A large share of households that are eligible for federally funded food assistance programs do not participate. This empirical regularity implies that eligible households do not know that they can participate in food assistance programs or that a utility cost is associated with participation. One part of this utility cost is that households that choose to participate must bear the associated time cost of filling out forms and obtaining required documents needed to verify eligibility. Rather than a one-time event, ongoing re-verification of eligibility is required as well as other time costs, which may include visits to the welfare office and inconvenience at the grocery store. Past research has also recognized the potential role of psychological costs (welfare stigma) as a deterrent of welfare program participation. Prior studies have shown that observable characteristics, which the researchers argue are associated with psychological costs, are negatively associated with program participation.
Although the existing literature agrees that there are both time costs and psychological costs, the only attempts to estimate the size of these costs have been to estimate them together in an all-encompassing welfare stigma term. The literature has found that together these costs are large. However, policymakers would benefit from knowing the degree to which psychological costs, rather than time costs, are responsible for the lack of participation in food assistance programs.
This study contributes to the literature by decomposing the costs associated with participation in the Food Stamp Program (FSP)—now the Supplemental Nutrition Assistance Program (SNAP)—and the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) into time costs and psychological costs. In addition, this study measures the change in psychological costs caused by FSP adoption of the electronic benefit transfer (EBT) system, which makes food stamp use at the grocery store much less observable to the public.
Identification of the psychological costs separate from the time costs of participation in WIC and FSP comes from a structural assumption about the nature of the psychological costs. The model is constructed around the assumption that the total psychological costs for participants do not increase with the number of programs in which the household participates. Time costs are specific to the program and thus accrue according to the specific programs in which the household participates. This means that, for a participant in one program, the marginal cost of participation in the other program is just the program-specific time cost; there are no additional psychological costs. However, with the adoption of EBT, participating in FSP is assumed to have a reduced level of psychological costs as well as reduced time costs (due to the EBT adoption and other reforms).
The data used in this study are two samples of female household heads from the Survey of Income and Program Participation (SIPP), the first from 1997 and the second from 2004. To alleviate concerns about joint labor supply decisions, the sample consists of working-age nonmarried women who are in households in which they are the sole decisionmaker. The parameters of the structural model are estimated using a simulated quasi-maximum likelihood estimation procedure that allows for heterogeneity in psychological costs, preference for leisure, and earning ability. Once estimated, the structural model is used to make predictions about outcomes of alternative scenarios, such as the manner in which the take-up rate would be affected if WIC were to adopt EBT.
One of the implications of the model is that the reduction in the psychological costs for FSP participants implies that WIC participation should fall among FSP participants. Simple regression analysis gives strong evidence in favor of this prediction of the model, indicating that the data are consistent with the structural assumptions of the model.
Estimation of the parameters of the model reveals that psychological costs are much more important than time costs in explaining the lack of participation in food assistance programs by eligible female-headed households. FSP adoption of the EBT system is estimated to have reduced average psychological costs by 31 percent, although there is a great deal of heterogeneity in the response. Using the model to simulate what would have happened if WIC had implemented the EBT system at the same time that FSP did leads to the finding that WIC take-up rates (participation rate among eligible female-headed households) would increase by 20 percent, with little change in eligibility. An important caveat is that this result assumes that the EBT system implemented for WIC is identical to that implemented for FSP, which may be impractical given the current denomination of WIC benefits in ounces rather than dollars.
Direct inquiries about this study to the Project Contact listed above.