Race,Class,and Gender and the Impact of Racial Segregation on Black-White Income Inequality

Abstract

African Americans have yet to achieve parity with whites in terms of income. A growing number of studies have identified several factors that have influenced the size of the racial gap, which has been found to vary by social class status and gender as well as across space. While most research has examined these factors separately, they may interact with each other in shaping racial inequality. Using an intersectional approach with a multilevel model, this study focuses on the impact of residential segregation and social class on racial differences in earnings for men and women. Findings indicate that (1) earning differences between African Americans remain after controls for socioeconomic status, gender, and other control variables; (2) racial differences increase with rising social class status; (3) segregation increases the disparity between African American and white males; and (4) among males only, segregation worsens the disparity that increases with rising social class.

Keywords

racial segregation racial inequality race relations racial stratification gender

African Americans have yet to achieve parity with whites in terms of income. The decline in the black-white income gap since the 1980s has been small (Semyonov and Lewin-Epstein 2009). A growing number of studies have identified several factors that have influenced the size of the racial gap in earnings. Of particular relevance to this study, the size of the racial gap in earnings has been found to vary by social class status, by gender, and across regions and localities (e.g., Beggs, Villemez, and Arnold 1997; Cohen 1998; Thomas 1993; Tomakovic-Devey, Thomas, and Johnson 2005). Common explanations for African American and white income differentials have focused on racial discrimination, racial differences in human capital, and spatial mismatch.

An important factor not usually discussed in the context of income inequality is racial segregation. African Americans are the most segregated ethnic group in the United States. Although black-white segregation has declined each decade since 1970, it remains quite high, with the index of dissimilarity score in most metropolitan areas being higher than 60 (Glaeser and Vigdor 2012). Segregation and the black-white earnings gap may be directly linked through processes associated with spatial mismatch, access to white social networks, and stigmas associated with people who live in “black neighborhoods.” Therefore, it is important to examine how variations in metropolitan segregation affect the black-white income gap.

Gender has also been neglected in most studies of race and income. Women earn significantly less money than men (Lips 2013). African American women experience the disadvantages associated with being both an African American and a woman (Belkhir and Barnett 2001; Browne and Misra 2003). White women and minorities are both less likely to be promoted than white men (Yap and Conrad 2009). By contrast, white men experience privileges associated with being white and a male, including access to larger social networks that provide job leads (McDonald 2011).

Although examining the impact of race, class, and gender separately is important and informative, an intersectional perspective can provide additional insight because of how race, class, and gender interact in shaping inequality (Browne and Misra 2003). While a few studies have examined how race, class, and gender jointly affect economic outcomes, fewer have examined the impact of residential segregation. This study addresses these important issues by examining the impact of segregation and social class on the earnings of African American and white men and women.

Race, Human Capital, and Income

One of the most widely used explanations of racial inequality is human capital. According to this view, labor market earnings reflect levels of productivity for individuals at different stages of their lives. Differences in productivity in turn result from deliberate self-investments in human capital such as education and on-the-job training (Becker 1975). Similarly, Wilson (1980) argued that young educated African Americans in the early 1980s had the same life chances as their white counterparts and that the growth of the African American middle class constitutes evidence that social class, rather than race, is the primary determinant of life chances for African Americans. Similarly, more recent research has concluded that there are no significant contemporary race differences in earnings, net of variation in individual human capital measures (e.g., Farkas et al. 1997; Neal and Johnson 1996).

Although the human capital perspective continues to be popular among social scientists, it has only limited empirical support when applied to racial inequality. Several empirical studies have seriously questioned the idea that education and other forms of human capital acquisition explain all racial differences in earnings (e.g., Cotton 1989, 1990; Landry 1987; Thomas 1993; Thomas and Horton 1992; Tomaskovic-Devey, Thomas, and Johnson 2005). This body of research focuses on how well middle-class blacks are doing compared with similar whites. These empirical studies generally conclude that race plays a significant and negative role in the labor force experiences of even middle-class blacks and that racial gaps widen across the career. Some authors (e.g., Cotton 1990; Thomas 1992; Thomas 1995; Thomas and Horton 1992) have even claimed that racial discrimination negatively affects the earnings of higher status blacks more than it does lower status blacks. For example, in regard to racial differences in personal income, Thomas (1992:23) found that African Americans who were more educated and had attained higher occupational status were worse off than less educated, lower status blacks when compared with similar whites. Similarly, Tomaskovic-Devey et al. (2005:82-83) concluded that

Blacks and Hispanics receive lower wages than whites not only because they have lower average levels and quality of education, but because the human capital they do possess is devalued in the labor market. The negative impact of this is cumulative over the work career and increases with education.

Race, Class, and Gender

A problem in the existing research is that it has focused almost exclusively on male employment or earnings (Hellerstein, Neumark, and McInerney 2008; Semyonov and Lewin-Epstein 2009). Women earn less than men even when they work the same jobs. This is true across all racial and ethnic groups as well as across all social class levels. However, there are some important variations in the degree of the gender disparities within racial-ethnic groups. For example, some studies have found that the racial gap in earnings between African American and white women disappears after controls for social class and related variables, while the racial gap for men persists (e.g., Thomas 1993). Therefore, it is important to take into account how race affects men and women differently.

