When and Where Residential Racial Segregation Matters for the Black Self-employment Rate

Abstract

Scholars debate whether residential racial segregation associates positively, negatively, or at all with the Black self-employment rate in the United States. This study engages that debate using data from the Integrated Public Use Microdata Series (IPUMS) 1980, 1990, and 2000 5% sample and 2006–2010 American Community Survey (ACS) 5-year sample. Specifically, I investigate the county-level association between residential racial segregation and the Black self-employment rate, and whether this association varied by region in 1980, 1990, 2000, and 2010. Using fractional logit models and net of control variables, I find that residential racial segregation positively predicts Black self-employment in the South. Implications for understanding how time and region condition Black self-employment opportunities are discussed.

Keywords

African Americans enclave economies region residential racial segregation self-employment

Introduction

Self-employment can increase Blacks’ social mobility (Oliver & Shapiro, 1995; Wingfield, 2008). However, Blacks have the lowest self-employment rate in the United States. In 2015, the Black self-employment rate was 3.6%, roughly one and a half times lower than the Asian (5.6%), Hispanic (6.4%), and White (6.9%) self-employment rates (Hipple & Hammond, 2016). Scholars agree that racial disparities in self-employment exist because of group differences in educational attainment, access to capital, and social networks (Bates, 1995; Lofstrom et al., 2014; Waldinger, 1995). Yet, scholars debate whether residential racial segregation contributes to racial disparities in self-employment. This debate endures because prior studies find that residential racial segregation associates positively (Boyd, 1996, 1998; Butler, 2005), negatively (Fairchild, 2008; Fischer & Massey, 2000), and sometimes not at all (Boyd, 1991) with the Black self-employment rate in the early and late 20th century.

Why is the evidence equivocal? On the one hand, residential racial segregation produces protected markets that shield Black self-employees from broader competition (Butler, 2005). On the other hand, residential racial segregation manifests institutional discrimination (Farley & Allen, 1987; Feagin, 2006; Massey & Denton, 1993) and concentrates Black disadvantage (e.g., poverty, joblessness, and violence). Du Bois’ (1899) The Philadelphia Negro describes this duality. In racially segregated Philadelphia in the late 19th century, Black self-employees flourished in protected markets that provided Black residents goods and services (e.g., grocers, barbers, and cemeteries). However, residential racial segregation also prevented Black self-employees from scaling up their businesses.

The present study argues the utility of protected markets shifts over time and across regions. Although de jure residential racial segregation protected the Black self-employment rate in the early 20th century (Butler, 2005), de facto residential racial segregation undermined the Black self-employment rate by the late 20th century (Boyd, 1991; Fairchild, 2008; Fischer & Massey, 2000). By the 1980s, however, regional shifts emerged that transformed the nature of residential racial segregation in the South. By introducing time and region into the debate, this study examines how both influence the association between residential racial segregation and the Black self-employment rate.

In what follows, I first describe ethnic and racial enclave economies, and then how residential racial segregation manifests institutional discrimination. Next, I outline residential racial segregation’s shifting effects on the Black self-employment rate over time. Third, I review how and why residential racial segregation differs by time and region, and why these differences matter for Black self-employment. Finally, I investigate the county-level association between residential racial segregation and the Black self-employment rate by region using data from the Integrated Public Use Microdata Series (IPUMS) 1980, 1990, and 2000 5% sample and 2006–2010 American Community Survey (ACS) 5-year sample (Ruggles et al., 2017).

Racism, Enclaves, and Residential Racial Segregation

Contemporary studies of the self-employment rate often ignore racism. Such color-blind approaches focus on residential ethnic segregation. They investigate the relationship between ethnic enclaves and self-employment. In this literature, residential ethnic segregation, among other factors (e.g., chain migration, labor market discrimination, and social networks, Light & Gold 2000; Zhou 2004), produces ethnic enclave economies. Ethnic enclave economies: (a) spatially cluster coethnic businesses, (b) contain financially interdependent coethnic businesses, and (c) employ coethnics (Light & Gold 2000). Thus, residential ethnic segregation clusters coethnics to create protected markets. Scholars in this area argue coethnic clustering structures self-employment outcomes.

Yet, racism influences all economic and social outcomes in the United States (Feagin, 2006; Valdez, 2011; Wingfield, 2008). Consequently, scholars marshal historical and contemporary data to argue residential racial segregation creates racial enclave economies (Butler, 2005; Wingfield, 2008). In racial enclave economies, laws and customs involuntarily cluster Black residents and self-employees. Laws and customs also restrict Blacks’ access to goods and services. As a result, Black-owned businesses and institutions emerge to provide those goods and services. For instance, Du Bois (1899) found that Black-owned businesses in Philadelphia’s Seventh Ward provided food, clothes, burial services, and other necessities to Philadelphia’s 40,000 Black residents. In addition, Black self-employees faced White terrorism. The 1921 Tulsa, OK, race riots razed Tulsa’s prosperous racial enclave economy (Butler, 2005). Such White terrorism is part of the historical record ignored by scholars who generalize from ethnic enclave economies to racial enclave economies.

Even though laws mandated desegregation in the 1960s, residential racial segregation persisted after the 1970s (Charles, 2003; Massey & Denton, 1993). Sociologists conclude that residential racial segregation reproduced and concentrated Black disadvantage in segregated Black neighborhoods (Farley & Allen, 1987; Massey & Denton, 1993). They also theorize it manifested institutional discrimination. For instance, Wilson (1987) argued deindustrialization transformed segregated Black neighborhoods. Specifically, it precipitated changes (i.e., unemployment, poverty, single-parent households, and middle-class out-migration) that isolated segregated Black neighborhoods. Isolation, in turn, stripped segregated Black neighborhoods of essential institutions and resources, and made them qualitatively unequal. Further, isolation compounded inequalities such as poverty, high unemployment, and low educational attainment (Massey & Denton, 1993; Sampson, 2013).

These compounded inequalities hurt segregated Black neighborhoods. For example, residential racial segregation exacerbated health disparities (LaVeist, 1993; Riley, 2018; Williams & Collins, 2001), wealth disparities (Oliver & Shapiro, 1995), the impact of predatory lending (Hwang et al., 2015), isolation from public transportation (McKenzie, 2013), unequal access to neighborhood businesses (Small & McDermott, 2006), and so on. The following sections describe evidence suggesting these compounded inequalities, alongside other factors, reduced the utility of racial enclave economies over time.

Residential Racial Segregation and the Black Self-employment Rate Over Time

Positive Association in the Early 20th Century

Racial enclave economies thrived in the early 20th century. The “golden years” of Black business (1919–1929), a decade of rapid Black business growth, coincided with the most extensive period of residential racial segregation in the United States (Marable, 1983). Durham, NC, and Tulsa, OK, represented prosperous racial enclave economies from this time period. For example, Durham, NC, was nicknamed the “The Wall Street of Negro America” (Butler, 2005, p. 176). Further, at this time, Black schools, hospitals, insurance companies, banks, and mutual-aid organizations provided goods and services that residential racial segregation made unavailable (Butler, 2005; Du Bois, 1899; Gold 2016). The Great Migration precipitated racial enclave economies in northern U.S. cities. Despite “escaping” Jim Crow, laws and customs clustered native-born Black migrants into isolated Black neighborhoods. As isolated Black neighborhoods grew in size, so did demand for retailers and niche market self-employees like hairdressers (Boyd, 1996, 1998; Gold, 2016).

In the early 20th century, Jim Crow laws and customs supported residential racial segregation and dictated where Blacks lived, worked, and shopped, which produced protected markets. Studies investigating this era addressed whether de jure residential racial segregation predicted the Black self-employment rate. In the late 20th century, after public accommodations were legally desegregated, residential racial segregation no longer structured Blacks’ access to goods and services, in theory. It did, however, continue to manifest institutional discrimination and concentrate Black disadvantage. As such, desegregation, institutional discrimination, and concentrated disadvantage converged to erode protected markets in segregated Black neighborhoods.

