Welfare Spending and Incarceration Rates

Interesting correlations both across countries and across U.S. states between welfare spending and incarceration rates. When welfare spending goes up, incarceration rates go down:

Crime doesn't pay, by Tom Clark, Commentary, The Guardian: ... Looking across 18 developed countries, David Downes and Kirstine Hansen find a striking, negative link between the share of national income devoted to the welfare state on the one hand, and the number behind bars on the other. All seven of the nations most given to incarceration have below-average welfare spending; and, all but one of the eight countries with the lowest prison population spend atypically heavily on welfare.

Criminologists have long seen the US as the archetypal penal state, with 2 million people - or 2% of the male labour force - locked up each day. In line with the general pattern, their mass jailing comes alongside austere welfare. But America is far from homogenous, with the substantial freedom that individual states enjoy over both punishments and income support giving rise to big variations on both counts. This provides another test..., and the authors point to evidence showing ... that the same negative relationship holds. There are counter-cases - Japan has few prisoners and modest welfare, while in Britain record levels of jailing have recently arrived in tandem with new Labour's extra welfare expenditure. But ... the ... overall pattern is clear, and has become clearer over the years.

The researchers establish that the link is statistically significant - i.e. unlikely to be explained by any chance quirk of the data - but they do not settle the question of why it arises, which is all-important for policy. One interpretation, which I read Downes and Hansen as preferring, is that more generous welfare cuts poverty and exclusion, and with it the risk of crime and incarceration. If that were right, governments could cut imprisonment by raising benefits. But in theory, it could also be that jails somehow cut welfare bills - for example, by taking out of society people who would otherwise be on benefits, a conclusion that might encourage some governments to lock more people up. ...[A] third account is also possible, one that sees both mass imprisonment and minimal welfare as flowing from some single underlying factor, such as a particular strain of political culture. ...

Only more research - and particularly, closer working between economists, welfare specialists and criminologists - will definitively unpick the striking relationship that Hansen and Downes have uncovered. Ensuring that this is done should be a priority for anyone concerned about either the rising prison population or broader questions of economic equality.

It's certainly possible that a third variable causes both, e.g. the same attitudes that lead to high incarceration rates might also cause low social spending, and other explanations for why the correlation might be spurious are possible. But if the relationship holds up and is causal, I wonder what the return, in terms of lowered incarceration costs, an extra dollar of spending on welfare programs would bring?