Race, Space, and Segregation

Residential segregation has a negative impact on the quality of life of African Americans. Living in a segregated urban area can have adverse effects on health because African Americans living in predominantly African American neighborhoods have poorer health facilities and greater exposure to environmental risks (Landrine and Corral 2009). In addition, residents of segregated, majority–African American neighborhoods are subject to suspicion from police and are much more likely than residents of integrated neighborhoods to be victim to police brutality and unwarranted stops (Brunson and Weitzer 2009; Weitzer 2000). African American’s homes in segregated metropolitan areas often fail to appreciate, or they even lose value (Flippen 2004; Kim 2003). Racial segregation may have negative consequences for African American employment opportunities. According to spatial mismatch theory, African Americans in segregated, inner-city neighborhoods are at a disadvantage in accessing employment opportunities (Jargowsky 1997; Massey and Denton 1993).

One thing that is clear from prior research is that the level of racial-ethnic discrimination and inequality varies across regions and localities. Many African Americans continue to live in the rural south, particularly in the Black Belt regions, which have been characterized by high levels of racial inequality and exploitation since slavery (Tomaskovic-Devey and Roscigno 1996). Although the historical pattern differs somewhat across regions, it is also the case that as the size of the minority population increases so do the levels of discrimination, segregation, and racial-ethnic inequality (Beggs et al. 1997; Blalock 1967; Cohen 1998; McCreary, England, and Farkas 1989; Tienda and Lii 1987).

Housing segregation, in particular, has been singled out as one of the most powerful local practices producing racial-ethnic inequality (Jargowski 1997; Massey and Denton 1993). Kirschenman and Neckerman (1991) report that employers often take residence into account when deciding not to hire otherwise qualified minority candidates. Those who live in clearly defined “black” areas may be seen as less desirable employees. In general, the literature points to local processes of racial competition and discrimination in employment and housing creating spatial variation in racial-ethnic labor market opportunity. Where jobs spread to suburban areas and blacks are disproportionately located in inner city areas due to segregation, blacks are less likely to be employed (Farley 1987).

Compared with non-Hispanic whites, Asians, and Latinos, blacks are much more spatially isolated from job opportunities (Stoll and Covington 2012). Stoll and Covington used census data and zip code business patterns to measure the imbalance between where people live and where the jobs are, for each racial group. The investigators found that the geographical isolation from jobs that blacks face as a result of racial segregation is a large contributor to the black-white gap in access to jobs. Using data from the National Longitudinal Survey of Youth, Howell-Moroney (2005) found that segregation has a negative impact on blacks’ educational attainment as well as probability of employment, net of educational attainment. Dickerson (2007) found that rates of employment for blacks are lower in the more segregated cities across the country, and, as segregation increases over time, rates of employment for blacks decrease, once again linking segregation with lower employment outcomes for blacks.

Other researchers disagree with the thesis of the spatial mismatch hypothesis. Using data from the 2000 census, Hellerstein et al. (2008) created education and race-specific measures of job density for each zip code. They found that blacks are not being hired for the jobs in the neighborhoods around them. The investigators concluded that a skills mismatch, and perhaps discrimination, rather than simply a spatial mismatch, work against blacks (Hellerstein et al. 2008). Weinburg (2000) also argued that black residential centralization—the tendency for blacks to live in neighborhoods close to the central business districts—has a much larger impact on the black-white employment gap than does segregation. Boustan and Margo (2009) examined spatial mismatch and decentralization by focusing on black postal employment and found that blacks gravitated to postal employment as other employers moved to suburban locations. However, Boustan and Margo also found that the link between segregation and postal employment declined between 1970 and 2000 and that spatial mismatch is not as important as it once was.

Much research has been conducted on trends in segregation and the possible causes and consequences of segregation, but relatively little research has explored the negative impact of segregation on the earnings of African Americans. African Americans who live in segregated neighborhoods may have a more difficult time networking with coworkers if they live far from each other. Networking can play a large role in determining who advances within a company and within an industry. White male networks provide twice as many job leads as nonwhite networks and on average are composed of higher status individuals (McDonald 2011). When white managers need to fill a position requiring a college degree, they are more likely to hire a white person when using their social networks to recruit (Braddock and McParkland 1987). Minority networks may be limited in helping blacks achieve high-wage positions because white males remain vastly overrepresented in managerial positions (Stainback and Tomaskovic-Devey 2009).

Black-white residential segregation can negatively affect blacks’ job outcomes because segregation reduces the opportunity for blacks to network with nonblacks. Green, Tigges, and Brown (1995) found that African Americans who live in segregated high-poverty neighborhoods had fewer college-educated individuals in their social circle. The more segregated a metropolitan area is, the more African American workers will need to rely on individuals in their neighborhoods for job leads. Stoll and Raphael (2000) found that racial segregation affects job search patterns, as job seekers tend to search areas within a limited distance of where they live. This limits the quality of jobs searched by blacks living in segregated neighborhoods, which in turn can limit the quality of job obtained. Elliott (1999) found that blacks living in segregated neighborhoods are more likely to work in segregated workplaces, typically predominantly “black” jobs that pay less than jobs with nonblacks. Together, the work of Elliott (1999), Stoll and Raphael (2000), and Green et al. (1995) suggests that blacks in more segregated metropolitan areas will have lower employment outcomes due in part to constraints related to black-white residential segregation.

