Introduction. Competition in the Promised Land: Black Migrants in Northern Cities and Labor Markets Leah Boustan November 2012

Transcription

1 Introduction Competition in the Promised Land: Black Migrants in Northern Cities and Labor Markets Leah Boustan November 2012 **Preliminary and incomplete** In 1938, Gunnar Myrdal, a Swedish economist and eventual Nobel prize winner, began his exploratory tour of the US South, in preparation for his multi-year study of American race relations. His project progressed during the early years of World War II, as the United States began positioning itself as the champion of liberty against the forces of fascism and totalitarianism. Published in 1944 under the title An American Dilemma, the volume argued that a contradiction lay at the heart of the American character. Even as the country adopted the role of defender of democracy, a group of its own citizens were disenfranchised and denied to the right to engage in the pursuit of happiness due to the color of their skin. Myrdal s travels afforded him a first-hand view of this near-caste system. At the time, 75 percent of black Americans lived in the South, the region in which racial restrictions on political and economic freedoms were most severe. Yet, Myrdal emphasized that systematic discrimination against a racial minority was not merely a southern problem but, rather, was a crisis that affected the entire nation. He anticipated that this moral conflict would only truly be confronted and perhaps resolved when residents of the North finally acknowledged the presence of this system of second-class citizenship, and their role in its perpetuation. Myrdal predicted that northern awareness of the Negro problem would be hastened by mass migration of poor black southerners to northern and western cities. Even before his study 1

2 began, 1.6 million blacks had already left the South to settle in industrial cities in the North and West, with especially large waves of out-migration during the years of World War I. In the decades after the report s publication, over 5 million of their fellow black southerners joined them. By 1970, for the first time since before the country s founding, the majority of black residents lived outside of the region where their ancestors had first been imported to work as slaves. The population of the typical northern or western city, where black residents were still a rarity in 1940, increased from 5 percent black in 1940 to 26 percent black by For these black migrants, the attraction of the North was twofold. First, the North held out a promise of social and political equality. In cities like Chicago and Philadelphia, blacks were not required by law to sit in the back of the bus or to queue up to drink from the water fountains marked for colored only. School buildings were not doubled, one for black children and the other for whites, but instead black and white children could attend school together (even if they rarely did). Black residents could express themselves at the ballot box and even elect members of their own community as representatives to city council or to Congress. 1 Just as importantly, the urban North offered a wider array of well-paid industrial jobs than did the primarily rural South and provided higher pay even in jobs like cook, porter or driver that blacks were able to secure in southern cities. The average male worker in the North and West earned 30 percent more than his southern counterpart, while the northern advantage reached 50 percent for the typical black worker. As Myrdal predicted, the economic benefits of 1 Chicago elected its first African-American member of congress in 1929 (Oscar Stanton De Priest). Since that time, De Priest s seat has been held by a string of black politicians, including Harold Washington who went on to become mayor of Chicago. Residents of the Harlem neighborhood of New York elected Adam Clayton Powell, Jr. to the House of Representatives in 1945; Detroit residents followed suit in 1955 by electing Charles Diggs. 2

3 migration to the North and West [were] a tremendous force in the general amelioration of the Negro s position (p. 200). Mass migration from a low-wage region has been an important force in black economic advancement and in the (partial) closure of gaps in black-white economic outcomes. Over the twentieth century, the ratio of black earnings to white earnings increased from 40 percent to 80 percent. Much of this change was concentrated in the decades of the 1940s and the 1960s, two periods of large black out-migration from the South. James Smith and Finis Welch (1989) conclude that mass migration from the low-wage South can account for 20 percent of the blackwhite convergence between 1940 and Upon arrival in the North, black migrants quickly caught up with and even surpassed the earnings of blacks who had been born and reared in the North. The experience of earlier waves of white immigrants to US cities suggested that, within a generation, southern blacks may have been able to close the economic gap with northern whites as well. After all, southern blacks were just the latest group of migrants to settle in northern cities, following waves of Irish and German and then Italian, Polish and Jewish arrivals from abroad. As Oscar Handlin, a prominent early historian of immigration to the United States, reasoned, black migrants would soon follo[w] the general outline of the experience of earlier [white] migrants, whereby immigrants who start out in low-paying jobs quickly moved up the occupational ladder, using their newfound savings to buy their own homes and provide education for their children (1959, p. 120). 2 2 Abramitzky, Boustan and Eriksson (2012) have recently revised the long-standing view that European immigrants started out with lower occupational standing than the native born but were able to successfully catch up within a single generation. Instead, they find that the average permanent immigrant from Europe had already arrived in the US with occupational skills similar to those of the native born and therefore experienced similar rates of occupational upgrading. 3