Negative Association in the Late 20th Century

Desegregation may have reduced the utility of racial enclave economies (Bates, 1997; Fairchild, 2008; Fischer & Massey, 2000). For example, it forced Black self-employees to compete with major corporations (e.g., chain pharmacies, grocers, and restaurants) that offer cheaper prices and larger inventories (Bates, 1997; Bonds, 2007). As a result, money flowed out of segregated Black neighborhoods after desegregation (Bates, 1993; Oakland et al., 1971; Schaffer, 1973). After desegregation, unfair lending also undercut self-employees’ economic power in segregated Black neighborhoods (Bates, 1993). Black-owned firms in minority neighborhoods received $39,564 less in loans than Black-owned firms in nonminority neighborhoods (Bates, 1993). Further, self-employment was less likely in high-poverty-segregated Black neighborhoods because concentrated poverty created unviable markets (Bates, 1993, 1997; Fischer & Massey, 2000).

Nonsignificant Association in the Late 20th Century

Despite evidence that residential racial segregation undermined the Black self-employment rate after desegregation, some late 20th century studies (e.g., Boyd 1991) find no statistically significant association between residential racial segregation and the Black self-employment rate. This noneffect suggests the “protected market theory may be inadequate to explain [contemporary] Black self-employment” rates (Boyd, 1991, p. 243). Specifically, by the early 1990s, demographic shifts altered the profile of Black self-employees (Bates, 1993, 1997). At the height of de jure residential racial segregation, Black self-employees in segregated Black neighborhoods owned small firms, held less than a high school diploma, and served Black clients. However, by the early 1990s, Black self-employees owned large firms, held college degrees, and served racially diverse clients (Bates, 1993, 1997). According to Bates (1993, 1997), Black self-employees left segregated Black neighborhoods for racially diverse neighborhoods and viable markets. As a result, the Black self-employment rate declined in some segregated Black neighborhoods and increased in some nonsegregated neighborhoods, which would attenuate residential racial segregation’s effect.

Region and Residential Racial Segregation From 1980 to 2010

By the late 20th century and early 21st century, regional shifts transformed segregated southern Black neighborhoods into places qualitatively different from nonsouthern ones. Of importance, these qualitative differences mean residential racial segregation’s association with Black self-employment may vary by region. The following sections describe how and why these qualitative differences emerged.

How Residential Racial Segregation Varies by Region

The nature of residential racial segregation varies by region. Historically, northern and midwestern cities had neighborhood residential racial segregation (i.e., Blacks and Whites lived in separate neighborhoods); southern cities had backyard residential racial segregation (i.e., Blacks and Whites lived in the same neighborhoods but in unequal spaces, Grigoryeva & Ruef, 2015; Logan & Martinez, 2018). Thus, the spatial form of residential racial segregation varied by region. As a result, Black–White neighborhood residential racial segregation was always low in the South (Farley & Frey, 1994; Grigoryeva & Ruef, 2015; Iceland et al., 2013; Logan & Martinez, 2018; Massey & Denton, 1993). For instance, from 1980 to 2000, U.S. cities with consistently high Black–White neighborhood residential racial segregation were in the Northeast and Midwest (i.e., Chicago, IL; New York, NY; Newark, NJ; Detroit, MI; Gary, IN; Milwaukee, WI; Logan & Stults, 2011). Further, Black hypersegregation (i.e., high levels of neighborhood residential racial segregation on four out of five segregation indices) occurred most often in the Northeast and Midwest (Charles, 2003; Massey & Denton, 1993).

Grigoryeva and Ruef (2015, p. 833) argued that “distinct forms of segregation may affect upward or downward mobility in different ways.” By the late 20th century, growth of segregated Black middle-class neighborhoods altered Black social mobility throughout the South (Lacy, 2007). For instance, residential racial segregation in the South concentrated Black disadvantage or advantage (Lacy, 2007). Consequently, some middle-class southern Blacks lived in racially segregated neighborhoods that resembled White middle-class neighborhoods (Lacy, 2007).

Why Residential Racial Segregation Varies by Region

By the 1980s, regional shifts began to transform segregated southern Black neighborhoods into places qualitatively different from their counterparts outside the South. For instance, although, Wilson (1987) argued deindustrialization transformed segregated Black neighborhoods, its economic repercussions (e.g., joblessness, low wages, skills mismatch, and poverty) were most severe throughout northeastern and midwestern cities (Wilson, 1987). This is partly because deindustrialization occurred alongside desegregation in the South. As a result, southern Blacks’ access to the formal labor market expanded. This change is evidenced by lower rates of Black joblessness, increased job stability, and higher wages (Branch & Hanley, 2011; Wilson, 1987). Consequently, extreme poverty likely never transformed segregated southern Black neighborhoods or decreased the Black self-employment rate in the South (Kolesnikova & Liu, 2010). Furthermore, the economy grew in southern cities, like Atlanta, GA, Charlotte, NC, Houston, TX, and Nashville, TN (Eckes, 2005; Guillory, 2010). This prosperity translated into better economic outcomes for southern Blacks (Frey, 2004; Hunt et al., 2008; Kolesnikova & Liu, 2010; Wilson, 1987), and probably protected segregated Black neighborhoods and the Black self-employment rate.

In addition, since the late 1970s, many Blacks return migrated to the South (Frey, 2004; Hunt et al., 2008; Pendergrass, 2013). Return migrants—disillusioned by the low quality of life in northeastern and midwestern cities—returned home. However, they returned home changed. As Stack (1996, p. xvii) argued, return migrants “ha[d] experienced other ways of life, [and came] home with new ideas, new energy, new skills, [and] new perspectives.” Importantly, they learned they would have to create the land of promise, prosperity, and opportunity they (and their parents) left the South for decades earlier. To this end, return migration was a “chance to start something new, to remake the South in a different image,” specifically a South that improved conditions for Blacks (Stack, 1996, p. xvii). Return migrants were also older and economically stable—a demographic profile for self-employees (Bates, 1997; Frey 2002, 2004; Hipple, 2010; Hunt et al., 2008). In the South, Black familial and social networks as well as segregated Black neighborhoods gained well-resourced residents eager to remake the South.

Although Marable (1983) hypothesizes segregated southern cities would have high Black self-employment rates after desegregation, studies do not consider whether region continued to matter into the late 20th and early 21st century. I would argue that regional shifts enhanced the utility of racial enclave economies in the South. Thus, while recent studies (see Boyd, 1991; Fairchild, 2008; Fischer and Massey, 2000) suggest that the utility of racial enclave economies disappeared over time, these same studies neglect how time and region may condition the association between residential racial segregation and Black self-employment.

Hypotheses

Consistent with prior studies (see Boyd, 1991; Fairchild, 2008; Fischer & Massey, 2000), I hypothesize that by 1980 the benefits of racial enclave economies disappeared nationwide. This is because desegregation, concentrated disadvantage, and demographic shifts undercut the utility of racial enclave economies. However, this nationwide trend may overshadow the influence that regional shifts (e.g., spatial patterns of residential racial segregation, diverging economies, and return migration) had in enhancing racial enclave economies in the South from 1980 to 2010. Therefore, I hypothesize the benefits of racial enclave economies emerged over time in the South. As a result, compared to 1980, residential racial segregation associated positively with the Black self-employment rate throughout the South in 2010.

Methods

Data

Data came from the IPUMS 1980, 1990, and 2000 5% sample and 2006–2010 ACS 5-year sample (Ruggles et al., 2017). The 1980–2010 IPUMS samples were clustered by household or dwelling and used a stratified sample design. Each sample includes sampling weights that adjust for the probability of selection, nonresponse, and noncoverage. The 1980, 1990, and 2000 5% samples were 1-in-20 national random samples of the U.S. population. They covered over 8 million respondents nationwide. Geographic areas with <100,000 residents were unidentifiable. The smallest identifiable geographic area was a county group or a Public Use Micro Area (PUMA), which consisted of counties or partial counties with at least 100,000 residents.

The 2010 ACS 5-year sample was a 1-in-20 national random sample of the U.S. population. It replaced the long-form Decennial Census. The 2010 ACS collected self-employment data, which were unavailable on the short-form 2010 Decennial Census. The 2010 ACS included all households and persons from the 1% 2006, 2007, 2008, 2009, and 2010 ACS samples. Again, geographic areas with <100,000 residents were unidentifiable. The smallest identifiable geographic area was a PUMA.