If spatial mismatch contributes to lower employment outcomes for African Americans because of limitations in their ability to commute, its impact may be more severe for African American women. Women tend to have shorter commutes than men; however, African American women commuting to suburban jobs tend to have the longest commutes (Johnston-Anumonwo 1997; McLafferty and Preston 1991). This situation may be even more difficult for African American women who balance both job and childcare responsibilities. Therefore, the impact of segregation on African American women’s earnings may be greater than the impact on African American men’s earnings (Wang 2008).

Research Questions

What is the impact of segregation on black-white income inequality? If African Americans in segregated metropolitan areas experience economic hardships that African Americans in less segregated places do not, then we would expect the black-white disparity in income to be greater in highly segregated places. Furthermore, the impact of segregation on racial disparities may be different for individuals of high socioeconomic status (SES) compared with low SES. The class stratification of African Americans means that there is no uniform “black experience” (Wilson 1980). Higher status African Americans may be insulated from the negative impact of racial segregation, and we would expect their incomes to be similar to whites’ incomes regardless of the level of segregation they experience.

Is the impact of segregation different for men and women? As stated earlier, much of the research has focused on racial differences between African American and white men. Part of this focus is due to the fact that the racial gap in income between African American and white men has historically been greater than the gap between African American and white women (e.g., Thomas 1993). However, given the unique challenges that African American women face, segregation may affect them differently.

Based on the previous research, several hypotheses will be tested:

Hypothesis 1. The African American–white income disparity will be greater for high-SES individuals.

Hypothesis 2. The African American–white income disparity will be greater in more highly segregated metropolitan areas.

Hypothesis 3. Segregation will be associated with greater African American–white income disparities among low-SES individuals than high-SES individuals.

Hypothesis 4. Segregation will be associated with greater African American–white income disparities among women than men.

This study addresses these questions by examining the impact of racial segregation and its interplay with social class and gender on the earnings of African Americans and whites.

Data and Methods

The data for this study come from the 2010 IPUMS Current Population Survey (CPS) merged with 2010 indices of dissimilarity scores for the largest 100 (over 500,000) metropolitan areas. Included in the analysis are African American and whites living in those metropolitan areas who worked full-time the previous year. The sample size is 33,528 (27,204 white and 6,324 African American). The dependent variable is income from wages and salaries the previous year.

The index of dissimilarity is used as a measure of black-white segregation. It measures the relative separation of groups across all the neighborhoods of a metropolitan area. Other independent demographic variables include race, sex, age, region, and SES. Race is recoded into a dummy variable where African Americans equal one and whites equal zero. Sex is also recoded into a dummy variable where female equals one and male equals zero. Age of the respondent (coded as actual years) and age-squared are used to estimate the age effect. Region of the interview is recoded into a dummy variable where South equals one and non-South equals zero. SES is the average z-score for education and occupational prestige. Educational attainment is represented by number of years of schooling. Occupational prestige is the 1989 Nakeo-Treas social ranking of occupations (Nakao and Treas 1994).¹

Summary Statistics

The mean income is $57,243.52. Across all of the metropolitan areas, the average dissimilarity index is 64.94, meaning that 64.94% of the black population would need to relocate to a majority white neighborhood in order to achieve perfect integration. The five most segregated metropolitan areas are Milwaukee, New York, Chicago, Detroit, and Cleveland. The five least segregated metropolitan areas are Modesto, Ogden, Albuquerque, El Paso, and Boise (Table 1).

Table 1.

Summary Statistics.

	Mean	Standard Deviation
Income	$57,243.52	$61,793.26
Non-Hispanic black	−.189	−.391
Female	−.44	−.50
Age, years	43.01	12.10
South	−.35	−.48
City	−.33	−.47
Socioeconomic status	−.00	−.87
Dissimilarity	64.94	8.90

We include a multilevel model to assess the impact of metropolitan-level segregation on wages for individuals. Specifically, in level 1, individual characteristics such as age, gender, SES, and race are included as predictors of wage income. Level 2 allows us to account for variations in wage income across different metropolitan areas as a function of race-related metropolitan-level differences. The level 2 variable is the index of dissimilarity (measuring the degree of black-white racial residential segregation in the metropolitan area).

The model is represented by the following equations. The individual-level model is:

y_{i j} = β_{0 j} + β_{k j} χ_{k i j} + ε_{i j},

where y_ij is wage income for the ith individual in metropolitan area j, β_0j is the intercept for metropolitan area j, χ_kij is the value of the kth variable for individual i in metropolitan area j, β_kj is the corresponding vector of coefficients, and ε_ij is the error term, assumed to be normally distributed with a mean of 0.

The intercepts from the individual-level model become the dependent variable in the metropolitan-level equation:

β_{0 j} = γ_{00} + Ζ_{l j} + υ_{0 j} .

β_0j is the intercept from the individual-level model, which represents the average wage income adjusted for individual attributes. γ₀₀ is the metropolitan-level intercept, Z_ij represents the dissimilarity score of metropolitan area j, and υ_0j is the error term specific to the metropolitan-level model, assumed to be normally distributed with a mean of 0.