4 We now know from hindsight that the optimistic predictions of those who, like Myrdal and Handlin, believed in the transformative power of mass migration were not borne out. Although the shift in the black population away from the low-wage South contributed to blackwhite convergence at the national level, the black-white earnings ratio in the North remained unchanged from 1950 to the present. Moreover, the residential isolation of northern blacks in neighborhoods often marked by high rates of poverty increased as the migration got underway, due both to growing segregation between city neighborhoods and to departures of urban white households from the city altogether. In 1940, the typical black household in a northern or western city lived in a neighborhood whose other residents were only one-third black, whereas by 1970 the typical black household lived in a neighborhood that was two-thirds black. A far cry from the expectations of southern migrants hoping to get ahead, the frustration of stagnant economic opportunities and isolated neighborhoods in the North culminated in a wave of urban unrest in the mid-1960s. Why did migration to the same industrial cities that had fostered generations of European migrants, many of whom were unlettered and hailed from places just as rural as the US South, result in such a different outcome for southern blacks? The usual explanation for these very different outcomes has two parts. First, black migrants are said to have arrived in the wrong place at the wrong time. European arrivals in 1880, 1890 or 1900 experienced four or five decades of American manufacturing ascendancy, while black newcomers in 1940 or 1950 enjoyed perhaps a decade or two of plentiful manufacturing jobs, before the American manufacturing sector was eclipsed by global competition. Second, although European Apparent convergence in previous studies are due to familiar biases associated with inferring assimilation patterns from a single cross-section of data. These issues are discussed in more detail in Chapter 2. 4

5 immigrants faced some discrimination in their adoptive home, they were soon able to blend in with the white majority, a feat that most black migrants, marked by the color of their skin, could not achieve. Racial barriers in the labor market and housing market were both more severe and more persistent than the obstacles faced by white ethnics. Black workers were restricted to the lowest-skilled, least pleasant factory positions and were prevented from buying homes or renting in white neighborhoods. In the labor market, each of these factors would represent a labor demand shift changes in industrial composition reduced the demand for low- and semi-skilled work overall, while racial segmentation specifically reduced the demand for black labor in anything but the lowest-paid positions. To these demand-side factors, this book adds the element of labor supply: I argue that the slow black economic advancement in the North can be explained, in part, by competition from the on-going black migration from the South. Moving North was a means of individual improvement for the migrants themselves; yet these newcomers had a negative externality on existing black residents. New migrants expanded the supply of black workers competing for the limited set of jobs open to black applicants and increased the demand for the limited set of apartments available in black neighborhoods. By doubling the black population in the North, migrants also intensified efforts among some white households to defend the boundaries of their segregated city neighborhoods, while encouraging others to leave the cities altogether for towns in the suburban ring. In other words, blacks not only had to contend with falling labor demand in northern labor markets, as manufacturing contracted and racial barriers held firm, but they also did so while confronting the growing labor supply of competing black workers from the South. In urban housing markets, an equal and opposite process was at work. The invisible but all too 5

