打印

作者

Ingrid Gould Ellen;Justin P. Steil;Jorge De la Roca

来源

CITY & COMMUNITY,Vol.15,Issue1,P.8-13

语言

英文

关键字

作者单位

Wagner School New York University;Department of Urban Studies and Planning Massachusetts Institute of Technology;Sol Price School of Public Policy University of Southern California

摘要

Researchers have vigorously debated the significance of the reductions in residential segregation by race that U.S. metropolitan areas have experienced. While some argue that we have witnessed the “end of the segregated century” (Vigdor and Glaeser 2012; Vigdor 2013), others highlight the persistence of high levels of segregation in many areas (e.g., Logan 2013). There has been far less debate about the relationship between segregation and access to opportunity in the 21st century. Yet such exploration is critical to a richer understanding of the significance of segregation. Almost two decades ago, David Cutler and Edward Glaeser (1997) found that in 1990 a one‐standard‐deviation reduction in levels of residential segregation would eliminate one‐third of the gap between whites and blacks in high school graduation rates, earnings, rates of single motherhood, and the likelihood of being simultaneously out of work and out of school. Cutler and Glaeser (1997) also found that only one‐third of the effects of segregation could be explained by exposure to less educated neighbors, distance from jobs, or parental background. Analyzing data from 1940 through 1980, however, William Collins and Robert Margo (2000) subsequently found that some of the negative socio‐economic effects of segregation identified by Cutler and Glaeser (1997) were actually a relatively recent development. For instance, Collins and Margo (2000) identified no significant relationship between segregation and the likelihood of employment or single motherhood between 1940 and 1970. These findings suggest that some of the negative effects of black–white residential segregation emerged primarily with the economic restructuring and dramatic neighborhood change of the 1970s (Wilson 1996). Together, these results indicate that the effects of segregation vary over time in relation to broader social, economic, and political developments. There are at least three key questions to explore about the effects of segregation in the 21st century. First, does segregation continue to constrain economic and social mobility for minority groups? Second, are the effects similar for the two largest non‐white groups in the United States, African Americans and Latinos, despite their different historical experiences of segregation? Third, and perhaps most importantly, how does segregation matter in the 21st century? In other words, what are the mechanisms through which racial or ethnic segregation contribute to inequality? We take each of these questions in turn. In recent work, we showed that higher levels of segregation continue to be associated with wide disparities by race in socio‐economic outcomes (Steil et al. 2015). Here, we focus on changes over time in those relationships. Specifically, Table 1 presents results from ordinary least squares regressions of four individual educational and labor market outcomes in 1990, 2000, and 2010 on contemporaneous levels of metropolitan area segregation in those three years (as measured by the dissimilarity index). In the first panel, the sample is restricted to native‐born whites and blacks between the ages of 25 and 30. We consider four outcomes: college graduation; not being idle (or, being either in school or the labor force); professional occupation; and log earnings. In addition to segregation measures, all regressions also include individual controls for age and gender, census region year indicator variables as well as a set of CBSA control variables, which include log population, log median household income, poverty rate, and the shares of the population that are black, Hispanic, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, and college‐educated. In order to ensure that we capture people who have lived in the metropolitan area for a substantial period of time, we exclude individuals from the 1990 and 2000 samples who lived in another CBSA five years ago and exclude those who lived in another CBSA one year ago from the 2010 sample. Table 1. Estimation of the Effect of Metropolitan Area Segregation on Individual Outcomes College Not Professional Log graduation idle occupation earnings Dependent variable: (1) (2) (3) (4) Black‐white DI 1990 0.022 0.001 0.030 0.054 (0.036) (0.021) (0.027) (0.098) Black‐white DI 2000 0.091 0.014 0.074 0.033 (0.040)**Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.020) (0.032)**Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.100) Black‐white DI 2010 0.236 0.049 0.125 0.088 (0.045)**Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.018)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.034)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.088) Black‐white DI 1990 × black 0.185 0.166 0.072 0.463 (0.044)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.028)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.029)**Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.102)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Black‐white DI 2000 × black 0.213 0.163 0.086 0.370 (0.051)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.029)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.034)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.114)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Black‐white DI 2010 × black 0.277 0.100 0.148 0.411 (0.061)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.036)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.041)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.119)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Observations 1,560,958 1,560,958 1,560,958 1,436,969 Number of CBSAs 204 204 204 204 R2 0.084 0.044 0.047 0.145 Latino‐white DI 1990 0.041 0.033 0.025 0.226 (0.032) (0.013)**Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.022) (0.080)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Latino‐white DI 2000 0.036 0.053 0.037 0.235 (0.049) (0.016)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.032) (0.091)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Latino‐white DI 2010 0.179 0.089 0.099 0.175 (0.064)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.018)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.041)**Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.099)*Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Latino‐white DI 1990 × Latino 0.243 0.179 0.173 0.671 (0.051)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.025)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.033)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.114)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Latino‐white DI 2000 × Latino 0.275 0.195 0.177 0.591 (0.057)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.025)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.037)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.127)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Latino‐white DI 2010 × Latino 0.362 0.141 0.224 0.639 (0.060)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.026)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.039)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. (0.140)***Coefficients are reported with robust standard errors in parenthesis, which are clustered by Core Based Statistical Area (CBSA). ***, **, and * indicate significance at the 1, 5, and 10 percent levels. In the top (bottom) panel, the sample is restricted to native‐born whites and blacks (Latinos) between 25 and 30 years of age. Individuals who lived in another CBSA five years ago (or one year ago in 2010) are excluded. All specifications have a constant term, a female indicator variable, age, and census region‐year indicator variables. The top panel includes a black indicator variable while the bottom panel includes ancestry‐group Latino indicator variables. Additional CBSA controls include log population, log median household income and shares of population that are black, Latino, Asian, over 65 years, under 15 years, unemployed, working in manufacturing, in poverty status, and with college degree. These controls are also interacted with a black or Latino indicator variable accordingly. di stands for dissimilarity index. Observations 1,522,096 1,522,096 1,522,096 1,402,104 Number of CBSAs 207 207 207 207 R2 0.084 0.051 0.044 0.136