The estimation sample included 301 Metropolitan Statistical Area (MSA) central counties (i.e., MSA counties that contain the largest urban area) with at least 1,000 Black (i.e., non-Hispanic Black) and Hispanic residents (see more details in Appendix A). I limited the sample to MSA central counties available in 1980 and 2010 with at least 1,000 Black and Hispanic residents for three reasons. First, because MSA central counties include the municipal boundaries of an MSA’s core urban area (e.g., Houston, TX is within Harris County, TX), MSA central counties provide a stable and consistent measure of large urban areas over time (see Frey & Zimmer, 1998). Second, segregation indices become unreliable for areas with small minority populations. Third, these counties most accurately capture the United States’ demography. Large urban areas today tend to be multiracial. Overall, these restrictions increase generalizability to large urban areas.

Measures

Dependent Variables

At the county level, the total number of Black self-employees divided by the total number of employed Black adults who are 25–64 years of age provided the Black self-employment rate. Although working age is defined typically as 16–64 years, I restricted the sample to respondents who were at least 25 years old to allow for college completion and entry into the labor force.

Independent Variables

Three indices measured residential racial segregation: (a) Black–White dissimilarity, (b) Black–White isolation, and (c) Black clustering. All three indices refer to non-Hispanic Blacks and non-Hispanic Whites. Similar to prior studies (see Boyd, 1991; Fischer & Massey, 2000), dissimilarity measured residential racial segregation. Of theoretical importance, however, isolation and clustering most accurately capture how residential racial segregation isolates Blacks (see Boyd, 1998; Fairchild, 2008; Hwang et al., 2015).

Dissimilarity captured the absence of evenness. Evenness exists when two groups are dispersed equally across a county. The dissimilarity index represents the proportion of residents within a county who must move to ensure Blacks and Whites are equally dispersed across a county (Massey & Denton, 1988). Isolation measured the probability a county’s Black residents would meet or interact with other Black residents (Massey & Denton, 1988). Isolation represents the degree to which residential racial segregation socially isolates Blacks from Whites. As calculated here (and typically), dissimilarity and isolation are aspatial measures of residential racial segregation. They obscure the spatial relationship between census tracts. To overcome this limitation, I calculated Moran’s I (Anselin, 2013), which evaluates if attributes are clustered, dispersed, or randomly distributed spatially. Here, Moran’s I measured if residential racial segregation clustered Black residents into census tracts separate from Whites within a county.

I used the Longitudinal Tract Database (LTDB; Logan et al., 2014) to calculate the residential racial segregation indices. Black–White dissimilarity and Black–White isolation were calculated in Stata 14.2 using Reardon’s seg package (Reardon & Townsend, 2018). I calculated Moran’s I using the spdep package (Bivand, 2019) in R (version 3.5.3) with 2010 TIGER/Line Shapefiles for census tracts merged with LTDB data (Manson et al., 2017). Moran’s I was based on queen contiguity weights. Appendix B reports Spearman correlations among the indices in 1980, 1990, 2000, and 2010. In 1980, the correlations ranged from 0.49 to 0.80. In 2010, the correlations ranged from 0.57 to 0.60. A county’s region was coded as 0 = nonsouth and 1 = south, and defined by U.S. census regions.

Control Variables

County-level controls included: (a) proportion of Black women, (b) proportion of Blacks with a college education, (c) proportion of Black immigrants, (d) Black poverty rate, (e) Black–White poverty ratio, (f) county’s proportion Black residents, (g) proportion of adults who are 25–64 years old, (h) total county square miles, and (i) population density per square mile. Prior studies show men, college graduates, and immigrants have higher rates of self-employment (Bogan & Darity, 2008; Boyd 1991; Fairlie & Meyer, 1996; Smith, 2005a, 2005b). Further, concentrated poverty creates unviable markets (see Bates, 1993, 1997; Fischer & Massey, 2000), hence I include controls for a county’s Black poverty rate and Black–White poverty ratio. Finally, a county’s population size, total square miles, and population density likely positively associate with its self-employment rate.

A county’s proportion of Black women measured the total number of Black women of all ages divided by the total number of Black residents. Proportion of Blacks with a college education measured the total number of Black residents 25 years and older with a college education divided by the total number of Black residents 25 years and older. Proportion of Black immigrants equaled the total number of Black immigrants divided by the total number of Black residents.

The Black poverty rate measured a county’s total Black residents below the national poverty threshold divided by the total number of Black residents. The Black poverty rate more accurately captures a county’s viable Black market compared to a county’s Black median income (Fischer & Massey, 2000). The Black–White poverty ratio measured the proportion of Black residents below the national poverty threshold divided by the proportion of White residents below the national poverty threshold. It captures concentrated Black disadvantage (Quillian, 2017).

The proportion of Black residents measured a county’s total number of Black residents divided by a county’s total number of residents. A county’s proportion of Black adults measured the total number of Black residents who were 25–64 years old divided by the total number of Black residents. Total square miles represented a county’s land area. Finally, population density per square mile measured the total number of residents per square mile of a county’s total land area.

Analytic Strategies

To investigate the county-level association between residential racial segregation and the Black self-employment rate across time and region, I estimated repeated cross-sectional regression models for 1980, 1990, 2000, and 2010. I treated the Black self-employment rate as a proportion. Therefore, I ran cross-sectional fractional logit models using fracreg in Stata 14.2 (see Papke & Wooldridge, 1996; Williams, 2018). These logit models explain a dependent variable ranging between zero and one. I assessed multicollinearity concerns with the variance inflation factor (VIF). VIF values were less than three, on average. Again, the estimation sample included 301 MSA central counties.

Results

Table 1 reports means or proportions and standard errors for the Black self-employment rate, residential racial segregation indices, region, and control variables in 1980, 1990, 2000, and 2010. From 1980 to 2010, a county’s average Black self-employment rate increased from 3% to 5%. Black–White dissimilarity declined from 0.60 in 1980 to 0.55 in 2010. From 1980 to 2010, Black–White isolation was about 0.40. However, Black clustering increased from 0.47 in 1980 to 0.54 in 2010. Roughly 38% of counties were in the South.

Table 1.

Means, Proportions, and Standard Errors for the Black Self-employment Rate, Residential Racial Segregation, Region, and Control Variables using the 1980, 1990, 2000, and 2010 Integrated Public Use Microdata Series.

	1980	1990	2000	2010
	Mean (SE) or proportion (SE)	Mean (SE) or proportion (SE)	Mean (SE) or proportion (SE)	Mean (SE) or proportion (SE)
Self-employment rate	0.03	0.04^a	0.05^ab	0.05^abc
Self-employment rate	(0.00)	(0.00)	(0.00)	(0.00)
Black–White dissimilarity (range: 0–1)	0.60	0.56^a	0.54^a	0.55^a
Black–White dissimilarity (range: 0–1)	(0.01)	(0.01)	(0.01)	(0.01)
Black–White isolation (range: 0–1)	0.40	0.38	0.39	0.40
Black–White isolation (range: 0–1)	(0.01)	(0.01)	(0.01)	(0.01)
Black clustering (−1 = dispersed; 1 = clustered)	0.47	0.52^a	0.54^a	0.54^a
Black clustering (−1 = dispersed; 1 = clustered)	(0.01)	(0.01)	(0.01)	(0.01)
Region (0 = nonsouth; 1 = south)	0.38	0.38	0.38	0.38
Region (0 = nonsouth; 1 = south)	(0.03)	(0.03)	(0.03)	(0.03)
Proportion of Black women	0.51	0.51	0.50	0.51
Proportion of Black women	(0.00)	(0.00)	(0.00)	(0.00)
Proportion with college education	0.11	0.14^a	0.16^ab	0.19^abc
Proportion with college education	(0.00)	(0.00)	(0.00)	(0.00)
Proportion of Black immigrants	0.03	0.05^a	0.07^ab	0.10^abc
Proportion of Black immigrants	(0.00)	(0.00)	(0.00)	(0.01)
Black poverty rate	0.26	0.27	0.23^ab	0.25^b
Black poverty rate	(0.00)	(0.01)	(0.00)	(0.00)
Black–White poverty ratio	3.30	3.37	3.01^ab	2.57^abc
Black–White poverty ratio	(0.06)	(0.06)	(0.06)	(0.05)
Proportion of Black residents	0.12	0.13	0.14	0.15^a
Proportion of Black residents	(0.01)	(0.01)	(0.01)	(0.01)
Proportion of Black adults (25–64 years)	0.41	0.48^a	0.50^ab	0.51^abc
Proportion of Black adults (25–64 years)	(0.00)	(0.00)	(0.00)	(0.00)
Total square miles	1,054.07	1,131.04	1,130.97	1,130.93
Total square miles	(102.00)	(99.38)	(99.37)	(99.34)
Population density per square mile	1,483.95	1,534.82	1,696.92	1,762.15
Population density per square mile	(271.17)	(250.95)	(294.53)	(301.32)
Sample size	301	301	301	301

Notes. Standard errors reported in parentheses underneath means. Estimation sample size in each year is 301.