6 palpable fences dividing white and black neighborhoods remained in place, creating a limited housing supply for black households. New black migrants entered cities in which blacks already faced restricted supplies of housing; their presence created even more demand for the already cramped and expensive apartments. White workers were relatively insulated from the competition of black arrivals in the labor market. In 1950, for example, the typical black migrant had six years of schooling, four fewer years than the typical white worker. Furthermore, black and white men with the same years of education were hired to complete very different tasks, both because of explicit racial barriers and because of the lower quality of each year of schooling received in a segregated black school in the South. Yet, the security that whites may have felt in the labor market did not carry over into the housing market. White households feared that, with continued in-migration, the implicit border between white and black areas would start to shift, changing their racial character of their neighborhoods. Some whites households chose to stay in the city and defend their communities, forming neighborhood associations to limit black entry through overt violence and intimidation or more subtle legal or social pressures. Yet many more households chose the less strident but equally consequential option of leaving the city altogether for newly-built and racially homogeneous neighborhoods in the suburban ring. The possibility of choosing flight over fight was an outcome of the specific historical moment, following World War II, in which black migration reached its apex. In these years, movement to the suburbs was facilitated by new housing construction on the suburban ring and by state and federal road building programs that enabled residents of these bedroom communities to quickly and easily commute by car to jobs in the central city. 6

7 Black in-migration increased the racial diversity of certain parts of the city, including predominately white neighborhoods near the expanding black ghetto. Yet many white households in the city lived in white enclaves at the greatest distance from the burgeoning black neighborhoods. In 1940, for example, the typical white household in the central city lived in a neighborhood that was 98 percent white. Black arrivals influenced the location choices of these protected white households as well by shifting the racial composition and, more importantly the income distribution of the city electorate. In choosing whether or not to move, households are concerned not only with the changing character of their immediate neighborhood but also with aspects of their municipality and local school district. Cities with more residents in poverty tend to levy higher property tax rates and to increase spending on services like police protection. This new mix of higher tax rates and higher public spending encouraged some residents to prefer a suburban location. Initially, the civic concerns that contributed to white flight were fiscal in nature. This focus changed to education with the court-ordered desegregation of many urban school districts in the North and West in the 1970s. Before desegregation, local schools reflected the demographic composition of the surrounding neighborhood, meaning that the typical white student in a northern or western city attended a school with predominately white peers. In the early 1970s, the Supreme Court ruled that school districts outside of the South could be obligated to counteract racial segregation even if the presence of de facto white schools and black schools was not due to an explicit segregationist policy. As a result, many students were reassigned to schools outside of their immediate neighborhood and were exposed to mixed-race classrooms for the first time. White households responded to these school desegregation orders by further deserting central cities. 7

8 The book makes a number of contributions to our understanding of the consequences of the Great Black Migration for the northern economy. The first set of findings pertain to the labor market. First, I show that migrants themselves benefited from the move to industrial cities, gaining around $8 billion a year (in 2000 dollars). This figure is based on an improved estimate of the economic return to migration based on a comparison of pairs of brothers in 1940, one of whom moved to the North, while the other remained in the South. Second, I find that, through competition for a limited set of jobs open to black workers, black migrants lowered the wages of existing black workers in the North by an aggregate $2.5 billion a year. I propose a new framework that I propose for estimating the effect of migrant arrivals on black and white workers in the North and West. This method also reveals that even similarly-skilled whites and blacks were not very substitutable inputs into northern production in the mid-twentieth century, a fact that is consistent with, and helps to quantify, the degree of racial segmentation otherwise observed in the northern labor market. Third, I use the style of counterfactual analysis typical of economic historians to ask how much faster would black-white earnings convergence have been in northern and western cities if not for the arrival of large numbers of black migrants from the South? I show that there would have been somewhat more economic convergence in the North if the migration had slowed but that blacks would still not have achieved economic parity with whites by 1970, even absent competition with migrants. The second set of findings pertain to the topics of cities and urban space. I provide the first convincing causal evidence that black migration contributed to white suburbanization and the growing residential segregation between cities and suburbs. I find that each black arrival to a central city led to more than two white departures, so that black in-migration ultimately contributed to urban population decline. The typical city received 50,000 new black residents 8

