Double Minority Status and Neighborhoods: Examining the Primacy of Race in Black Immigrants’ Racial and Socioeconomic Segregation

Abstract

Sociologists have long viewed spatial assimilation as a measure of minorities’ socioeconomic progress. While assimilation increases as socioeconomic status (SES) improves, blacks remain more highly segregated than any other race/ethnic group. I use the locational attainment model to determine whether black immigrants—like their U.S.-born counterparts—are highly segregated. This paper broadens the segregation literature by determining: (1) black immigrant segregation patterns after controlling for individual-level characteristics, (2) the extent to which segregation varies by location, and (3) if racial segregation has the same socioeconomic consequences for U.S.- and foreign-born blacks. I find that black immigrants face high racial and socioeconomic segregation in mainly Caribbean settlement areas. However, black immigrants in all but two predominantly African settlement areas experience no segregation. Essentially, I find that there is a great deal of diversity in black immigrants’ segregation patterns stemming from differential treatment in the housing market based on African immigrants’ higher SES and/or African immigrants’ residential choices. Results in the two outlier African settlement areas (Minneapolis and Washington, D.C.) suggest that entry visa may play an important role in black segregation.

INTRODUCTION

Over the years, many scholars have lamented the United States’ ongoing residential segregation and its role in reinforcing social inequalities (Dwyer 2010). While improvements in socioeconomic status (SES) reduce segregation for blacks, they are still more segregated from whites than any other race/ethnic group at all SES levels (Iceland et al. 2005; Iceland and Wilkes 2006). Blacks who live in segregated neighborhoods suffer higher poverty and more inequality (Ananat 2011; Rosenbaum and Friedman 2001) and have fewer amenities (Bayer and McMillan 2005; Zenk et al. 2005). Immigrants now make up nearly 10 percent of blacks (Kent 2007) and have begun to redefine blackness in the United States. Therefore, investigating black immigrant segregation offers a new way to understand the causes and consequences of racial segregation.

Several authors suggest that black immigrants’ ethnicity gives them an advantage over U.S.-born blacks in dealings with whites (Bashi Bobb and Clarke 2001; Foner 1985; Waters 1994). If this holds true, black immigrants may improve their residential outcomes by highlighting their ethnicity to distinguish themselves from U.S.-born blacks (Freeman 2002). Black immigrants may then navigate the housing market differently than U.S.-born blacks, who—despite recent segregation declines—still suffer higher levels of segregation than any other group (Adelman 2004; Dawkins 2004). Thus, black immigrant segregation patterns may mirror the moderate racial segregation of Asian and Hispanic immigrants (Charles 2003)—groups whose SES plays a larger role in their segregation patterns than it does for U.S.-born blacks (Darden and Kamel 2000; Iceland and Wilkes 2006; Logan et al. 1996). Yet, research thus far finds that black immigrants are at least as segregated from whites as U.S.-born blacks (Crowder 1999; Cutler et al. 2008; Freeman 2002, 2005; Iceland and Scopilliti 2008). However, there are three major gaps in the existing literature that may call these conclusions into question.

First, measures of segregation used in previous work do not control for individual-level characteristics. Most studies use traditional measures like the dissimilarity index and these analyses risk flawed conclusions on the patterns and processes of segregation because they use aggregate-level data to discuss segregation, an individual-level process (Alba and Logan 1993). Second, despite growing evidence that immigrants are now moving to places with little immigration history (Camarota and Keeley 2001), black immigrant segregation research has only been conducted at the national level or focused on traditional immigrant settlement areas. Thus, it is unclear whether new destinations have higher rates of black immigrant segregation due to hostility toward new race/ethnic groups, or lower rates of segregation because there is no established ethnic community.

Finally, while researchers examined black immigrants’ racial segregation, none investigated whether it has the same socioeconomic consequences as for U.S.-born blacks. Ethnic networks and family reunification (Farrell 2008) lead black immigrants to form ethnic enclaves (Crowder 1999; Freeman 2002). Based on relative preferences for foreign-born over U.S.-born blacks, staying distinctive in ethnic communities may also be highly important to black immigrants (Crowder 1999). Due to their greater resources—African immigrants are more educated¹ and Caribbean immigrants earn more than U.S.-born blacks (Corra and Kimuna 2009; Kalmijn 1996)—racial segregation may not lead to socioeconomic disadvantage for black immigrants.

I address these gaps in earlier work with the locational attainment model (which controls for individual-level factors) to determine black immigrants’ racial and socioeconomic segregation in the top African and Caribbean settlement areas. I use American Community Survey (ACS) data to answer three questions: First, are black immigrants racially segregated from whites after controlling for individual-level characteristics? Second, does racial segregation have the same socioeconomic consequences for black immigrants as for U.S.-born blacks? Finally, do black immigrants’ segregation patterns vary by location? Overall, I find that, like the results for Hispanic and Asian immigrants (Charles 2003), there are meaningful national origin differences in segregation patterns, showing that blanket explanations do not adequately define black segregation in the United States.

BACKGROUND

A large body of sociological literature attempts to explain ongoing segregation. In this section, I examine common explanations—spatial assimilation and place stratification theories, ethnic communities, and neighborhood preference—and offer a less discussed explanation: the impact of settlement area on minority and immigrant segregation.

