The Food Desert Paradox: Evaluating Grocery Store Interventions as Economic Policy for Obesity Prevention in Rural America

Abstract

An estimated 70% of adults in the United States are overweight or obese, contributing to nearly $1.4 trillion in annual health and economic costs. One commonly cited explanation is the prevalence of food deserts, areas with limited access to affordable, nutritious food, as a major contributor to poor health outcomes. However, research on the food desert theory and related government interventions remains inconclusive—often constrained by narrow settings, short timelines, or design limitations. This thesis addresses those gaps by presenting a long-term panel analysis of rural Kansas counties from 2010 to 2021. Using fixed effects regression models, I examine the relationship between residence in food deserts and obesity, utilizing data from the American Community Survey and the Kansas Behavioral Risk Factor Surveillance System. I implement a series of difference-in-differences (DiD) regressions to evaluate whether changes in grocery store access translate into measurable health improvements in obesity and food insecurity. The results are validated through pre-treatment trend analysis, placebo tests, alternative treatment thresholds, and fixed effects sensitivity checks. Overall, I find that while store closures significantly increase food inaccessibility, store openings do not meaningfully improve access or reduce obesity. The fixed effects model reveals a significant but almost negligible association between food access and obesity, with physical inactivity emerging as a stronger predictor across all specifications. These results raise concerns about the efficacy of current public spending aimed at reducing food deserts through new grocery store development. Rather than focusing narrowly on geographic access, policy efforts should expand toward behavioral, cultural, and lifestyle interventions that more directly address the root causes of obesity.