The Supplemental Nutrition Assistance Program (SNAP) is the largest of the USDA’s programs. It lifts more Americans out of poverty than any other means-tested program besides Medicaid. Yet, there are important gaps in our understanding. States have substantial autonomy in administering the program including setting income limits and deciding what assets are counted when determining eligibility, but there is only limited understanding of how these policies affect participation. In addition, and given well-recognized challenges in studying SNAP, there are few studies that have been able to identify a plausibly causal effect of participation, limiting the evidence base at a time when there are renewed calls to restrict the program.
Considering these limitations, this study makes two important contributions by 1) investigating if it is possible to categorize state SNAP policies approaches, and – if so – whether these approaches are related to SNAP take-up, and 2) using these policies as instruments in and instrumental variables (IV) analysis to identify plausibly causal effects of participation on household food insecurity.
Methods
The USDA’s State SNAP Policy Database records information on monthly variation in policies from 1996 to 2016. From this database, I selected 19 policies, including those examples listed above and others such as whether the state requires fingerprinting of recipients or allows them to certify eligibility by phone. I combined these policies into a standardized mean index (α=0.89). I also conducted factor analysis with varimax rotation to identify common policy approaches among states.
These monthly data were merged with the 2001-2016 December Current Population Survey (CPS), the cross-sectional household survey used to generate national estimates of food insecurity. The CPS asks about SNAP receipt for each month in the past year as well as household food insecurity. The analytic sample consisted of n=3,334,800 household-month observations linked to pertinent monthly policy data. I assessed the impacts of state policies on participation after controlling for household characteristics and state and year fixed effects. To estimate the effects of SNAP participation, I used IV analyses with the same controls and the mean index of policies as the exogenous instrument.
Results
Results point to five distinct state SNAP policy approaches: (1) using broad-based categorical eligibility to expand eligibility; (2) allowing certification via telephone and use of state call centers; (3) promoting access for non-citizens; (4) lenient recertification periods; and (5) minimizing administrative burden. Except for promoting access for non-citizens, all other policy approaches were significantly associated with increases in household SNAP participation. IV models found that past-year SNAP participation was related to a 0.41 decrease in the probability of any household food insecurity but not to very low food security.
Conclusions
Findings from this study suggest that states adopt a number of approaches to SNAP policy and that participation in SNAP leads to large reductions in household food insecurity. Amidst ongoing policy debates about the size and scope of SNAP, this is an important and timely finding that contributes to a scant evidence base and highlights SNAP’s effectiveness in meeting its core goal.