The Impact of on Racial Inequality in in the Contemporary U.S. South

Heather A. O’Connell, University of Wisconsin, Madison

Despite Civil Rights legislation, racial inequality persists, especially in the context of poverty. This study advances the literature on racial inequality and the Southern legacy of

slavery by examining slavery’s relationship with inequality in poverty. I analyze county-level Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 U.S. Census data using regression and spatial data analysis techniques. I find the 1860 slave concentration is related to contemporary black-white inequality in poverty, independent of contemporary demographic and economic conditions, racialized disparities and racial threat. My research suggests the importance of slavery for shaping existing U.S. racial inequality patterns. Insights derived from this research, including the formulation of legacy as a place-based, continuous phenomenon that is distinct from racial threat, provide the basis for future research on legacy’s mechanisms.

Racial inequality persists despite improvements in racial minorities’ and eco- nomic standing. Although racial minorities have faced more favorable legal conditions since the 1960s, differences in social and economic success by race continue. It is critical that we understand the sources of racial inequality in the United States given direct efforts to ameliorate such disparities. Disadvantage is especially pronounced when comparing black poverty rates to those of non-Hispanic whites (Schiller 2008). In 2000, the average county disparity in the black poverty rate relative to the non-Hispanic white poverty rate is greater in the South than in the non-South (2.70 vs. 2.56, p < .05). However, the South is not homogenous. On average, the Southern black poverty rate was nearly 170 percent greater than the corresponding non-Hispanic white poverty rate (2.70), yet black disadvantage is not uniformly high across the region (see Figure 1). I advance the racial inequality literature by investigating the role of the legacy of slavery in generating black-white inequality in the U.S. South. Previous research on Southern poverty sug- gests spatial variation in slavery is negatively related to contemporary county poverty rates (Levernier and White 1998), but research has yet to examine slavery’s relationship to racial disparities in poverty despite slavery’s inextricable link to race in the U.S. context. I follow previous research by focusing on the South because the regional concentration of slaves within the South suggests slavery had the greatest direct impact on southern local areas.1

This was partially supported by the University of Wisconsin-Madison Center for Demography and Ecology center grant, #R24 HD047873. I would also like to thank Katherine Curtis, Mara Loveman, Mustafa Emirbayer, Amelia Karraker, Matt Desmond, Julia Goldberg, Adam Slez, Tim Smeeding, Ruth Lopez-Turley and Jack DeWaard for all their comments and support. Direct correspondence to Heather O’Connell, Department of Sociology, University of Wisconsin, Madison, 1180 Observatory Drive, Madison, WI 53706-1393. E-mail:[email protected].

© The Author 2012. Published by on behalf of the University of North Carolina at Social Forces 90(3) 713–734, March 2012 Chapel Hill. All rights reserved. For permissions, please e-mail: [email protected]. doi: 10.1093/sf/sor021 Advance Access publication March 26, 2012 714 • Social Forces 90(3)

Figure 1. Black-White Inequality in County Poverty in the United States South, 2000 Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021

< 1 (min = 0) 1 - 2.5 2.5 - 3.5 > 3.5 (max = 21.05) Note: < 1 represents a white disadvantage.

The importance of this work lies in emphasizing the added utility of consider- ing legacy when studying racial inequality. In so doing, I draw out the theoretical implications of what this analysis means for how we understand slavery’s contem- porary impact. I accomplish this goal by simultaneously examining several major explanations for county-level racial inequality, including local historical attachment to slavery.

Explanations for Spatial Variation in Black-White Inequality in Poverty The extant literature suggests four sources of Southern spatial variation in the black- white poverty gap: demographic and economic conditions, racial threat, wealth accu- mulation, and the legacy of slavery. First, extensive work on the spatial distribution of overall poverty (i.e., the poverty rate for the total population) has indentified key correlates of poverty (e.g., Voss et al. 2006). Established poverty correlates include metropolitan status, industrial composition and economic conditions. These factors are important for explaining overall county poverty rates, yet they may not explain why there is greater racial inequality in poverty in some places vs. others. Additional explanations need to be considered. Second, the racial threat literature has documented a positive relationship between a minority collective’s relative size in a place and its disadvantage compared to the non-Hispanic white group (e.g., Beggs, Villemez and Arnold 1997; Blalock 1956, 1957; Cohen 1998, 2001; Kornrich 2009; Tigges and Tootle 1993). Blalock’s (1967) Impact of Slavery on Poverty in the Contemporary U.S. South • 715 theoretical work on the topic suggests as the threat to the majority posed by a minority increases, so does the majority’s motivation to maintain its privileged posi- tion. Feelings of threat induce discriminatory behaviors on the part of the majority toward the minority. In this case, the threat of losing limited economic resources to blacks translates into an even more limited job market for blacks. To the extent to which large minority populations increase the majority’s desire to protect available resources, such as jobs, blacks will be more economically in places Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 where there is a relatively large black population compared to places where there are fewer blacks. Third, research related to wealth accumulation suggests wealth improves a per- son’s social and economic status, including political influence and purchasing power (Gilbert, Sharp and Felin 2002; Oliver and Shapiro 1995; Shapiro 2004). Wealth may be important for keeping households out of poverty because of the advantaged opportunities it provides in terms of neighborhood location and (Oliver and Shapiro 1995; Shapiro 2004). Wealth is not evenly distributed along racial- ized lines, such that blacks are at a distinct disadvantage relative to non-Hispanic whites (Shapiro 2004). Racial inequality in wealth accumulation patterns is one plausible explanation for the spatial distribution of racial inequality in poverty in the U.S. South. Finally, demographic and economic conditions, racial threat and racial disad- vantage in wealth accumulation are potentially important factors for explaining county-level racial inequality in poverty; however, an additional body of literature requires consideration, that of the legacy of slavery. Research on the legacy of slavery suggests slavery is related to a distinct contemporary social structure (Levernier and White 1998; Vandiver, Giacopassi and Lofquist 2006). A legacy of slavery social structure is theorized to be infused with remnants of slavery, including its ideology and practices. Researchers suggest that “historical conditions have become imprinted on the social and legal traditions of states, or even areas within a state.”(Vandiver et al. 2006:20) The social and legal traditions of a place with a historical connec- tion to slavery stem from an ideology that ranked whites above blacks. Drawing from this understanding of slavery’s ideology, I extend previous work by suggesting slavery’s social structural legacy differentially affects life outcomes to the disadvan- tage of blacks through adherence to social expectations of blacks’ racial inferiority. Economic disadvantage may exist due to discrimination within the labor market, or through lower quality education, which translates into lower . This is not to imply that all individuals living in the South are inherently racist. In fact, the legacy of slavery and should be considered as distinct concepts. Racism pervades all regions of the United States. The argument presented in the extant legacy literature focuses on the social milieu of places, which suggests the legacy of slavery affects the guidelines by which individuals structure their behaviors not individual attitudes. The legacy of slavery literature suggests the utility of this concept for understanding the spatial distribution of racial inequality in poverty, yet the causal pathway through time is unclear. Slavery officially ended 150 years ago. However, despite its dissolution, 716 • Social Forces 90(3) it has shaped social, economic and political agendas since its inception in the United States. Coercive systems of social and economic control in the U.S. South, such as sharecropping, lynching and Jim Crow laws, developed in response to the social and economic void created by the abolition of slavery (see Royce 1985; Shifflett 1982; Tolnay and Beck 1995). As in the period from 1870 to 1960, slavery may continue to affect contemporary Southern through the replacement of previous coercive systems with modified institutions that remain rooted in an exploitative slavery ideol- Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 ogy. Contemporary institutions, such as the education system, may in turn generate contemporary black-white economic disparities. Although this work does not empiri- cally identify legacy’s mechanisms, I take four steps toward this goal by defining its dimensions. First, I focus on legacy from a place-based perspective. Second, I challenge the binary conceptualization of legacy that pervades the literature by suggesting it is present in places to varying degrees according to historical slave concentration. Third, I discuss why a historical measure of slavery is preferable to the contemporary concentra- tion of blacks as an indicator of the legacy of slavery. Finally, I articulate the potential linkages between slavery and contemporary black-white disparities. It is necessary to first understand the shape of legacy before research can explore its mechanisms. The dimensions laid out here provide the essential foundation for more in-depth work on the legacy of slavery.

