Decline of the London murder

The city's streets are safer than any time for decades - perhaps centuries

Magazine illustration of the discovery of one of Jack the Ripper’s victims. From Museum in Docklands...

One of the most potent myths in British public life is that crime is spiralling out of control. Nigel Farage and Robert Jenrick talk breathlessly about “lawless Britain”. Knife attacks, gangs, TikTok brawls: the imagery is vivid, the anecdotes endless. I’ve written before about how NHS data on patients treated for violent crime and knife attacks shows a multi-year low for the UK. But there is a monthly dataset that tells an extraordinary story: London murders. As the below chart shows, they now stand at the lowest since monthly data started being published in 2003 - and lowest murder rate since annual data started almost 30 years ago.

The ‘London Has Fallen’ meme on social media bears as much relation to reality as Gotham City does to New York. The actual streets of our capital city are safest for decades, if not centuries. Here’s the data.

Murder is the hardest, most reliable crime metric. Other crime has been going to in London: mobile phone theft, watch theft. But other crime is going down fast: like stabbings. Another reliable dataset is NHS hospital data for people admitted for ‘assault by sharp object’: lower than any time since this data started being released 13 years ago. Manchester has not been so lucky.

Over the last few years, police have also to log every instance of a gun being fired. It’s a short dataset but is trending down - and is hard to reconcile with the ‘crimewave’ narrative.

London has become a test case for political reality-stretchers: can they tweet out isolated incidents (especially with a non-white criminal) and portray the capital as a lawless dystopia when the figures suggest otherwise? How can the city be so dangerous if it now be experiencing what is probably the lowest rate of murders - as a share of population - than any time since records began?

I say ‘probably’ because the deeper you go into London’s murder data, the murkier it becomes: a statistical cold case stretching back through layers of bureaucracy, boundary changes and shifting definitions of what, exactly, counts as a “murder.” In a way, it’s the most reliable crime figure: bodies are always counted. But what counts as murder isn’t always the same.

A murder mystery in the data

The modern series begins in Dec03, when London recorded 216 killings for the previous 12 months. Back then, Ken Livingstone was mayor, Tony Blair was warning about the “respect” agenda. Today that rolling total is below 100. That’s a fall of more than half, even as London’s population has grown by nearly two million.

That is, in any rational sense, a public-safety triumph. But to say “lowest since records began” doesn’t tell the full story. The real question is: lowest since when? Because 2003 isn’t the beginning of the story of London murder: it’s just the point where the spreadsheets start.

To go back further, you leave the world of PDFs and open data and step into the archive room: literally, in some cases. The trail leads to National Archives boxes marked MEPO 20: the Metropolitan Police Registers of Murders and Violent Deaths, 1891 to 1966.

The archaeology of body counts

From those registers you could, in principle, count each homicide per year for the Metropolitan Police District — that’s almost all of Greater London, but not quite. The City of London, the Square Mile, keeps its own police force, so you have to add those numbers separately to get a proper London total.

After 1966 the MEPO registers stop, and you must switch to the Home Office Criminal Statistics annuals: thick printed volumes, bound in blue, produced by HMSO each year. There, among pages of burglary and theft tables, you’ll find one line labelled “homicide: murder, manslaughter, infanticide.”

In other words, not just murder - the legal term for intentional killing - but all forms of homicide. That’s fine: homicide is the standard measure for long-run trends, because prosecutions can reclassify cases from one category to another.

Add together the Metropolitan Police District and the City of London for each year, and you can reconstruct London’s annual homicide totals for the late 1960s through to the 1980s. After that, you can check your sums against the Home Office Homicide Index — a definitive database that allows analysis by region, including London, from the early 1980s onward.

The data story, then, would have to come in three acts:

• 1891–1966: the handwritten murder registers (MEPO 20)

• 1967–1989: the Home Office Criminal Statistics volumes.

• 1990–present: digital records and the ONS/Homicide Index.

Add them up, clean them, and — in theory — you can build a continuous series for London murder going back more than a century. But then you encounter….

The perils of definition

What counts as London? The Metropolitan Police District is not the same as Greater London. The Met’s boundaries were fixed in Victorian times, long before London government was reorganised in 1965. If you want a modern comparison, you need to be clear whether you’re talking about the MPD + City (the policing geography) or the London region (the statistical geography used by the Office for National Statistics). They’re similar, but not identical.

Then there’s the question of what, exactly, you’re counting. “Murder” and “homicide” aren’t the same. Murder requires intent; homicide also includes manslaughter and infanticide. Some tables are based on offences initially recorded by police, others on offences currently recorded after investigations and reclassifications. Those reclassifications can change history: the Shipman case alone added scores of retrospective deaths to national figures in the early 2000s.