9 from 1940 to 1970; I estimate that an inflow of this magnitude would reduce urban population by 16 percent, an effect as large as building one new highway through the city center. Results in this section are based on aggregate data, such as population flows and housing prices, for the universe of 70 large and mid-sized cities in the North and West. This focus on the typical city complements the many historical case studies of neighborhood change in particular cities in a number of ways. First, I demonstrate that the pattern by which black arrivals encourage white departures is widespread, afflicting not only paradigmatic cities like Chicago and Detroit but occurring in cities of all sizes and in all regions. Second, I can tease apart the effects of in-migration from the role of other factors, like economic growth and industrial change, that tend to coincide with (and perhaps attract) migrant arrivals. Third, my focus on aggregate outcomes like population flows and housing price uncovers patterns that can be missed in more standard data sources, like newspaper reports or personal accounts. In particular, vocal white efforts to defend their neighborhoods from black infiltration are easily captured in individual records; yet, quiet relocations of households from city to suburb only become apparent when widening one s lens to the city level. Fourth, I emphasize that, beyond concerns about the racial composition of their immediate neighborhoods, white households worried about and responded to the racial and income composition of the city as a whole. This new attention to city-wide dynamics is made possible by the availability of housing price data down at the block level. This price data allows me to compare trends in similar neighborhoods on either side of the city-suburban border as the population and school policy changes on the city side of the border. The possibility that new black migrants may have harmed existing black residents in the North has been suggested previously by two prominent sociologists, Stanley Lieberson and William Julius Wilson. Both scholars were interested in explaining how blacks failed to get 9

10 ahead in the same cities that had nurtured white immigrant groups just a generation before. They argued that, ironically, white immigrants benefited from the tightening of the border in the Immigration Act of 1924; blacks, in contrast, suffered from the fact that the Mason-Dixon line, the informal border between the North and the South, could not be closed to new in-migrants. Lieberson and Wilson propose a number of channels by which new arrivals can disrupt a nascent immigrant community. First, although small numbers of an ethnic or racial group may go unnoticed, sizable numbers of newcomers raise the level of ethnic and/or racial consciousness on the part of others, often negatively, encouraging the dominant group to redouble their efforts to protect their economic privilege (Lieberson, 1980, p. 380). In my context, the expanding geographic boundaries and growing salience of black enclaves can help explain the phenomenon of white flight. Other mechanisms pertain most readily to the labor market. For example, a continual flow of rural newcomers will reduce the average skill level in an immigrant community, thereby reinforc[ing] stereotypes and negative dispositions that affect all members of the group (p. 380). This type of statistical discrimination will affect existing residents if employers cannot or do not bother to distinguish between earlier and later arrivals in a migrant flow. In addition, barriers that block entry into certain occupations will ensure that members of a racial or ethnic group tend to compete most heavily with each other in the labor market, a hypothesis that I very much share (although Lieberson s model of the labor market as a job queue is quite different from the standard economic framework adopted here). Although neither scholar provides concrete evidence linking migrant numbers to poor black outcomes, Wilson goes so far as to say that the flow of migrants [may be] the most important single contributor to the varying rates of urban racial and ethnic progress in the twentieth-century United States (Wilson, 1987, p. 33). 10

11 The process by which recent arrivals compete with previous migrants in the labor and housing markets is not unique to the black experience. Jewish immigrants, who were initially clustered in the garment industry, welcomed their countrymen but also worried about overcrowding in their occupational niche (Kahan, 1978; Aldrich and Waldinger, 1990). In more recent years, the swell of migrants from Mexico and Central America generated competition and lowered wages for other Hispanic immigrants, who hold a similar set of jobs in gardening, housekeeping, construction and restaurant work, but had little effect on the economic standing of the native-born (Ottaviano and Peri, 2006). New waves of immigrants compete with their fellow countrymen because they arrive with similar skills and settle in the same neighborhoods. However, the force of this intra-group competition will vary across ethnic groups. The stronger the discriminatory barriers that members of the immigrant community face in the labor and housing market, the more difficult it is for existing migrants to respond to this competitive pressure by switching occupations or by moving out of the old neighborhood, and therefore the more concentrated the force of this competition will be. The potential for southern black inflows to threaten the economic standing of existing black residents was noted as early as the 1920s, when W.E.B. Du Bois cautioned that a great reservoir of [southern black] labor could reduce black waves and generate tension between blacks and whites over residential space. This sense of rivalry can explain the ambivalence with which blacks in the North greeted subsequent arrivals; as James Grossman has put it, the existing black community seemed to welcome [newcomers from the South] [but] they didn t seem to open-arm welcome them (1991, p. 139). Reflecting on his own experience as a new arrival in the North, one migrant remembers initially [receiving] a chilly reception from many longtime 11

