On Sunday I came across references to an academic/think tank paper that generated headlines like ‘Jail DOES reduce crime’ (The Sun) and ‘Longer prison terms really do cut crime, study shows’ (The Guardian). The research paper was published on Monday, and so unlike some social research stories in the media, we can actually go back to the original paper and see whether the politicised claims made bear any relation to the actual research.
However, my first thought was ‘so what?’ It would be astonishing if longer prison terms didn’t cut crime: after all, the more people are in prison at any one time, the fewer are out with the opportunity to commit crime. This would be true whether or not the people inside have been convicted of crime, and we also don’t know which prisoners would ‘go straight’ or not. Keeping people in prison for longer reduces crime, but keeping specific people inside may have the same effect. That’s also true when thinking about any deterrent effect: longer sentences might deter some but not others.
Indeed, although this story is used to say Ken Clarke is wrong, this would only really be the case if he’d argues that increasing sentences (or policing, for that matter) didn’t reduce crime. What he actually said was that increasing sentences has less impact on crime than the economy. So when the Guardian says that the research disagrees with Clarke, they put it thus:
‘The findings tend to support the thrust of policies followed by the last Labour government, which increased funding to the police and concentrated on the roughly 100,000 persistent offenders responsible for a high proportion of crime. This approach increased the prison population, but it also led to reductions in overall levels of crime. By contrast, the current justice secretary, Kenneth Clarke, often questions the relationship between criminal justice policy and the level of crime and suggests that economic factors may be just as, if not more, important.’ (The Guardian)
Amazingly, the research that Civitas commissioned (Bandyopadhyay, 2012) actually agrees with Clarke. For three of the four crimes studied, real earnings have a greater effect than sentencing or detection by police.
Furthermore, as usual in academic papers, there are lots of caveats that don’t end up in the newspaper versions. These include:
‘For some convicted robbers, a longer sentence acts as additional incapacitation and a potential deterrence from future crime. For others, being incarcerated for longer tends to reinforce, rather than reduce, criminal behaviour. For those offenders, an alternative disposal to custody might be more appropriate for reducing crime. In other words, an across-the-board increase in sentence length will not be optimal.’
‘… prison can have very different effects on different offenders. This suggests that an optimal policy will target repeat and serious offenders for long sentences while using alternatives to custody for other offenders.’
This is hardly the total endorsement of increasing sentences that the papers seem to have seen in or repeated from the Civitas press releases.
At this point, though, I’d like to take the author of the original paper to task. He’s an economist, but strangely didn’t think about the costs of any increase in policing or sentences for any given effect. And one of the most important costs is implicit in these caveats. We may know that we can decrease the number of burglaries by increasing sentences, but we don’t know whose sentences to increase. A deterrent effect on criminals-to-be might work with a couple of high profile long sentences, or publication of higher averages. The deterrent or incapacitation of prisoners due to be released may be overkill: half of the prisoners kept inside might not have reoffended anyway but we don’t know which. Like any punishment, one cost to society is the punishment of people we wouldn't want to punish - the false positives - and here there may be needless harm done to a person because we didn't know that they were in the 'unlikely to re-offend' category.
But the most glaring omission is the financial cost, surely of interest to an economist. Is any proposal worth it for the reduction in crime? Using the figures in his paper we find that increasing burglary sentences by a third would reduce burglaries by an estimated 21,000. But this is from 963,700, a huge number… your chances of being burgled will fall by a couple of percent. And the cost to society? I think there are about 5000 burglars in prison at any one time, so they’d do an extra 25000 months of prison time, a cost of about £83million, not including the cost of new prisons. So each burglary saved would cost the taxpayer £4000. Is it worth it, given that the average burglary claim is £1400?
Surely the question to ask is whether this £83million is money well spent. We could reduce burglary by 21,000 or employ 4,000 nurses. Indeed, the existence of these 4,000 nurses might increase health which might then reduce crime, who knows. Everything has costs, there are no magic bullets, and governments should choose policies on the basis of costs as well as benefits. At the same time, it would be helpful if newspapers didn’t reduce everything to the binaries of either it works or doesn’t, it’s caused by this or it isn’t: real life is complex, with multiple causes, multiple and unseen effects.