Association for Public Policy Analysis & Management

The benefits of a “healthy diet” are firmly established, but the ability to follow established dietary-guidelines differs significantly. While personal choices do matter, the local food environment plays an increasingly greater role in the ability to access the right foods. The literature is abundant on the differential food landscape based on socio-economic characteristics of a neighborhood. However, the underlying pathways by which residential segregation impacts the food environment have not received attention in previous research.

This study looks at the varied effect of residential segregation on the food environment in terms of both access and quality, variety of food available.  Using the Dallas-Fort Worth Metropolis as an example case, I develop a composite food environment score by assessing the factors of density and proximity, along with type of food outlet, thereby providing an in-depth, comprehensive view of both the food environment and residential segregation down to the granularity of the census block group.

We start with the hypothesis that increased residential segregation impacts the mean/median food score of a neighborhood. Regression results show that this hypothesis cannot be disproved. Both residential segregation by race, as well as residential segregation by income significantly affect the food environment.

We then test the hypothesis that increased residential segregation is associated with increased variance in the food environment. The results show that this hypothesis cannot be disproved either. The level of residential segregation in a neighborhood affects not just variability in the overall availability of food, but especially appears to impact the variability in the availability of healthy food.

Finally, in addition to examining between-neighborhood differences in the food scores, variance function regression (VRF) is used to simultaneously model the means and variances, and the net effects of residential segregation on the food score. The variance function model therefore describes not just where the food score will fall on average, but how far the food score will fall from this average value, given residential segregation and other neighborhood characteristics.

Overall we can conclude that there is a clear and strong positive relationship between residential segregation and the variability in the food environment. While residential segregation by race does have a positive effect, residential segregation by income has a more distinct, positive, and stronger impact on the variability in the food environment, particularly when considering access to healthy food.

Identifying the particular aspects of residential segregation that impact the food environment, allows us to address effective policies to combat these imbalances. By shedding light on the particular aspects of residential segregation that impact the food environment, this study fits into both the debate over the consequences of segregation, and the debate over effective food retail zoning and accessibility