Legacy of Slavery Perspectives: Individual vs. Place The legacy of slavery concept takes multiple forms in the extant literature. Broadly, these forms can be classified as either individual-based or place-based perspectives. In the debate on the influence of slavery on contemporary black family formation and delinquency, the underlying perspective assumes the legacy of slavery is a charac- teristic of blacks, including their cultural norms (e.g., Cross 2003; Moynihan 1965; Patterson 1998; Wilson 1987). This represents the individual-based perspective on legacy. Drawing from this perspective, one would expect slavery to affect how descen- dants of slaves think and behave, which would then, for example, affect the prevalence of delinquency among contemporary blacks (for a counter argument to this thesis see Cross 2003). An alternative view of legacy suggests it is a characteristic of place (e.g., James 1988; Levernier and White 1998; Roscigno and Bruce 1995; Ruef and Fletcher 2003; Vandiver et al. 2006). Although both of these perspectives provide productive avenues of study, there are important benefits to focusing on the impact of slavery on places. First, this perspective suggests a new way of analyzing the effect of slavery on contem- porary society in light of recent arguments against the individual perspective (Cross 2003). Second, it allows for the direct measurement of slavery, which provides a means of estimating legacy (as imperfectly as it does) unlike in discussions of an individual- level legacy where a legacy is largely assumed. Third, the place-based legacy of slavery perspective acknowledges the fact that blacks, or more precisely slaves, were not the only group, or social object, to be affected by the institution of slavery. Slavery affected Impact of Slavery on Poverty in the Contemporary U.S. South • 717 the dynamics between racialized collectives, not just one collective. Therefore, the legacy of slavery would manifest at the place level in the form of, for example, norms or social knowledge structuring interaction between blacks and whites. These features of the place-based perspective prove elucidating when confronted with a range of county- level black-white inequality within the U.S. South. Given my interest in understanding why some places have greater black-white inequality than others, I draw solely from the place-based legacy perspective. Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021

Place-Based Legacy of Slavery Research Few studies on the continued importance of slavery employ the place perspective. I review three important exceptions that have shown the usefulness of conceptualizing legacy as a characteristic of place (Levernier and White 1998; Ruef and Fletcher 2003; Vandiver et al. 2006). My review of these studies also underscores the need for and uniqueness of the present research. In a study of individual migration outcomes (i.e., migration to another Confederate state or outside of the Confederate states) and individual, intra-gen- erational occupational mobility, Ruef and Fletcher (2003) find evidence that ante- bellum social structures are reproduced after the abolition of slavery. The data for their analysis are from free-black and slave gathered by the WPA Federal Writer’s Project. They find a black individual’s status while slavery was legal is largely reproduced in the economic conditions following the abolition of slavery. However, occupational status did improve significantly for those who left their state of residence, particularly for those who migrated to states outside the Confederacy. This evidence supports their argument regarding the reproduction of an antebellum social structure in places with a historical tie to slavery and their argument that social structure affects blacks’ economic outcomes. It is worth note that this par- ticular study is of slavery’s relationship with blacks’ economic outcomes in the late 1800s, a relatively short temporal distance from the abolition of slavery compared to the year 2000. While Ruef and Fletcher (2003) examined outcomes from a short time period after slavery’s abolition, other research has established a relationship between a of slavery and contemporary outcomes. For example, Vandiver et al. (2006) analyze the relationship between slavery and modern legal executions. The aim of their study is to better understand the relationship between slavery and the death penalty. To do so the researchers refine identified slave state categories to reflect states without slavery in 1860, Confederate states, slave states not in the Confederacy, and slave . They analyze the distribution of all executions carried out between 1977 and 2005 across each of the above categories. Slave states, regardless of Confederate status, carried out the vast majority of the legal executions in the United States from 1977 through 2005. They conclude that slavery legitimates harsh punishment in the modern era, perhaps via the development of cultural attitudes that accept the death penalty as a way to maintain social order. Even though Vandiver et al. discuss the differential impact of 718 • Social Forces 90(3) such a culture on blacks vs. whites, they do not study racial inequality per se. However, critically, their study suggests historical legacies, particularly slavery’s, are related to the spatial distribution of contemporary inequality outcomes. Finally, Levernier and White (1998) conducted the only study to analyze the rela- tionship between slavery and poverty, which was a study explaining overall poverty rates. Their focus was on the extent to which the institutionalized culture of a place adds to our understanding of county-level family poverty rates. They employed Bartley’s Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 (1990) plantation belt definition, which identifies areas dependent on a plantation as defined in the 1910 U.S. Census, to delineate where distinct institution- alized cultures would have developed and endured. Analyzing Georgia’s county-level poverty rates for 1990, they conclude there are institutionalized cultural factors related to the plantation belt that affect poverty rates within the region. Unexpectedly, they report a negative relationship between poverty and a county’s plantation belt status. Although poverty rates are higher in the plantation belt, a county’s overall family poverty rate would be higher if it were located outside of the region and had the same racial and demographic characteristics.

