Association for Public Policy Analysis & Management

Recent research has documented the effectiveness of SNAP participation in increasing food expenditures, reducing food insecurity, and to a lesser extent, improving the nutritional quality of diets and general health of participants. However, far less is known on how purchasing behaviors, food security, or health outcomes differ by benefit levels among SNAP participants. Further, even in existing research, deriving causal estimates is complicated by misreported and endogenous benefit levels.

In this study, we propose to use household purchasing scanner data matched to administrately verified program participation as part of the National Household Food Acquisition and Purchase Survey (FoodAPS) to examine how food insecurity and diet quality differ according to SNAP benefit levels. FoodAPS provides a rich set of quality measurements ideal for research on SNAP and food purchases. We aim to answer this question by applying two distinct methodological approaches. First, we examine how SNAP benefit levels are associated with food insecurity status and diet quality across similar households employing multivariate regression analyses and matching estimators. Second, we utilize quasi-experimental variation in per-capita SNAP benefits to estimate causal effects of additional benefits on food security status and diet quality among SNAP households with children.

To address our first question we will construct a sample of all SNAP participating households in FoodAPS. Given the comprehensive set of observable characteristics in the FoodAPS data, we will employ multivariate regressions and matching estimators to examine how food insecurity and diet quality compare across similar SNAP households, but with varying levels of benefits. To address our second question, we address endogeneity of SNAP participation and benefit amounts by employing a unique identification strategy that exploits the fact that children of SNAP households automatically qualify for free in-school meals. As such, we will use share of school-age children coupled with extended school breaks (when meals are not available) to examine exogenous variation in a household’s per-capita SNAP benefit availability on food insecurity and diet quality. We will also attempt to explore the potential for non-linearities in the effect of SNAP benefits by exploiting differences in household compositions.

The findings of our study will shed light on whether current benefit levels are sufficient to ensure SNAP households with children are food secure and are able to purchase healthy foods, particularly during extended school breaks when food insecurity rates are higher and when diet related health outcomes for children suffer the most. Equally important, our results will provide evidence on the exact nature of the foods available to children from SNAP households when they are on and off school meals.