The take-up of social benefits

It is generally agreed that people do not take up benefits if the costs outweigh the benefits, but after many years of research, we still have relatively little insight into precisely what types of costs matter most, and what types of measures are most likely to reduce them. A few general conclusions can be drawn, however. First, take-up is enhanced by automatic or default enrollment and lowered by administrative barriers, although removing individual barriers such as reducing the length of forms or increasing the number of offices that process forms does not necessarily have much effect, suggesting that one must address the whole bundle. Second, although it may be impossible to devise a definitive test of the "stigma hypothesis," it seems clear that stigma cannot be the only cost facing participants. Other, more concrete types of transaction costs are probably a good deal more important to most people than stigma or lack of information.5 Third, although people generally have means-tested programs in the United States in mind when they discuss take-up, low take-up is also a problem in many non-means-tested social insurance programs and in other countries. Historically, economists have paid much attention to rules about eligibility and virtually no attention to how these rules are enforced or made known to participants. This review suggests that the marginal return to new data about these features of programs is likely to be high in terms of understanding take-up. Anecdotal evidence suggests that there is a great deal of variation in the ways that similar types of programs are implemented both within and across countries, and this variation could be exploited to identify the most important barriers to participation. For example, some states implemented SCHIP as an extension of their Medicaid programs, while others created separate standalone programs in order to reduce the stigma associated with receiving public insurance. To my knowledge, the difference has not been exploited to investigate the "stigma hypothesis." About half of the 109 Food Stamp Program offices surveyed in a recent U.S. Department of Agriculture study of program access provided services such as extended hours, whereas a small number of programs required applicants to attend a series of meetings before they were even permitted to sign their application forms (Gabor et al. 2003). Procedures such as requiring third-party verification of income are not standardized across locations either, and could easily explain variation in take-up across areas. More systematic collection and analysis of this type of data would add a great deal to the study of take-up. In an era of social experiments, it might also prove useful to consider experimental manipulations of factors thought to influence take-up. For example, it might be possible to design an outreach program that would directly test the hypothesis that take-up is influenced by information exchange among members of social networks. Similarly, parameters such as application procedures, recertification intervals, payments for community enrollment assistance, and incentives to service providers to give application assistance could be varied across areas in order to study their effects. This paper was prepared for a conference in honor of Eugene Smolensky held at Berkeley, December 12 and 13, 2003. The author thanks Alan Auerbach, Jeffrey Biddle, David Card, Sheldon Danziger, Irving Garfinkel, Robert Haveman, John Quigley, and conference participants for helpful comments. Princeton's Center for Health and Well-Being provided financial support. Graciana Rucci provided excellent research assistance.