Theoretical and Analytical Contributions Despite slavery’s inextricable link to race in the United States, none of the aforemen- tioned studies examine slavery’s relationship with racial inequality. In addition, to my knowledge, only one study has analyzed the relationship between slavery and poverty, and that was a study of overall poverty rates (Levernier and White 1998). Overall poverty includes all racialized categories. Racial inequality in poverty is the ratio of the poverty rates for two racialized categories. It is important to study inequality in poverty in addition to overall poverty rates because the processes behind inequality may be distinct from those contributing to overall poverty. In fact, where overall poverty rates are higher one might expect racialized differences in poverty rates to be smaller. Whites may be more likely to suffer from poverty in places with a higher overall poverty rate compared to places with a low poverty rate, thus decreasing the disparity between white and black poverty rates. Examining only total poverty ignores the possibility slavery is differentially related to racialized poverty rates. In addition to analyzing racial disparities, this research is distinct in a second pivotal way. Previous studies have measured the legacy of slavery with slave or Confederate state identifiers (Ruef and Fletcher 2003; Vandiver et al. 2006) or with a plantation belt binary variable (Levernier and White 1998). These measures do not capture the extent to which a place was dependent on slavery. Capturing the depth of slavery in a place is important because the degree to which slavery was practiced may have implications for the extent to which it had lasting effects on that place. To examine the relation- ship between inequality and legacy I propose a measure of legacy based on the 1860 concentration of slaves in a county unit. This modification adds to the legacy of slavery literature by examining a new measure of the legacy of slavery. This is an important move because it has implications for how we understand legacy. Impact of Slavery on Poverty in the Contemporary U.S. South • 719

In sum, the present research extends the extant literature by empirically examining the relationship between slavery and a contemporary outcome – the black-white pov- erty gap – and by employing an innovative measure of the legacy of slavery. Although a continuous measure of the legacy of slavery has not been employed in the previous literature, the extant rhetoric suggests where there was a large concentration of slaves there remains a social structure that negatively affects blacks’ life outcomes (Levernier and White 1998; Vandiver et al. 2006). Therefore, I anticipate a positive relationship Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 between the historical slave concentration and contemporary racial inequality in poverty within a place, such that blacks are disadvantaged relative to whites.

Concentration of Blacks: Why a Historical Operationalization is Necessary An alternative operationalization of the legacy of slavery has been employed in the poverty literature: the population concentration of blacks (Adeola 1999; Swanson et al. 1995). The rationale in previous research for measuring the legacy of slavery in this way is the strong geographic correlation between where blacks live today and where slavery was once practiced. Such an approach avoids the issue of attaching historical data to contemporary units and potentially addresses concerns regarding the extent to which slavery was important to a place. However, the large amount of black migration that has occurred since 1860 both out of and within the South (Tolnay 2003) compromises the accuracy of this estimate of legacy. The concentration of blacks as an operationalization of the legacy of slavery also has a conceptual drawback. The contemporary black concentration operationalization is the same measure employed in the racial threat literature. Employing the same opera- tionalization for two distinct concepts negates examination of the relationship between the two concepts and makes it difficult to discern which conceptual understanding is correct.2 Disentangling the independent relationships of the legacy of slavery and racial threat is further complicated by the fact that both sets of literature predict a positive relationship with racial inequality. The historical operationalization of the legacy of slavery introduced in the present study allows researchers to discern between racial threat and the legacy of slavery.

Alternative Explanations of Racial Inequality in Poverty Rates To provide a rigorous test of the extent to which the legacy of slavery contributes to understanding racial inequality in poverty in the contemporary South, I examine the impact of slavery on inequality independent of several alternative explanations. I consider the demographic and economic structure of a place, racial threat and racial disparities in wealth accumulation.

Demographic and Economic Structure Greater inequality in poverty is likely to exist in places where there is a greater range of economic opportunities because there is a larger difference between the high and low 720 • Social Forces 90(3) ends of the spectrum. This potential inequality will be translated into racial inequality to the extent that one racial group is concentrated at one end of the spectrum relative to the other group. A place with little variation in job opportunities would not be expected to manifest as much racial inequality even if both places support similar social structures. Therefore, it is important to take place characteristics into consideration when examining racial inequality.

Research has shown opportunity structures are different in metropolitan counties Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 than non-metropolitan ones (Levernier 2003; Lichter 1989). Metropolitan status may serve as a proxy for the range of options, where a greater range is more open to inequality (metropolitan counties) than a limited range found in smaller, more isolated places (non-metropolitan counties). Similarly, places con- taining high-income industries, namely FIRE (finance, insurance and real estate industry), may have greater inequality because those places support a wider range of than places without high paying jobs. The opposite could be posited for the , manufacturing, services and other professional industries; a higher concentration of these industries represents a compact income range, which would be negatively related to inequality. It is also likely that a place with higher unemploy- ment rates will positively correlate with racial inequality because the excess labor force suppresses the wages of low-income earners, who are more likely to be black ( 2000; Lichter 1989). I expect the income distribution of a place, as represented by the proportion of households with a middle-class income, to be negatively related to racial inequality in poverty. Where the proportion of households in the is higher there are smaller proportions of households at the extreme ends of the income distribution. A larger middle class represents a less polarized income distribution and therefore fewer opportunities for racial inequality to arise. I have chosen a measure that focuses specifi- cally on the middle of the distribution over a for two reasons. First, this measure allows me to identify the proportion of households that are securely out of poverty. My interest in poverty specifically rather than income levels more generally makes the percentage well above poverty more important than the distance between extreme incomes. Indeed, there may be near income equality in a place where everyone is poor. A Gini coefficient for overall income inequality is better suited for an analysis of racial inequality in income than of poverty. Second, in qualitative research, a promi- nent middle class has been shown to mediate between the social interests of the and low-income classes to the benefit of the poor (Duncan 1999). My analysis provides a large-scale, quantitative test of such a buffer relationship within a racialized context. Other variables often included when analyzing overall county poverty rates could also explain why the degree of racial inequality in poverty in certain places is higher or lower. Higher rates of single-mother households are expected to be positively associ- ated with inequality in poverty because blacks have a higher rate of single motherhood than whites and are, therefore, disproportionately affected by the economic restric- tions and related to single motherhood. Alternatively, the proportion of families with children is not expected to affect black-white inequality, nor is overall Impact of Slavery on Poverty in the Contemporary U.S. South • 721 county educational attainment, because these variables equally affect the black and white groups.