12 black residents who feared the newly arrived blacks would jeopardize their tenuous position (quoted in Trotter, 1985, p. 115). Migrant inflows were not only a source of competition for existing black residents; migrant arrivals also served as a customer base, audience or constituency for some (typically well-educated) blacks in the North. The expanding pool of applicants for factory positions hurt the job prospects of the existing black workforce, but the same crowds, when patronizing black churches, entertainment venues and business, could generate a livelihood for black preachers, musicians, professionals, and politicians. In general, those selling services to the black community, whether they be religious, legal, or political in nature, had the most to gain from the growing numbers. Often these service providers were among the most educated members of the community and the most likely to express their views in print. Perhaps as a result, historians have tended to emphasize the positive consequences of the migration. James Gregory (2005) and Thomas Sugrue (2008), for example, have recently highlighted the role that northern migrants played in the birth of the Civil Rights movement, while Kusmer (XX) lauds the creation of distinctive black institutions, including black newspapers, self-help societies and political organizations, in US cities following mass in-migration (p. 458, 471). I should note that the emphasis in this book on the competitive pressure that migrants exerted in northern labor and housing markets is quite distinct from the once-common argument that migrants imported a maladaptive southern culture to the North. In this view, black migrants harmed northern black communities through a process of contagion rather than one of competition. This idea was first voiced by black social reformer Sadie T. Mosell in the 1920s and echoed by eminent sociologist E. Franklin Frazier in the 1930s, who wrote that masses of ignorant, uncouth and impoverished migrants changed the whole structure of the Negro 12

13 community (1939, p. XX). Gilbert Osofsky (1966) picked up this theme in his account of the making (and unmaking) of Harlem. Most recently, this idea has been revived by Nicholas Lemann who declared that black sharecropper society was the equivalent of big-city ghetto society today in many ways; migrants, therefore, contributed to inner-city problems by transplanting their culture of out-of-wedlock childbearing, limited education and casual violence (Lemann, 1991, p. 31). Yet the notion that southern migrants slowed economic progress by introducing a widely-emulated, yet counter-productive, set of social norms (in modern parlance, by generating negative peer effects ) is not consistent with the fact that black southern migrants out-earned northern-born blacks, despite having fewer years of formal schooling, a fact that has been well-documented by others and that I review in detail in Chapter 2. My argument unfolds over five chapters. The first chapter places the black migration to the North into its southern context. I demonstrate that mass migration to northern cities continued a trend of rising mobility rates within the South in the late nineteenth century. The high cost of being the first migrant to a new destination dampened black inter-regional migration before World War I, despite large wage gaps between the North and South. Mass migration to the North began in the 1910s, spurred by a triple shock to the regional wage gap. Northern wages rose with heightened labor demand in industrial cities during World War I and a temporary freeze on immigration from Europe, the typical labor supply in northern factories. At the same time, southern wages fell as the boll weevil, an agricultural pest, destroyed much of the cotton crop. Once the migration began, its numbers swelled rapidly, rising through the 1940s and 1950s, peaking and then falling thereafter. The remainder of the book takes place in northern cities, the receiving areas for this southern migrant flow. Chapter 2 profiles these agents of change: who were these black migrants 13