SPATIAL ASSIMILATION

Spatial assimilation theory predicts that as minority groups establish themselves in the labor market, they use their income to gain better housing by moving to less racially segregated neighborhoods with more advantages and amenities (Alba and Logan 1993). For immigrants, spatial assimilation also relies on cultural assimilation (usually measured by increased English ability, time in the United States, and U.S. citizenship). If this theory explains segregation patterns, controlling for individual-level factors will greatly reduce group-level differences in neighborhoods (Rosenbaum and Friedman 2001).

There is evidence to support spatial assimilation. Low SES minorities are more segregated from whites than their high SES counterparts (Adelman 2005; Iceland et al. 2005; Iceland and Wilkes 2006). Among immigrants, measures of cultural assimilation strongly affect where Hispanics live, but seem unimportant for Asians (Fong and Shibuya 2005; Freeman 2000; Logan et al. 1996). Black immigrants’ experience of cultural assimilation and residential outcomes largely matches that of Asian immigrants: Citizenship makes little difference to black immigrant segregation (Freeman 2002), and there is inconsistent evidence of the effect of English ability on segregation (Freeman 2002; Logan and Alba 1993). Thus, black immigrants’ segregation patterns should match those of Asian immigrants. However, the impact of SES complicates matters. Controlling for group-level SES reduces blacks’ segregation levels less than those of Asians and Hispanics (Darden and Kamel 2000; Iceland and Wilkes 2006; Logan et al. 1996). Foreign-born blacks are much more racially segregated than other immigrants and at least as segregated from whites as U.S.-born blacks (Crowder 1999; Cutler et al. 2008; Freeman 2002, 2005; Iceland and Scopilliti 2008). As a result, spatial assimilation theory may not describe black immigrants’ segregation patterns well.

LOCATION

The relationship between cultural assimilation and Asians’ residential segregation may be due, in part, to recent immigrants settling directly in new settlement areas (Logan et al. 1996). Location may also play a role in black immigrants’ segregation patterns: One-fourth of black immigrants’ main settlement areas are not traditional settlement areas as defined by Hall et al. (2011). Yet, previous research has not addressed whether location impacts black immigrants’ segregation patterns.

New destinations have their own dynamics partly stemming from different segregation histories (Frey and Farley 1996). New settlement areas in the south and west have less of a history of racial discrimination and, thus, lower levels of black segregation (Iceland et al. 2013). The racial histories of new settlement areas may also give immigrants more scope to define their positions within the local racial hierarchy (Waters and Jimenez 2005). Some studies show that immigrants have lower rates of segregation in new destinations than in traditional settlement areas (Fischer 2003; Park and Iceland 2011), while other researchers find that immigrant segregation is higher in new destinations (Fischer and Tienda 2006; Lichter et al. 2010).

Among Asians (Park and Iceland 2011) and Hispanics (Lichter et al. 2010), segregation differences by settlement area are largely the result of population size, showing that minority threat is a factor in their segregation. Where immigrant in-migration led to large populations, social boundaries heightened, leading to more discrimination and higher rates of segregation (Blalock 1967). Based on these findings, black immigrant segregation may be lower in new than in traditional settlement areas. However, black immigrant segregation will be higher where there is a large coethnic population.

While conclusions of segregation research in new destinations have been mixed (Hall 2013), immigrant segregation patterns in traditional settlement areas largely follow spatial assimilation theory (Park and Iceland 2011; Singer 2004). Yet, due to their race, black immigrants’ experiences living in these places may be quite different from Asians and Hispanics. Traditional settlement areas are, on average, large metro areas with sizeable U.S.-born black populations—two characteristics linked to black–white segregation (Iceland et al. 2013). Indeed, many traditional immigrant settlement areas have hyper-segregation between blacks and whites (Massey and Tannen 2015). Because these places have long histories of racial discrimination and tighter housing markets (Logan et al. 2004; South and Crowder 1997), black immigrants may not be able to use their ethnicity to gain better housing. Instead, black immigrants may face high rates of segregation like their U.S.-born counterparts.

PLACE STRATIFICATION THEORY

U.S.-born blacks have the highest segregation in both new and traditional settlement areas. This has largely been attributed to place stratification, which argues that minorities are sorted based on their relative standing in society, limiting their ability to live in the same areas as comparable whites (Alba and Logan 1993). While acknowledging the influence of SES on minority access to less segregated communities (Crowder et al. 2006), this theory posits that blacks cannot translate economic success to integrated neighborhoods due to their place at the bottom of the U.S. racial hierarchy.

Blacks receive the lowest returns on economic attainment (Freeman 2000) because structural constraints limit their housing choices (Rosenbaum and Argeros 2005). These structural constraints—for example, steering toward certain areas (Galster and Godfrey 2005) and more frequent mortgage application rejection after controlling for SES (Dawkins 2005; Kim and Squires 1995)—lead to a racial hierarchy in the housing market. This leaves blacks most segregated (Iceland and Scopilliti 2008; Park and Iceland 2011) and provides evidence for place stratification as an explanation for black immigrant segregation.

ETHNIC COMMUNITY

Because place of residence is an important determinant of quality of life (Freeman 2002), access to racially integrated neighborhoods plays a large role in cultural and socioeconomic assimilation. However, black immigrants’ high rates of racial segregation may not have the same socioeconomic implications it does for U.S.-born blacks. Instead, segregation in ethnic communities may have benefits. Ethnic communities foster group-specific networks providing access to work opportunities (Bayer et al. 2005) and support for group-specific businesses or community organizations (Waldfogel 2003). Immigrants may, therefore, choose to live in ethnic communities (Logan et al. 2002) rather than living there out of economic necessity. Due to their SES, if black immigrants are racially segregated in an ethnic community, they will not face the socioeconomic segregation of ethnic enclaves (Light and Gold 2000) and suffered by racially segregated U.S.-born blacks (Ananat 2011; Rosenbaum and Friedman 2001).