Significantly different (p < .05) from the county mean or proportion in 1980 using a two-tailed Tukey’s test for pairwise comparisons.

Significantly different (p < .05) from the county mean or proportion in 1990 using a two-tailed Tukey’s test for pairwise comparisons.

Significantly different (p < .05) from the county mean or proportion in 2000 using a two-tailed Tukey’s test for pairwise comparisons.

Black women represented more than half the Black population in the counties analyzed. By 2010, the proportion of Blacks with a college education (0.19) almost doubled and the proportion of Black immigrants (0.10) more than tripled from 1980. From 1980 to 2010, the Black poverty rate remained constant around 26%. However, the Black–White poverty ratio declined from 3.30 in 1980 to 2.57 in 2010. On average, a county’s proportion of Black residents increased from 12% in 1980 to 15% in 2010. From 1980 to 2010, the proportion of Black adults in these counties increased from 0.41 to 0.51. The average total square miles increased from 1054.07 in 1980 to 1130.93 in 2010. On average, a county’s population density per square mile increased from 1483.95 residents to 1762.15 residents. County means and proportions reported in 2010 were significantly different from 1980, excluding Black–White isolation, region, proportion of Black women, Black poverty rate, total square miles, and population density.

Table 2 presents estimates (i.e., logits and standard errors) from fractional logit models linking residential racial segregation (i.e., Black–White dissimilarity, Black–White isolation, and Black clustering) and the Black self-employment rate in 1980, 1990, 2000, and 2010. Model 1 estimates the bivariate relationship between each residential racial segregation index and the Black self-employment rate. Residential racial segregation significantly predicted the Black self-employment rate in 1980, 1990, and 2010. Specifically, Black clustering predicted 0.35 decreases in the logits (dy/dx = −0.01, p < .05) of the Black self-employment rate in 1980. Considering the average county in the estimation sample, the predicted Black self-employment rate was 1.05% lower at Black clustering’s highest observed value, compared to its lowest observed value. Similarly, Black–White dissimilarity predicted 0.55 decreases in the logits (dy/dx = −0.02, p < .05) of the Black self-employment rate in 1990, and Black–White isolation predicted 0.26 decreases in the logits (dy/dx = −0.01, p = .05) of the Black self-employment rate in 2010.

Table 2.

Estimates From Fractional Logit Regressions of the Black Self-employment Rate on Residential Racial Segregation Using the 1980, 1990, 2000, and 2010 Integrated Public Use Microdata Series.

	1980	1990	2000	2010
	Model 1	Model 1	Model 1	Model 1
Black–White dissimilarity (range: 0–1)	−.22	−.55*	−.17	−.16
Black–White dissimilarity (range: 0–1)	(.26)	(.24)	(.16)	(.19)
Constant	−3.30***	−2.91***	−2.94***	−2.78***
Constant	(.17)	(.15)	(.10)	(.11)
Log pseudolikelihood	−41.90	−48.97	−56.05	−62.81
Akaike information criterion (AIC)	87.81	101.94	116.10	129.61
Black−White isolation (range: 0–1)	−.14	−.15	.03	−.26*
Black−White isolation (range: 0–1)	(.16)	(.14)	(.11)	(.13)
Constant	−3.37***	−3.16***	−3.05***	−2.77***
Constant	(.09)	(.08)	(.06)	(.07)
Log pseudolikelihood	−41.90	−49.00	−56.05	−62.78
Akaike information criterion (AIC)	87.80	102.00	116.10	129.57
Black clustering (−1 = dispersed; 1 = clustered)	−.35*	−.04	−.01	−.14
Black clustering (−1 = dispersed; 1 = clustered)	(.16)	(.15)	(.13)	(.15)
Constant	−3.27***	−3.20***	−3.03***	−2.80***
Constant	(.10)	(.09)	(.08)	(.10)
Log pseudolikelihood	−41.88	−49.01	−56.05	−62.80
Akaike information criterion (AIC)	87.77	102.01	116.10	129.61

Notes. Robust standard errors reported in parentheses underneath logits. Estimation sample size equals 301.

*p < .05 ** p < .01 *** p < .001 (two-tailed tests).

Table 3 reports estimates (i.e., logits and standard errors) from fractional logit models linking residential racial segregation (i.e., Black–White dissimilarity, Black–White isolation, and Black clustering) and a county’s region with the Black self-employment rate, adjusting for control variables in 1980, 1990, 2000, and 2010. For each residential racial segregation index, two models are shown: Model 1 includes a county’s region and the control variables. Model 2 introduces a statistical interaction between a county’s region and each residential racial segregation index.

Table 3.

Estimates From Fractional Logit Regressions of the Black Self-employment Rate on Residential Racial Segregation, Region, and Control Variables Using the 1980, 1990, 2000, and 2010 Integrated Public Use Microdata Series.

	1980		1990		2000		2010
	Model 1^a	Model 2^a	Model 1^a	Model 2^a	Model 1^a	Model 2^a	Model 1^a	Model 2^a
Black–White dissimilarity (range: 0–1)	.36	−.10	.02	−.49	.06	−.24	.23	−.12
Black–White dissimilarity (range: 0–1)	(.32)	(.45)	(.31)	(.44)	(.22)	(.27)	(.23)	(.29)
Region (0 = nonsouth; 1 = south)	.13	−.42	.16*	−.48	.11	−.30	.05	−.42
Region (0 = nonsouth; 1 = south)	(.07)	(.32)	(.07)	(.28)	(.06)	(.19)	(.06)	(.22)
Region × dissimilarity		.89		1.14**		.77*		.88*
Region × dissimilarity		(.49)		(.44)		(.33)		(.36)
Constant	−3.77***	−3.56***	−3.21***	−3.06***	−4.98***	−4.82***	−2.95**	−2.75*
Constant	(.74)	(.80)	(.78)	(.76)	(.72)	(.71)	(1.12)	(1.13)
Log pseudolikelihood	−41.75	−41.73	−48.89	−48.86	−55.91	−55.90	−62.65	−62.64
Akaike information criterion (AIC)	107.49	109.46	121.77	123.72	135.82	137.80	149.31	151.28
Black−White isolation (range: 0–1)	.44	.29	.35	.29	.27	.21	.04	−.01
Black−White isolation (range: 0–1)	(.28)	(.30)	(.24)	(.26)	(.19)	(.19)	(.23)	(.23)
Region (0 = nonsouth; 1 = south)	.11	−.12	.17**	.05	.11*	−.09	.03	−.21
Region (0 = nonsouth; 1 = south)	(.06)	(.14)	(.06)	(.15)	(.05)	(.11)	(.05)	(.12)
Region × isolation		.52*		.30		.51*		.61**
Region × isolation		(.25)		(.28)		(.21)		(.22)
Constant	−3.56***	−3.49***	−2.91***	−2.91***	−4.74***	−4.66***	−2.87*	−2.87*
Constant	(.81)	(.81)	(.83)	(.83)	(.72)	(.72)	(1.20)	(1.19)
Log pseudolikelihood	−41.74	−41.72	−48.88	−48.88	−55.91	−55.90	−62.66	−62.64
Akaike information criterion (AIC)	107.47	109.45	121.76	123.75	135.81	137.79	149.31	151.28
Black clustering (−1 = dispersed;1 = clustered)	−.20	−.29	.13	.15	.13	−.04	.10	−.14
Black clustering (−1 = dispersed;1 = clustered)	(.16)	(.21)	(.19)	(.28)	(.12)	(.15)	(.14)	(.19)
Region (0 = nonsouth; 1 = south)	.08	−.03	.18**	.19	.12**	−.09	.04	−.28
Region (0 = nonsouth; 1 = south)	(.07)	(.15)	(.07)	(.18)	(.05)	(.16)	(.05)	(.15)
Region × clustering		.26		−.04		.40		.62*
Region × clustering		(.25)		(.29)		(.25)		(.26)
Constant	−3.86***	−3.83***	−3.20***	−3.20***	−4.93***	−4.95***	−2.83*	−2.78*
Constant	(.74)	(.74)	(.77)	(.76)	(.72)	(.71)	(1.12)	(1.13)
Log pseudolikelihood	−41.75	−41.74	−48.88	−48.88	−55.91	−55.90	−62.66	−62.63
Akaike information criterion (AIC)	107.49	109.49	121.77	123.77	135.82	137.80	149.31	151.26

Notes. Robust standard errors reported in parentheses underneath logits. Estimation sample size equals 301.