14 and how did they fare in their new home? I review the extensive literature and provide some new evidence on the questions of migrant selection, migrant assimilation and the returns to migration in the Great Black Migration. Previous work has shown that blacks who left the South were more educated than those who remained, particularly in the first years of the migration. Perhaps due to this positive selection, migrants were quickly able to out-earn northern-born blacks in the northern economy. By moving from the low-wage South to the prosperous North, black migrants earned, on average, a percent return after World War II and even more in the 1910s and 1920s. I use novel Census linking techniques to compare the family background of migrants who moved to the North in the 1920s and 1930s and non-migrants who remained in the South. I find that migrants were no more likely than non-migrants to have been raised by literate parents or by fathers with high-status occupations. That migrants attended more years of school than nonmigrants despite hailing from similar backgrounds suggests that they may have received additional education in anticipation of moving North. Census links also allow me to contrast pairs of brothers, one of whom moved to the North while the other remained in the South. Consistent with the neutral selection of the migrant flow, the return to migration among brother pairs, which controls for selection at the household level, is similar to the return to migration estimated in the full population. The fact that southern migrants appear to out-earn northern-born blacks within a few years of arrival has been put forward as evidence that southern migrants were positively selected on unobservable dimensions, such as motivation and perseverance. Yet, I demonstrate that the rate of migrant assimilation in the labor market has been overstated by methods that compare recent arrivals to the long-standing migrant pool at a point in time. This cross-sectional technique 14

15 will overstate the rate of assimilation when either the initial (positive) selection of migrant arrivals falls over time or when migrants who end up returning to the South are negatively selected (or both). I instead follow two cohorts of migrant arrivals by combining Census data with the Racial Attitudes in 15 Cities survey conducted in the 1960s and find XX. Chapter 3 asks how this migrant flow, distinguished by its sheer number rather than by its particular talent, affected the existing black workforce in the North. I present a framework to estimate the effect of southern migrant arrivals on existing workers of both races. This method, adopted from the economics of immigration literature, divides the receiving economy (here: the North and West) into skill groups by education and experience level. Each skill cell receives a different labor supply shift of black and white workers due to southern migration in the decades between 1940 and If blacks and whites are used interchangeably in the labor market, one would expect that black migrants would affect the wages of similarly-skilled white and black workers in the North to the same degree. Instead, I find that black arrivals lowered the wages of existing black workers but had little effect on similarly-skilled whites. This pattern implies that men with similar education levels and years of work experience were used differently in the production process on the basis of race. The observed racial stratification in the labor market is consistent with and provides quantitative precision to historical accounts of racial barriers whereby blacks were confined to the least desirable factory positions and were unable to ascend into supervisory roles or to enter occupations governed by crafts unions. Racial divisions in the northern labor market also reflect differences in the quality of schools attended by the average black and white worker, and thus in the true level of skill embodied in each reported year of education. 15

16 Chapter 3 concludes by documenting that, if not for the large migration flow after 1940, blacks would have experienced some economic convergence with northern whites, although would not have reached complete parity. In this way, Wilson may be right that open migration from the South represents the most important single contributor to slow black economic advancement in the North but some explanatory power remains for demand-side forces like the decline in the manufacturing sector. Chapters 4 and 5 turn to the effects of black migration on city neighborhoods and on the racial segregation of urban space. In 1940, black migrants would have arrived into northern metropolitan areas in which 55 percent of the white population lived in the city center. By 1970, only 30 percent of white metropolitan population remained. Over the same decades, the dissimilarity index, a measure of residential segregation that indicates the share of households in the city that would need to switch neighborhoods in order for each neighborhood to reflect the racial composition of the city as a whole, rose from moderate (dissimilarity = 0.55) to severe (dissimilarity = 0.80). Recent work has shown that blacks who live in segregated metropolitan areas have lower educational attainment and earnings than those who live in more integrated areas. Segregation occurs both across neighborhoods in the city and between the city and the suburb. Using a unique dataset of individual Census records attached to neighborhood characteristics in 1970, I demonstrate that living in a majority-black neighborhood reduces high school graduation and increases the likelihood of living in a female-headed household, even after controlling for the median income and poverty rate of the neighborhood. Segregation that occurs across jurisdiction lines has the added cost of weakening the municipal tax base. Blacks arrived in metropolitan areas in which the processes of suburbanization were already underway. Rising incomes, new housing construction after World War II, and state and 16

