The ‘Smarter Sanctions’ report produced by Policy Exchange has prompted many responses from those working in the field. David Webster has published his response (you can download it here), and Professor Lawrence Mead of New York University has sent this open message to the author, Guy Miscampbell:

“Guy Miscampbell,

I last spoke at Policy Exchange in 2010.

I have read your report on “Smarter Sanctions” with close attention. This a very important subject, central to welfare reform.

Your analysis taught me quite a bit about the current British system. I was surprised how many recipients of FSA or ESA were referred for sanctions, and also how many were exonerated. In American TANF, I suspect, both numbers would be proportionately smaller.

Your conclusion that sanctions are in some ways too severe, and in others too lenient, strikes me as reasonable–within the assumptions you make and the data you have. But these assumptions require more attention. You assume that sanctions are the key to enforcing conditionality, that their logic is economic, and that the current system requires more complexity.

Based largely on American experience, I take the other view on these questions. I think the key to enforcing work in welfare is making clear to benefit claimants up front that they must work to get aid. One has to demand serious work effort immediately on going on benefit, and even beforehand. Many states in America demand serious work search even as applications are being processed, and they pay no benefits until they see proof of it. Britain should to the same. It’s immensely harder to get claimants to look for work after they are on benefit, compared to before. If they see a serious work test up front, they will get their own jobs quickly or simply avoid aid. Sanctions are also necessary but secondary.

In my early studies of welfare work programs, I found that offices that sanctioned many recipients did worse at moving them into jobs than those that sanctioned fewer. That was because they has failed to convey a clear work expectation more informally. The effective offices sanctioned less often because they made work demands clearer at the outset, so fewer recipients fell foul of them.

The chief way to improve conditionality in Britain is to make a closer connection between benefit eligibility and work demands up front, in the administrative process. My sense from talking to British officials (including one Job Centre in Liverpool in 2010) is that those two processes are still far apart. People claim benefit and then they need do nothing about work for some weeks. I have repeatedly urged DWP to research the implementation of work tests in the Job Centres, but in vain. Policy Exchange ought to take this on.

As to the essence of conditionality, it’s not economic. It’s about authority. What gets recipients to work in America is chiefly being told clearly that they have to work. Economic payoffs are secondary. Incentives have little to do with whether the poor go to work. After all, living on benefit and not working is usually contrary to economic sense in the first place. We cannot assume that the sort of small sanctions you are talking about will much affect behavior, if the penalties are noticed at all.

Because authority is important, conditionality can have wide effects beyond welfare. You focus only on recipients who are actually referred to sanction. But one effect of serious sanctions is that other claimants take work more seriously–and others avoid going on aid at all. The latter effect is what Americans call diversion. It’s the largest and most important effect of work reform–and it occurs mostly off caseload. This is the main force that drove caseloads down in the American reform, although sanctions also counted. JSA has also seen a sizable caseload fall, so there has probably been diversion in the UK too. That should be counted in favor of sanctions.

As to making sanctions more complex, I see your good intentions. Possibly the yellow cards you propose would help re-engage some clients who have drifted apart. But in general, I’d argue for simplicity, not complexity. The sanction process you describe is already witheringly complex, with many steps and levels. Perhaps that is necessary to justify the regime to parliament, or public opinion. But it’s completely beyond the recipients. The first rule in sanctioning is to make penalties immediate and clear. The current system fails to do that, and your changes would make it more opaque, even if it generated fewer mistaken referrals. (By the way, the card systems you mention as precedents, such as the American SNAP, are not equivalent, because they don’t require regular visits to the agency). Few recipients would get the point, and many would perhaps not even realize they had been penalized, as is often true now.

American practice has gone just the other way. Before 1996, federal policy forbade state to penalize more than the welfare mother who refused to work; she was deleted from the grant, but benefits continued for her children. Few apparent effects resulted. So in 1996, states were allowed to close the entire grant, and most did so. That got through to the recipients, with dramatic effects on the caseload. Sanctions are commonly imposed quickly. Recipients may appeal, but the process is much less elaborate than the one you describe.

As to data, it is surprising how little you can say, chiefly because the information from DWP is so limited. It makes a big difference why cases leave aid who are “reserved” or “cancelled.” If they leave for work, that’s a positive outcome. If they are hassled off the rolls without work, it’s less positive. What do recipients and staff say? It appears that DWP or outside researchers have not done the “leaver” studies that followed reform in the US. These pursued those who had left the rolls to find out why they left and how they had fared since. The findings were largely reassuring, although not entirely so. Most of the leavers had taken jobs, although a minority had not done so.

In part, I take a different view of these issues than because our recipients are different from yours. They are mostly single mothers, their grants are lower, and they are more disadvantaged than is typical in Britain. That is why more directive programs seems needed to reach them, and why economic sanctions alone have little effect. For your caseload, which is typically better off and more calculating, incentives probably matter more, and more recipients may grasp–and game–the existing penalty system. Nonetheless, I still think administrative changes are the way to go. For that, above all, you need more research on the Job Centres. One cannot reason simply from data released from DWP, as everyone in Britain seems to do. You need to go out there and lay hands on the institutions. I’m willing to consult on how to approach that. I made that offer to Neil O’Brien before he left Policy Exchange, and I repeat it to you.”

Professor Lawrence Mead, New York University