Mr Kenneth Clarke was quite right to say that short prison sentences are not effective but, with the practised lack of logic of a man who has spent far too long in politics for the good of his own mind, he has drawn precisely the wrong conclusions from it. His error will cause much unnecessary suffering.

There are indeed many arguments against short sentences. The recidivism rate after such sentences are completed is very high. They pose large administrative costs on the prison system. They do not reassure victims that the suffering or loss inflicted upon them by criminals has been taken seriously by the state. They discourage and demoralise the police, who labour mightily, if mainly bureaucratically, to procure a conviction for very little result. They promote intimidation of witnesses.

Remember, I'll been walking the same streets as you in six weeks' time

has deterred many a crucial witness (criminals having absolutely no doubt about the value or effectiveness of deterrence).

In the sub-culture from which many criminals live, they are a badge of honour rather than a mark of shame. And if rehabilitation were one of the aims of prison - personally, I think it is a very minor one - short sentences would do nothing to contribute to that aim.

But it is quite wrong to suppose that if something is not very effective it has no effect at all. Short prison sentences are ineffective by comparison with long ones, but that is not to say that they are ineffective by comparison with no prison sentences at all.

It is a fact that a large proportion of crimes are committed by a relatively small number of people. It is not unusual for career criminals to commit a hundred or more offences a year. Therefore, keeping them in prison for six weeks, say, prevents the commission of 12 crimes. Of course, if they were kept in prison for four years, 400 crimes would be prevented. But it is better to prevent 12 crimes than no crimes at all.

The only possible argument against the logical response to a high recidivism rate after short sentences being a lengthening of sentences is that something more effective than imprisonment exists: but it doesn't. Community sentences (except if used very selectively indeed) are quite useless.

Let us examine the ways in which our grossly dishonest official class, of which Mr Clarke is perhaps a victim, has sought to persuade us otherwise. Success is often measured by the percentage of "completion" of a community order, for example of probation. But from the law-abiding public's point of view this is irrelevant.

What is important is how many crimes people on a community order commit while they are on it or soon afterwards. And the fact is that the recidivism rate, as measured by the re-conviction rate, is very nearly as high as that of imprisonment after short sentences, despite the fact that the latter are probably slightly worse to begin with, which is why they were sent to prison. (Almost all prisoners are graduates of community sentences.)

It is easy to show that huge numbers of crimes are committed by people already on bail or serving community sentences: so many, in fact, that they represent a considerable proportion of all recorded crimes. There are about 150,000 people on probation at any one time. At least 50,000 of them a year, and probably more, are re-convicted. Furthermore, the reconviction rate used in measuring outcome does not measure re-re-convictions, that is to say those who are convicted more than once, of whom there are many.

Now the conviction rate for all recorded crime in Britain is a shade over one in twenty. If the criminals on bail or serving community sentences are typical of criminals as a whole (there might be some criminals who never get caught and some who always get caught, but assuming a normal distribution, most will fall somewhere between these two extremes), then at least 1,000,000 and probably more than 2,000,000 crimes are committed annually by people on bail or serving community sentences. Thus at least 20-40 per cent of recorded crime is committed by people serving the kind of community sentences that Mr Clarke wishes to extend.

If we add in the numbers of crimes committed by people who have just served short prison sentences were added to this, we can see that it is highly likely that, if long prison sentences were imposed upon criminals as a matter of course (bearing in mind the need to take account of exceptions, extenuating circumstances and so forth), it is likely that the recorded crime rate would decline by more than fifty per cent, irrespective of any deterrent effect such a policy would exert. In reality, the effect would be even greater.

Why does something so obvious escape our political and official class? The late Lord Bauer wrote:

When nonsense shows systematic bias, it probably reflects the pursuit of unacknowledged objectives which often have political or emotional bases.

It is so in this case.

Those who argue against imprisoning criminals think they are being kind to those poorer than they because most criminals are poorer than they. What they forget is that most of the victims are also poorer than they, and furthermore that the class of victim is much larger than the class of perpetrator (if it were not, poverty would be in itself a marker of criminality, which it is not).

So those who argue for the non-incapacitation of criminals are arguing for victimisation of the poor, whether they intend so to argue or not. Of course, it might be possible to argue that what Mr Clarke is trying to do is to ensure that the costs of crime remain where they arise, ie among the poor, and not transfer them to the middle classes.

An unfortunate effect of this is that it de-legitimises the state, one of whose indisputable tasks is to keep the peace. A state that wilfully fails to do so for a large section of its population can hardly be regarded as legitimate: a reason, perhaps, why so many people do not vote.

If I were a Marxist, I might add as another possible explanation the need of lawyers for clients and criminologists for subject matter. The class of lawyers and academic criminologists has increased greatly in the last few decades, and crime is one way of ensuring them employment (under- or unemployed lawyers and academics are very dangerous). The last thing, therefore, that lawyers and criminologists as a class would want is a decrease in crime.

But I am not a Marxist.

First published at Social Affairs Unit.