This study includes both two- and three-way interactions. Examining the race and socioeconomic status interaction will help determine whether the racial differences in the income vary by socioeconomic status, and examining the Black × Dissimilarity interaction will help determine whether they vary by segregation. Examining the three-way Black × SES × Dissimilarity interaction will reveal whether race, SES, and segregation in combination affect income, addressing the third hypothesis. Each of the independent variables were centered (i.e., subtracting the mean from each score of the variable in question) prior to creating interaction terms.

Results

We estimated four models for all respondents with income as the dependent variable (Table 2). Model I includes all of the key variables without any interaction terms. Model I includes the demographic variables including race (black) and gender (female), age, urbanity (city), and region (South), SES and metropolitan Dissimilarity Index score in the model. To test the second hypothesis, Model II includes all of the variables from Model I and adds the Black × SES interaction terms. To test the third hypothesis, Model III includes the Black × Dissimilarity interaction term. Model IV adds the two-way Black × SES × Dissimilarity interaction term. To test the fourth hypothesis that segregation affects black women differently than black men, Table 3 reproduces the complete analysis from Table 2 but only includes males, and Table 4 does the same analysis but includes only females.

Table 2.

Unstandardized Coefficients for Hierarchical Linear Model (HLM) Regression on Annual Wage Income (Full Sample).

	1	2	3	4
Black	−18.044***(1.373)	−20.287***(1.407)	−16.317***(1.633)	−18.07***(1.699)
Female	−31.392***(1.001)	−31.143***(1.001)	−31.36***(1.001)	−31.106***(1.001)
Age	1.082***(.041)	1.08***(.041)	1.083***(.041)	1.081***(.041)
Age-squared	−.074***(.003)	−.074***(.003)	−.074***(.003)	−.074***(.003)
South	−3.096(2.041)	−3.147(2.03)	−3.339(2.044)	−3.357⁺(2.032)
City	−4.36***(1.136)	−4.856***(1.137)	−4.222***(1.138)	−4.75***(1.139)
SES	47.4***(.59)	47.339***(.589)	47.392***(.59)	47.436***(.592)
Dissimilarity	.258**(.096)	.256**(.095)	.288**(.097)	.297**(.097)
Black × SES interaction		−10.884***(1.498)		−8.903***(1.801)
Black × Dissimilarity interaction			−.25⁺(.128)	−.305*(.135)
Black × SES × Dissimilarity two-way interaction				−.285⁺(.149)
Constant	230.169***(1.455)	230.698***(1.448)	230.01***(1.456)	230.445***(1.45)
Percentage of Level 1 variation explained	22.86%	22.97%	22.86%	22.98%
Percentage of Level 2 variation explained	64.10%	64.82%	64.20%	64.96%

Notes: SES = socioeconomic status. The dependent variable is the square root of wage income. Standard errors are in parentheses.

p ≤ .05 (one-tailed), *p ≤ .05 (two-tailed), **p ≤ .01, ***p ≤ .001

Table 3.

Unstandardized Coefficients for Hierarchical Linear Model (HLM) Regression on Annual Wage Income, Males.

	1	2	3	4
Black	−23.313***(2.184)	−27.363***(2.314)	−20.528***(2.556)	−23.255***(2.743)
Age	1.144***(.063)	1.138***(.063)	1.144***(.063)	1.14***(.063)
Age-squared	−.093***(.004)	−.093***(.004)	−.093***(.004)	−.093***(.004)
South	−2.971(2.614)	−3.092(2.606)	−3.278(2.617)	−3.352(2.6)
City	−9.844***(1.706)	−10.324***(1.707)	−9.662***(1.708)	−10.165***(1.708)
SES	50.926***(.866)	50.417***(.871)	50.9***(.866)	50.547***(.874)
Dissimilarity	.262*(.116)	.26*(.116)	.24*(.117)	.237*(.116)
Black × SES interaction		−12.637***(2.389)		−9.335***(2.832)
Black × Dissimilarity interaction			−.423*(.202)	−.608**(.221)
Black × SES × Dissimilarity two-way interaction				−.505*(.235)
Constant	224.216***(1.356)	224.24***(1.351)	224.391***(1.357)	224.442***(1.347)
Percentage of Level 1 variation explained	22.11%	22.22%	22.11%	22.23%
Percentage of Level 2 variation explained	54.86%	55.49%	55.17%	56.29%

Notes: SES = socioeconomic status. The dependent variable is the square root of wage income. Standard errors are in parentheses.

p ≤ .05, **p ≤ .01, ***p ≤ .001.

Table 4.

Unstandardized Coefficients for Hierarchical Linear Model (HLM) Regression on Annual Wage Income, Females.