NEIGHBORHOOD PREFERENCE

The ethnic community model focuses on immigrants’ preferences to live in same-ethnicity areas. Neighborhood preferences also play a role in U.S.-born blacks’ residential outcomes; however, preferences of out-group members largely shape their patterns of segregation. Although blacks rarely want to live where they are less than 50 percent of the population, there is little evidence that their preferences explain black segregation rates (Clark 2009). Segregation patterns are instead shaped by whites’ preferences. Whites are most likely to avoid areas with black residents (Friedman et al. 2005): An increase in the black composition of a neighborhood leads to a decrease in whites’ likelihood of moving there, while a similar increase in the Asian or the Hispanic composition of a neighborhood does not (Emerson et al. 2001).

Although research consistently shows that whites’ avoidance of minority neighborhoods stems from negative stereotypes (Bobo and Zubrinsky 1996; Charles 2000; Farrell 2008), it may also be due to differences in neighborhood knowledge. Whites and blacks both use their social networks—who are largely the same race as themselves—to navigate the rental and housing markets (Krysan 2008). The racial composition of the social network plays a key role in an individual’s local knowledge. The geography of segregation structures familiarity with communities (Bader and Krysan 2015): As the percentage of blacks in an area increases, whites become less likely to know about that neighborhood (Krysan and Bader 2009). Because whites have little knowledge of neighborhoods with substantive black populations—and negative perceptions of black neighborhoods in general—if whites do not find a home in largely white neighborhoods, they are unlikely to look in places with more than a token percentage of blacks (Bader and Krysan 2015).

The different segregation histories of traditional and new settlement areas may result in cross-racial neighborhood knowledge differences. Therefore, comparing residential outcomes in new and traditional settlement areas adds theoretical insight into the processes of immigrant incorporation (Waters and Jiménez 2011) and racial segregation. It also introduces more variation in the structural and contextual characteristics of immigrants’ destinations, reflecting the United States’ new demography (Marrow 2005). In this paper, I explore black immigrants’ racial and socioeconomic segregation patterns in traditional and new settlement areas. In doing so, I address the following hypotheses:

1. If black immigrant segregation follows spatial assimilation theory, they will not face racial or socioeconomic segregation after controlling for SES.

2. Place stratification theory and neighborhood preference predict that immigrants’ and minorities’ neighborhood choices are limited by discrimination in the housing market—place stratification limits choices through structural discrimination and neighborhood preference does so through individual-level discrimination. If these patterns explain black immigrant segregation, foreign-born blacks will be racially and economically segregated after controlling for individual-level characteristics.

3. The ethnic community model states that immigrants choose racial segregation in ethnic communities due to its economic advantages. Hence, black immigrants living in ethnic communities will not face socioeconomic segregation.

4. If location plays a significant role in segregation patterns, black immigrant segregation will vary based on immigrant destination. In traditional destinations—where there is a longer history of racial discrimination—black immigrants’ segregation patterns will match those of U.S.-born blacks and be higher than those of black immigrants in new settlement areas. However, black immigrant segregation rates will vary within new destinations. In new destinations with large coethnic populations, black immigrants will face segregation rates similar to those in traditional settlement areas. Segregation in new settlement areas with smaller black immigrant populations will be lower.

DATA AND METHODS

To measure black immigrant segregation in the United States, I use pooled 2006–2010 ACS data. Individual-level data were accessed through the Public Use Micro Sample (Ruggles et al. 2010) and aggregate (census-tract) data through American FactFinder (U.S. Census Bureau 2012). I use pooled data due to the small black immigrant population. The sample for both the census-tract and individual-level data contains men and women who live in the top 10 settlement areas for black African immigrants and the top 10 settlement areas for non-Hispanic black Caribbean immigrants (16 metro areas in total). I focus on adults (age 18+) because black immigrants’ residential segregation patterns are nearly exclusively determined by where adults live: In 2010 only 8 percent of black immigrants were under 18.

While analyses are possible at the national level, it would be unclear what black immigrant segregation means due to their small population. To provide substantive results to interpret black immigrant segregation, I analyze segregation separately for 16 metro areas: Atlanta, Boston, Chicago, Dallas, Ft. Lauderdale, Hartford, Connecticut, Houston, Los Angeles, Miami, Minneapolis, New York, Orlando, Philadelphia, San Francisco, Washington, D.C., and West Palm Beach.² Atlanta, Minneapolis, Orlando, and Philadelphia are new destinations as they are former, emerging, reemerging, or preemerging gateways (Hall et al. 2011). The rest are traditional settlement areas.

I use Alba and Logan’s (1992) locational attainment model to measure and analyze how black immigrants’ individual-level characteristics affect the places they live. This model is a method for measuring segregation that avoids the risks of ecological inference and multicollinearity (Alba and Logan 1993)—drawbacks of traditional measures (Charles 2003) used in previous analyses of black immigrant segregation. The dependent variable is based on census tract-level data and I calculate the correlation between dependent and independent variables from tract-level data. I estimate correlations among independent variables from the individual-level data. I then combine the correlations from the tract-level and individual-level data into a common correlation matrix. This correlation matrix does not match people from the tract-level data set to the same people in the individual-level data. Instead, the locational attainment model uses the correlations from individual and aggregate data to form a combined matrix estimating the regression models.

