Association for Public Policy Analysis & Management

The objective of this study is to provide evidence of the likely effects of restricting food types eligible under the Supplemental Nutrition Assistance Program (SNAP) on child and adult participants’ food choices. We are developing a structural food demand system model that accounts for the effects of SNAP participation and benefit amount on household demand for various food categories differentiated by nutritional quality. We will then use demand parameters estimated from the FoodAPS data to simulate the effects on purchase quantity, calories, and nutrients of several intervention strategies. FoodAPS is the only recent data set that collects detailed information on household food purchases and acquisition and SNAP participation status. This feature makes FoodAPS the only data source for the proposed research. Potential intervention strategies include a) restricting purchases of sugary drinks and/or other low-nutrition foods with SNAP benefits; b) implementing pricing interventions, such as subsidizing nutritious foods or taxing low-nutrition foods; and c) changing the SNAP benefit amount.

Whether a SNAP restriction on sugary drinks might reduce total sugary drink consumption depends crucially on the degree to which SNAP benefits are substitutable with cash income in the minds of SNAP participants. Evidence suggests SNAP participants have a significantly higher propensity to spend out of SNAP benefits than out of cash income on at-home food in the aggregate. In a review article, Fraker (1990) concluded that the estimated effect of a dollar increase in SNAP allotment on total household food-at-home spending is 2 to 10 times the estimated effect of a dollar increase in cash income. Fraker’s conclusion is further supported by findings from studies that used more recent data (e.g., Lin et al., 2010; Beatty & Tuttle, 2012). However, no previous study has examined whether SNAP participants have a higher propensity to spend SNAP benefits on low-nutrition foods than on nutritious foods or vice versa, which we will address in this study.

By analyzing survey data on household-level food purchases under a system of structural food demand equations controlling for self-selection into SNAP, we gain key insights into how the SNAP program interacts with social, economic, and environmental factors to affect household food decisions. By recovering the structural parameters of participants’ food demand and understanding the causal pathway(s) of SNAP effects on food choices, we will be able to simulate the effects and relative merits of SNAP food restrictions and other interventions on participants’ food purchases in terms of quantity, calories, and key nutrients recommended by the Dietary Guideline for Americans to increase or limit.