	1	2	3	4
Black	−12.845***(1.582)	−13.77***(1.59)	−12.03***(1.909)	−12.948***(1.945)
Age	.993***(.05)	.997***(.05)	.994***(.05)	.997***(.05)
Age-squared	−.05***(.003)	−.051***(.003)	−.05***(.003)	−.051***(.003)
South	−1.51(2.168)	−1.531(2.161)	−1.677(2.18)	−1.652(2.174)
City	2.315⁺(1.375)	1.823(1.376)	2.388⁺(1.378)	1.861(1.38)
SES	42.512***(.747)	42.875***(.749)	42.512***(.747)	42.927***(.753)
Dissimilarity	−.258*(.1)	−.253*(.1)	−.248*(.101)	−.246*(.1)
Black × SES interaction		−9.273***(1.762)		−8.368***(2.133)
Black × Dissimilarity interaction			−.114(.15)	−.107(.153)
Black × SES × Dissimilarity two-way interaction				−.127(.176)
Constant	194.083***(1.141)	194.112***(1.137)	194.167***(1.147)	194.18***(1.143)
Percentage of Level 1 variation explained	22.43%	22.59%	22.43%	22.58%
Percentage of Level 2 variation explained	56.29%	56.93%	56.25%	56.86%

Notes: SES = socioeconomic status. The dependent variable is the square root of wage income. Standard errors are in parentheses.

p ≤ .05, **p ≤ .01, ***p ≤ .001

In Model I, shown in Table 2, blacks have lower wages than whites, and females have lower wages than males. Wages increase with age. Living in the south does not have any impact on wages, net of city intercept in the multilevel model. Individuals living in a city have significantly lower wages than individuals living in the suburbs. SES has a large, positive impact on wages. In addition, dissimilarity has a positive impact on wages. The more racially segregated an area is, the higher the incomes for the average wage earner. In Model II, interaction terms have been added for race and SES. The Black × SES interaction variable is negative and statistically significant. High-SES blacks earn significantly less than would be expected based on SES or race alone. For blacks, as SES increases, the wage disparity between them and their white counterparts increases.

The interaction term for race and segregation in Model III is not statistically significant. However, in Model IV, a three-way interaction term is added, Black × SES × Dissimilarity, and then the Black × Dissimilarity interaction becomes statistically significant. Blacks who live in segregated areas have significantly lower incomes than would be expected based on race or SES alone. The models presented are evidence that segregation uniquely affects high-SES African Americans.

In Table 3, the same series of models are estimated for only the men in the sample. With only men in the model, all of the interaction terms are statistically significant. For black men in particular, living in an area with high levels of segregation negatively affects their wages compared with white men. In Table 4, the same models are estimated including only females. For the models including only females, the city variable is not significant in the full model, and the Black × Dissimilarity interaction term is not significant. So the combination of race, SES, and dissimilarity has a negative impact on black men and not on black women. In fact, living in a segregated metropolitan area has a positive impact on wages for women, but that positive impact is not exclusive to black women. The Black × SES interaction term is still significant, so as black women attain higher levels of SES, the wage disparity between them and their white peers increases.²

Figures 1 and 2 show how predicted wages increase as SES increases for each race-gender group in the least and most segregated metropolitan areas. Figure 1 shows the relationship in the least segregated metropolitan areas (25th percentile or lower dissimilarity score), while Figure 2 shows the same relationship in the most segregated areas (75th percentile or higher dissimilarity score). The key difference is that the income disparity between high-SES white males and high-SES black males is much larger in the most segregated metropolitan areas. Further, as SES increases, the income disparity grows even more. The slope of the line is much steeper for white males in highly segregated metropolitan areas. Also important is that high-SES African American males in highly segregated metropolitan areas lose their predicted wage advantage over white women.

This is further evidence that human capital deficiencies cannot account for the black-white wage disparity and that there are other factors at work limiting blacks’ incomes as they acquire more education and obtain professional positions (Table 5).

Table 5.

Beta Coefficients for Wage Income for Men and Women in Least Segregated (25th Percentile or Lower) and Most Segregated (75th Percentile or Higher) Metropolitan Areas.

Variables^a	Least Segregated Metropolitan Areas	Most Segregated Metropolitan Areas
Men
Black	−.056***	−.091***
	(–10,867.41)	(–17,072.57)
SES	.299***	.337***
	(21,909.50)	(27,962.14)
Black × SES	−.033	−.064***
	(–7,215.53)	(–13,888.25)
Constant	66,384.61	71,651.01
Adjusted R²	.142	.180
Women
Black	−.057***	−.064***
	(–6,685.47)	(–6,693.57)
SES	.315***	.355***
	(15,832.33)	(19,419.91)
Black × SES	−.034	−.055***
	(–4,601.58)	(–6,572.10)
Constant	45,824.52	51,249.24
Adjusted R²	.142	.157

Notes: SES = socioeconomic status. Unstandardized regression coefficients are in parentheses.

Control variables are age, age-squared, south, city.

p ≤ .05, **p ≤ .01, ***p ≤ .001

Hypothesis 1 was supported. The income disparity between African Americans and whites was greater for high-SES men and women. The results provide qualified support for hypothesis 2. The African American and white income disparity was greater in more segregated metropolitan areas, but only for males. Hypotheses 3 and 4 were not supported. Segregation did not more negatively affect the incomes of low-SES African Americans. In fact, segregation widened the income gap between African American and white males. This gap grew with increasing SES. Segregation did not significantly affect the incomes of African American women.

