Food Stamp Program Costs and Error Rates, 1989-2001

by Christopher Logan and William Rhodes

Evidence is strong that, beginning in 1995, an increase in reported certification-related costs per Food Stamp Program (FSP) household contributed to reduced error rates. This report presents the results of a study of trends in FSP administrative costs and errors from 1989 to 2001. It describes the trends and composition of FSP administrative costs. It also presents a multivariate regression analysis of the relationship of reported certification costs to FSP error rates (including overpayments, underpayments, and incorrect eligibility decisions). The report presents alternative models that relate a composite case error rate to certification effort per FSP household, caseload characteristics, the implementation of welfare reform, and short certification periods. The results imply that, in the period after the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, States on average had to spend more effort on certification-related activities than in previous years to achieve a given level of accuracy. The models predict that, if a State's FSP certification budget is fixed and the number of FSP households increases, the effort per FSP household will fall and error rates will rise, all other things equal.

This study was conducted by Abt Associations Inc., and the University of Georgia under a cooperative research contract with USDA’s Economic Research Service (ERS) Food and Nutrition Assistance Research Program (FANRP): contract number 43-3AEM-2-80099 (ERS project representative: Ken Hanson). The views expressed are those of the authors and not necessarily those of ERS or USDA.

Keywords: Food Stamp Program, FSP, food assistance, caseloads, FSPQC, FNS Quality Control (QC) public-use microdata, Food Assistance and Nutrition Research Program, FANRP

In this publication...

• Entire Report