Supplemental Nutrition Assistance Program and Childhood Obesity in the U.S : Evidence from the National Longitudinal Survey of Youth 1997

Using the National Longitudinal Survey of Youth 1997, this paper employs difference-in-difference propensity score matching to examine whether the Supplementary Nutrition Assistance Program (SNAP) contributes to childhood obesity. We find no statistically significant SNAP effect among the 12- to 20-year-old participants when controlling for selection bias and more accurately defining the treatment and comparison groups. The results are robust to various robustness checks including redefining the treatment and comparison groups by excluding those who previously enrolled in the SNAP, using an alternative treatment definition based on SNAP benefits received, using different specifications of the propensity score equation, and employing different estimation techniques (covariate matching and inverse probability weighting). This study differs from previous research in the three major aspects. First, we carefully examine the intensity of SNAP participation (full-time vs. part-time) and the amount of SNAP benefits received for one-, two-, and three-year durations. Second, we focus on the change in the BMI or the obesity status rather than the level and control for the pre-treatment BMI to avoid the confounding effects of the time-invariant factors. Third, instead of making parametric assumptions on the outcomes, we employ a variety of semi-parametric estimators to control for the selection bias of program participation. The results show that the SNAP is not responsible for the higher prevalence of obesity among adolescents of low-income households. Proposed SNAP changes such as more frequent benefit distribution and a focus on fresh fruits and vegetables are likely to be ineffective in reducing childhood obesity, although they might encourage healthy dietary practices among SNAP participants