Measuring the Effect of Supplemental Nutrition Assistance Program Participation on Food Insecurity Using a Behavioral Rasch Selection Model

This paper examines the relationship between Supplemental Nutrition Assistance Program (SNAP) participation and food insecurity using data from the 2001-2008 Current Population Survey Food Security Supplement (CPS-FSS). A behavioral Rasch selection model is proposed and estimated using four subsamples of low-income households: unmarried parent households, married parent households, all-elderly households, and other adult-only households. The model is identified using exogenous changes in state-level polices related to SNAP. The results indicate that SNAP has a strong ameliorative effect on food insecurity for married parent, all-elderly, and other adult-only households, while SNAP continues to be associated with greater food hardships for unmarried parent households. Participating in SNAP reduces the probability of food insecurity by 22.4 percent for other adult-only households, 18 percent for all-elderly households, and 17 percent for married parent households.