This study examines the impact that participation in the Food Stamp Program has on household food insecurity using data from the Panel Survey of Income Dynamics. Two strategies are used to identify the causal effect of the program. First, endogenous treatment effect models are estimated using state-level errors in payments of benefits as instruments. Additionally the impact of losing benefits due to a government decision on the food insecurity of program participants is examined. The paper finds that program participation lowers food insecurity by at least 18%. Copyright 2010, Oxford University Press.