17 federal road building combined to encourage households to live at greater distance from the city center. However, in making location decisions, white households also responded to the arrival of black migrants, whose presence as (potential) neighbors, as members of the city electorate and as users of the public schools reduced the attraction of city living. I show that, in the universe of 70 large and mid-sized northern cities, each black arrival is correlated with two white departures. However, in-migration to a city rises with economic prosperity; economic fortunes in the North could be a third factor that prompt both black in-migration and white moves to the suburban ring. Using the logic of instrumental variables analysis, I suggest that cities received migrant in-flows unrelated to their own fortunes when their traditional southern sending states were undergoing economic change, particularly during episodes of agricultural mechanization. The white response to black arrivals driven by southern change is equally strong. By my estimate, the typical city would have experienced a 16 percent decline in urban population due to this migration-induced white flight, an out-flow equivalent in size to the estimated response to one new interstate highway. The majority of white households in the central city lived in predominately white neighborhoods far from a black enclave. Chapter 5 analyzes the motivates of these households, many of whom left the central city despite experiencing few local interactions with black neighbors. I suggest that, for these households, flight was prompted by changes to city finances and local public goods as the municipal electorate and enrollment in city schools because increasingly black, and thus increasingly poor, over time. The first part of the chapter argues that we can learn about the demand for suburban residence by studying the relative prices of city and suburban housing. In general, housing prices provide useful information about the value of location-specific goods, such as ocean views, that are implicitly traded through the housing 17

18 market. By this logic, higher housing prices for housing units with a particular attribute (say, a suburban location) signals that the value of this feature to the typical homeowner. I then introduce an original dataset of housing prices from the Census of Housing collected along 100 municipal borders that divide cities from suburbs. Residents on either side of these borders live in predominately white neighborhoods whose racial composition changes little as central cities receive large inflows of black migrants. The housing stock in these adjacent neighborhoods was also built at a similar time and at a similar quality level. Yet, as in-migration increased the black population share in the central city, the housing prices on the city side of the border fell relative to their suburban neighbors. This decline in housing prices cannot be attributed to local factors, given that the neighborhoods on either side of the border were so uniform. Instead, before 1970, this willingness to pay via the housing market to avoid a racially-diverse municipality can be entirely explained by the correlation between changes in a city s racial composition and its poverty rate. Municipalities with more impoverished residents tend to have higher property tax rates and more spending per capita on non-educational services, two local policies that residents seek to avoid. After 1970, with the advent of court-ordered desegregation in northern and western cities, race itself began playing a role in this civic form of white flight. In cities that faced court-orders to desegregate, residents were no longer assured that their children could attend local public schools whose racial composition reflect the demographics of the surrounding area. As a result, some white households that had remained in the central city up to that point moved to the suburbs, a pattern that is reflected in falling housing prices on the city side of borders in districts subject to desegregation. By the 1970s, the black migrant flow out of the South slowed substantially and, in the subsequent three decades, blacks have experienced net in-migration to the South. In the 18

19 postscript, I discuss this population reversal and the legacy of the earlier migration for black economic progress today. In moving southward since the 1980s, blacks joined a national movement toward Sunbelt cities. As the South finally caught up with the rest of the country in economic terms for the first time since the Civil War, southern cities now boast wages on par with their northern counterparts and offer their residents far lower housing prices. However, black migration to the South may have another, more racially-specific meaning, finally severing the long-standing association in the black community between economic progress and moving North. Blacks in the North now face a larger earnings gap relative to whites than do blacks in the South and northern cities remain far more segregated than their southern counterparts. The American dilemma that Myrdal once stressed could not considered a southern problem alone is now more firmly rooted in the old industrial cities of the North than in the formerly agricultural region of the South. 19