London isn't 'propping up' falling crime in England and Wales

"Nationally crime is falling ... yet the true picture is being masked by a dramatic fall in the number of recorded offences in London"

The Times, 7 July 2014

The Times and Daily Mail reported last week that official police crime figures were expected to show London propping up the rest of the country. The figures are now out, and they don't.

London's Deputy Mayor for Policing is the original source. In a speech last month he was already making predictions about what the new figures would say:

"In fact I predict that the next quarterly ONS statistics for recorded crime will demonstrate that London's performance is propping up the rest of the country — where I note that some forces are now seeing crime rise.

"Without this progress in London, it might not be possible for Home Office Ministers to laud a national crime reduction this year."

The Deputy Mayor is referring to police recorded crime figures, although Home Office ministers could just as well point to the more reliable Crime Survey figures published alongside these to show that crime is falling. While the Metropolitan police are recording among the sharpest falls in crime, it isn't the case that it's propping up the rest of the country, since more than two thirds of other police forces also recorded falls in crime.

In addition, the police figures are considered unreliable and aren't a good measure for comparison across the country.

Jumping the gun and getting it wrong

It's possible, as the newspapers and the Deputy Mayor have shown, to get hold of police crime figures before they become official statistics like the ones published today. That's because they come from local police forces, who may release the data for their own areas beforehand.

But the police's figures are taken from live databases which are constantly in flux and don't have the same quality checks as provided by the Home Office when it compiles them. For instance some crimes can later be downgraded ('no crimed') if an investigation reveals no offence had actually occurred.

In practice, this has led to inaccurate reporting. There were unusual references to shoplifting in North Wales having increased by 22% since last year. The latest figures show shoplifting there up 9%, and it's difficult to see where this could have come from.

London leading, but not reliably

The Met Police - which the Deputy Mayor oversees - can certainly boast one of the largest annual falls in recorded crime. There was a 9% fall in recorded crime in the Met area since last year, a fall beaten only by Northamptonshire Police which saw a 13% fall.

But this doesn't mean that the Met necessarily saw the biggest fall in crime. Police figures are currently considered unreliable and were stripped of their quality mark earlier this year, following evidence suggesting their recording practices weren't up to standard.

An obvious example of this is Kent police. It saw the largest annual increase of any force in recorded crime this year: up 11%. This doesn't mean Kent is suddenly the crime hotspot of England as well as its garden. The force has notably improved its recording practices following a recent inspection by HM Inspectorate of Constabulary, and this is considered a likely explanation for the rise.

So even though the Met's in the leading pack, it's not in a well timed race.

London not the linchpin for falling crime

Even though the Met area (and the City of London Police as well in fact) has recorded many fewer crimes than last year, it's not the linchpin for falling crime nationally. More than two thirds of police forces oversaw falls in recorded crime. If you take the Met police out of the equation, recorded crimes would still have fallen overall across the 43 forces.

The only areas in which recorded crime is significantly rising - Kent excepted - aren't even geographic: they are fraud offences recorded by the recently launched Action Fraud. They rose by 80% since last year - on the face of it an enormous increase. Again, this is more a reflection of better recording and more reporting: fraud could be increasing anyway, but these figures aren't good enough to show that.

There's also an argument that recorded crime figures should take into account population changes - i.e. the crime rate. Running all of the analysis again using these figures produces a similar story.

So the claims are a bad presentation of bad data.