Built on Division: A Regression Analysis of County Segregation and Food Insecurity As a Structural Outcome

Background and Purpose
Food insecurity in the United States disproportionately affects communities of color, reflecting the historical and ongoing impact of structural racism. Residential segregation, an outcome of the legacy of discriminatory housing and zoning policies, continues to shape the social and economic landscape of counties across the country. This analysis explores whether racial segregation at the county level predicts food insecurity levels, controlling for county median income, rural/urban status, and demographic composition. The research question guiding this study is: Does county-level racial segregation predict food security status, controlling for socioeconomic and demographic covariates? It is hypothesized that counties with higher racial segregation will be more likely to experience higher levels of food insecurity across racial groups.

Methods
To examine the relationship between residential segregation and food insecurity, this analysis utilized a multinomial logistic regression model. The dependent variable was categorical county-level food insecurity, derived from Feeding America’s "Map the Meal Gap" dataset. The key independent variable was the county-level racial segregation index compiled by the National Institutes of Health (NIH) Office of Minority Health and Health Disparities. This index, ranging from 0 (complete integration) to 100 (complete segregation), measures the proportion of either Black or White residents who would need to relocate to achieve integration. Counties were grouped into three segregation categories: low (<38), medium (38–61), and high (>61). Covariates in the model included median income, rural/urban classification, and racial and ethnic demographic percentages. The multinomial logistic regression compared the likelihood of being in higher food insecurity categories based on segregation levels while adjusting for these factors.

Results
The multinomial logistic regression revealed a significant association between racial segregation and food insecurity. Specifically, counties in the highest segregation category (index > 61) were 3.7 times more likely to fall into a higher food insecurity category compared to counties in the lowest segregation category. The model demonstrated a good fit to the data with a pseudo r² value of 0.423. These results suggest that segregation is a robust predictor of food insecurity even when accounting for income and demographic differences.

Conclusions and Implications
The findings underscore the enduring influence of structural racism on access to food and other social determinants of health. High levels of racial segregation significantly increase the likelihood of food insecurity at the county level for all racial groups. These results point to the need for policy and planning interventions that directly address the spatial legacy of segregation and its effect on the food system. Food policy must move beyond individual or household-level solutions to consider the role of place-based structural inequities. Future research should extend this work by examining historical patterns of segregation, disinvestment, and land use, and by incorporating counties with low Black populations that were excluded from this analysis due to data limitations.