Regional Differences in Affluent Black and Affluent White Residential Outcomes

Abstract

This study compares the residential outcomes of affluent black and affluent white households using data from the 1990 and 2000 censuses and pooled data from the 2005–2009 American Community Survey. Results indicate that affluent black households are highly segregated from their white economic peers. Furthermore, affluent black households live in neighborhoods of lower average quality compared to affluent white households. Affluent black households are least segregated from affluent white households in the South, but the greatest equality in neighborhood‐quality outcomes occurs in the West. The South, however, shows the greatest improvement in both average neighborhood quality for affluent black households and a substantial reduction in affluent black–affluent white segregation over the entire study period. The authors find that place stratification theory better describes the residential geography of affluent black households than does spatial assimilation theory.

Key words:

• affluence
• neighborhood quality
• residential segregation
• South

Note: We would like to thank Andrew T. Carswell, David Kaplan, and the anonymous reviewers for their thoughtful feedback for improving this paper. Any errors are our responsibility alone. Portions of this manuscript were presented at the Race, Ethnicity and Place VII conference in Ft. Worth, Texas in October of 2014.

Note: We would like to thank Andrew T. Carswell, David Kaplan, and the anonymous reviewers for their thoughtful feedback for improving this paper. Any errors are our responsibility alone. Portions of this manuscript were presented at the Race, Ethnicity and Place VII conference in Ft. Worth, Texas in October of 2014.

Notes

Note: We would like to thank Andrew T. Carswell, David Kaplan, and the anonymous reviewers for their thoughtful feedback for improving this paper. Any errors are our responsibility alone. Portions of this manuscript were presented at the Race, Ethnicity and Place VII conference in Ft. Worth, Texas in October of 2014.

1. Please see Logan (2005) for a response to this critique.

2. The problem is that a rental cost variable would serve to both define affluence and then evaluate neighborhood quality.

3. Some households categorized as affluent in low cost‐of‐living metropolitan areas would not be considered affluent in high cost‐of‐living metropolitan areas using our approach. Defining affluence using a fixed approach, however, is also problematic. For example, a household with a yearly income of $100,000 per year can arguably buy into a “better quality” neighborhood in a low‐cost metropolitan area than in a high one. To the extent that “all real estate markets are local,” we suggest that a contextual approach better accounts for any variation in neighborhood‐quality outcomes associated with differences in cost‐of‐living. Furthermore, we suggest our method estimates affluence more conservatively than prior “fixed approach” methods. For example, in 2000, 18.3 percent and 11.7 percent, respectively, of all black households in our sample would be considered affluent using the poverty and median‐income methods compared to 8.7 percent using our method.

4. For example, even within the same metropolitan area, an affluent household earning $250,000 a year arguably has greater resources to buy into different neighborhoods than does one earning $100,000 per year. Because black households have lower incomes, on average, than white households, we might expect to see some degree of affluent black–affluent white residential segregation and differences in neighborhood‐quality outcomes due to these differences. See Reardon and colleagues (2015) for an alternative approach that explicitly treats income as a continuous measure.

5. Results are available by authors upon request.

6. Research suggests single‐race and mixed‐race households may have different neighborhood outcomes (Holloway and others 2005) and part of the recent reduction in black‐white segregation is due to the presence of mixed‐race households (Ellis and others 2007; Ellis and other 2012).

7. The correlation coefficient for the 2000 black‐white Dissimilarity Index computed with persons versus households is high (r = 0.98).

8. Prior research has used census tract data to create composite measures of neighborhood quality (Darden and Kamel 2000; Darden 2004; Darden and others 2010) and neighborhood disadvantage (Sampson, Rauenbush, and Earls 1997; Sharkey 2014). Other noncensus indicators—such as parks, walkability, and crime—arguably also factor into neighborhood quality, but were not available in the dataset.

9. Chi‐Square tests indicate that black‐white distributions are statistically significantly different for all time periods.