A combined matrix is possible because all relevant means, standard deviations, and sample sizes are known (Hanushek and Jackson 1977) and because individual-level data are a sample from which tract-level data are constructed, making the data sets consistent (Alba et al. 2000). Like Alba and Logan (1992), I take the means and standard deviations from the tract-level data because they are based on the larger sample, providing more efficient estimates. However, the N used to calculate measures of statistical significance (sample size in the metro area) came from the individual-level data. This is the more conservative approach but, more importantly, it is also most appropriate, because the interest lies in the analysis of individual rather than aggregate processes. The resulting regression model uses the following equation:

$$Y_{ij}=\alpha +{\beta }_1{X}_{1ij}+{\beta }_2{X}_{2ij}\dots +{\epsilon }_{ij}.$$

The subscript j represents the census tracts within each metro area and—based on the correlation matrix that provides a relationship between individual- and aggregate-level data—the subscript i represents individuals within each census tract. In this analysis, there are three dependent variables: the percentage of the census tract that is non-Hispanic, U.S.-born white³ (white); the percentage that completed at least four years of college; and the percentage that is high income. High-income individuals are defined as those with an income at least three times the poverty threshold. All three dependent variables, specified as Y_ij, are aggregate- rather than individual-level characteristics; therefore, they are assumed to be constant across i for any value of j.

I include race/nativity as an independent variable that includes whites (the reference group) and U.S.- and foreign-born blacks.⁴ Race is self-reported on the ACS and, because racial classification systems have at least some flexibility, immigrants’ choice of racial identity depends on their experiences in the United States (Itzigsohn et al. 2005). Although black immigrants may identify themselves based on their culture to distance themselves from the racial stigma associated with blackness in the United States, they still identify as black (Benson 2006). Consequently, black immigrants’ racial self-identification should largely match the American system of racial stratification, which categorizes people based on skin color alone (Alba and Nee 1997).

Black immigrants in this analysis are not disaggregated by Hispanic ethnicity or sending region (i.e., Africa or the Caribbean) because these data are not available at the tract level given the small size of these populations. However, analyses still focus on non-Hispanic black immigrant segregation due to (1) the definition of major black immigrant settlement areas and (2) racial self-identification among Hispanics. Major black immigrant settlement areas are defined here based on the settlement patterns of African- and Caribbean-born non-Hispanic blacks. In addition, Hispanics are likely to reject the usual notion of race in the United States (Itzigsohn and Dore-Cabral 2000), leading to fewer people identifying as black regardless of skin tone (Landale and Oropesa 2002). The individual-level data show that analyses focus on non-Hispanic black immigrants. Approximately 90 percent of black immigrants identify as non-Hispanic, and Hispanics comprise no more than 20 percent of black immigrants in any metro area. Percentage Hispanic is highest in Los Angeles, San Francisco, and Miami (approximately 19 percent) and lowest in Washington, D.C., West Palm Beach, Minneapolis, and Atlanta (5 percent or less).

Along with race/nativity, I also include individual-level characteristics that play a role in minority groups’ segregation. In addition to age and marital status, I include educational attainment, income to poverty ratio, English ability, and naturalized citizenship as independent variables. If the coefficients for independent variables measuring individual-level SES are significant and sizable, this will support the spatial assimilation model.

RESULTS

DESCRIPTIVE STATISTICS

Table 1 presents the descriptive statistics for each metropolitan area using tract-level data. This table shows that while whites were the majority in the nation in this survey period, they are the majority in only half of the metro areas in this analysis. However, percentage white in a metro area neighborhood varies by race/nativity.⁵ Whites are the majority population in their neighborhood in all metro areas except Miami, where whites only make up 13 percent of the total metro area population. Conversely, whites only make up the majority of U.S.- and foreign-born blacks’ neighborhoods in Minneapolis, where whites make up 81 percent of the metro area.

TABLE 1.

Descriptive Statistics of Metropolitan Area by African-to-Caribbean Ratio, 2006–2010 Pooled Aggregate ACS Data, American FactFinder

	# of Tracts	% White	% U.S.-Born (USB) Black	% Foreign-Born (FB) Black	% College+	% High Income	% Speaks English at least Very Well
Predominantly Caribbean Black Immigrant Population (African-to-Caribbean Ratio <0.5)
Boston	1,267	72.8	3.4	2.7	36.6	64.1	90.1
Ft. Lauderdale	361	42.1	11.9	12.5	25.0	53.5	83.2
Hartford	429	74.0	5.5	2.2	30.0	64.7	92.7
Miami	511	12.6	11.0	5.8	23.9	41.5	59.3
New York	2,426	30.8	13.6	10.7	32.4	51.9	75.8
Orlando	389	54.7	10.6	4.1	28.1	51.0	87.8
West Palm Beach	332	59.8	8.3	6.7	29.4	55.5	86.0
African-to-Caribbean Ratio between 0.5 and 1.5
Atlanta	915	51.6	27.8	3.4	32.0	56.9	91.0
Philadelphia	1,317	65.1	17.9	1.6	30.0	60.5	93.4
Predominantly African Black Immigrant Population (African-to-Caribbean Ration >1.5)
Chicago	2,010	51.6	16.2	0.8	32.0	59.2	85.2
Dallas	912	51.2	14.2	1.3	29.4	54.6	84.2
Houston	946	39.9	16.2	1.6	25.9	51.7	79.7
Los Angeles	2,323	25.8	5.1	0.9	26.2	48.8	69.5
Minneapolis	771	81.0	4.0	2.2	34.6	64.7	94.4
San Francisco	406	42.3	4.1	0.4	46.0	65.7	78.7
Washington, D.C.	1,367	48.7	20.8	4.4	42.9	71.2	88.8

