Does the Neighborhood Food Environment Influence the Relationship between Supplemental Nutrition Assistance Program Participation and the Food Purchase Decisions and Weight Status of Low-Income Women?

Purpose: This paper examines whether the neighborhood availability of food retail and food service establishments (the “neighborhood food environment”) plays a role in explaining the associations between Supplemental Nutrition Assistance Program (SNAP) participation and the food purchase decisions and weight status of low-income women. Data: The project used restricted-access data on low-income women from the 2007-2008 National Health and Nutrition Examination Survey combined with 2007-2008 ZIP Code Business Patterns data on the ZIP Code-level availability of food retail and food service establishments. Methods: Ordinary Least Squares models of body mass index (BMI), obesity, and food purchase decisions (monthly food spending and where and how frequently food was purchased) were estimated that included SNAP participation and a large set of other individual characteristics. These models were then re-estimated after the addition of measures of the neighborhood food environment and interactions of these variables with SNAP participation. Results: SNAP participants tended to live in neighborhoods that had a higher density of “small or quick” establishments (small grocery stores, convenience stores, and fast-food restaurants) than low-income non-participants. These neighborhood differences lead to a higher predicted BMI and a greater likelihood of obesity for SNAP recipients relative to low-income non-participants. Estimates of the association between SNAP participation and weight status that exclude contextual variables are likely to be biased. However, the paper also finds that SNAP participation remained positively and significantly associated with BMI even after controlling for neighborhood contextual variables. SNAP recipients who lived in neighborhoods with a higher density of “small or quick” establishments allocated a significantly higher percentage of their monthly food budget to eating out. This suggests that SNAP benefits free up cash resources for non-grocery spending and that these additional resources are more likely to be allocated to eating out when the neighborhood density of “small or quick” establishments rises. Given that meals eaten away from home tend to have more calories than meals eaten at home, this may be a potential mechanism for the positive association found between SNAP participation and BMI for low-income women.