Racial Threat As previously mentioned, where there is a greater degree of racial threat there is likely to be greater racial inequality, such that the minority collective is disadvantaged relative Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 to the majority (Blalock 1967). Threat is estimated by the relative size of the minor- ity population. Therefore, I expect a positive relationship between the contemporary concentration of blacks and racial inequality in poverty. However, it is important to note the relationship between relative size, threat and subsequent inequality is not necessarily linear (Blalock 1967). Blalock (1967) suggests the positive relationship will either decrease or increase at the highest concentrations depending on the source of threat leading to the observed inequality. A decreasingly positive relationship corresponds to an economic threat interpretation of the propor- tion black relationship. Conversely, an increasingly positive relationship implicates political threat. Alternatively, a negative curvilinear relationship could be the result of a minority group developing a means of protection from majority discrimination where there is a large minority concentration. I capture any nonlinearities in the relationship using the proportion black squared.

Racial Inequality in Wealth Accumulation Swanson et al. (1995) and Shapiro (2004) suggest contemporary poverty and inequal- ity are related to blacks’ historical disadvantage in accumulating wealth. Black wealth accumulation was stunted both by slavery and its successors, sharecropping, Jim Crow laws and overt economic discrimination against blacks. Blacks were unable to begin accu- mulating wealth until a much later date compared to non-Hispanic whites. This wealth disadvantage is a result of the successive coercive systems replacing slavery and could be conceptualized as one of slavery’s legacies. It is a key alternative to the direct impact of the social structural legacy of slavery on contemporary disparities discussed in previous research (Levernier and White 1998; Ruef and Fletcher 2003; Vandiver et al. 2006). Historical discrimination, as embodied by slavery, sharecropping and Jim Crow, affects the present generation’s outcomes via the wealth earlier generations passed on. Wealth from parents, for example, has a large impact on an adult child’s ability to buy a house and on the location of that house (Shapiro 2004). The link between neigh- borhoods and public schools ensures the quality of education, and therefore future job outcomes, of children within historically disadvantaged categories are affected by wealth accumulation. To be clear, both education and occupation are important for understanding who is poor. Homes are a major form of wealth in the United States (Oliver and Shapiro 1995; Shapiro 2004). I propose a measure of homeownership rates to represent wealth accu- mulation.3 A variable expressing a county’s ratio of the black and white homeownership rates measures the extent to which blacks have been disadvantaged in accumulated 722 • Social Forces 90(3) wealth relative to whites in that county. As the ratio increases, the black collec- tive’s wealth position improves relative to whites and the black collective would have more wealth on which to rely in hard times; therefore, I expect a negative relationship between the homeownership ratio and racial inequality in poverty.

Data Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 I limit the analysis to counties within the U.S. Census-defined South, excluding Oklahoma, which was not a state in 1860. Given its status in 1860, data on Oklahoma was not collected at the county level, which precludes it from this analysis. Focusing solely on the South is ideal for several reasons. First, the South provides the most rigor- ous context in which to disentangle the concepts represented by the concentration of blacks and 1860 slavery because of the high concentration of both within the region. Second, limiting analysis to the South permits comparison with previous studies on the legacy of slavery that only include places located there (e.g., Levernier and White 1998; Swanson et al. 1995). I use county-level, summary file 3 U.S. Census data from 2000 and population data from the 1860 U.S. Census to examine the relationship between a county’s structural characteristics and its racial inequality in poverty.4, 5 The county is appropriate because it closely approximates a person’s social and political environment (Irwin 2007). In addition, an extensive body of literature on place poverty has focused on the county unit (e.g., Adeola 1999; Friedman and Lichter 1996; Levernier and White 1998; Voss et al. 2006; White, Voss and Long 2008).

Measures Details of the included variables are described in Table 1. All variables are from the 2000 U.S. Census except metropolitan status, which is from the Economic Research Service 2003 Rural-Urban Continuum Codes.6 There have been few attempts to create a middle-class variable in the quantitative poverty literature, but an income range of 200 and 400 percent of a family matches well with census percentiles suggested to encompass the middle class (Lichter, Qian and Crowley 2005). I used the poverty threshold for a four-person family with two children as the base from which I calculated the $33,790 to $67,580 range. The census income categories necessitated rounding to the final $33,000 and $64,000 range. The spatial distributions of the focal independent variables are presented in Figure 2 to depict their interrelationship and geographic location.

Data Limitations The historical data source only (and understandably) includes data on counties that existed in 1860. County creation and deletion patterns are such that using either 1860 or 2000 boundaries exclusively does not address the issue because some counties exist at both time points but with varying geographic areas. Spatially consistent units are Table 1: Variables with Measurement Descriptions Dependent Variable Measurement Black-White Inequality in Poverty Black poverty rate/ the non-Hispanic white poverty rate Independent Variables Legacy of Slavery Number of slaves in 1860/ total population in 1860 Racial Threat: proportion black Number of blacks/ total population proportion black squared Proportion black * proportion black Relative Wealth Accumulation Black homeownership rate/ non-Hispanic white homeownership rate Metropolitan Status Codes 1-3 from the Economic Research Service Rural-Urban Continuum Codes are metropolitan counties (coded 1) and codes 4-9 are non-metropolitan counties (coded 0) Impact ofSlavery onPovertyintheContemporaryU.S.South • 723 Rate Individuals unemployed/ the civilian labor force Agricultural Industry Individuals employed in agriculture, forestry, fishing or hunting, and mining/ the civilian labor force Manufacturing Industry Number of individuals employed in durable and non-durable manufacturing/ the civilian labor force FIRE Industry Number of individuals employed in finance, insurance and real estate/ the civilian labor force Other Professional Industry Number of individuals employed in education, , public administration and other professional services/ the civilian labor force Service Industry Number of individuals employed in arts, entertainment, recreation, accommodation and services, retail and wholesale trade/ the civilian labor force Education Number of individuals 25 or older with a high school education or higher/ the number of individuals 25 or older Families with Kids Number of families with children under 18/ the total number of families Single Mother Families Number of families with children headed by a single woman/ the total number of families Middle Class Number of households with an income between $33,000-64,000/ the total number of households

Note: Variables are measured in 2000 unless otherwise specified. Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 September 29 on guest by https://academic.oup.com/sf/article/90/3/713/2235802 from Downloaded 724 • Social Forces 90(3)

Figure 2. The Spatial Distribution of the Focal Independent Variables

Proportion of Slaves, 1860 < .10 (min = 0) .10 - .25 .25 - .50 > .50 (max = .91) Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021

Proportion of blacks, 2000 < .10 (min = 0) .10 - .25 .25 - .50 > .50 (max = .86)