Models adjust for a county’s: (a) proportion of Black women, (b) proportion with college education, (c) proportion of Black immigrants, (d) Black poverty rate, (e) Black–White poverty ratio, (f) proportion of Black residents, (g) proportion of Black adults, (h) total square miles, and (i) population density per square mile.

*p < .05 **p < .01 ***p < .001 (two-tailed tests).

In Model 1, none of the three residential racial segregation indices significantly predicted the Black self-employment rate in 1980, 1990, 2000 or 2010. Region did not significantly predict the Black self-employment rate in 1980. By 1990, region predicted 0.16 (dy/dx = 0.006, p < .05) to 0.18 (dy/dx = 0.006, p < .05) increases in the logits of the Black self-employment rate across each residential racial segregation index. To put it another way, the Black self-employment rate was 0.6% higher in southern counties compared to nonsouthern counties for each residential racial segregation index. By 2000, region predicted 0.11 (dy/dx = 0.005, p < .05) and 0.12 (dy/dx = 0.005, p < .05) increases in the logits of the Black self-employment rate for Black–White isolation and Black clustering, respectively. By 2010, region became a nonsignificant predictor of the Black self-employment rate.

In Model 2, region moderated the relationship between Black–White isolation and the Black self-employment rate in 1980. Specifically, compared to counties outside the South, southern counties had 0.52 (dy/dx = 0.02, p < .05) higher logits of the Black self-employment rate when Black–White isolation increased. Considering an average county in the estimation sample, at Black–White isolation’s highest observed value in 1980 the predicted Black self-employment rate was roughly 1.36% higher in southern counties compared to nonsouthern counties (see row 1, column 1 of Figure 1).

Figure 1.

Statistical interactions of Black–White dissimilarity, Black–White isolation, Black clustering, and a county’s region

Over time, region emerged as a significant moderator for Black–White dissimilarity. Specifically, compared to counties outside the South, southern counties had 1.14 (dy/dx = 0.04, p < .05) and 0.77 (dy/dx = 0.03, p < .05) higher logits of the Black self-employment rate when Black–White dissimilarity increased in 1990 and 2000, respectively. Considering an average county in the estimation sample, at Black–White dissimilarity’s highest observed value in 1990 and 2000 the predicted Black self-employment rate was roughly 2% higher in southern counties compared to nonsouthern counties (see row 1, column 2 and row 2, column 1 of Figure 1). Similarly, region predicted 0.51 (dy/dx = 0.02, p < .05) increases in the logits of the Black self-employment rate when Black–White isolation increased, in 2000. Again, for an average county in the study, at Black–White isolation’s highest observed value the expected difference in Black self-employment between southern and nonsouthern counties was 2% (see row 2, column 2 of Figure 1). From 1980 to 2000, residential racial segregation associated both positively (Black–White dissimilarity and Black–White isolation) and not at all (Black clustering) with the Black self-employment rate in the South.

By 2010, this association was consistent across all three residential racial segregation indices. Notably, a county’s region moderated the relationship between each residential racial segregation index and the Black self-employment rate. Compared to their counterparts, southern counties had 0.88 (dy/dx = 0.04, p < .05) higher logits of the Black self-employment rate when Black–White dissimilarity increased (see row 3, column 1 of Figure 1). Likewise, in southern counties, Black–White isolation and Black clustering predicted 0.61 (dy/dx = 0.03, p < .05) and 0.62 (dy/dx = 0.03, p < .05) increases in the logits of the Black self-employment rate, respectively (see row 3, columns 2 and 3 of Figure 1). For the average county in this study, at Black–White dissimilarity, Black–White isolation, and Black clustering’s highest observed values the predicted Black self-employment rate was 1.77%, 1.93%, and 1.65% higher in southern counties compared to nonsouthern counties, respectively. These results suggest that residential racial segregation associated positively with the Black self-employment rate in the South by 2010.

Discussion

Motivated by an ongoing debate (see Boyd, 1996, 1998; Butler, 2005; Fairchild, 2008; Marable, 1983), this study investigated the county-level association between residential racial segregation and the Black self-employment rate by region in 1980, 1990, 2000, and 2010. Three indices measured residential racial segregation: (a) Black–White dissimilarity, (b) Black–White isolation, and (c) Black clustering. Overall, county-level associations between residential racial segregation and the Black self-employment rate were conditional upon time and region. These conditional associations support claims that context matters (Fischer & Massey, 2000).

The present results indicate that residential racial segregation did not correlate with the Black self-employment rate in 1980, 1990, 2000, or 2010, net of region and the control variables. However, in the South, Black–White isolation associated positively with the Black self-employment rate in 1980. By 2010, all three residential racial segregation indices associated positively with the Black self-employment rate in the South (see Figure 1). Overall, the results indicate that the utility of racial enclave economies emerged over time in the South.

It is important to note that the segregation indices measure residential racial segregation differently, even though they are correlated (see Appendix B). Broadly, high levels of Black–White dissimilarity represent the uneven distribution of Blacks and Whites. It weakly captures whether residential racial segregation isolates Blacks. High levels of Black–White isolation represent Blacks having few interactions with Whites, and high levels of Black clustering represents Blacks’ geographic separation from Whites. Of theoretical importance, Black–White isolation and Black clustering most accurately capture how residential racial segregation isolates Blacks. Therefore, considering Black–White isolation and Black clustering shows that Blacks’ social and physical separation from Whites emerged as a predictor of Black self-employment in the South in 2010. Consequently, my results suggest the nature of Blacks’ geographic separation from Whites shifted over time in the South. For instance, despite evidence showing that Black–White dissimilarity and isolation are declining (see Iceland et al., 2013; Logan & Stults, 2011; Table 1), Black neighborhoods became increasingly separate from White ones by 2010 (see Table 1). I speculate, Blacks’ growing separation from Whites made goods and services inaccessible, and Black self-employees emerged to provide them in the South. Similarly, Blacks’ growing separation from Whites likely increased Black autonomy in the South. For instance, LaVeist (1993) finds Black political power is high in segregated cities. The present results may corroborate that residential racial segregation similarly increases Black economic autonomy.

Although the present findings suggest that residential racial segregation facilitates protected markets, the story is more complicated. Note that Boyd (1991) claims that the protected market theory poorly explains the contemporary Black self-employment rate. And as Small and McDermott (2006, p. 1716) write, “…businesses do not rely exclusively on customers’ pockets.” Consequently, in addition to protected markets, I conclude regional shifts (see Bates, 1993, 1997; Reynolds et al., 1995; Small & McDermott, 2006) shaped residential racial segregation’s association with the Black self-employment rate. Although Sampson (2013) argued residential racial segregation creates “qualitatively unequal neighborhoods,” “the ghetto” does not define neighborhood disadvantage across regions (Small, 2008; Small & McDermott, 2006). Segregated Black neighborhoods do not uniformly experience isolation. Nonetheless, segregated Black spaces are uniformly under-resourced (Adelman, 2004; Alba et al., 2000; Pattillo, 2005; Small & McDermott, 2006). Therefore, I argue by the late 20th century and early 21st century regional shifts concentrated advantage in segregated southern Black neighborhoods, and southern Blacks used their relative advantage to provide basic goods and services to under-resourced neighborhoods in urban counties. Outside the South, extreme concentrated disadvantage prevented this effort.