U.S.-born blacks are less than 10 percent of the population in 6 of the 16 metro areas. Yet, U.S.-born blacks never live in neighborhoods where the weighted average percentage black is less than 10 percent (not shown). Despite shared racial characteristics, percentage U.S.-born black in a neighborhood is always lower for black immigrants than U.S.- born blacks—ranging from 3 percent (Boston) to 34 percent (Chicago) smaller U.S.-born black concentration. It is important to note vast differences in racial makeup, as neighborhoods with more U.S.-born blacks are often poorer and more segregated (Lichter et al. 2012).

As expected, the foreign-born black proportion of each metro area is quite small. Black immigrants make up less than 5 percent of the population in three quarters of their top settlement areas. However, foreign-born blacks are a large portion of the population in New York (10.7 percent) and Ft. Lauderdale (12.5 percent)—numbers that are very close to those of the U.S.-born black population in these places. The percentage foreign-born black varies by African-to-Caribbean ratio in the metro area; the percentage of the population that is foreign-born black is generally higher where Caribbean immigrants are the majority of the black immigrant population.

Overall, the 16 metro areas are very similar in terms of SES. In most places, the percentage with a college education is approximately 30 percent, with the highest proportions with at least a college degree (over 40 percent) in Washington, D.C., and San Francisco. The concentration of the college educated is higher in white neighborhoods than black neighborhoods (not shown). Black immigrants are more likely to live in neighborhoods with college-educated individuals in mainly African settlement areas; in these places, concentration of the college-educated is higher in black immigrants’ than in U.S.-born blacks’ neighborhoods. Given the link between education and income, it is unsurprising that places with large college-educated populations have the most high-income earners.

The SES of individuals in each metro area is presented in Table 2 and shows that the education and income levels of U.S.-born whites and blacks are similar across metro areas, but this is not true of foreign-born blacks. In the metro areas where the black immigrant population is mainly Caribbean, the percentage with a college degree is closer to that of U.S.-born blacks (15–19 percent) than whites (33–50 percent). In most of these places (Ft. Lauderdale, Hartford, Miami, New York, and Orlando), the proportion of black immigrants with a college degree is lower than that of U.S.-born blacks. The proportion of black immigrants with a college degree or higher is closer to that of U.S.-born whites in metro areas with mainly African black immigrant populations. Regardless of settlement area type, African immigrants are more likely than Caribbean immigrants to have a college education in all metro areas except Minneapolis.

TABLE 2.

Descriptive Characteristics of Individuals in Each Metropolitan Area by African-to-Caribbean Ratio, 2006–2010 Pooled ACS Data, PUMS

	%College+					%High Income
	White	USB Black	FB Black			White	USB Black	FB Black
			Total	African	Caribbean			Total	African	Caribbean
Predominantly Caribbean Black Immigrant Population (African-to-Caribbean Ratio < 0.5)
Boston	47.14	22.95	21.86	28.65	17.33	75.86	49.92	48.57	46.46	46.24
Ft. Lauderdale	33.53	18.51	19.05	49.59	18.17	66.89	42.73	39.88	52.89	38.82
Hartford	40.25	14.82	16.50	32.39	13.11	76.67	48.32	47.75	44.37	47.98
Miami	44.50	14.20	15.66	48.21	14.13	70.69	37.94	33.19	44.44	30.79
New York	50.63	18.99	20.66	31.11	18.05	73.45	45.57	51.94	42.61	51.24
Orlando	31.46	17.46	19.19	30.97	17.38	64.56	38.67	35.26	35.71	33.83
West Palm Beach	37.04	15.63	14.88	56.76	13.56	68.46	36.31	34.25	60.00	33.17
African-to-Caribbean Ratio between 0.5 and 1.5
Atlanta	39.48	24.24	32.87	38.70	26.76	71.65	48.63	46.74	40.69	48.30
Philadelphia	34.86	14.78	28.39	35.71	21.85	73.89	43.82	49.35	43.02	52.73
Predominantly African Black Immigrant Population (African-to-Caribbean Ratio >1.5)
Chicago	42.52	17.46	36.39	41.56	26.28	75.93	42.24	46.51	42.07	49.89
Dallas	40.51	19.19	39.69	39.74	33.80	73.75	44.27	46.66	43.41	52.90
Houston	38.73	18.60	39.36	46.02	28.60	74.20	44.25	49.09	44.66	55.99
Los Angeles	42.84	19.96	34.45	46.61	26.12	71.73	48.78	53.39	53.86	56.90
Minneapolis	35.81	15.40	24.12	22.93	37.25	73.16	36.33	32.04	30.53	62.50
San Francisco	59.96	22.08	40.19	35.64	44.44	79.28	48.05	57.64	50.00	63.64
Washington, D.C.	56.72	26.52	36.02	37.66	28.11	85.55	66.23	59.85	54.44	67.24

Despite differences in educational attainment among black immigrant settlement areas, there is little variation in the proportion with a high income across metro areas: The proportion of black immigrants that are high income is very similar to that of U.S.-born blacks (see Table 2). The percentage of African and Caribbean immigrants who are high income mirrors results of previous research showing that, while Caribbean-born blacks are not as well educated as the African-born, they earn more (Butcher 1994; Corra and Kimuna 2009; Darity et al. 1996). Africans are more likely than Caribbeans to have a high income in only four metro areas (Boston, Ft. Lauderdale, Miami, and Orlando), but the difference is less than 5 percent in Boston and Orlando.