Conclusions

This study examines the impact of segregation and social class on the earnings of African American and white men and women. Important findings indicate that (1) earning differences between African Americans remain after controls for SES and demographic variables, (2) racial differences increase with rising social class status, (3) segregation increases the disparity between African American and white males, and (4) the increasing racial disparities with rising social class are more acute for males in more segregated areas.

Controlling for SES, age, region, and urbanity and related variables does not eliminate racial disparities. In fact, as SES rises, so do racial disparities. This supports the idea that racial differences in income are not caused by African Americans lacking human capital. Rather, these differences are caused by difficulties that African Americans face in translating their human capital to income in the labor market. If African Americans are not receiving the same financial benefits as their white counterparts for their education and skills, then the more of such skills they acquire, the further they will fall behind similar whites. This pattern was evident for both men and women. The findings support studies that suggest social closure provides some explanation for racial disparities in income. Because more income and status are at stake at the high-SES end, whites fight harder to protect their privilege. Also, high-SES whites have more resources with which to resist African Americans’ efforts to secure a portion of the advantages that come with more human capital.

The initial findings indicated that segregation increased the racial disparities in income and the increase in disparities with SES. However, further analysis revealed that this pattern was specific to males. The income disparity between African American and white women was not affected significantly by segregation. However, white women were similar to white males in that their wages rose more steeply with rising SES than for either African American men or women. Also, because of the increased difficulties African American males have converting their high levels of SES into income in more highly segregated metropolitan area, the predicted earnings of high-SES white women are nearly identical to those of similar African American males in these areas. This finding supports previous research (e.g., Stainback and Tomaskovic-Devey 2012) which documents the improving access of white women to traditionally white male-dominated professional jobs that has outpaced the movement of both African American men and women into such jobs. However, the most striking finding was that the racial and gender advantage that white males experienced increased with rising SES and became more pronounced in more highly segregated metropolitan areas.

An alternative explanation of the results is suggested by the competition literature, which, as stated earlier, posits that as the size of the minority population increases, so does the level of discrimination (Beggs et al. 1997; Blalock 1967; Cohen 1998; McCreary et al. 1989; Tienda and Lii 1987). Because including percentage black in models that included dissimilarity caused multicollinearity problems, we estimated models (not presented here) using the percentage of black residents as the level 2 variable instead of dissimilarity in order to see whether it produced similar results. In those models, none of the interactions that we found with dissimilarity were significant with percentage black. The negative impact of race on wages did not increase as percentage black increased. This was true for the male, female, and combined samples.

Why segregation increases the racial disparities for males is a difficult question to answer with this analysis; however, previous research suggests some possible explanations. One explanation is that there is elevated racial fear based on stereotypes of African American males in highly segregated metropolitan areas. African American males have been stereotyped as lazy, irresponsible, criminal, violent, and prone to substance abuse. Segregated metropolitan areas may be environments in which these views can be nurtured and sustained. Because integrated metropolitan areas allow for more daily interracial contact, they may provide more experiences that challenge stereotypes. If this is the case, then even highly educated African American males may be affected by the “ghetto” stereotype because middle-class status cannot shield them from being associated with the negative images.

Another possible explanation is that African American males tend to pursue jobs that are more affected by segregation. These would be jobs in which gatekeepers and mentors are predominately white males. African American males who live in highly segregated metropolitan areas may find their employment and mobility opportunities significantly reduced because of restricted access to white male–dominated networks. If this is the case, then segregation makes this problem more acute for middle-class African American males.

Limitations

This study has several limitations. First, the problem of overcontrol is inherent in these types of analyses. When accessing the relative impact of race, we have to control for variables (e.g., education, occupation, etc.) that themselves are affected by race. The result is that we have probably underestimated the impact of race by removing some of its effects. Therefore, the results are a conservative estimation, and the true impact of race is probably greater than what is presented here.

Second, a study of this nature has limited ability to identify the specific mechanisms that produce the findings presented here. More research, perhaps qualitative, is needed to identify the income-inhibiting factors for African Americans that are more prevalent in more highly segregated areas than in less segregated areas. Also of interest is determining what mechanisms particularly advantage white males and disadvantage African American males in segregated environments.

A third limitation is our focus on the impact of black-white segregation on the earnings of employed African Americans and whites. African Americans experience significantly more racial segregation than any racial or ethnic group in the United States. Nevertheless, it may be instructive for future research to focus on the racial segregation of Latinos, Asians, and Native Americans. Residential segregation may have a similar impact on the incomes of these groups. African Americans are more likely to be unemployed than whites and are therefore excluded from the analysis. Because if this, the African Americans in this analysis are somewhat of a select group. The impact of segregation on short-term and long-term unemployment is beyond the scope of this study.

Footnotes

Acknowledgements

Special thanks to Anna Manzoni, Jeff Leiter, Donald Tomaskovic-Devey, and Steve McDonald, who commented on earlier versions of this paper.

Notes

Author Biographies

Melvin E. Thomas is an Associate Professor in the Department of Sociology and Anthropology at North Carolina State University. His research focuses on racial and social class inequalities and how they affect the lives of minorities in the United States. He earned a PhD in Sociology from Virginia Polytechnic Institute and State University.

