A Spatial Examination of Racial-Ethnic Population Patterns in Metro Atlanta Using Bayesian Conditional Autoregressive Models

After 1990, population movements in the U.S. experienced significant shifts as immigrants settled in new destination cities, wealthier residents moved into city centers, and suburbs experienced increased racial diversity and poverty. The American South, with its sprawling suburban landscape, stands at the confluence of these trends as migrants flocked to its metro areas. Still, it remains a region that is under researched. To better understand the conditions of the growing population of color in the South, this work asks three questions: (1) from 1990 to 2019 where did populations live? (2) Where were they growing? And (3) how did this changed over time? To examine the residential patterns of Whites, Blacks, Asians, and Latinxs in the Atlanta metro area, this work contributes a spatial stock-flow model. I add an intrinsic conditional autoregressive spatial parameter (Besag et al Annals of the Institute of Statistical Mathematics 43(1):1–20, 1991) to a stock-flow framework examined across three time periods: 2000, 2010, 2015–2019. This work finds that the racial-ethnic composition within the City of Atlanta boundaries remains one that is predominantly Black and White as Asians and Latinx largely settle outside of city boundaries alongside expressways. However, the Bayesian models also demonstrate that populations of color also have the highest growth rates in Whiter neighborhoods though growth in these neighborhoods do not necessarily point to increases in neighborhood education and income. And rather than inner-city gentrification or White flight into inner-suburbs, there is a movement of Whites to the outskirts of metro Atlanta. For all models, there remains a persistent positive effect of new housing for population growth and concentration across all racial-ethnic groups.