Prison Works

FRIDAY, JULY 16, 2010

Prison Works

More good news

As regular readers will know, Tyler is one of those New Jerusalem deniers. And when it comes to Labour's good news crime stats, he's always been particularly sceptical (eg see this blog).

Yesterday we got the final instalment of those stats - the annual update of crime for 2009/10 as recorded by the police, alongside the results of the British Crime Survey (ie the crime opinion poll). And - surprise, surprise - both showed another fall. Naturally the BBC is very pleased:

"The number of crimes committed in England and Wales has fallen to its lowest level since records began in 1981...

...The statistics show that crime has fallen by 43% since 1997, leaving Labour able to claim that its level is lower now than when Tony Blair first entered Downing Street."

Oh, no wait... that second sentence only appears in the BBC's sister publication (the Grun). Anyway, the left is delighted, especially since it supports their argument that we don't need to build any more prisons. Crime is already plummeting they say, so why waste money on prisons? The cash would be much better spent on restorative arts projects for rapists.

But whatever the BBC and the Grun say, our splendid new Home Secretary clearly shares our scepticism. Theresa says:

'Any reductions in crime are welcome, but we know these statistics offer a partial picture about the true level of crime. And there are many offences, including antisocial behaviour, which are not always reported or fully recorded, but which ruin too many lives...

We are determined to restore trust in crime statistics and are currently considering how they should be collected and published in future. We are working with the UK Statistics Authority and others to consider this carefully.'

Good. We look forward to getting stats that at least vaguelyapproximate our actual experience of rising crime (see previous blog, andthis blog for an account of the most recent criminal outrage to hit Tyler Towers).

But just for a moment, let's suspend our disbelief. Let's for once listen to the mellifluous sermons of Bishop Snow and the Rev Easton, and let's assume the existing stats are the unvarnished truth. Let's assume that crime reallyhas fallen by 43% since 1997. What does that tell us?

Well, the obvious, jump-off-the-page, shout-in-your-face, message is that Michael Howard was right - prison really does work.

Let's take a look. The following chart shows how measured crime has fallen since 1995, on Labour's preferred British Crime Survey measure. It also shows how the number of prison places (strictly, the prison population) has risen over the period. For ease of comparion, both series are shown in index form, with 1995=100.

As we can see, as prison numbers have increased, measured crime has fallen (the gaps in the early years reflect gaps in the reported BCS numbers).

Let's look at the statistical relationship more closely. The following chart shows a simple regression of each year's crime numbers against the same year's prison numbers:

Now for those of you who never did get round to reading Teach Yourself Statistics, those regression results indicate a very close relationship between the two series. A very close relationship indeed. An R² of 0.88 is equivalent to a correlation coefficient of 0.94 - as near as damn it, a perfect fit.

And specifically, what our estimated regression equation says is that each additional prisoner is associated with 275 fewer crimes. In other words, prison works. Bigtime.

What? What's that you say? Correlation does not imply causation?

Well, no, to be sure, you must be right.

Yes. Yes indeed. Well remembered.

I mean that extraordinarily close relationship could be caused by all sorts of other things.

Like... ummm... well, maybe the police have become so much more efficientthat that they are preventing crime before it even happens. Could that be it?

Hmmm... no... that can't be right... not our police with their plummeting detection rates and their social worker Chief Constables.

No, it must be... aahhh... yes, better burglar alarms and Krooklocks on cars.

And yes, there must be some truth in that. Mustn't there?

Wellll... you really can't help spotting that those crime stats went on rising right up to the very point where M Howard got his massive prison programme going (see chart at top). Surely burglar alarms and Krooklocks were already improving long before then. How do you explain the sudden break?

Of course, it is possible that all those sermons from the BBC and C4 pulpits have finally done the trick. That the sinners hath repentethed, and the lions hath laid down with the lambs.

Yes, that's possible.

Although it doesn't yet seem to have happened round Tyler's way.