Black Homeownership Relative to White Homeownership, 2000 < .10 (min = 0) .10 - .30 .30 - 1 > 1 (max = 4.13) Note: < 1 represents a black disadvantage. Impact of Slavery on Poverty in the Contemporary U.S. South • 725 critical because the data from both time periods need to reflect the character of the same place; a unit encompassing a larger area in 1860 than in 2000 may include an area that would qualitatively shift the character of the place in 2000. There are two avenues for addressing this issue. The primary approach is to aggregate the spatial units to create spatially consistent units across time. A second method for attaching historical data is to allocate data to contemporary units based on some crite- ria. The former approach has some negative analytical consequences. First, aggregation Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 can lead to large spatial units that provide little social meaning. Second, models based on aggregated units may suffer from more multicollinearity issues than models of the original spatial units. Therefore, I present results from models using contemporary county boundaries with the historical data attached through a redistribution process. I applied population weights from 2000 to the 1860 data for counties that would have been clustered. For example, if County A has a 2000 population size of 100 and is in cluster 1, which has a cluster population size of 400, then County A’s population weight is .25. One fourth of the total 1860 and slave populations would be reallocated to County A. This approach ignores population concentration shifts across counties within a cluster, yet it is preferable to assuming slaves were evenly distributed across space (Downey 2006). An even distribution assumption is particularly problematic for slave data because plantations, where the bulk of the slave population resided, were clustered in agriculturally fertile places. Assuming an even distribution could skew the results by misallocating the slave population. This is not to say the assumption of constant population concentrations within a county cluster is beyond reproach. As a check on the validity of these results, I repeat the analysis using aggregated units. Such an analysis is an appropriate method for examining the extent to which the assumption of constant population concentrations affects my results because in the aggregated analysis the historical data values are unaf- fected by assumptions about population distribution. To identify counties involved in the creation, deletion or expansion of another county I compared the 1860 county area to that in 2000 using shapefiles manipulated in ArcGIS. The criterion for clustering county units is based on whether a county’s area changes by more than 5 percent between the two time points. All counties involved in the significant area exchange are combined in a cluster. The raw data are correspond- ingly summed to reflect the geographic aggregation process.7 Data Analysis Approach I use ordinary least squares regression and spatial data analysis techniques to esti- mate the relationship between racial inequality in poverty and the legacy of slavery. Spatial data analysis techniques are necessary because the data are geo-referenced and poverty research has demonstrated the presence of spatial autocorrelation (Voss et al. 2006). Spatial autocorrelation is a type of correlated residuals: it affects regression results through artificially low standard errors, which can result in a higher probabil- ity of type I error than expected given the p . I examine each model for spatial autocorrelation. 726 • Social Forces 90(3)

Although I add the sets of variables into the model separately, I only present the final model to focus on the main results related to legacy. The steps taken to reach the final model are as follows. First, I include only the standard correlates of poverty. Second, I add each primary independent variable in turn to the baseline model, wealth accumulation, racial threat and legacy. Third, I examine the contribution of each of the poverty covariates. Those that do not affect model fit or the other variables’ coefficients are excluded from the final model. Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021

Results The Moran’s I statistic reported for the final model suggests spatially autocorrelated residuals do not affect the OLS results (see Table 2). The initial spatial autocorrela- tion (Moran’s I = .15, p < .001) is largely explained by the demographic and economic characteristics included in the baseline model (Moran’s I = .03, p < .05). This suggests spatial autocorrelation in racial inequality in county poverty rates is a result of spatial heterogeneity. The similarity of one county’s racial inequality with its surrounding level of inequality is explained by the similarity of these counties’ structural factors. The Moran’s I statistic is not significant in the final model (Moran’s I = .02). However, for reassurance, I ran a spatial error model. The further corrected model provides substantively analogous results to those provided by the OLS regression; therefore, I report the results from the simpler OLS model. Variables that did not contribute to understanding racial inequality were excluded from the final model and therefore do not appear in the presented table (metropolitan status, unemployment, the other professional industry, the service industry, educational attain- ment and families with children). The results for the remaining standard poverty covariates

Table 2: Unstandardized OLS Estimates from a Model Explaining County Black-White Inequality in Poverty for the U.S. South, 2000 Variable Sβ .E. Constant - .32 .47 Legacy of slavery .64* .30 Proportion black 1.83* .86 Proportion black squared .79 1.85 Homeownership ratio - .51 .36 Agriculture 3.36*** .88 Manufacturing - 1.16* .55 FIRE 13.02*** 2.91 Single-mother families 5.53** 2.04 Middleclass 3.72*** .92 Number of units 1,336 R-squared .15 Moran's I .02 Note: *p < .05, **p < .01, ***p < .001 Impact of Slavery on Poverty in the Contemporary U.S. South • 727 conform to expectations, with the exception of the sign of the agricultural industry and the middle-class variables. These results suggest racial inequality is greater in counties with larger proportions of the labor force in agriculture and with larger proportions of households with a middle-class income. Given the racially stratified history of Southern agriculture it is not surprising agriculture is positively related to racial inequality. One potential explanation for the positive sign for the middle class is that there is a larger share of whites in the middle class than blacks. The reliance on income to iden- Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 tify the relative size of the middle class limits my ability to test the protective effect of a prominent middle class (Duncan 1999). Cynthia M. Duncan’s work suggests there is more to defining who is a member of the middle class than income; those with middle- class incomes must distinguish themselves and their interests from the elite class in order to have an impact on local inequality. Alternative data sources are required to examine the extent to which Duncan’s findings regarding the role of the middle class in reducing inequality generalize to racial inequality.

Focal Independent Variables The slave concentration coefficient is positive and significant. Racial inequality in poverty is greater in counties where there was a larger concentration of slaves compared to Southern counties with a weaker connection to slavery. Because racial threat and wealth accumulation are both included in the model, these results suggest slavery has an independent impact on black economic net of contemporary factors. A separate analysis shows that once other structural characteristics are controlled, slave concentration is not related to contemporary white poverty rates; however, the corresponding black poverty rates are significantly, positively related to 1860 slave con- centrations. The legacy of slavery is related to contemporary inequality in poverty rates through higher black poverty rates. I find the same results in my sensitivity analysis using aggregated units (not shown). The proportion of the population that is black is significantly and positively related to black disadvantage. Yet, the nonlinear specification for racial threat is not significant. This indicates the relationship is linear rather than decreasingly or increasingly positive as suggested by Blalock (1967). Importantly, black concentration’s relationship is inde- pendent of the positive slave association. This suggests the contemporary concentration of blacks and the historical concentration of slaves represent two separate concepts. Finally, racial inequality in homeownership is not significantly related to inequality in poverty. However, in analyses conducted without the proportion of blacks squared, homeownership’s negative coefficient attains significance at the p = .05 level. Future research should consider the relative importance of a nonlinear term compared to racial inequality in wealth accumulation. My results suggest the nonlinear form of black concentration’s relationship to disadvantage could be explained by greater minority wealth accumulation in places with a large concentration of co-minorities compared to places with smaller concentrations. Black wealth may be an important buffer against discrimination by whites.8 728 • Social Forces 90(3)

