« Sir Ringo! | Main | The Glass Ceiling »

December 06, 2006

Murders on Probation

Interesting number here:

Criminals under the supervision of the probation service have been convicted of nearly 100 murders and more than 500 other serious violent and sexual offences, including rape, over the past two years, according to official figures out yesterday.

100 over two years, eh? The actual number of murders in the country is somewhere in the 800 to 1,000 level isn't it ( 737 per year apparently)? Per year? So those on probation are responsible for 5% (to be accurate, 6.7%) of all murders?


Broken down, the statistics show that people on probation were convicted of 60 murders between 2004/05 and there were convictions for a further 38 the following year.

This is convictions.

The final number may be even higher as more than 250 cases are still waiting to come to trial.

Now we can play any number of games we like with these numbers. What's the clear up rate for murder? 80%? 20 %? I have no idea at all to be truthful. How many of those 250 cases will result in convictions? (They amount to fully 12.5% of all murder cases in the UK from those two years...17% to be accurate)). How many accused of murder, charged with it and in fact did it do not get convicted? (Must be some as justice isn't perfect but whether it's 0.01% or 50% I have no way of knowing).

It's Dean Baker over in the US who keeps insisting that economic numbers have to be put in a way that people intuitively understand.

So how about trying it with these crime numbers? Somewhere between 5% (those already convicted) and what, 25% (those convicted, on trial and those wrongly found not guilty or not even charged?) of murders are committed by those on parole?

As the article says, yes, that does sound like a good reason to reform the probation service. At the very least actually.

December 6, 2006 in Law | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Murders on Probation:


There's a few errors here. A very minor one is that you didn't take off those who were find guility but didn't do it. A more major one is that you've mistook the 250 cases waiting to come to trial as being murder cases, when the article doesn't say that (and given the magnitude I would be surprised if it were so).

Tim adds: Fair comment. I think I did point out that we could play games with these numbers, didn't I?

Posted by: Matthew | Dec 6, 2006 10:24:33 AM

One other minor point is tat I think you should also compare the murder rate of those on parole with those not, as if, say, it was no higher than these murders are only additional insofar as not locking everyone up lets murders get away with it.

I'ts in fact a lot higher. If it's 60 and 38 over two years, so say 50 a year, and there's 150,000 people on parole, then its 1/3,000. The nation's murder rate is going to be something like 700/50m = 1/70,000, so it's 23 times higher for those on parole.

Hence the "extra" murders caused by those on parole is about 47, not 50.

What is important, I think, is whether there is a consistent pattern in what the offenders were on parole for. If they were all paroled for violent crime, then the murder rate would be very high.

Posted by: Matthew | Dec 6, 2006 10:40:03 AM

It's important, I think, to be clear about whether we're discussing people who're under the supervision of the probation service because they're serving a community sentence for something that the court didn't think merited custody or they're being supervised because they've been released on licence from a custodial sentence.

I'd also want to know more about the circumstances of both the murderers and the murders; if, for example, someone's released on licence half way through his sentence for burglary and goes on to kill his partner during a domestic dispute (which most murders in the UK are, AFAIK) then it's not immediately clear how the probation service could have stopped him or that making burglars in general serve the extra six or eighteen months of their sentences and then releasing them without supervision would necessarily avoid such situations.

Posted by: Not Saussure | Dec 6, 2006 11:16:05 AM

One number by itself does not tell you anything about whether the factors influencing that number should be "reformed". You need a comparison to the relevant alternatives. The correct comparison may not be to the general population at large.

Assume that murders are committed largely by some certain fraction of society (defined by factors other than that they have committeed murders). You need not sign on to any theory of race, class, or politics to observe that murders are mostly the product of ________. (In the US, it's abusive men and people involved in the drug trade; I don't know what it is in the UK.) The reason why such people commit murders is not the point - only that we can predict that a large percentage of murders will come from that group. Assume also that, for every murderer incarcerated, a potential murderer takes their place - someone takes over their part of the drug trade, becomes involved in their criminal gang, grows up in their shitty neighborhood, or whatever.

If those assumptions are correct, then the right comparison is between convicted murderers who have gone through the criminal justice system and "potential murderers" (people from the same walks of life that produce most murderers) who have not. 5% of murders are committed by the former, but 75%, 80%, 90%, or whatever are committed by the latter. Then you can ask whether, for someone "at risk" for violent crime, whether having "gotten it out of your system", been incarcerated, gotten older, been to whatever degree rehabilitated, and being on parole is more or less of an indicator of likely future criminality than merely being part of the same community without prior history. That ratio (crimes by the former to crimes by the latter) is certainly going to be a lot closer than if you throw into the "potential" pool the millions of people whose life circumstances are simply never going to put them in a situation in which a murder is even an imaginable outcome; it's possible the ratio will tip the other way (in favor of the parollees).

At any rate, pretending that every member of society is equally likely to commit a murder, and then comparing those who have been caught to those who will simply never be in any such danger, is unrealistic - whether you believe those differences are genetic, economic, social, or what have you.

Posted by: Kevin T. Keith | Dec 6, 2006 4:05:45 PM