Richard Moye is an assistant professor of Sociology in the Department of Behavioral Sciences at Winston Salem State University. He earned his PhD in Sociology at Temple University in 2011. His research focuses on patterns of racial residential segregation, explaining trends in shifts in segregation patterns, and on the consequences of residential segregation for overall racial inequality in American society.

References

Becker

Gary

. 1975. Human Capital. New York: National Bureau of Economic Research.

Beggs

John

Villemez

Wayne

Arnold

Ruth

. 1997. “Black Population Concentration and Black-white Inequality: Expanding the Consideration of Place and Space Effects.” Social Forces 76(1):65–97.

Belkhir

Jean A.

McNair

Bernice

. 2001. “Race, Gender and Class Intersectionality.” Race, Gender & Class 8:157–74.

Blalock

Hubert M.

Jr.

1967. Toward a Theory of Minority-group Relations. New York: Wiley.

Boustan

Leah Platt

Margo

Robert A.

2009. “Race, Segregation, and Postal Employment: New Evidence on Spatial Mismatch.” Journal of Urban Economics 65(1):1–10.

Braddock

Jomills Henry

McPartland

James

. 1987. “How Minorities Continue to Be Excluded from Equal Opportunities: Research on Labor Markets and Institutional Barriers.” Journal of Social Issues 43(1):5–39.

Browne

Irene

Misra

Joya

. 2003. “The Intersection of Gender and Race in the Labor Market.” Annual Review of Sociology 29:487–513.

Brunson

Rod K.

Weitzer

Ronald

. 2009. “Police Relations with Black and White Youths in Different Urban Neighborhoods.” Urban Affairs Review 44(6):858–85.

Cohen

Phillip

. 1998. “Black Concentration Effects on Black-white and Gender Inequality: Multilevel Analysis for U.S. Metropolitan Areas.” Social Forces 77(1):207–29.

10.

Cotton

Jeremiah

. 1989. “Opening the Gap: The Decline of Black Economic Indicators in the 1980s.” Social Science Quarterly 70:803–19.

11.

Cotton

Jeremiah

. 1990. “The Gap at the Top: Relative Occupational Earnings Disadvantages of the Black Middle Class.” The Review of Black Political Economy 18(3):21–38.

12.

Dickerson

Nikki T.

2007. “Black Employment, Segregation, and the Social Organization of Metropolitan Labor Markets. Economic Geography 83(3):283–307.

13.

Elliott

James R.

1999. “Social Isolation and Labor Market Insulation: Network and Neighborhood Effects on Less-Educated Urban Workers.” The Sociological Quarterly 40(2):199–216.

14.

Farkas

George

England

Paula

Vicknair

Kevin

Kilbourne

Barbara Stanek

, 1997. “Cognitive Skill, Skill Demands of Jobs, and Earnings Among Young European American, African American, and Mexican American Workers.” Social Forces 75:913–40.

15.

Farley

John E.

1987. “Disproportionate Black and Hispanic Unemployment in U.S. Metropolitan Areas: The Roles of Racial Inequality, Segregation and Discrimination in Male Joblessness.” American Journal of Economics and Sociology 46(2):129–50.

16.

Flippen

Chenoa

. 2004. “Unequal Returns to Housing Investments? A Study of Real Housing Appreciation among Black, White, and Hispanic Households.” Social Forces 82(4):1523–51.

17.

Glaeser

Edward

Vigdor

Jacob

. 2012. The End of the Segregated Century: Racial Separation in America’s Neighborhoods, 1890–2010. New York: Manhattan Institute for Policy Research.

18.

Green

Gary P.

Tigges

Leann M.

Browne

Irene

. 1995. “Social Resources, Job Search, and Poverty in Atlanta.” Research in Community Sociology 5:161-182.

19.

Hellerstein

Judith K.

Neumark

David

McInerney

Melissa

. 2008. “Spatial Mismatch or Racial Mismatch?” Journal of Urban Economics 64(2):464–79.

20.

Howell-Moroney

Michael

. 2005. “The Geography of Opportunity and Unemployment: An Integrated Model of Residential Segregation and Spatial Mismatch.” Journal of Urban Affairs 27(4):353–77.

21.

Jargowsky

Paul A.

1997. Poverty and Place: Ghettos, Barrios, and the American City. New York: Russell Sage Foundation.

22.

Johnston-Anumonwo

Ibipo

. 1997. “Race, Gender, and Constrained Work Trips in Buffalo, NY, 1990.” The Professional Geographer 49(3):306–17.

23.

Kim

Sunwoong

. 2003. “Long-term Appreciation of Owner-occupied Single-family House Prices in Milwaukee Neighborhoods.” Urban Geography 24(3):212–31.

24.

Kirschenman

Joleen

Neckerman

Kathryn M.

1991. “‘We’d Love to Hire Them, But . . .’: The Meaning of Race for Employers.” The Urban Underclass 203:203–32.

25.

Landry

Bart

. 1987. The New Black Middle Class. Berkeley: University of California Press.

26.

Landrine

Hope

Corral

Irma

. 2009. “Separate and Unequal: Residential Segregation and Black Health Disparities.” Ethnicity & Disease 19(2):179–84.

27.

Lips

Hilary M.

