The study develops a steady state model that breaks down the overpayment error rate into intake portion of the caseload and three probabilities: the intake error rate, the error introduction rate, and the error correction rate. The error rates are estimated using 1996-2000 data from FNS's Quality Control (QC) files. The effect of a shorter recertification period on the overpayment error rate is analyzed in terms of its effect on the three component error probabilities. This analysis is conducted after stratifying the caseload into households with earnings and households without earnings. In addition, a semi-structured interview will be conducted with the Food Stamp Program directors and the food stamp quality control directors in three selected States that have chosen to implement shorter recertification period for some or all of their caseloads. Measures will be collected on how the change has affected the tradeoffs between maintaining payment accuracy and pursuing other objectives such as promoting program participation and restraining the growth of administrative costs.