To assess the extent to which the null homeownership finding is sensitive to the form of wealth examined, I tested an alternative measure of disadvantage in accumu- lating wealth – land ownership. The extant literature on landownership in the U.S. South has highlighted both its economic and social importance (e.g., Gilbert et al. 2002; Gilbert, Wood and Sharp 2002). However, it is also not significantly related to racial disparities in poverty and does not affect legacy’s relationship with inequality

(not shown).9 Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 Discussion A place’s connection to slavery is important for understanding its contemporary social and economic conditions. Within the U.S. South, blacks are at a greater disadvan- tage relative to whites, with respect to poverty, in places where there is a stronger historical connection to slavery compared to places with a weaker connection to slav- ery. Significantly, this relationship is independent of key contemporary conditions suggested to explain racial inequality, including alternative conceptualizations of the legacy of slavery. My results demonstrate the utility of considering the concentration of slavery in addition to local demographic and economic characteristics, racial inequality in wealth accumulation and racial threat when studying racial inequality in poverty and racial disparities in general. In order for research on the legacy of slavery to move forward, a clear definition of its character is needed. I make three points regarding the treatment of legacy and discuss potential pathways through which the legacy of slavery may affect contem- porary black-white disparities. First, my results support the conclusion that slavery has affected places and that legacy can be understood as a place-based phenomenon. However, my results do not refute the possibility that slavery also affects group char- acteristics. Both understandings of legacy are plausible and are not mutually exclusive; there are multiple legacies of slavery. For example, although I focus on slavery’s impact as a social structural legacy, slavery’s effect on racial inequality can also be conceptual- ized as the cumulative result of blacks’ long history of disadvantage in accumulating wealth. My county-level results suggest wealth does not fully mediate the concentration of slaves’ relationship and therefore does not encompass the entirety of slavery’s legacy. Yet, wealth accumulation patterns may still be important for elucidating an individual- level relationship between slavery and poverty. Additional research connecting both individuals and places to slavery would be needed to disentangle the theoretically independent processes. Second, suggest legacy exists in varying degrees rather than wholly in some places and not at all in others. Therefore, a continuous measure rather than a binary variable indicating slave state or plantation belt status is necessary to accurately reflect the nature of the concept. Third, I demonstrate that racial threat and legacy are distinct concepts. This find- ing has methodological and theoretical implications. Methodologically, it suggests that although the contemporary concentration of blacks has been used as a proxy for the legacy of slavery, it should be reserved for analyzing threat. Theoretically, it Impact of Slavery on Poverty in the Contemporary U.S. South • 729 suggests a need to distinguish between the concepts. Unlike legacy, threat is sensitive to contemporary population changes, such that threat can decrease if the proportion of blacks declines. I propose legacy is distinct from threat because its effect on a place is comparatively constant over time. Its effect may decline as time progresses but that effect will not oscillate; I expect the trend would be monotonic. In-migration is one explanation for a decreasing effect of legacy over time that is consistent with an understanding of slavery’s legacy as a social structure that is embed- Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 ded in local institutions. Since at least the 1960s, but particularly since 1970, migration flows have shifted such that the U.S. South now receives more migrants than it loses. Although primary white in-migration to the South has declined since the 1990s, it remains substantial (Hunt, Hunt and Falk 2008). In addition, starting in the 1990s, Hispanics have immigrated or re-located to Southern states at increasing rates (Lichter and Johnson 2009). The accumulation of in-migrants, potentially with social structural experiences distinct from those had primarily in the South, may reshape the social structure and related institutions suggested to exist by the legacy of slavery literature. The continuous in-flow of new residents does not, by itself, negate the possibility that a social structure based on the assumption of black inferiority explains the relation- ship between slavery and contemporary racial inequality. Such a structure could persist in the face of in-migration particularly if it is built in to local institutions. Indeed, research on racial attitudes suggests convergence among in-migrants towards the aver- age attitudes of their new residence (Glaser and Gilens 1997). Qualitative research also provides some evidence that new residents conform to community norms and expectations (Duncan 1999). Therefore, the in-flow of migrants could have a limited impact on the local structure. More work is needed to understand the extent to which in-migration has weakened the legacy of slavery. Finally, although not directly addressed by my analyses I articulate two potential pathways through which slavery might be linked to the contemporary period. First, slavery might be associated with contemporary black-white inequality because local areas previously dependent on slavery continue to be dependent on exploitative eco- nomic systems. Future research should examine the historical trajectory of in local areas, including contemporary economic structures, to assess the extent to which slavery was replaced with subsequent exploitative economic systems as suggested in previous historical work (Royce 1985; Shifflett 1982). Second, the legacy of slavery may indirectly affect economic disparities through its influence on educational and political institutions. The distribution of resources across and within schools in local areas with high historical concentrations of slaves may be such that blacks receive a lower quality education than whites and are therefore less qualified in the labor market. In addition, differential access to political power could affect local blacks’ control over other aspects of their lives, such as their ability to bargain with employers for higher wages and their voice in the distribution of local resources (see Duncan 1999). Regardless of the mechanism, my results suggest researchers and policy makers concerned with racial inequality in poverty within the South should focus on places with a higher historical concentration of slaves. Controlling for contemporary 730 • Social Forces 90(3) factors related to poverty, I have demonstrated these places have a greater degree of black disadvantage than places that had lower concentrations of slaves. Although policy cannot change the history of local areas, counties with a history of slavery are not doomed to high levels of black-white inequality in poverty. However, my results sug- gest different or additional policies would be needed to address black-white economic disparities in historically high-slave areas. To the extent to which a legacy of slavery affects inequality in poverty through contemporary social structure or an exploitative Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 economic structure, racial inequality can be reduced by dismantling or reconfiguring those structures. This is not to say such a task would be easy, but it would be more feasible than changing history. Research on the mechanisms linking slavery to racial inequality is needed to develop specific recommendations for poverty policy. In addition, research could extend this field by studying legacy’s relationship with racial inequality among social outcomes besides poverty and other minority collectives relative to whites. Poverty is but one outcome that is affected by the legacy of slavery; its relationship, or lack thereof, with other outcomes will enhance our understanding of how legacy operates. Similarly, studying other minority collectives that did not experi- ence slavery as blacks did contributes to an understanding of the extent to which legacy is a generalized conception of racial hierarchy compared to a specifically white-black phenomenon. Finally, research should assess the impact of Northern slavery on con- temporary black-white inequality in Northeastern local areas. Because slavery took a different form in this region compared to in the South (Berlin 1998) legacy’s historical pathway and resultant mechanisms may also differ. Although many questions remain, the current study provides a foundation on which researchers can build to address those questions and advance the study of racial inequality and the legacy of slavery.

Notes 1. Slaves comprised 35 percent of the 1860 southern population. In non-southern counties with a slave population, slaves comprised only 8 percent of the population by 1860. It is important to emphasize that my distinction between the South and non-South does not mean the non-South, particularly the Northeast, was unaffected by the history of slavery. The distinction reflects the different trajectories and therefore consequences of slavery (for a discussion of slavery’s varied history see Berlin 1998). 2. One possible way to distinguish between racial threat and legacy, not explored in the current study, is threat’s relationship with change in the minority population. Threat can be operationalized as either the concentration of blacks or the change in that collective’s population size (Blalock 1956), while legacy, as it is operationalized in the extant literature, is only measured by the concentration of blacks. 3. It is unclear at what point owning a home may be considered a source of wealth. The strictest definition would suggest a house’s worth depends on how much of the mortgage has been paid and the house’s value. Unfortunately, the SF3 Census data do not allow for such a fine-grained definition. However, because I am concerned with racial inequality in wealth, rather than wealth itself, the crude measure employed in this study is likely a conservative estimate. I include the whole range of homeownership, from new homeowners who may never see a return on their investment to homeowners who have Impact of Slavery on Poverty in the Contemporary U.S. South • 731

paid off their mortgages on million-dollar houses, as if the positions were equal. Blacks are overrepresented among those who will lose their homes and own houses of less value compared to whites (Boehm and Schlottmann 2004), which biases my estimate of wealth inequality toward underestimating blacks’ disadvantage relative to whites. 4. SF3 values were used for the black population estimates, except for in Montgomery, AR; Elliott, Jackson, Owsley and Magoffin, KY; Graham County, NC; Unicoi, TN; Armstrong, Brooks, Hamilton, Hansford, Ochiltree, Sterling, Throckmorton and Zapata, TX; and Clay, Tucker and Wirt, WV. For these counties I used the SF1 count of blacks. Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 The population in these counties is small enough to be missed by the census sample used for SF3 data. Whereas the SF1 Census data are based on surveys sent to every identified household (the short form), the SF3 is based on a random sample of households (the long form). Although it is not incorrect to employ the SF1 proportions for all counties, I have chosen to use them only when necessary. I did so because nearly all of the variables in the analysis are based on SF3 data. Sensitivity analyses suggest my results are unaffected by taking this approach. The following counties have no identified black population in either the SF1 or SF3: Kenedy, King, Loving and Terrell, TX. 5. Data on black poverty were not available for Montgomery, AR; Elliott, Jackson and Owsley, KY; Graham, NC; Unicoi, TN; Hamilton, Ochiltree, Sterling, Throckmorton and Zapata, TX; and Tucker and Wirt, WV. There was no reported black population in the SF3 data, which contains information on poverty status, even though a black population was reported in the SF1. It is likely that data swapping, employed by the U.S. Census to maintain confidentiality, explains why the remaining counties in Note 4 did not require black poverty estimation (see chapter 8 of the Summary File 3, 2000 Census of Population and Housing: Technical Documentation). I estimated black poverty for these counties using a series of fitted regressions (for more details see White et al. 2008). Results excluding these counties are substantively equivalent to those observed when including counties with imputed black poverty values. 6. See http://www.ers.usda.gov/Data/RuralUrbanContinuumCodes/. 7. My cluster results are consistent with those from the Horan-Hargis County Longitudinal Template (1995). However, the template excludes county changes from 1990 through 2000 because it had not been updated for the year 2000. In addition, it solely addresses county changes based on the creation or deletion of counties while neglecting boundary changes that result in a significant transfer of land between counties. Therefore, I independently created time-constant units to ensure data refer to the same area across time. Adam Slez, Katherine Curtis and I are in the process of publishing an article detailing the differences between the Horan-Hargis template and the aggregation approach employed in this research. 8. The variance inflation factor values for the final model suggest it is affected by highly interrelated independent variables. I ran sensitivity analyses to examine the extent to which my results are affected by existing multicollinearity. The results are robust. They remain the same when the indicated variables (black concentration and homeownership) are excluded from the model. 9. I measure racial inequality in landownership from 1997 race-specific Agricultural Census data. The variable is constructed in the same way as the homeownership measure. In earlier versions of this article I also examined the role of the racial state in explaining inequality in poverty and slavery’s relationship with inequality in poverty. Following James (1988), I constructed a variable of black disenfranchisement relative to whites using voter registration data. Due to data limitations I was only able to conduct this analysis for a 732 • Social Forces 90(3)

subset of the South, including Alabama, Georgia, Louisiana, North Carolina and South Carolina. These results suggest disparities in voter registration have no effect on slavery’s relationship with inequality and are not significantly related to county variation in racial inequality in poverty.

References

Adeola, Francis O. 1999. “Minority Occupational Structure and Poverty: Empirical Evidence Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 from Louisiana Parishes.” Sociological Spectrum 19(2):163-99. Bartley, Numan V. 1990. The Creation of Modern Georgia. University of Georgia Press. Beggs, John J., Wayne J. Villemez and Ruth Arnold. 1997. “Black Population Concentration and Black-White Inequality: Expanding the Consideration of Place and Space Effects.” Social Forces 76(1):65-91. Berlin, Ira. 1998. Many Thousands Gone: The First Two Centuries of Slavery in North America. Belknap Press of Harvard University Press. Blalock, Jr., Hubert M. 1967. Toward a Theory of Minority Group Relations. Wiley. ______. 1957. “Percent Non-White and Discrimination in the South.” American Sociological Review 22(6):677-82. ______. 1956. “Economic Discrimination and Negro Increase.” American Sociological Review 21(5):584-88. Boehm, Thomas P., and Alan M. Schlottmann. 2004. “The Dynamics of Race, Income, and Homeownership.” Journal of Urban 55(1):113-30. Brown, Cliff. 2000. “The Role of Employers in Split Labor Markets: An Event- Structure Analysis of Racial Conflict and AFL Organizing, 1917-1919.” Social Forces 79(2):653-81. Cohen, Phillip N. 2001. “Race, Class, and Labor Markets: The White and Racial Composition of U.S. Metropolitan Areas.” Social Science Research 30(1):146-69. ______. 1998. “Black Concentration Effect on Black-White and Inequality: Multilevel Analysis for U.S. Metropolitan Areas.” Social Forces 77(1):207-29. Cross, Jr., William E. 2003. “Tracing the Historical Origins of Youth Delinquency and Violence: Myths and Realities about Black Culture.” Journal of Social Issues 59(1):67-82. Downey, Liam. 2006. “Using Geographic Information Systems to Reconceptualize Spatial Relationships and Ecological Context.” American Journal of Sociology 112(2):567-612. Duncan, Cynthia M. 1999. Worlds Apart: Why Poverty Persists in Rural America. Yale University Press. Friedman, Samantha, and Daniel T. Lichter. 1998. “Spatial Inequality and Poverty among American Children.” Population Research and Policy Review 17(2):91-109. Gilbert, Jess, Gwen Sharp and M. Sindy Felin. 2002. “The Loss and Persistence of Black-Owned Farms and Farm-Land: Review and Implications.” Southern Rural Sociology 18(2):1-30. Gilbert, Jess, Spencer Wood and Gwen Sharp. 2002. “Who Owns the Land? Ownership by Race/Ethnicity.” Rural America 17(4):55-62. Glaser, James M., and Martin Gilens. 1997. “Interregional Migration and Political Resocialization: A Study of Racial Attitudes under Pressure.” Public Opinion Quarterly 61(1):72-86. Horan, Patrick M., and Peggy G. Hargis. 1995. “County Longitudinal Template, 1840-1990.” Ann Arbor, MI: Inter-university Consortium for Political and Social Research. DOI: 10.3886/ICPSR06576. Impact of Slavery on Poverty in the Contemporary U.S. South • 733

Hunt, Larry L., Matthew O. Hunt and William W. Falk. 2008. “Who is Headed South? US Migration Trends in Black and White, 1970-2000.” Social Forces 87(1):95-119. Irwin, Michael D. 2007. “Territories of Inequality: An Essay on the Measurement and Analysis of Inequality in Grounded Place Settings.” Pp. 85-109. The Sociology of Spatial Inequality. Linda M. Lobao, Gregory Hooks and Ann R. Tickamyer, editors. State University of New York Albany Press. James, David R. 1988. “The Transformation of the Southern Racial State: Class and Race Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 Determinants of Local-State Structures.” American Sociological Review 53(2):191-208. Kornrich, Sabino. 2009. “Combining Preferences and Processes: An Integrated Approach to Black-White Labor Market Inequality.” American Journal of Sociology 115(1):1-38. Levernier, William. 2003. “An Analysis of Poverty in the American South: How are Metropolitan Areas Different from Nonmetropolitan Areas?” Contemporary Economic Policy 21(3):372-82. Levernier, William, and John White B. 1998. “The Determinants of Poverty in Georgia’s Plantation Belt: Explaining the Differences in Measured Poverty Rates.” American Journal of Economics and Sociology 57(1):47-70. Lichter, Daniel T. 1989. “Race, Employment Hardship, and Inequality in the American Nonmetropolitan South.” American Sociological Review 54(3):436-46. Lichter, Daniel T., and Kenneth M. Johnson. 2009. “Immigrant Gateways and Hispanic Migration to New Destinations.” International Migration Review 43(3):496-518. Lichter, Daniel T., Zhenchao Qian and Martha L. Crowley. 2005. “ among Racial Minorities and Immigrants: Explaining Trends and Differentials.” Social Science Quarterly 86(Supplement):1037-59. Moynihan, Daniel P. 1965. “The Negro Family: The Case for National Action.” Washington, DC: Office of Policy Planning and Research, U.S. Department of Labor. Oliver, Melvin L., and Thomas M. Shapiro. 1995. Black Wealth/White Wealth: A New Perspective on Racial Inequality. Routledge. Patterson, Orlando. 1998. Rituals of Blood: Consequences of Slavery in two American Centuries. Civitas/Counterpoint. Roscigno, Vincent J., and Marino A. Bruce. 1995. “Racial Inequality and Social Control: Historical And Contemporary Patterns in the U.S. South.” Sociological Spectrum 15(3):323-49. Royce, Edward. 1985. “The Origins of Southern Sharecropping: Explaining Social Change.” Current Perspectives in Social Theory 6:279-99. Ruef, Martin, and Ben Fletcher. 2003. “Legacies of American Slavery: Status Attainment among Southern Blacks after Emancipation.” Social Forces 82(2):445-80. Schiller, Bradley R. 2008. The Economics of Poverty and Discrimination. Pearson Education. Shapiro, Thomas M. 2004. The Hidden Cost of being African American: How Wealth Perpetuates Inequality. Oxford University Press. Shifflett, Crandall A. 1982. Patronage and Poverty in the Tobacco South: Louisa County, Virginia, 1860-1900. University of Tennessee Press. Swanson, Louis E., Rosalind P. Harris, Jerry R. Skees and Lionel Williamson. 1995. “African Americans in Southern Rural Regions: The Importance of Legacy.” Pp. 103-18. Blacks in Rural America. James B. Stewart and Joyce E. Allen-Smith, editors. Transaction. Tigges, Leann M., and Deborah M. Tootle. 1993. “ and Racial Competition in Local Labor Markets.” Sociological Quarterly 34(2):279-98. 734 • Social Forces 90(3)

Tolnay, Stewart. 2003. “The African American ‘Great Migration’ and Beyond.” Annual Review of Sociology 29(1):209-32. Tolnay, Stewart, and E.M. Beck. 1995. A Festival of Violence: An Analysis of Southern Lynchings, 1882-1930. University of Illinois Press. Vandiver, Margaret, David Giacopassi and William Lofquist. 2006. “Slavery’s Enduring Legacy: Executions in Modern America.” Journal of Ethnicity in Criminal Justice 4(4):19-36. Voss, Paul R., David D. Long, Roger B. Hammer and Samantha Friedman. 2006. “County Downloaded from https://academic.oup.com/sf/article/90/3/713/2235802 by guest on 29 September 2021 Child Poverty Rates in the US: A Spatial Regression Approach.” Population Research and Policy Review 25(4):369-91. Wilson, William J. 1987. The Truly Disadvantaged: The Inner , the , and Public Policy. University of Chicago Press. White, Katherine J. Curtis, Paul R. Voss and David D. Long. 2008. “Spatial Variation in the Structural Correlates of Child Poverty in the United States.” Paper presented at the annual meeting for the Population Association of America, April 17-19, , LA.