LOCATIONAL ATTAINMENT ANALYSES

Tables 3 and 4 present the locational attainment model results predicting racial segre-gation (percentage white⁶) and socioeconomic segregation (percentage with a college degree or higher and percentage high income), respectively.⁷ Coefficients should be in-terpreted as a type of exposure index (Logan et al. 1996): the probability that members of the group have residential contact with the group in the dependent variable (Alba and Logan 1993) relative to whites. Each row is a separate regression analysis for each metro area.

TABLE 3.

Locational Attainment Models Predicting % White in Census Tract (ref. White)

	U.S.-Born Black	Foreign-Born Black	R 2	N
Predominantly Caribbean Black Immigrant Population (African-to-Caribbean Ratio < 0.5)
Boston	−37.05***	−30.30***	0.35	131,220
Ft. Lauderdale	−25.93***	−28.54***	0.27	64,747
Hartford	−30.84***	−36.91***	0.30	27,840
Miami	−3.21***	12.59***	0.17	82,681
New York	−26.15***	−8.75***	0.33	299,276
Orlando	−17.31***	−13.23***	0.20	79,762
West Palm Beach	−25.96***	−12.87***	0.34	55,861
African-to-Caribbean Ratio between 0.5 and 1.5
Atlanta	−33.63***	−27.42***	0.35	168,817
Philadelphia	−42.88***	−34.08***	0.48	171,683
Predominantly African Black Immigrant Population (African-to-Caribbean Ratio > 1.5)
Chicago	−37.26***	−4.29***	0.41	268,859
Dallas	−17.91***	13.27***	0.37	137,189
Houston	−17.84***	0.23	0.35	159,018
Los Angeles	−8.18***	12.33***	0.37	342,695
Minneapolis	−21.76***	−15.68***	0.22	81,622
San Francisco	−15.88***	−1.66	0.23	66,602
Washington, D.C.	−32.78***	−28.36***	0.30	197,564

Note: Regression models run separately for each metropolitan area. The race/nativity variable also includes U.S.- and foreign-born Asians in all models. Results for these groups are available upon request. In addition to race/nativity, all models control for educational attainment, income to poverty ratio, marital status, English ability, age category, and citizenship. Results of full models are presented in the Online Appendix.

^*p < 0.05

^**p < 0.01

^***p < 0.001.

TABLE 4.

Locational Attainment Models Predicting % College+ and % High Income in Census Tract (ref. White)

	% College+			% High Income
	U.S.-Born Black	Foreign-Born Black	R 2	U.S.-Born Black	Foreign-Born Black	R 2	N
Predominantly Caribbean Black Immigrant Population (African-to-Caribbean Ratio < 0.5)
Boston	−4.39***	−3.47***	0.21	−9.50***	−5.95***	0.36	131,220
Ft. Lauderdale	−9.01***	−11.78***	0.26	−12.18***	−14.79***	0.28	64,747
Hartford	−3.83***	−2.64***	0.22	−9.88***	−8.30***	0.35	27,840
Miami	−9.66***	−7.36***	0.26	−7.88***	−6.07***	0.29	82,681
New York	−8.04***	−1.32***	0.30	−5.12***	1.58***	0.34	299,276
Orlando	−4.44***	−2.21***	0.17	−7.91***	−4.68***	0.21	79,762
West Palm Beach	−9.47***	−4.47***	0.27	−12.74***	−3.02***	0.30	55,861
African-to-Caribbean Ratio between 0.5 and 1.5
Atlanta	−2.33***	−0.18	0.22	−9.06***	−3.43***	0.29	168,817
Philadelphia	−12.80***	−7.84***	0.23	−17.38***	−11.12***	0.42	171,683
Predominantly African Black Immigrant Population (African-to-Caribbean Ratio > 1.5)
Chicago	−7.86***	22.14***	0.33	−12.95***	8.21***	0.39	268,859
Dallas	−2.58***	0.76***	0.43	−2.38***	26.21***	0.53	137,189
Houston	−0.83***	20.02***	0.43	−3.11***	17.26***	0.50	159,018
Los Angeles	2.43***	25.27***	0.68	2.56***	28.14***	0.57	342,695
Minneapolis	−3.47***	−1.77***	0.16	−11.0***	−9.69***	0.25	81,622
San Francisco	−8.55***	5.70***	0.24	−10.85***	0.78	0.27	66,602
Washington, D.C.	−10.09***	−2.23***	0.26	−4.26***	−1.74***	0.27	197,564

Note: Regression models run separately for each metropolitan area. The race/nativity variable also includes U.S.- and foreign-born Asians in all models. Results for these groups are available upon request. In addition to race/nativity, all models control for educational attainment, income to poverty ratio, marital status, English ability, age category, and citizen ship. Results of full models are presented in the Online Appendix.

^*p < 0.05

^**p < 0.01

^***p < 0.001.

The results of racial segregation analyses (Table 3) show that in new—Atlanta, Minneapolis, Orlando, and Philadelphia—and traditional settlement areas (all other metro areas), foreign-born blacks’ segregation patterns vary by the composition of the black immigrant population. Focusing on the results for metro areas with mainly Caribbean black immigrants and metro areas where there is no dominant black immigrant group (the top and middle panels), I find that black immigrants are highly segregated from whites in all but one metro area (Miami). In these places, black immigrant racial segregation patterns largely follow those of U.S.-born blacks. Because high segregation rates remain after controlling for SES, these results support place stratification theory.