Limitations

I investigated the county-level association between residential racial segregation and the Black self-employment rate by region at four time points after desegregation, an era neglected in prior studies. Results indicated that in 2010 region conditioned residential racial segregation’s association with the Black self-employment rate. This finding represents a novel contribution to a debate about residential racial segregation’s main effects. Still, this study has limitations.

First, the Black self-employment rate was operationalized as the proportion of Black self-employees. It does not capture a self-employee’s business type, sector, size, or earnings. Therefore, this study cannot investigate whether or how residential racial segregation associates with different types of Black self-employment. For instance, do Black self-employees remain confined to vulnerable service sectors that supply basic goods and services? Or, are Black self-employees well represented in robust markets like construction, finance, and insurance (Bates, 1997; Thornton, 1999)? Additionally, information on business locations was unavailable. Therefore, I cannot address whether residential racial segregation predicts where Black self-employees locate their businesses. The present analyses assume that Black self-employees cluster in segregated Black neighborhoods to serve under-resourced neighborhoods.

Next, this study investigated legal self-employment. Unexplored here is the relationship between residential racial segregation and illegal self-employment—a relationship that some scholars theorize is positive (Gold 2010; Sassen, 2001; Venkatesh, 2009). Addressing the relationship between residential racial segregation and illegal self-employment can reveal how Blacks leverage their skills and capital for survival (Matlon, 2016; Shahida et al., 2017; Valdez, 2011; Wingfield, 2008).

Future Directions

This study provides a roadmap for future research involving residential racial segregation, the Black self-employment rate, and region. First, I offer evidence that segregated southern Black neighborhoods are qualitatively different from ones outside the South. Future studies should examine the mechanisms behind these regional differences. For instance, in the estimation sample, the Black–White poverty ratio was higher in the South, compared to outside the South in 1980. By 2010, however, the Black–White poverty ratio was lower in the South, relative to outside the South. Although these results suggest that segregated southern Black neighborhoods became qualitatively different over time, further research in this area should explicitly investigate the regional shifts that protected segregated Black neighborhoods and the Black self-employment rate.

Second, new studies in this area should explore whether racism motivates Black self-employment (see Bento & Brown, 2021). Racism embeds Blacks in unequal structures that undermine their economic power. At the same time, Blacks leverage their skills, capital, and social resources to make do within unequal structures (Matlon, 2016; Valdez, 2011; Wingfield, 2008). For instance, in the South, Black–White income disparities are lower among Black self-employees (median income: $12,000) and White self-employees (median income: $15,500), than among Black paid employees and their White peers (analyses not shown, 2006–2010 ACS 5-year sample). These results suggest self-employment may help Blacks circumvent labor market discrimination. However, does racism motivate Black self-employment for other ends? Is it leveraged to assert social value or self-worth (Matlon, 2016)? Is it leveraged for resistance to racial oppression (Herring, 2004; Marable, 1983)? These and related research questions are worth pursuing.

Next, I did not explore the correlation between residential racial segregation and the Black immigrant self-employment rate. Since 1980, the foreign-born population has grown in the United States. Today, foreign-born Blacks are 9% of the Black population and are expected to comprise 17% of the Black population by 2060 (Anderson, 2015). This demographic shift is relevant because immigrants have high self-employment rates. Relatedly, in ethnically and racially diverse cities, residential racial segregation occurs between multiple groups (Logan, 2013). Scholars do not know how other patterns of residential racial segregation (e.g., Black-Hispanic dissimilarity or Hispanic clustering) influence the Black self-employment rate. Are multigroup measures of residential racial segregation important predictors of the Black self-employment rate? Further, in cities where immigrant groups live in segregated Black neighborhoods, do they use their financial and human capital to fill the need for goods and services? Future research should attend to these questions.

Conclusion

Time and region are critical when investigating the county-level association between residential racial segregation and the Black self-employment rate. It appears that residential racial segregation associates positively with the Black self-employment rate in the South in the early 21st century. Still, residential racial segregation remains insidious with far-reaching implications for Black individuals and neighborhoods. Residential racial segregation divides neighbors, schoolmates, and families. It manifests institutional discrimination and concentrates Black disadvantage. The fact that Black self-employees leverage residential racial segregation is not a positive outcome. Black self-employees are making do within unequal structures, not undoing those structures. In the end, residential racial segregation creates economic opportunities while exposing the persistence of racism and the durability of economic inequality.

Footnotes

Appendix A

The final estimation sample consisted of 301 MSA central counties. This is because counties are a stable measure of large urban or metropolitan areas over time. However, the smallest identifiable geographic area in the 1980, 1990, and 2000 5% and 2010 5-year IPUMS samples were county groups or PUMAs. County groups and PUMAs are socially irrelevant geographic areas created by the U.S. Census Bureau. For instance, large metropolitan areas, like New York City, consist of several county groups or PUMAs; small rural areas consolidate several counties into one county group or PUMA. Consequently, county-level demographic data (e.g., a county’s Black self-employment rate) were unavailable. As a solution, I conducted county group-to-county crosswalks and PUMA-to-county crosswalks to calculate county-level demographic data. Using the University of Michigan’s Population Studies Center’s (PSC) crosswalk utility (Population Studies Center, 2010), crosswalks connected county groups and PUMAs to counties with an equivalency file. The equivalency file used county group or PUMA population counts to calculate county-level demographic data.

For the 2010 PUMA-to-county crosswalk, I used the PUMA-to-county widget (aff2c), and corresponding equivalency file provided by the University of Michigan’s PSC (Population Studies Center, 2010). For 1980, 1990, and 2000, no publicly available widget or equivalency file existed, to my knowledge. Therefore, I used the PSCs 2010 PUMA equivalency file as a guide to construct 1980 county group, and 1990 and 2000 PUMA equivalency files.

For the 1980 county group-to-county and 1990, 2000, and 2010 PUMA-to-county crosswalks, I matched the corresponding equivalency file to the 1980, 1990, and 2000, 5% and 2010 5-year samples. Then I estimated each year’s respective county-level demographic counts for race as well as a county’s total population of women, adult population size, population with a college degree, immigrant population size, and aggregate employment characteristics by race. Analyses were weighted using person weight (PERWT) to obtain population representative statistics, and then aggregated to the county group or PUMA. Next, the crosswalk disaggregated these demographic counts at the county level. The end result was county-level demographic counts, by race, for all identifiable U.S. counties in 1980, 1990, 2000, and 2010, excluding Puerto Rico. To ensure validity, I checked crosswalk results for a county’s total population and total population by race to within ±5% of the 1980, 1990, and 2000 Decennial Census and 2006–2010 ACS 5-year data.

I obtained a county’s Black poverty rate, Black–White poverty ratio, population density per square mile, and total square miles from the 1980, 1990, and 2000 Decennial Census and 2006–2010 ACS. After completing the 1980 county group-to-county crosswalk, there were 3,137 identifiable counties. After completing the 2010 PUMA-to-county crosswalk, there were 3,143 identifiable counties.

County boundaries change occasionally. I accounted for all significant boundary changes reported by the U.S. Census Bureau between 1980 and 2010 (U.S. Census Bureau, n.d.b). For 1980 and 2010, partial and full county annexations were consolidated into joint counties. For instance, in 2007, Newport News, VA, partially annexed York County, VA. Therefore, both counties were combined into Newport News-York County, VA, in 1980 and 2010. I combined fully and partially annexed counties to ensure county boundaries remained consistent across time. There was one exception. In 2001, Broomfield County, CO, was created from four partially annexed counties (Adams County, Boulder County, Jefferson County, and Weld County). I dropped these four counties, because I believed four consolidated counties would provide an invalid measure of local self-employment rates and residential racial segregation levels. After consolidation, 3,131 counties for 1980 and 3,137 counties for 2010 remained. I restricted the sample to 1,100 counties that comprised the 366 MSAs that existed in 2010 (U.S. Census Bureau, n.d.a).

Appendix B

See Table b1.