2013. The Gender Pay Gap: Challenging the Rationalizations. Perceived Equity, Discrimination, and the Limits of Human Capital Models. Sex Roles 68(3-4):169–85.

28.

Massey

Douglas S.

Denton

Nancy A.

1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press.

29.

McCreary

Lori

England

Paula

Farkas

George

. 1989. “The Employment of Central City Male Youth: Nonlinear Effects of Racial Composition.” Social Forces 68(1):55–75.

30.

McDonald

Steve

. 2011. “What’s in the “Old Boys” Network? Accessing Social Capital in Gendered and Racialized Networks.” Social Networks 33(4):317–30.

31.

McLafferty

Sara

Preston

Valerie

. 1991. “Gender, Race and Commuting Among Service Sector Workers.” The Professional Geographer 43(1):1–15.

32.

Nakao

Keiko

Treas

Judith

. 1994. “Updating Occupational Prestige and Socioeconomic Scores: How the New Measures Measure Up.” Sociological Methodology 24:1–72.

33.

Neal

Derek A.

Johnson

William A.

1996. “The Role of Premarket Factors in Black-white Wage Differences.” Journal of Political Economy 104:869–95.

34.

Semyonov

Moshe

Lewin-Epstein

Noah

. 2009. “The Declining Racial Earnings Gap in United States: Multi-level Analysis of Males’ Earnings, 1960–2000.” Social Science Research 38(2):296–311.

35.

Stainback

Kevin

Tomaskovic-Devey

Donald

. 2009. “Intersections of Power and Privilege: Long-term Trends in Managerial Representation.” American Sociological Review 74(5):800–20.

36.

Stainback

Kevin

Tomaskovic-Devey

Donald

. 2012. Documenting Desegregation: Racial and Gender Segregation in Private-sector Employment Since the Civil Rights Act. New York: Russell Sage Foundation.

37.

Stoll

Michael A.

Covington

Kenya

. 2012. “Explaining Racial/Ethnic Gaps in Spatial Mismatch in the US: The Primacy of Racial Segregation.” Urban Studies 49(11):2501–21.

38.

Stoll

Michael A.

Raphael

Steven

. 2000. “Racial Differences in Spatial Job Search Patterns: Exploring the Causes and Consequences.” Economic Geography 76(3):201–23.

39.

Thomas

Melvin

. 1993. “Race, Class, and Personal Income: An Empirical Test of the Declining Significance of Race Thesis, 1968–1988.” Social Problems 40(3):328–42.

40.

Thomas

Melvin

. 1995. “Race, Class, and Occupation: An Analysis of Black and White Earning Differences for Professional and Non-professional Men, 1940–1990.” Pp. 139–56 in Research in Race and Ethnic Relations, edited by Rutledge

Greenwich, CT: JAI Press.

41.

Thomas

Melvin

Herring

Cedric

Horton

Hayward D.

1994. “Discrimination over the Life Course: A Synthetic Cohort Analysis of Earnings Differences between Black and White Males, 1940–1990.” Social Problems 41(4):608–28.

42.

Thomas

Melvin

Herring

Cedric

Horton

Hayward D.

1995. “Racial and Gender Differences in Returns from Education.” Pp. 239–53 in Race and Ethnicity in America: Meeting the Challenge in the 21st Century, edited by Thomas

Washington, DC: Taylor and Francis.

43.

Thomas

Melvin

Horton

Hayward D.

1992. “Race, Class, and Family Structure: The Case of Family Income.” Sociological Perspectives. 35(3):433–50.

44.

Tienda

Marta

Lii

Ding-Tzann

. 1987. “Minority Concentration and Earnings Inequality: Blacks, Hispanics, and Asians Compared.” American Journal of Sociology 93:141–65.

45.

Tomaskovic-Devey

Donald

, 1993. Gender and Racial Inequality at Work: The Sources and Consequences of Job Segregation. Ithaca, NY: ILR Press.

46.

Tomaskovic-Devey

Donald

Thomas

Melvin

Johnson

Kecia

. 2005. “Race and the Accumulation of Human Capital across the Career: A Theoretical Model and Fixed-effects Application.” American Journal of Sociology 111(1):58–89.

47.

Tomaskovic-Devey

Donald

Roscigno

Vincent J.

1996. “Racial Economic Subordination and White Gain in the U.S. South.” American Sociological Review 61:565–89.

48.

Wang

Qingfang

. 2008. “Race/Ethnicity, Gender and Job Earnings across Metropolitan Areas in the United States: A Multilevel Analysis.” Urban Studies 45(4):825–43.

49.

Weinberg

Bruce A.

2000. “Black Residential Centralization and the Spatial Mismatch Hypothesis.” Journal of Urban Economics 48(1):110–34.

50.

Weitzer

Ronald

. 2000. “Racialized Policing: Residents’ Perceptions in Three Neighborhoods.” Law & Society Review 34:129–55.

51.

Wilson

William Julius

. 1980. The Declining Significance of Race. Chicago: University of Chicago Press.

52.

Yap

Margaret

Konrad

Alison M.

2009. “Gender and Racial Differentials in Promotions: Is There a Sticky Floor, a Mid-level Bottleneck, or a Glass Ceiling?” Industrial Relations 64:593–619.