Access versus Choice: testing the “food desert” construct in Champaign, IL

How does one’s food environment affect food purchase decisions? Food access has received significant political and academic attention, particularly under the Obama administration. Existing literature on food access and “food deserts” mainly focuses on geographic distance or the retail of fresh fruits & vegetables versus fast food within a neighborhood to determine and identify inequitable access. In this paper I attempt to develop an endogenous measure of food access by asking how geographic placement of food retail affects food expenditure, particularly of fruits & vegetables. I use novel data on 886 households matched to food prices from a census of geocoded food retailers in Champaign County to approach this question from two perspectives. I first estimate the household’s share of grocery expenditures allocated to fresh, frozen, and canned fruits & vegetables versus other grocery items. I then use data on a person’s residence and geocoded data on food retail locations in Champaign County to test for relationships between retailer proximity, and the share of expenditure on fruits & vegetables. The next perspective uses a choice experiment to measure the tradeoff among store characteristics that determine where a consumer shops. The demand estimation reveals how much fruits & vegetables a person is actually consuming, while the choice experiment reveals whether that individual is constrained in their consumption by their existing characteristic set of stores. I find that while proximity to a grocery store is positively correlated with healthier food consumption, policy response should focus on improving store quality and product quality to induce behavioral change. I further find policy response should be cognizant of endogenous locational sorting which may require alternative means to improve health other than changing the food geography.