Acknowledgments

The author thanks Tony N. Brown, James R. Elliott, Elizabeth Roberto, and members of the Racism and Racial Experiences (RARE) Workgroup and SPACE Co. at Rice University for their critical feedback on early manuscript drafts and perpetual encouragement.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Asia Bento

References

Adelman

R. M.

(2004). Neighborhood opportunities, race, and class: The Black middle class and residential segregation. City & Community, 3(1), 43–63. https://doi.org/10.1111/j.1535-6841.2004.00066.x

Alba

R. D.

Logan

J. R.

Stults

B. J.

(2000). How segregated are middle-class African Americans? Social Problems, 47(4), 543–558. https://doi.org/10.2307/3097134

Anderson

(2015). A rising share of the U.S. Black population is foreign born. Pew Research Center. Retrieved October 20, 2018, from http://www.pewsocialtrends.org/2015/04/09/a-rising-share-of-the-u-s-black-population-is-foreign-born/

Anselin

(2013). Spatial econometrics: Methods and models. Springer.

Bates

(1993). Banking on Black enterprise: The potential of emerging firms for revitalizing urban economies. University Press of America.

Bates

(1995). Self-employment entry across industry groups. Journal of Business Venturing, 10(2), 143–156. https://doi.org/10.1016/0883-9026(94)00018-P

Bates

(1997). Race, self-employment, and upward mobility: An illusive American dream. Woodrow Wilson Center Press.

Bento

Brown

T. N.

(2021). Belief in systemic racism and self-employment among working Blacks. Ethnic & Racial Studies, 44(1), 21–38. https://doi.org/10.1080/01419870.2020.1737170

Bivand

R. S.

(2019). SPDEP: Spatial dependence: Weighting schemes, statistics and models. R package version 1.1–3, Retrieved from http://CRAN.R-project.org/package=spdep

10.

Bogan

Darity

Jr. (2008). Culture and entrepreneurship? African American and immigrant self-employment in the United States. The Journal of Socio-Economics, 37(5), 1999–2019. https://doi.org/10.1016/j.socec.2007.10.010

11.

Bonds

(2007). Looking beyond the numbers: The struggles of Black businesses to survive: A qualitative approach. Journal of Black Studies, 37(5), 581–601. https://doi.org/10.1177/0021934705280306

12.

Boyd

R. L.

(1991). A contextual analysis of Black self-employment in large metropolitan-areas, 1970–1980. Social Forces, 70(2), 409–429. https://doi.org/10.2307/2580246

13.

Boyd

R. L.

(1996). The great migration to the north and the rise of ethnic niches for African American women in beauty culture and hairdressing, 1910–1920. Sociological Focus, 29(1), 33–45. https://doi.org/10.1080/00380237.1996.10571070

14.

Boyd

R. L.

(1998). Residential segregation by race and the Black merchants of northern cities during the early twentieth century. Sociological Forum, 13(4), 595–609. https://doi.org/10.1023/A:1022843300237

15.

Branch

E. H.

Hanley

(2011). Regional convergence in low-wage work and earnings, 1970–2000. Sociological Perspectives, 54(4), 569–592. https://doi.org/10.1525/sop.2011.54.4.569

16.

Butler

J. S.

(2005). Entrepreneurship and self-help among Black Americans: A reconsideration of race and economics (Revised ed.). SUNY Press.

17.

Charles

C. Z.

(2003). The dynamics of racial residential segregation. Annual Review of Sociology, 29(1), 167–207. https://doi.org/10.1146/annurev.soc.29.010202.100002

18.

Du Bois

W. E. B.

(1899). The Philadelphia Negro: A social study. University of Pennsylvania.

19.

Eckes

(2005). The south and economic globalization, 1950 to the future. In J.

Cobb

Stueck

(Ed.), Globalization and the American south (pp. 36–65). University of Georgia Press.

20.

Fairchild

G. B.

(2008). Residential segregation influences on the likelihood of Black and White self-employment. Journal of Business Venturing, 23(1), 46–74. https://doi.org/10.1016/j.jbusvent.2006.10.010

21.

Fairlie

R. W.

Meyer

B. D.

(1996). Ethnic and racial self-employment differences and possible explanations. The Journal of Human Resources, 31(4), 757–793. https://doi.org/10.2307/146146

22.

Farley

Allen

W. R.

(1987). The color line and the quality of life in America. Russell Sage Foundation.

23.

Farley

Frey

W. H.

(1994). Changes in the segregation of Whites from Blacks during the 1980s: Small steps toward a more integrated society. American Sociological Review, 59(1), 23–45. https://doi.org/10.2307/2096131

24.

Feagin

J. R.

(2006). Systemic racism: A theory of oppression. Routledge.

25.

Fischer

M. J.

Massey

D. S.

(2000). Residential segregation and ethnic enterprise in U.S. metropolitan areas. Social Problems, 47(3), 408–424. https://doi.org/10.2307/3097237

26.

Frey

W. H.

(2002). Metro magnets for minorities and Whites: Melting pots, the new sunbelt, and the heartland. Research Report No. 02-496. 2 2002., Population Studies Center (PSC). Retrieved August 28, 2018 from https://www.psc.isr.umich.edu/pubs/abs/1319

27.

Frey

W. H.

(2004). The new great migration: Black Americans’ return to the south, 1965–2000. Brookings. Retrieved April 28, 2018, from https://www.brookings.edu/research/the-new-great-migration-black-americans-return-to-the-south-1965-2000/

28.

Frey

W. H.

Zimmer

(1998). Defining the city and levels of urbanization. Research Report No. 98–423. 9, Population Studies Center (PSC).

29.

Gold

S. J.

(2010). The store in the hood: A century of ethnic business and conflict. Rowman and Littlefield.

30.

Gold

S. J.

(2016). A critical race theory approach to Black American entrepreneurship. Ethnic & Racial Studies, 39(9), 1697–1718. https://doi.org/10.1080/01419870.2016.1159708

31.

Grigoryeva

Ruef

(2015). The historical demography of racial segregation. American Sociological Review, 80(4), 814–842. https://doi.org/10.1177/0003122415589170

32.

Guillory

(2010). A changing south: The 2008 elections illuminate demographic and economic trends. The Review of Black Political Economy, 37(3–4), 183–189. https://doi.org/10.1007/s12114-010-9071-1

33.

Herring

(2004). Open for business in the Black metropolis: Race, disadvantage, and entrepreneurial activity in Chicago’s inner city. The Review of Black Political Economy, 31(4), 35–57. https://doi.org/10.1007/s12114-004-1009-z

34.

Hipple

S. F.

(2010). Self-employment in the United States. Monthly Labor Review, 17–32. https://www-jstor-org-443.web.bisu.edu.cn/stable/monthlylaborrev.2010.09.017

35.

Hipple

S. F.

Hammond

(2016). BLS Spotlight on statistics: Self-employment in the United States. Federal Publications.

36.

Hunt

L. L.

Hunt

M. O.

Falk

W. W.

(2008). Who Is headed south? U.S. migration trends in Black and White, 1970–2000. Social Forces, 87(1), 95–119. https://doi.org/10.1353/sof.0.0099

37.

Hwang

Hankinson

Brown

K. S.

(2015). Racial and spatial targeting: Segregation and subprime lending within and across metropolitan areas. Social Forces, 93(3), 1081–1108. https://doi.org/10.1093/sf/sou099

38.

Iceland

Sharp

Timberlake

J. M.

(2013). Sun belt rising: Regional population change and the decline in Black residential segregation, 1970–2009. Demography, 50(1), 97–123. https://doi.org/10.1007/s13524-012-0136-6

39.

Kolesnikova

Liu

(2010). A bleak 30 years: Economic progress slim for Black men from 1970 to 2000. Retrieved August 28, 2018, from https://www.stlouisfed.org/publications/regional-economist/july-2010/a-bleak-30-years-for-black-men-economic-progress-was-slim-in-urban-america

40.

Lacy

(2007). Blue-chip Black: Race, class, and status in the new Black middle class. University of California Press.

41.

LaVeist

T. A.

(1993). Segregation, poverty, and empowerment: Health consequences for African Americans. The Milbank Quarterly, 71(1), 41–64. https://doi.org/10.2307/3350274

42.

Light

I. H.

Gold

S. J.

(2000). Ethnic economies. Emerald Group Publishing.

43.

Lofstrom

Bates

Parker

S. C.

(2014). Why are some people more likely to become small-businesses owners than others: Entrepreneurship entry and industry-specific barriers. Journal of Business Venturing, 29(2), 232–251. https://doi.org/10.1016/j.jbusvent.2013.01.004

44.

