooperation between police and Carnival organisers rather than sheer weight of numbers. That’s worth holding onto — it suggests an alternative to saturation policing is at least being tested, however belatedly.
But it’s worth asking why it’s taken sixty years and immense community pressure to get even this far. And it’s worth asking why “hundreds of arrests” still does so much heavier lifting in the British imagination than “0.03% of two million people” — or “two million people celebrated peacefully” — ever will. That imbalance didn’t start with this year’s headlines, and on current form, it won’t end with them either.