These findings contrast sharply with those for places with mainly African black im-migrant populations (bottom panel: African-to-Caribbean ratio greater than 1.5). Black immigrants in these places face little to no segregation from whites—evidence of spatial assimilation theory. In two metro areas (San Francisco and Houston), there is no signif-icant difference between whites’ and foreign-born blacks’ exposure to whites. In Dallas and Los Angeles, black immigrants are more likely than even whites to live near whites after controlling for SES (the coefficients are positive and significant).

Black immigrants in only two of the mainly African settlement areas—Minneapolis and Washington, D.C.—are more than 10 percent less likely than whites to live near whites. The racial segregation of black immigrants in Minneapolis and Washington, D.C., may be due to (1) the racial composition of the native-born and (2) the largest sending coun-tries of the African immigrant population. Washington, D.C., has a large U.S.-born black population and the presence of U.S.-born minorities has been linked to higher rates of immigrant segregation (Frey and Farley 1996). Washington, D.C., has also long been a destination for Ethiopian immigrants. Today, Washington, D.C., is Ethiopian refugees’ third largest settlement area (Chacko 2003) and Ethiopian immigrants make up one-quarter of African immigrants in the metro area. In Minneapolis, Ethiopian (and Somali) immigrants make up nearly half of black immigrants. Minneapolis is the largest Somali community in the United States due to a mix of initial refugee settlement and consider-able secondary migration of refugees (Forrest and Brown 2014).

While it is impossible to control for entry visa status because the census does not collect these data, the Yearbook of Immigration Statistics provides entry visa information for those who received legal permanent residency. Between 2005 and 2010, just under one-third of Ethiopians and over 90 percent of Somalis who became permanent residents came to the U.S. as refugees (Department of Homeland Security 2006, 2007, 2008, 2009, 2010, 2011). Black immigrants’ likely visa status is important in explaining segregation patterns observed in Washington, D.C., and Minneapolis because, on arrival, refugees have lower average incomes than other immigrant groups (Cortes 2004). Together, the U.S.-born racial composition and black immigrants’ refugee status may explain the differences between Washington, D.C.’s, and Minneapolis’s black immigrant segregation patterns and other metro areas with mainly African populations.

Table 4 shows exposure to individuals who have completed at least four years of college and high-income individuals in each census tract. Focusing first on exposure to the college educated (shown on the left), it is clear that U.S.-born blacks are significantly less likely than whites to live near those with at least a college education in all but one metro area (Los Angeles) after controlling for individual-level SES. Unlike the relative consistency of U.S.-born blacks’ coefficients, black immigrants’ exposure to those with at least a college degree varies dramatically by the composition of the foreign-born black population. In metro areas with a mainly African black immigrant population (bottom panel), black immigrants are either significantly more likely to live near the college educated than whites or, at most, 2 percent less likely (Minneapolis and Washington, D.C.).

As in the racial segregation analyses, black immigrants’ segregation levels in mainly African black immigrant settlement areas follow spatial assimilation theory. By contrast, black immigrants live apart from the college educated in all places with mainly Caribbean black immigrant populations and one of two places with no dominant black immigrant group. Black immigrants’ exposure to the college educated relative to whites ranges from 12 percent less likely in Ft. Lauderdale to 1 percent less likely in New York, and half of these places have a negative coefficient that is five or higher. In these metro areas, black immigrants are also racially segregated, indicating that when black immigrants are racially segregated, they face similar socioeconomic consequences to U.S.-born blacks.

Exposure to high-income individuals (shown on the right) follows very similar patterns to exposure to the college educated. U.S.-born blacks are significantly less likely than whites to live near high-income individuals in all metro areas. Here again, foreign-born blacks in areas where the black immigrant population is mainly African (bottom panel) are significantly more likely than whites to live near high-income people. Minneapolis is the exception once again—segregation levels mirror those of U.S.-born blacks. Black immigrants in Minneapolis may have different economic segregation patterns due to their jobs. The most common work for black immigrants, like U.S.-born blacks, is farming.

CONCLUSIONS

Unlike Asian and Hispanic immigrants, any group racially defined as black faces barriers to accessing neighborhoods of similar quality to those of whites (Charles 2003). Even as income and education increases, blacks remain highly segregated from whites and live in neighborhoods with lower SES than comparable whites (Alba et al. 2000). Because this pattern has been consistent over time, researchers largely explain black segregation using place stratification theory. Although research thus far finds similar segregation patterns among U.S.- and foreign-born blacks (Crowder 1999; Cutler et al. 2008; Freeman 2002, 2005; Iceland and Scopilliti 2008), my results show that this is not true across metro areas. Black immigrant segregation patterns vary by location, but there is no dichotomy between traditional and nontraditional settlement areas as I proposed. Instead, it is based on the make-up of the black immigrant population in the metro area. Black immigrants face high racial and socioeconomic segregation that remains after controlling for SES in mainly Caribbean settlement areas—patterns consistent with place stratification theory. However, controlling for SES reduces or eliminates racial and socioeconomic segregation in mainly African settlement areas, which is consistent with spatial assimilation theory.