Logan

J. R.

(2013). The persistence of segregation in the 21st century metropolis. City & Community, 12(2), 160–168. https://doi.org/10.1111/cico.12021

45.

Logan

J. R.

Martinez

(2018). The spatial scale and spatial configuration of residential settlement: Measuring segregation in the postbellum south. American Journal of Sociology, 123(4), 1161–1203. https://doi.org/10.1086/694652

46.

Logan

J. R.

Stults

B. J.

(2011). The persistence of segregation in the metropolis: New findings from the 2010 Census. Retrieved August 28, 2018, from https://s4.ad.brown.edu/Projects/Diversity/Data/Report/report2.pdf

47.

Logan

J. R.

Stults

B. J.

(2014). Interpolating U.S. decennial census tract data from as early as 1970 to 2010: A longitudinal tract database. The Professional Geographer, 66(3), 412–420. https://doi.org/10.1080/00330124.2014.905156

48.

Manson

Schroeder

Van Riper

Ruggles

(2017). IPUMS national historical geographic information system: Version 12.0 [database]. University of Minnesota. Retrieved April 30, 2018, from https://doi.org/10.18128/D050.V12.0

49.

Marable

(1983). How capitalism underdeveloped Black America: Problems in race, political economy, and society. Haymarket Books.

50.

Massey

D. S.

Denton

N. A.

(1988). The dimensions of residential segregation. Social Forces, 67(2), 281–315. https://doi.org/10.2307/2579183

51.

Massey

D. S.

Denton

N. A.

(1993). American apartheid: Segregation and the making of the underclass. Harvard University Press.

52.

Matlon

(2016). Racial capitalism and the crisis of Black masculinity. American Sociological Review, 81(5), 1014–1038. https://doi.org/10.1177/0003122416658294

53.

McKenzie

(2013). Neighborhood access to transit by race, ethnicity, and poverty in Portland, OR. City & Community, 12(2), 134–155. https://doi.org/10.1111/cico.12022

54.

Oakland

W. H.

Sparrow

F. T.

Stettler

H. L.

(1971). Ghetto multipliers: A case study of Hough. Journal of Regional Science, 11(3), 337–345. https://doi.org/10.1111/j.1467-9787.1971.tb00264.x

55.

Oliver

M. L.

Shapiro

T. M.

(1995). Black wealth/White wealth: A new perspective on racial inequality. Routledge.

56.

Papke

L. E.

Wooldridge

J. M.

(1996). Econometric methods for fractional response variables with an application to 401(k) plan participation rates. Journal of Applied Econometrics, 11(6), 619–632. https://doi.org/10.1002/(SICI)1099-1255(199611)11:6<619::AID-JAE418>3.0.CO;2-1

57.

Pattillo

(2005). Black middle-class neighborhoods. Annual Review of Sociology, 31(1), 305–329. https://doi.org/10.1146/annurev.soc.29.010202.095956

58.

Pendergrass

(2013). Perceptions of race and region in the Black reverse migration to the south. Du Bois Review-Social Science Research on Race, 10(1), 155–178. https://doi.org/10.1017/S1742058X13000088

59.

Population Studies Center (2010). Creating county-level statistics from Public Use Microdata Areas (PUMAs). University of Michigan Population Studies Center Institute for Social Research. Retrieved April 30, 2018, from https://www.psc.isr.umich.edu/dis/census/Features/puma2cnty/

60.

Quillian

(2017). Segregation as a source of contextual advantage: A formal theory with application to American cities. RSF: The Russell Sage Foundation Journal of the Social Sciences, 3(2), 152–169. https://doi.org/10.7758/rsf.2017.3.2.07

61.

Reardon

S. F.

Townsend

J. B.

(2018). SEG: Stata module to compute multiple-group diversity and segregation indices. Retrieved April 30, 2018, from https://EconPapers.repec.org/RePEc:boc:bocode:s375001

62.

Reynolds

P. D.

Miller

Maki

W. R.

(1995). Explaining regional variation in business births and deaths: U.S., 1976–88. Small Business Economics, 7(5), 389–407. https://doi.org/10.1007/BF01302739

63.

Riley

A. R.

(2018). Neighborhood disadvantage, residential segregation, and beyond—lessons for studying structural racism and health. Journal of Racial & Ethnic Health Disparities, 5(2), 357–365. https://doi.org/10.1007/s40615-017-0378-5

64.

Ruggles

Genadek

Goeken

Grover

Sobek

(2017). Integrated public use microdata series: Version 7.0 [dataset]. University of Minnesota. Retrieved April 30, 2018, from https://doi.org/10.18128/D010.V7.0

65.

Sampson

R. J.

(2013). Great American city: Chicago and the enduring neighborhood effect. University of Chicago Press.

66.

Sassen

(2001). The global city: New York, London, Tokyo. Princeton University Press.

67.

Schaffer

R. L.

(1973). Income flows in urban poverty areas: A comparison of the community income accounts of Bedford-Stuyvesant and Borough Park. Lexington Books.

68.

Shahid

M. S.

Rodgers

Williams

C. C.

(2017). Evaluating the participation of an ethnic minority group in informal employment: A product of exit or exclusion? Review of Social Economy, 75(4), 468–488. https://doi.org/10.1080/00346764.2016.1269941

69.

Small

M. L.

(2008). Four reasons to abandon the idea of ‘the ghetto’. City & Community, 7(4), 389–398. https://doi.org/10.1111/j.1540-6040.2008.00271_8.x

70.

Small

M. L.

McDermott

(2006). The presence of organizational resources in poor urban neighborhoods: An analysis of average and contextual effects. Social Forces, 84(3), 1697–1724. https://doi.org/10.1353/sof.2006.0067

71.

Smith

D. T.

(2005a). Developing entrepreneurship among African Americans: The effects of urban residence. Race, Gender & Class, 12(1), 152–168.

72.

Smith

D. T.

(2005b). Developing self-employment among African Americans: The impact of household social resources on African American entrepreneurship. Economic Development Quarterly, 19(4), 346–355. https://doi.org/10.1177/0891242405279693

73.

Stack

C. B.

(1996). Call to home: African-Americans reclaim the rural south. Basic Books.

74.

Thornton

P. H.

(1999). The sociology of entrepreneurship. Annual Review of Sociology, 25(1), 19–46. https://doi.org/10.1146/annurev.soc.25.1.19

75.

U.S. Census Bureau (N.d.a). Historical delineation files. United States Census Bureau. Retrieved June 30, 2020, from https://www.census.gov/geographies/reference-files/time-series/demo/metro-micro/historical-delineation-files.html

76.

U.S. Census Bureau (N.d.b). Substantial county changes 1970-present. Retrieved April 23, 2018, from https://www.census.gov/geo/reference/county-changes.html

77.

Valdez

(2011). The new entrepreneurs: How race, class, and gender shape American enterprise. Stanford University Press.

78.

Venkatesh

(2009). Off the books: The underground economy of the urban poor. Harvard University Press.

79.

Waldinger

(1995). The ‘other side’ of embeddedness: A case-study of the interplay of economy and ethnicity. Ethnic & Racial Studies, 18(3), 555–580. https://doi.org/10.1080/01419870.1995.9993879

80.

Williams

(2018). Analyzing proportions: Fractional response and zero one inflated beta models. Retrieved August 2, 2018, from https://www3.nd.edu/∼rwilliam/stats3/FractionalResponseModels.pdf

81.

Williams

D. R.

Collins

(2001). Racial residential segregation: A fundamental cause of racial disparities in health. Public Health Reports, 116(5), 404–416. https://doi.org/10.1093/phr/116.5.404

82.

Wilson

W. J.

(1987). The truly disadvantaged: The inner city, the underclass, and public policy. University of Chicago Press.

83.

Wingfield

A. H.

(2008). Doing business with beauty: Black women, hair salons, and the racial enclave economy. Rowman and Littlefield.

84.

Zhou

(2004). Revisiting ethnic entrepreneurship: Convergencies, controversies, and conceptual advancements. The International Migration Review, 38(3), 1040–1074. https://doi.org/10.1111/j.1747-7379.2004.tb00228.x