I know I’ve already talked about ‘deservingness’ among benefit claimants – but if UK newspaper and lobby groups keep making claims about it, then I’m just going to have to keep writing about it. Just after Christmas there was a rush of newspaper stories about the numbers of benefit claimants with criminal records, led by the Daily Mail’s typically careful headline, ‘£2bn cost of handouts to underclass is revealed for the first time’. This comes only months after police data generated headlines that 40% of the London rioters were claiming a benefit.
But is this really true? And if it is, does that really mean that a large number of benefit claimants are a criminal underclass?
What’s new, what’s not
The headlines were prompted by a new ‘ad-hoc statistical release by the Department of Work and Pensions (DWP), which linked administrative data on benefit claims, tax returns and criminal records. This showed that “51% of offenders sentenced or cautioned in England and Wales in the year ending Nov 2010 claimed one of the main out-of-work benefits at some point in the month before the sentence”. This varied by benefit type, with 24% claiming Jobseekers Allowance (JSA), and 22% claiming incapacity benefits [p21]. In comparison, the percentage of rioters claiming out-of-work benefits was actually slightly lower than you might expect (more of which below).
What’s most striking about this, though, is that these figures simply aren’t new. There was over a month’s gap between the statistics being released, and them being picked up by journalists. And when I actually saw the estimates before the headlines in Nov, I thought they weren’t worthy of special comment. For example, a 2005 survey of prisoners had already found that 62% had claimed a benefit in the year before custody (35% JSA, 15% incapacity) – figures that were actually slightly higher than in the (better) data here. And more generally, there’s a well-known connection between unemployment and crime, as much US urban sociology has shown.
It’s not really surprising that unemployment and crime are linked. Being unemployed raises the incentive to commit some crimes, while being a convicted criminal makes it harder to find a job – particularly given that half of crime is committed by people who already have a criminal record (p10). And moreover, people with wider disadvantages are more likely to be in both groups. 60% of new prisoners have literacy problems (p6), while a 2005 survey of prisoners found 46% had no qualifications. Moreover, the same survey found that 29% had received emotional, sexual or physical abuse as a child (2010 Compendium:118).
The scale of ‘benefit criminality’
What really caused the headlines, though, were the new estimates about the share of benefit claimants that have criminal records – or as the Sun screamed out, “one in three on dole is a criminal”. The release was clearly motivated to create this kind of finding, by presenting the % of claimants (p16) rather than just the patterns of benefit claims among those leaving prison. And this where it gets a bit more complicated.
Now, I don’t want to critique most of the methods of the study – data linkage is actually a complex business, and the study is interesting enough to merit it’s own session at the Royal Statistical Society next month. But there’s a big leap between those who feature on the Police National Computer (PNC) and people who can be called ‘criminals’.
I’ve struggled to find out the data I’d like on the PNC, but here’s what I can piece together. The PNC is a huge database – with some privacy concerns – of people who have been cautioned or found guilty of a ‘recordable offence’. The DWP report notes that this excludes less serious ‘summary offences’, but around one in five court fines are included (p8), while the list of recordable offences includes drunkenness, failing to provide a breath test, begging, nuisance calls and prostitution. Cautioning – given for first-time or more minor offences in less serious categories, and while rarer than non-custodial sentences, is twice as common as people being sent to prison – is impossible to tell apart.
For the purposes of this report, anyone who has been cautioned or sentenced for any of these offences in the past 10 years is a ‘criminal’. This means that 4.2m people are – by this study – members of the ‘criminal underclass’ (p7)! The Daily Mail says these figures “lay bare the degree to which an ‘underclass’ that drifts in and out of criminal activity is using state handouts to bolster its income, while often continuing a life of crime.” Yet obviously we know nothing about whether this is a ‘life of crime’, or a single caution for drunkenness ten years previously, which is hardly the same thing.
So how many people claim benefits from a ‘life of crime’? We don’t know. From the report, we know that 1 in 20 out-of-work benefit claimants had at least one spell in prison in the past 10 years – 250,000 people, one-fifth of the total number of benefit claimants on the PNC without being in prison. I don’t think we know what share of ex-prisoners this is – lifetime incarceration rates are not a simple matter to calculate. The figures for the US in Pettit & Western 2004’s classic study (free article here) are that 3% of White and 20% of Black men born 1965-69 had done time in prison by their 30s, but then the US incarceration rate is about five times that of the UK rate…
Another way of estimating this might be the number of benefit claimants who are about to be sent to prison – that is, they’ve been found guilty of a serious crime whilst claiming benefits. We know that 102,000 people are sent to prison every year, and the new DWP report finds that around half of them were claiming out-of-work benefits at the time. So another estimate of the number of ‘criminal benefit claimants’ is the relatively small number of 50,000, compared to the big number in the headlines of 1.3m – and even this will include many people who have done something seriously wrong, but who aren’t by any means a ‘criminal underclass continuing a life of crime’.
All of which is to say: the research itself seems good quality, and has valuable parts to it. But the interpretations of it are dubious. And if I was going to guess, I’d say that this was the Government’s intention all along.