The low levels of segregation in most metro areas with mainly African populations could be due to ecological explanations, individual-level characteristics, or some combination of the two. It is possible that there is something about a metro area having a larger African population that leads those who might discriminate against blacks to treat them differently. Based on Africans’ high SES, the majority group may assume that these immigrants will contribute to, rather than detract from, the SES of their neighborhood, allowing African immigrants more freedom to move into the neighborhoods of their choice. Comparing the segregation patterns of U.S.- and foreign-born blacks in hyper-segregated metro areas with mainly Caribbean and mainly African populations indicates that this may be the case. Boston and Chicago—mainly Caribbean and mainly African settlement areas, respectively—were both places with hyper-segregation between blacks and whites in 2010 (Massey and Tannen 2015). In Boston, the educational characteristics of both African and Caribbean immigrants were similar to those of U.S.-born blacks. In Chicago, African immigrants (the majority of the black immigrant population) were more than twice as likely as U.S.-born blacks to have a college education (Table 2).

Prior to controlling for SES, black immigrants in Boston had approximately the same percentage white in their neighborhoods as U.S.-born blacks (43 percent)—approximately half the percentage white in a neighborhood as that of a white individual. After controlling for SES, segregation declines but black immigrants still exhibit racial and socioeconomic segregation patterns similar to those of U.S.-born blacks. In Chicago, by contrast, controlling for SES has a much larger impact. Before controlling for SES, the percentage white in a black immigrants’ neighborhood is again half that of whites; however, they are only 4 percent less likely than whites to live with other whites after controlling for SES (Table 3). Because black immigrants were just as likely as the U.S.-born to have high incomes in both places, these results are consistent with previous work finding that education is more critical than income in creating integration (Clark and Blue 2004).

At the individual level, African immigrants may behave differently or make different residential choices than Caribbean- or U.S.-born blacks. Black immigrants in places with mainly Caribbean black immigrant populations are more highly segregated, but as with Hispanic immigrants (Flippen 2001), this may facilitate homeownership. Percent home-owner among the sample population is consistent with previous work (Tesfai 2016) showing that homeownership is higher among Caribbean than African immigrants in all metro areas. The difference between these groups is higher in mainly African settlement areas⁸: In all seven of these metro areas, the difference between Africans and Caribbeans ranges from 11.7 to 31.1 percent, while the difference between groups is 10 percent or higher in only three of the seven mainly Caribbean settlement areas. African immigrants in mainly African settlement areas may be forgoing homeownership in racially segregated areas to live in integrated neighborhoods with more amenities. When Africans do own their homes, they own homes of similar value to whites (Tesfai 2016), providing further evidence that Africans make different housing choices than other black groups.

Two metro areas do not follow these patterns: Minneapolis and Washington, D.C. Although black immigrants in these places are mainly African, they still face high rates of racial segregation that are resistant to controlling for SES. Some of their segregation may be due to minority threat. Not only does Washington, D.C.’s, population have a relatively small proportion of whites (less than half of the population), but it also has the highest proportion of U.S.-born blacks of any metro area in this analysis (Table 1). The finding of black immigrants’ high segregation from whites in Washington, D.C., is consistent with previous work finding that black immigrants, specifically Ethiopians, are more likely to live in neighborhoods with significant black populations (Chacko 2003). Washington, D.C.’s, overall population characteristics may explain black immigrants’ high racial segregation rates in this metro area.

Black immigrants’ segregation may also reflect their own population size. Of the black immigrant settlement areas that are mainly African, it is only in Minneapolis and Washington, D.C., that black immigrants make up more than 2 percent of the population. Higher segregation in these places is consistent with previous research. Asian immigrants’ segregation is higher in places where their population is larger (Lichter et al. 2010; Park and Iceland 2011) because white out-migration becomes more likely as the immigrant population increases (Crowder et al. 2011). As the African immigrant population grows, future research should investigate whether segregation rates in mainly African settlement areas increase to rates similar to those in mainly Caribbean settlement areas where black immigrants make up 2 percent or more of the population.

The socioeconomic impact of racial segregation in Minneapolis and Washington, D.C., may reflect the educational composition of black immigrants in the metro area. A lower proportion of black immigrants have college degrees in Washington, D.C., and Minneapolis than nearly any other African settlement area (Table 2)—likely a consequence of the African immigrant composition in these places. Unlike other metro areas where the African population is largely made up of West Africans (Ghanaians and Nigerians) who mainly enter the country as economic and family migrants, Washington, D.C., and Minneapolis have large East African refugee populations. Minneapolis is a target of refugee resettlement programs, while Washington, D.C.’s, participation in refugee resettlement is relatively low (Brown et al. 2007). However, considerable secondary migration of African refugees occurs in both places (Brown et al. 2007; Forrest and Brown 2014). Refugees have much lower English proficiency than the average for immigrants from their region (Zong and Batalova 2017) and secondary migration is most common among refugees who do not speak English or do not have marketable skills for the U.S. labor market (Weine et al. 2011). Black immigrants’ socioeconomic segregation in Washington, D.C., and Minneapolis may, therefore, reflect the characteristics of the immigrants who choose to live there.

One limitation of this study is that I cannot differentiate between the individual and ecological explanations for the patterns observed. Qualitative research would be well suited for investigating whether the link between a mainly African black immigrant population and low rates of segregation is due to individual or ecological factors. Future research should also address whether segregation patterns stay the same as the black immigrant population (particularly the African-born) grows or if the increased immigrant population in these areas leads to changes in black immigrants’ social positions and increases in their segregation levels. Given the relationship between neighborhoods and school quality for the children of immigrants, these findings also indicate the necessity of research on the later life outcomes of second-generation black immigrants.