Even the counting year matters. The police used calendar years for decades, then switched to financial years for some series. When building a long-run record, you have to choose one and stick with it.

All of this makes London homicide data less like a tidy Excel chart and more like an archaeological dig: layer upon layer of definitions, boundaries, and bureaucratic sediment that has to be carefully brushed away before the underlying pattern appears.

What the fragments tell us

Piecing together those fragments, the emerging picture is this. London’s homicide count rose through the post-war decades, peaking in the early 2000s. Since then, despite occasional spikes, it has trended sharply downwards.

In 2003, the Met recorded over 200 homicides. By 2012, during the London Olympics, the figure had dipped below 100. In 2018 it ticked up to 166, leading to headlines about London having a higher murder rate than gun-strewn New York. But even that was far below the early 2000s peak. And now, in 2025, the twelve-month rolling total has dipped below 100 - and the murder rate, per capita, fell to the lowest recorded.

Most crime is unrecorded and police are working to close this gap. The irony is that, as they get better, the trend looks worse: an improvement in reporting would seems to show crime surging. So how to adjust for this? One is the 25,000-household Crime Survey of England & Wales, regarded as the gold standard. But Nigel Farage has declared he has no faith in this survey (which contradicts his crime-is-surging narrative).

That’s why it helps to enlist a third, more reliable source: NHS hospital data. Anyone admitted for assault, or wounding by a sharp object (ie, knife) is recorded as such.The Crime Survey asks people if they have been the victim of crime. It excludes surging shoplifting (not counted as crime against a person) but shows that London has become less dangerous not just overall but relative to the rest of the country. Like all capital cities, it’s riskier - but the risk premium is closing.

You can see the trend. This suggests we are likely back to — or below — the murder counts of the 1970s. The MEPO registers indicate that mid-century London averaged between 120 and 160 homicides a year, depending on classification. If that’s right, the present figure is indeed the lowest in living memory — and quite possibly the lowest since before the Beatles.

To confirm it, someone would need to do the legwork: tally each MEPO 20 entry, cross-check with City of London Police archives, and compile the Criminal Statistics tables from 1967 onwards. That would take a few weeks with a scanner and a spreadsheet. But the trend is already clear enough to say: whatever else is going wrong in Britain, the capital’s murder rate is going right.

Men in medieval London were armed; brawls often ended up in bloodshed. Picure and more from Cambridge University Violence Research Centre,

A brief history of London murder

Memoirs from late medieval London testify the danger in the city’s streets, squares, markets and wharfs. Men there were armed. Pepys’ diaries testified as to the constant risk of robbery and assault. Data reconstruction efforts suggest that in the 1600s, London’s murder rate was about 20 times what it is today. The introduction of the Metropolitan Police calmed things and the postwar decades - 1950s and early 1960s - were remembered for stability. Even then, murder totals seem to have hovered around a 100 a year: similar to today’s number, but from a much smaller city.

In the 1970s and 1980s, murder numbers began to rise, paralleling increases in other violent crimes. The reasons are complex: demographic change, heroin and gang culture, looser policing after the Scarman reforms. But by the 1990s and early 2000s, murder had become the defining urban fear.

Then something changed. Partly it was policing — the Met’s focus on knives and hotspots, the use of CCTV, DNA, and data analytics. Partly it was medicine: victims of assaults now survive wounds that would once have been fatal. Partly it was demography: the youth bulge that drove crime in the 1990s had passed. Forensics make it harder to get away with murder: a fragment of the killer’s DNA on the victim can be enough to convict, even 20 years later. And also, it’s the city itself. London has gentrified, aged, and, for all its tensions, grown richer. Safer. More liveable. And, like all major cities, with CCTV everywhere. A summary of the forces behind murder, from drugs to honour-killing, can be found in this magnificent Home Office briefing.

Reconstructing the full London murder data series means scanning a century’s worth of Criminal Statistics tables and transcribing them. It’s not impossible, just tedious. Once done, London’s homicide trend from 1891 could be plotted, to determine at a glance whether today’s rate really is the lowest in history. Perhaps a student looking for a decent school project might do so. It would be a service to history — and a useful corrective to the doom narratives.

It matters because perception drives policy. If the public believes London is sinking into lawlessness, we will demand policies suited to a collapsing city: more prisons, harsher sentences, fewer liberties. If, instead, we accept that we are living through a remarkable long-term decline in lethal crime, we might ask a different question: what has gone right? I’ve never seen any estimates for London - in medieval, Regency or modern-day - murder putting it at below 1 per 100,000 population. Yet that’s where we are now.

The city of Sherlock Holmes, Jack the Ripper and the Krays, home of gangland myth and tabloid panic, has always had a morbid fascination with murder. But look past the headlines and you find a quieter and hugely important truth: the city is winning the battle over violent crime. It’s not a bad time to be a Londoner.