Saturday, February 21, 2009

Tax policy colloquium on Yoram Margalioth's "Employing Statistical Stigma as a Welfare Ordeal"

Last Thursday, we discussed Yoram Margalioth's "Employing Statistical Stigma as a Welfare Ordeal." In the welfare literature, stigma, leading to low take-up of benefits by eligible individuals, has been viewed as purely a bad thing, to be minimized, not only because it inflicts a bad experience on benefit claimants but because it discourages claiming by intended beneficiaries. The paper argues that a special type of stigma (whether or not that's the right name for it) can have desirable properties as a screening device, helping to focus claiming of benefits on the people who really ought to get them, thus permitting greater aid to those people.

Traditional stigma, in the paper's account, is the feeling of being a failure because one is unable to support oneself. As a matter of empirical description, I felt the paper overly viewed this as something purely internal, i.e., if I'm a prospective claimant I hate feeling like a failure, but exposure to others doesn't make it any worse. I would think that, as a common psychological matter, bad as it is to feel like you have failed, letting others know makes it far worse. Better to lick one's wounds in private.

The paper agrees that traditional stigma appears to have no good sorting functions in the form of discouraging claimants who ought to be ineligible but aren't given the imperfection of screening via eligibility rules. But it argues that another type, which it called "statistical stigma," can serve as a helpful screen. Suppose we have a benefit program for people with low earning ability but that some who don't need the benefits manage to con their way in. But suppose that people in the neighborhood both (a) have greater information about ability levels than the government can hope to get in screening for eligibility, and (b) will dislike false claimants, either directly for cheating in a government program or on the view that this is the evidence of broader dishonesty that might infect the claimant's other social dealings. Then being known locally as a claimant would tend to have positive screening characteristics, because those who truly were high-ability would tend to get stigmatized as suspected cheaters more, all else equal.

The point holds intellectually within its defined terms, and skepticism at the session focused mainly on the question of how well it applies to real world settings. We felt it was generally most likely to play a role in cases such as parking in disabled spaces in the shopping center. Among the issues potentially challenging its broader applicability were those of heterogeneity (in neighborhood values, in claimant temperament, etc.) and of traditional stigma's potentially outweighing statistical stigma in various welfare-type settings.

No comments: