Exogenous Policy,Racial Avoidance,and the Qualified Relevance of Macroeconomic Change to Metropolitan Inequality

Macroeconomic Change and Education

Social scientists’ suspicion that labor market changes were responsible for the educational status of Black residents of central cities increased in the latter quarter of the 20th century. As evidence of this realization, Anyon’s (1997) account of structural change argued that the deterioration of schools in cities like Newark, New Jersey, was “intimately linked to economic transformations of the city—and to federal and state policy as well as to local and national corporate decision making” (p. 156). While there were nuanced accounts that also found fault with the educational policy decisions of school leaders (K. M. Neckerman, 2010), a pervasive understanding emerged within research that “fundamental changes in the society and economy—particularly changes in the social ecology of cities and the structure of urban labor markets—have shaped the development of urban education” (Kantor & Brenzel, 1992, p. 279).

The effects of the labor market transformation were most acute in the large cities of the Midwest and Northeast where Black families were in greatest number (Jargowsky, 1997; Sugrue, 1996; Wilson, 1987, 1996). Residents of those cities saw higher wage manufacturing jobs eliminated by technological advances and globalization, and less-skilled jobs in their vicinity relocated to the urban fringe—both developments, in effect, reducing residents’ employment options and wages (Kain, 2004; Kasarda, 1989; Stoll, 2005). Federal policy reportedly assisted this transformation through transportation projects that subsidized the suburbanization of both jobs and economically able families, leaving families with incomes too low to relocate trapped in the inner city (Wilson, 1996). Subsequently, the number of predominantly Black “ghetto poverty tracts” with poverty concentrations of at least 40% doubled between 1970 and 1990 (Jargowsky, 1997; Wilson, 1996).

For some, the crystallization of these economic factors relegated racial discrimination to a position of less concern. As Dr. Vivian Henderson observed, “it’s as if racism having put Blacks in their economic place stepped aside [italics added] to see changes in technology and changes in the economy destroy that place” (Wilson, 1997b), while Wilson (1978) noted that the importance of race in determining the life chances of Black people had declined, owing to a shift of racial antagonisms from the economic to the sociopolitical realm. Whether this shift had occurred or not, racial marginalization in the sociopolitical context of schooling would seem especially consequential to life outcomes because schools remain society’s primary arena for life- and work-skill development. It would therefore be erroneous to artificially separate the influence of racial marginalization from the structure of the labor markets and the economy.

Nonetheless, sociologists such as William Julius Wilson hypothesized there were more troubling direct educational implications of the economic transformation. He observed that in neighborhoods with high joblessness, “teachers become frustrated and do not teach and children do not learn. A vicious cycle is perpetuated through the family, through the community and through the schools” (Wilson, 1987, p. 57). In his foreword to Anyon’s (1997) text, Wilson (1997a) reiterated that the “cumulative effects of economic and political decisions in the larger urban context have, over time, severely constrained the ability and actions of current actors in central-city schools, including their efforts to achieve meaningful school reform” (p. ix). Despite Wilson’s contention that the impact of racial tension had been relegated to the sociopolitical context in which I contend schools function, accounts of where and how race functions within central-city schools were largely unarticulated in the work of urban researchers. This has left the impression that race relations may have been and continue to be a concern secondary to economic forces, even within educational systems and institutions.

Convergence and the Limits of Macroeconomic Arguments

The marginal attention given to the influence of race relations in research would have been less concerning if a focus on economic change was, by itself, wholly informative about educational stratification. Yet, as I have pointed out elsewhere (Johnson, 2017), the macroeconomic transformation perspective has appeared especially incapable of the task in light of data from several national sources. In the National Assessment of Educational Progress (NAEP), for example, as the proportion of Black individuals residing in areas with poverty concentrations of 40% or higher climbed more than 30% by 1990, their average NAEP proficiency scores surprisingly increased relative to White Americans. In reading, the initial 44 point difference between Black and White performers in 1971 dwindled to 29 points in 1988. In math, inequality in test scores declined from 35 points in 1973 to 27 points in 1990 (Mullis et al., 1991).

Test-score convergence was not merely a case of success at the top of the Black educational distribution. Miller (1995) reports that “fewer than nine-in-ten African American 9-year-olds reached the Simple Arithmetic Facts level in 1978, but virtually all did so in 1990” (p. 55). Similarly, Jencks (1991) finds that from 1971 to 1988, reading among Black students rose in every skill level from basic (82.0%–97.1%) to the advanced (0.3%–1.9%) (p. 70). Throughout the time frame marked by urban decay, Black students’ scores actually increased and test-score gaps decreased at a rate based on which social scientists projected the end of the reading gap in just 25 years and the math gap a few decades later (Jencks & Phillips, 1998).

As NAEP trend assessments consist of repeated cohort sample estimates rather than multiple observations of one sample over time (see Tate, 1997, for a discussion of NAEP estimation), these trend assessments should not be regarded as precise measures of periodic change in the size of the test-score gap, but as a series of independent estimates to gauge the general direction of gaps in successive years. These trends nonetheless mirror others found in data that do not have similar limitations and include other relevant social dimensions. For example, not only had the nation witnessed racial convergence in test performances, but social class inequality (as measured by parents’ educational level) within and across racial groups appears to have declined as well. The National Center for Education Statistics Youth Indicators’ trends in school leaving showed that the dropout rate for Black students fell from 28% in 1970 to 15% in 1988; that it declined more among children whose parents had not completed high school than among children with more-educated parents; and that it had declined more among Black students than among Whites whose parents had the same amount of schooling (Jencks, 1991, pp. 64–67). These convergences within and across parental education levels not only appeared for rates of early school departure, but in students’ average mathematics proficiency scores as well. From 1978 to 1990, for instance, the test scores of Black youth whose parents had not finished high school experienced the largest increase, 23 points, while White Americans only experienced a four-point increase. The test-score gap was not merely converging between Black and White students within the same social class, but also between Black students of lower and White students of higher social classes.

These convergences are made even more remarkable by the precarious socioeconomic status (SES) of Blacks at that time. For instance, the low 6.4% Black unemployment rate that was achieved immediately after the War on Poverty increased to 10% by 1972 and remained in the double digits throughout the years marked by test-score convergence (Jencks, 1993). In fact, the Black unemployment rate peaked at 21.2% during the Reagan administration, and the ratio of Black to White unemployment rose to 2.57 by January 1989. To put this in perspective, Black nationwide unemployment during the Great Recession of 2008 reached a much lower rate of 12.7% and a ratio of 1.63 (Bureau of Labor Statistics, 2015). In contrast to statements that the life chances of Blacks were becoming more closely linked to the emerging economic class structure of society, these test-score trends and economic realities led Jencks (1991) to conclude that social class background mattered less among Blacks than among Whites of that time, and less in relation to Blacks in previous times.

As inharmonious as race and social class trends in education may seem to the macroeconomic perspective, one must acknowledge that Black trends in testing would not include only the poor, nor would the performance trends of the poor only include Black individuals. Advocates of the macroeconomic perspective would correctly point out that the trends of individuals who are socially located at the intersection of race, poverty, and urbanicity are their primary concern. Residential location provides the window through which the test scores of this specialized population become easiest to view. In our largest cities, for example, less than 2% of non-Hispanic Whites were found living in high-poverty areas in 1980 (Wilson, 1987, p. 49), making the estimation of trends between high- and low-poverty neighborhoods a crude but acceptable delineator of race and social class. Using this lens has shown that the macroeconomic perspective remains at odds with NAEP trends. From 1970 to 1990, the trend in math test scores of test takers in disadvantaged urban areas improved significantly, by roughly 15 points at age 9, 20 points at age 13, and 12 points at age 17. These gains reduced test-score disparities between the residents of advantaged and disadvantaged urban areas from 38 to 30 points for 9-year olds, by 22 points for 13-year olds, and by 16 points at age 17 (Mullis et al., 1991).

In sum, Black–White educational disparities converged within and between social classes, in multiple academic subjects, in graduation rates and college enrollments, and between advantaged and disadvantaged neighborhoods—suggesting that the nation had, from 1970 to 1990, the highest achieving (relative to other groups) Black and disadvantaged young people it ever had. ¹ While it is possible that test-score convergence could have been greater if urban labor markets had not constrained efforts to reform inner-city educational institutions, the evidence hardly suggests they encouraged an educational decline. With the inapplicability of economic explanations at the neighborhood and individual level, test-score convergence remains a remarkable event awaiting plausible explanation.

Neighborhood Racial Isolation as an Explanation of Disequilibrium and Convergence

Wilson’s (1997b) view that the macroeconomic transformation had changed institutional ghettos, previously maintained by race, to “jobless ghettos” sustained by the economy was not shared by all. Other scholars may have agreed that changes in urban labor markets had occurred, yet they still believed that racism and discrimination structured opportunity in urban areas (Galster & Hill, 1992; Massey & Denton, 1993; Wacquant, 1997). Wacquant (1997), for example, argued that the notion of the ghetto was diluted and its racial foundation obfuscated by designating it “an urban area of widespread and intense poverty” (p. 341). A racial basis of urban life was highlighted perhaps in greatest detail by Massey and Denton (1993), who observed that extremely high levels of neighborhood racial segregation were realities in most cities after 1970 and, in the largest of them, it was not unusual for Black–White dissimilarity indices to exceed .80 (Massey and Denton, 1993, p. 146). A dissimilarity index of that magnitude meant that 80% of an urban area’s population would need to move to achieve a spatial equilibrium of races. High levels of racial segregation and the lower incomes of Black families, according to Massey and Denton, built concentrated poverty into the structure of Black neighborhoods whether a representative proportion of higher income Black families lived in them or not. Without racially segregated neighborhoods, they convincingly argued, the neighborhoods in which Black families lived would not have become areas of concentrated poverty as macroeconomic changes occurred.

As an early detector of fundamental changes in the distribution of jobs, Kain (1992, 2004) might have also quibbled with the contention that race was not a key part of the urban decline thesis. He suspected that, well into the 1980s, and perhaps as late as 1992 (Kain, 1992, p. 437), residents of the ghetto were paying a premium for housing due to the limits that racial discrimination in suburban housing markets placed on their residential options. This promoted the possibility that the incomes of inner-city Black families had not trapped them behind ghetto walls as much as widespread exclusion from suburban housing markets. It is unknown to what extent this dynamic affected the residential options for families in poverty, given that most markup studies focus on housing prices instead of the rental agreements that poor families are most likely to have. This research void leaves unknown if those with lower incomes were paying more in rent within central cities than they would have been paying in White suburbs, and at which household income level, if any, would residential markups become unrelated to lowered rates of outmigration. ² Without this information, we can only speculate whether the recent movement of lower income families to the suburbs would have begun decades sooner had it not been for racial discrimination.

Nonetheless, to the extent that urban residents were able to overcome economic barriers to relocation, racial ones further reduced their chances of successfully doing so. Even with assistance from the government to settle in White suburbs, few Black families with low incomes were able to step over the suburban color line. In fact, the U.S. Department of Housing and Urban Development’s (HUD) efforts to “disperse” less-advantaged people of color throughout metropolitan areas in the Fair Share Housing and Scattered Site Housing Programs failed (Goetz et al., 2005) or were delayed due to suburban resistance, as seen in the Baltimore site of the tenant-based mobility program, Moving to Opportunity (Goering & Feins, 2003). Far from a situation in which race simply stepped aside and watched labor market changes destroy Black neighborhoods, racial avoidance and exclusion established housing market dynamics that had direct bearing on the concentration of poverty. Racial avoidance and exclusion inhibited the spatial equilibrium that would have otherwise resulted from the outmigration of Black families willing to rent elsewhere and held them in central cities to shoulder the burdens of economic change.

In response to the debilitating influence that structural change had on schooling and reform as Wilson and Anyon posited, Massey and Denton (1993) claimed that racial isolation in neighborhoods fostered oppositional orientations toward schooling, and ultimately, less academic success among Black students. These contentions, however, have not been convincingly supported by quantitative research, in that few studies of neighborhood racial isolation find that it significantly lowered education outcomes (Benson & Borman, 2010; Card & Rothstein, 2005; Johnson, 2010), and recent experimental studies have tended to associate Black students’ educational outcomes with neighborhood SES rather than race (Johnson, 2012; Rickford et al., 2015). Moreover, Massey and Denton’s assertions appear incongruent with education trends because neighborhood rates of Black–White exposure have not varied significantly enough to explain relatively dramatic educational changes. Logan and Stults (2011) have shown that Black–White exposure rates were very low, indexed at 35, and changed by no more than four percentage points from 1950 to 2010. All else being equal, the stagnation of interracial exposure should have held the test-score gap somewhat steady after 1970.

An Unintended Policy Effect: The Sociopolitics of Racial Avoidance

Ironically, political economic accounts of schooling have yet to accommodate the potential of an exogenous political event, such as desegregation, to unhinge the deterministic linkage between macroeconomic change and educational behavior. A consequence of this inattention to desegregation is the lack of realization that its social consequences could reverberate upward throughout levels of the social ecosystem. The greater presence of Black students in desegregating schools, for example, had not merely encouraged an exodus of Whites from those schools or public schools altogether; it also held implications for metropolitan residential patterns, as changing schools often included a residential change for families. To better understand policy influences on residential choice, research has adjusted desegregation effects on family mobility to account for total population growth and made comparisons to other areas without desegregation plans to rule out the possibility that White school enrollment declines were merely a result of their falling birthrate, growth in populations of color, or preexisting mobility trends. These recent studies show that an exodus of White families followed the implementation of desegregation plans large enough to offset a third of a district’s reduction in segregation, a 20% decline within just 2 years of implementation (Reber, 2005). Although this policy reaction could certainly be described as White flight, we must remember that a number of population realities such as the great migration and suburbanization of jobs—factors independent of schools—facilitated White flight. Regarding schools, however, White family mobility decisions are more appropriately defined as racial avoidance, meaning a decision to exit neighborhoods or schools made in response to the growing possibility of interracial contact. Distinguishing between White flight and racial avoidance is supported in research documenting that, in non-southern school systems, roughly “two-thirds of the aversion to desegregation plans was due to the introduction of racially integrated classrooms and associated changes in peer quality” (Boustan, 2012, p. 86).

An exodus of White families from desegregating schools not only resulted in more racially homogeneous schools, but it also transformed the next level of social aggregation, the school district. Coleman (1975) and Clotfelter (1979) noted that larger White district enrollment declines were associated with declines in segregation. Their early observations were only a hint of the transformation to come. While Clotfelter (2004) found that, in 1970, between-district segregation represented less than 4% of all segregation in the most segregated metropolitan areas, the comparable statistic reached 84% in 2000. In other words, racial avoidance has yielded not only segregated schools in contemporary times, but also highly segregated school systems (Johnson, 2014).

Beyond the school system, metropolitan areas were also transformed by desegregation. From 1970 to 1990, for instance, the number of two-parent families with children fell by 1.5 million in central cities and increased by 1.3 million in the suburbs. School quality—not jobs—was among the top two reasons these families gave for leaving inner cities (Briggs, 2005; U.S. Department of Housing and Urban Development, 1997). ³ There are several reasons to believe that racial desegregation and White exposure to Black students in school were the dimensions of school quality instigating Whites’ school and neighborhood selection. First, Clotfelter (2004) has shown that metropolitan areas with the fastest White enrollment declines tended to be those with the highest non-White enrollment and with the greatest opportunity for avoidance (as measured by areas near which Whites could move that had at least 10% fewer non-Whites). Second, White mobility rates were even higher under mandatory desegregation plans. Consequently, exposure to Black students rather than other groups, he found, was associated with the greatest racial avoidance. Third, avoidance of interracial exposure in schools may have remained an important phenomenon in the 1990s (Clotfelter, 2004). Its lingering effects on the spatial organization of metropolitan areas in the North extended in part from the fact that cities, such as Chicago, did not fall under desegregation consent decrees until 1980. In sum, the last decade of research has produced mounting evidence that desegregation had lasting benefits for individuals, but far-reaching consequences for the demography of schools, districts, metropolitan areas, and their neighborhoods.

Racial Avoidance: Macroeconomic Explanations of Metropolitan Inequality Reconsidered

It is in school selection trends animated by racial avoidance that we find several reasons to question the primacy given to macroeconomic explanations of urban life. First, an inseparable companion of racial avoidance was capital flight, that is, higher or two-income families relocating from urban areas to the suburbs. Given their stronger association with capital, dramatic White departures from an area could leave it economically disadvantaged if it did not have a larger-than-average share of the Black middle class. Wilson (1987) might have agreed that this point applies to his analysis of Chicago neighborhoods. He noted that by 1980, at least 12 of the 26 areas with significant growth in poverty concentrations had experienced a “net minus migration” equaling a loss of roughly 185,000 non-Blacks (p. 50). Nearly a decade later, Wilson (1996) argued that changes in racial composition could not have been responsible for the emergence of the new urban poor on account that five of the 14 Chicago neighborhoods that became areas with poverty rates exceeding 40% by 1990 were already majority Black before 1970. But that has left nine other neighborhoods, and the majority of these new areas of concentrated poverty in Wilson’s analysis that had greater racial diversity and therefore economic resources in 1970 than they had in 1990. About these neighborhoods, one could conclude that if White families seeking a certain racial mix in their children’s schools had not left, they would not have become areas of concentrated poverty for Black families even if the spatial and skills mismatch had affected their employment in the way Wilson claims. In other words, concentrated poverty is what might arise when predominantly White schools disappear, not merely “when work disappears.”

Second, White racial avoidance in schooling likely had secondary impacts that undermined the economic base of cities independent of macroeconomic change. Boustan’s (2012) analysis, for example, estimated that desegregation plans resulted in a 6% decline in non-southern urban housing prices and rents relative to neighboring suburbs (p. 86). As her analysis is expanded to include the entire city rather than just urban-suburban bordering neighborhoods, Boustan found that relative city housing prices declined by an additional 14.2 percentage points in cities subject to court-ordered desegregation over the 1970s, compared to a much smaller 2.2% decline in cities that were not (pp. 96–97). Avoidance-induced White outmigration was related to other urban problems, including a 17% reduction in the pace of housing renovation and a reduction in the urban tax base, which she shows decreased by as much as 7.7 percentage points (Boustan, 2012). A tax-base reduction of this magnitude would have had widespread ramifications for the economic well-being of cities and could have amounted to, in the case of schools, a $130 to $162 reduction in tax revenue per pupil if federal financial support to schools had not been increased during this time period (p. 103). Hence, the economic transformation of central cities related to race-related family mobility has far-reaching economic consequences.

Third, racial avoidance in school selection brought about its own variety of highly consequential mismatches. As Clotfelter (2004) pointed out, residing in a racially desirable school district likely entailed longer commutes to work because “good schools” were not located where White bread-winners worked and no longer in the neighborhoods where Black and low-income families lived. From this we can conclude that changes in the spatial organization of White labor attachment (e.g., commute times) can be directly related to school selection, not just the distribution of job growth. As school-choice policies expanded, the educational commute time of Black students likely increased due to the fact that fewer desirable schools were located where they lived. Likewise, the skills mismatch should be more appropriately recast as an “educational mismatch” given that racial avoidance confined failing schools to the geographic areas where Black families lived and where effective schools were essential in order for children living there to defy the odds. In preparing Black youth for a life of low-wage work, ghetto schools maintained by White exclusivity manufactured the skills mismatch within a transformed regional economy. Therefore, the ghetto remained institutional in nature despite a growth in joblessness, being characterized and sustained by the concentration of failing educational institutions (Briggs, 2005).

Fourth, White avoidance of desegregated schools not only encouraged the concentration of poverty among the educationally disadvantaged, but it also assisted the geographic concentration of economic advantage within White areas. While urban-poverty theorists were bringing attention to a growing number of Black high-poverty neighborhoods, the concentration of poverty was actually being outpaced by the geographic concentration of advantage (Massey, 1996). It is true that from 1970 to 1990, the neighborhood sorting index (i.e., measure of economic spatial segregation) increased among all major racial/ethnic categories and increased the most for Blacks. However, the era of relatively higher rates of economic sorting among Black families failed to close the spatial distance between them and Whites, from whom they have been, and today remain, the most separated of all racial/ethnic groups (Reardon et al., 2015). Concentrated affluence has fortified neighborhoods and schools against Black families’ access, providing a barrier of exclusivity for enclaves of White residential privilege that had previously been secured by suburban housing discrimination. Researchers have therefore questioned whether concentrated affluence has been a greater contributor to the test-score gap than has the concentration of poverty (Johnson, 2013; Owens, 2018; Quillian, 2014).

Research into suburban housing premiums, an acceptable indicator of an area’s SES, implies that some of this growth in concentrated affluence is related to desegregation policy. In non-southern metropolitan areas where the central city was required to desegregate, research has shown that suburban housing price premiums increased over the 1970s by an additional 5.8 percentage points in comparison to areas without desegregation orders (Boustan, 2012, p. 96). As modest as this percentage difference may seem, this change in housing premiums was only the portion directly related to desegregation orders, not the total share due to White flight. It is also worth considering that the areas to which desegregating districts were compared were experiencing White demographic shifts of their own, and concomitant increases in suburban housing premiums, owing to the growing presence of children of color within urban public schools. About this, Boustan (2010) estimated that the White share of suburban growth in non-southern regions would have been 20% lower in the absence of Black migration to non-southern cities.

In sum, White racial avoidance has provided a connection between the exogenous event of desegregation and a rising metropolitan spatial disequilibrium of race and class. Contrary to standing narratives of urban change within metropolitan areas, this analysis suggests that desegregation policy intensified a “racism of racial avoidance”—a process of strategic movement within and between neighborhoods and schools, in response to a sociopolitics of race and policy actions, to achieve a clannish coalescence of Whiteness, political autonomy, and residential privilege. Through this power of geographic and group membership exclusion, stronger trends toward convergence were replaced with more uncertain trends in educational inequality and a nearly complete categorical separation of the races in U.S. schools.