The increasing rate of obesity in the U.S., particularly among low-income households, necessitates a thorough understanding of the relationship between obesity and in-kind federal benefits such as the Supplemental Nutrition Assistance Program (SNAP), formerly the Food Stamp Program. However, when examining this relationship, the existing literature often ignores evidence that respondents frequently misreport their participation in SNAP. This paper studies the impact of such misreporting on the estimated average treatment effect (ATE) of SNAP participation on adult obesity. Our analysis also synthesizes the current empirical techniques available for estimating ATEs in light of misreported treatment participation, adopting a range of parametric analyses as well as nonparametric bounds. The results highlight the inherent bias of common point estimates when ignoring misreporting, with treatment effects from instrumental variable methods exceeding the nonparametric bounds by over 200% in some cases. Accounting for misreporting, the estimated effects of SNAP participation on obesity are largely inconclusive. We find a slight negative effect of SNAP participation on the probability of being overweight, but the results specific to gender remain inconclusive due to the high rates of misreporting, particularly among men.
