Does welfare reduce poverty? That might seem like a stupid question. Welfare is basically giving money to the badly-off. Of course, there are lots of glitches in the system which make it less efficient and effective than it could be and limit its potential (see the ASI's latest paper for more on that). But it still targets those in need reasonably well, even if it could do better. However this may not be true for all groups. Harvard's George Borjas, among the world's experts on the economics of immigration, finds in a new working paper that in one specific case welfare did not increase recipient incomes and reduce poverty.

The Personal Responsibility and Work Opportunity Reconciliation Act of 1996, the major Clinton-era bipartisan welfare reform, slashed federal spending on welfare benefits, cutting especially immigrant eligibility for major payments. Many, but not all, states decided to cushion their immigrant populations from the blow, making this a natural experiment. We can look at the difference between immigrant populations in states that did and didn't cushion them to discover whether cutting hit those who used to get it, or whether it induced extra waged labour to make up the gap?

In practice Borjas finds that the entirety of the loss in reduced welfare benefits is made up by extra earnings from working in the labour market. In fact, more than the entirety is made up, and cutting welfare actually reduced poverty for the most-affected immigrant groups.

In Borjas's words:

The evidence presented in this paper strongly suggests that, at least in terms of officially measured poverty rates in immigrant families, the welfare state is not a panacea. For these families, welfare contributes to poverty.

Now, I am sceptical as to whether this is widely applicable. There may well be differences between immigrant groups and natives of similar socioeconomic status—for example, first generation immigrants are widely perceived as having a stronger work ethic. They may also have lower savings to run down, credit to run up, or family ties to rely upon. It may be more credible that the authorities will let them endure real economic hardship—whereas the hardship of others may be more visible and salient for voters and welfare authorities. So this finding probably does not apply to native welfare recipients.

These points are why the ASI, and myself, favour a generous but simple welfare system: the negative income tax. But the finding is certainly interesting. In at least some cases, welfare does not reduce poverty, and may even increase it.