Do declines in residential segregation mean stable neighborhood racial integration in metropolitan America? A research note

Abstract

Between 1980 and 2000, residential segregation between non-Hispanic whites and blacks, as measured by the index of dissimilarity, declined by 8.7 percentage points. How this decline reflects stable racial integration at the neighborhood level remains to be seen. Using the Neighborhood Change Database, which provides 1980, 1990, and 2000 census-tract level data in 2000 boundaries, the descriptive analyses conducted here reveal three findings. First, the declines found in white–black segregation are due to increases in multiethnic neighborhoods rather than mixed white-and-black neighborhoods. Second, the decline in segregation appears not to be due to the fact that multiethnic neighborhoods are forming and remaining stable over time. Instead it is due to the increase in the number of these neighborhoods at one point in time. Finally, when focusing explicitly on the stability of mixed-race neighborhoods, such neighborhoods are less stable when examined over two decades as compared to one decade.

Introduction

Between 1980 and 2000, the level of residential segregation between non-Hispanic whites and blacks, as measured by the index of dissimilarity, declined by 8.7 percentage points, from 72.7 to 64.0 (Iceland et al., 2002). Although the current average level of segregation is considered to be in the “high” range of segregation scores (Massey and Denton, 1993, p. 20), it is evident that some progress, albeit slow progress, has been made with respect to residential integration between whites and blacks.¹ In metropolitan areas with fewer blacks, the progress has been more substantial. For example, in metropolitan areas where blacks comprised between roughly 6 and 10 percent of the population in 2000, the level of segregation or dissimilarity score dropped by 10.2 percentage points, from 71.5 to 61.3 between 1980 and 2000 (Iceland et al., 2002). Nevertheless, these metropolitan areas also have indices of dissimilarity considered to be in the “high” range of segregation.

How these trends in the aggregate have translated into residential integration within particular neighborhoods and the stability of such integration remains to be seen. Residential segregation and residential integration are not static phenomena. Instead, they are built upon the mobility or immobility of individuals, particularly whites, within the metropolis. While a neighborhood may appear to be integrated at any given point in time, the stability of that residential integration over several decades is uncertain. Perhaps the decline in residential segregation witnessed in recent years is due to the rise in stable, mixed white-and-black neighborhoods. On the other hand, perhaps it is instead attributable to an increase in the share of neighborhoods that are integrated at one point in time rather than over the course of a few decades. Studying the stability of integrated neighborhoods is important in understanding recent declines in segregation, but surprisingly, the topic has been under-researched in the urban sociological literature, particularly using the most recent data available (Charles, 2003). Another critical reason to study the stability of such neighborhoods is because integrated neighborhoods afford minorities access to relatively whiter neighborhoods, which have more amenities, better schools and job opportunities, and in general more connections to society’s opportunity structure (Logan and Alba, 1993a, Logan and Alba, 1993b, Logan and Schneider, 1984, Massey and Denton, 1993).

To my knowledge, there are just two studies that have addressed this issue using 1990 and 2000 census data (Fasenfest et al., 2004, Rawlings et al., 2004).² The other studies on this topic use data from 1990 and earlier censuses (e.g., Clark, 1993, Denton and Massey, 1991, Ellen, 2000, Lee and Wood, 1991, Nyden et al., 1997, Nyden et al., 1998). For the 10 largest metropolitan areas, Fasenfest et al. (2004) find that nearly half (47 percent) of mixed white-and-black neighborhoods in 1990 remained that way by 2000; they find that 71 percent of mixed white-and-other neighborhoods in 1990 remained that way by 2000.³ Rawlings et al. (2004) examine neighborhood racial stability in 69 of the largest metropolitan areas in which the black population is the dominant minority group and Hispanics comprise 20 percent or less of the metropolitan population. They find that about 80 percent of mixed-majority white neighborhoods (where blacks constitute 10–50 percent of the neighborhood’s population) in 1990 remain the same by 2000, although the most stable neighborhoods within this group are those in which the share of the black population is closest to 10 percent.⁴ Taken together, these recent studies have expressed optimism regarding the extent to which racially diverse neighborhoods remain stable and integrated.

However, it remains to be seen whether the declines in residential segregation witnessed in the past two decades have translated into stable residential integration between whites and blacks or whether such integration is a short-lived phenomenon. The main problem with the existing literature, with the exception of Ellen (2000), is that the stability of integrated neighborhoods is examined only over a decade-long period. Ellen (2000) finds that just under 57 percent of neighborhoods that were integrated in 1970 remained that way by 1990. Based upon these findings, Ellen suggests that racial integration is more stable than before and the prospects for future integration are good. However, a main limitation with her research is that she defines integrated neighborhoods as those with populations that are 10–50 percent black. As a result, the majority of the residual population may not necessarily be non-Hispanic white. Thus, the extent to which truly white–black integrated neighborhoods are stable over a two-decade period is unclear from her study.

Another limitation of the studies on neighborhood racial change/stability relates to the generalizability of the findings. Fasenfest et al. (2004) examine data only for the 10 largest metropolitan areas in the United States. These areas house just about 30 percent of the nation’s metropolitan minorities. The research by Rawlings and colleagues (2004) is based on data from 69 of the largest metropolitan areas where Hispanic population shares are 20 percent or less. As a result of this restriction, data from major metropolitan areas, such as New York, Los Angeles, and Miami are not included in their analysis. Although not available in the paper, it is likely that the share of the metropolitan minority population within the 69 metropolitan areas in their analysis is less than that found in the study by Fasenfest and colleagues (2004). Previous research on this topic that has used older data is also limited to either case studies (e.g., Clark, 1993, Maly, 2000, Modarres, 2004, Nyden et al., 1997, Nyden et al., 1998, Smith, 1998) or a subset of metropolitan areas in which the minority population is less than two-thirds of the total minority population (Ellen, 2000). Examining areas that house a larger share of the metropolitan minority population is essential in determining the stability of racially integrated neighborhoods as it exists in recent decades.

The goal of this paper is to take the existing research a step further and focus explicitly on the stability of racially integrated neighborhoods between 1980 and 2000 in order to see whether the decline in segregation at the aggregate has reflected stable, racial integration at the neighborhood level. Using data from the Neighborhood Change Database (NCDB), I document the percent of mixed-race neighborhoods in 1980, 1990, 2000. Second, I determine the impact that time has on examining the longevity of racially integrated neighborhoods. More specifically, I document the percent of neighborhoods that remain integrated over a decade-long period as compared to over two decades.