Applying for and Staying on the Supplemental Nutrition Assistance Program in South Carolina

David Ribar 2, Christopher Swann1
1University of North Carolina at Greensboro, USA, 2IZA, Germany

This research examines how households apply to and maintain participation in the Supplemental Nutrition Assistance Program (SNAP) using case records from South Carolina. Program take-up is a major issue for assistance programs. Previous micro-analyses of SNAP participation using administrative data have focused on the length of assistance spells once households have joined the program but have overlooked the processes by which households begin receiving benefits. The paper examines electronic case management records from South Carolina covering the period from 1996-2007. We use these to estimate joint models of application resolutions and subsequent participation spells for those with accepted applications. Jointly modeling these processes allows for the possibility that successful applications constitute a non-random sample of all applications. The model of application resolutions is specified as a multinomial logit model where the possible resolutions are acceptance and denial due to three different reasons. Each outcome depends on the past application behavior and other observable case characteristics. For cases with successful applications, the resulting participation spell is modeled using a discrete-time, competing-risk hazard model. We distinguish between FSP exits that are the result of missed recertifications, financial ineligibility, incomplete or missing information, or other reasons. The estimation results indicate that a household's application and participation history affect its subsequent application success and program tenure. Additionally, the results indicate that unobserved characteristics that affect application resolutions also affect subsequent participation spells. The results are consistent with “positive selection” into participation.

View full paper