The Impact of Increased SNAP Benefits on Childhood Obesity: Examining the Natural Experiment Resulting from the 2009-2013 ARRA SNAP Benefit Increase

Childhood obesity is a public health concern of epidemic proportions in the United States. The US federal and state governments have numerous programs in place to improve nutritional and health outcomes; these initiatives are often intertwined with poverty assistance. One such program is SNAP (the Supplemental Nutrition Action Program), which is meant to provide the necessary supplementary income for families at or below 130% of the federal poverty threshold to afford a nutritious diet. A 2009-2013 increase of 13.6% in SNAP benefits as part of the ARRA (American Recovery and Reinvestment Act) provided a perfect natural experiment with which to investigate a potential causal relationship between the amount of SNAP benefits and childhood obesity prevalence in low-income populations. Using 2003-2016 data from NHANES (the National Health and Nutrition Examination Survey), I calculated cross sectional childhood obesity prevalence for each NHANES cycle in children and adolescents ages 2-19. Using these statistics for SNAP recipients and a near-eligible comparison population with subgroups by age, gender, and race/ethnicity, I ran difference-in-differences analyses in both R and Stata to examine the impact of the ARRA SNAP increase on childhood obesity rates. Trend differences between SNAP recipients and the comparison population in the overall 2-19 age group were not statistically significant. In subgroup analyses, the regressions of boys’ data and the 4-10 age group were the only two which approached statistical significance: in these groups, SNAP recipients were less likely to be obese than their non-SNAP low-income counterparts. In fact, for all but the 2-4 age group, I found SNAP participation to be negatively correlated with obesity prevalence. However, none of these relationships were statistically significant (not even in boys or the 4-10 age group), with many having p-values above 0.5. Given these results, I conclude that while there are indications that the ARRA SNAP benefit increase may have had positive effects on childhood obesity in certain groups, the high level of standard error in these data prevent such interpretation. I conclude that the ARRA SNAP benefit increase did not impact obesity rates in children ages 2-19 to a statistically significant level, a result which held for all age, gender, and racial/ethnic groups analyzed. As such, my policy recommendations involved investment in research regarding the expansion of SNAP-Ed and the indefinite extension of the COVID-19 P-EBT