The law largely has overlooked one of the most important sociological developments of the last half century: a sharp decline in residential segregation. In 1970, 80 percent of African Americans would have had to switch neighborhoods for blacks to be spread evenly across the typical metropolitan area. By 2010, this proportion was down to 55 percent and was continuing to fall. Bringing this striking trend (and its causes) to the attention of the legal literature is my initial aim in this Article.

My more fundamental goal, though, is to explore what desegregation means for the three bodies of civil rights law—housing discrimination, vote dilution, and school segregation—to which it is tied most closely. I first explain how all three bodies historically relied on segregation. Its perpetuation by housing practices led to disparate impact liability under the Fair Housing Act. It meant that minority groups were “geographically compact,” as required by the Voting Rights Act. And it contributed to the racially separated schools from which segregative intent was inferred in Brown and its progeny.

I then argue that all of these doctrines are disrupted by desegregation. Fair Housing Act plaintiffs cannot win certain disparate impact suits if residential patterns are stably integrated. Nor can claimants under the Voting Rights Act satisfy the statute’s geographic compactness requirement. And desegregating homes usually result in desegregating schools, which in turn make illicit intent difficult to infer.

Lastly, I offer some tentative thoughts about civil rights law in a less racially separated America. I am most optimistic about the Fair Housing Act. “Integrated and balanced living patterns” are among the statute’s aspirations, and it increasingly is achieving them. Conversely, I am most pessimistic about the Voting Rights Act. One of its objectives is minority representation, which is threatened when min­orities are politically distinctive but spatially dispersed. And a mixed verdict seems in order for school desegregation law. Rising residential integration eventually should produce rising school integration. But it has not done so yet, and even when it does, this improvement may not reach schools’ other racial imbalances.


Two generations ago, in the wake of rioting that scarred dozens of American cities, the Kerner Commission issued its landmark report on urban unrest.1 The report warned darkly of high and rising racial segregation. “To continue present policies,” it intoned, “is to make permanent the division of our country into two societies: one, largely Negro and poor, located in the central cities; the other, predominantly white and affluent, located in the suburbs.”2 One generation ago, a pair of prominent sociologists, Professors Douglas Massey and Nancy Denton, penned another highly influential work on racial separation. American Apartheid3 traced the ways in which public policy produced what the authors termed “hypersegregation,”4 and argued that it was “the key structural factor[ ] responsible for the perpetuation of black poverty.”5

In the legal academy, the conventional wisdom is that little has changed since the Kerner Report and American Apartheid. The drafter of the preeminent treatise on housing discrimination law asserts that “the United States continues to be characterized by high levels of racial segregation.”6 Another housing expert comments that “the failure to stem racial residential segregation has helped it to deepen, widen, and become seemingly intractable.”7 A recent amicus brief signed by dozens of housing scholars declares that “[r]esidential racial segregation across the United States remains pervasive.”8 Summing up the literature, Professor Michael Maly observes, “The volume of research on the extent of segregation . . . makes it difficult to believe that integrated neighborhoods even exist.”9

But the conventional wisdom is wrong. In fact, a great deal has changed over the last two generations—so much that sociologists are now churning out works with titles like The Waning of American Apartheid?10 and The End of the Segregated Century.11 Take the most common measure of segregation, which represents the share of group members who would have to switch neighborhoods for the group to be spread evenly across a metropolitan area. This metric peaked at about 80 percent for African Americans in 1970. But it has since sunk to roughly 55 percent, the same value, more or less, as in 1910.12 Or consider another popular index of segregation, which captures the makeup of the community of the typical group member. In 1970, the average black lived in a neighborhood that was about 60 percent more black than her metropolitan area as a whole. But this figure has since dropped to roughly 30 percent, or approximately the same level as in 1920.13 Almost all of the rise in segregation that took place during the twentieth century thus has been reversed.

What accounts for this impressive (and underappreciated) trend? One factor is the decline in housing discrimination by both public and private parties. Overtly segregative governmental policies are now rare,14 and according to a series of studies by the Department of Housing and Urban Development (HUD), discriminatory acts by real estate professionals have fallen in frequency as well.15 Another explanation is the increased willingness of whites to live in integrated areas. In 1976, for instance, only 50 percent of Detroit-area whites said they would consider moving to a community that was one-fifth black.16 By 2004, this proportion had surged to 79 percent.17 And still another cause is the spectacular population growth of nonblack minorities, in particular Hispanics and Asian Americans. These groups now seem to serve as “buffers” that enable whites and blacks to live together in durably diverse neighborhoods.18

My initial aim in this Article, then, is to bring to the legal literature’s attention the recent sociological findings about the shifts in, and sources of, segregation. It is time for the stylized facts that have long guided thinking about these topics to be updated. My more fundamental goal, though, is to explore what the decline in segregation means for the law itself. At least three bodies of civil rights doctrine—involving housing discrimination, vote dilution, and school segregation—are closely connected to racial groups’ residential patterns. For each of these areas, I show how the existence of segregation historically has supported the imposition of liability and aggressive remedies. I then argue that desegregation is reshaping the legal landscape and making key doctrinal elements harder to establish. Lastly, I offer some tentative thoughts about the role of civil rights law in a less racially separated America.

Start with housing discrimination, which is banned at the federal level by the Fair Housing Act19 (FHA). The FHA is tied to segregation in several ways. First, the Supreme Court has held repeatedly that plaintiffs have statutory standing if they live in areas that are segregated or in danger of becoming segregated.20 The deprivation of the “social and professional benefits of living in an integrated society” is a cognizable injury.21 Second, segregated residential patterns are useful evidence in FHA cases brought pursuant to a disparate treatment theory. They help to demonstrate the discriminatory intent of, say, housing authorities that limit low-income projects to minority-heavy areas. And third, as the Court recently confirmed,22 one type of disparate impact claim available under the FHA is that certain practices “have the effect of perpetuating housing segregation in a community.”23 Both actual segregation levels and the levels that would have arisen but for the challenged practices are crucial to such a claim.

All of these aspects of FHA doctrine are destabilized by desegregation. For example, plaintiffs do not have standing (at least on this basis) if they live in areas that are integrated and likely to remain so. They do not suffer the harm of segregated living recognized by the FHA. Similarly, it is more difficult to establish discriminatory intent in the absence of segregated residential patterns. Without them, plaintiffs cannot benefit from the presumption that parties intend the foreseeable consequences of their actions.24 And segregative impact may not even be a viable theory of liability in a stably integrated region. It founders on both the lack of existing segregation and the improbability of demographic change.

Next take racial vote dilution, which refers to policies that diminish minorities’ electoral influence without disenfranchising them outright, and which is prohibited federally by the Voting Rights Act25 (VRA). In a key decision construing the VRA’s core provision, the Court held that in order to state a claim, minority populations must be “geographically compact,26 and there must be racial polarization in voting.27 Geographic compactness is almost a synonym for geographic segregation. The criterion is satisfied only by minority groups that are densely concentrated in discrete areas. Racial polarization is related to segregation as well, only methodologically rather than substantively. It is easier to estimate the share of each racial group that supports a given candidate if there exist precincts occupied almost exclusively by each group’s members. These “homogeneous precincts” make the analysis more tractable.28

Again, desegregation unsettles the doctrine. If minority populations are residentially integrated, then they cannot comply with the compactness requirement imposed by the Court, meaning that there cannot be liability under the VRA. If a jurisdiction nevertheless encloses a dispersed minority group within a single district, then the district probably violates the constitutional ban on racial gerrymandering.29 Race is the only justification for this kind of constituency, but it is not a permissible one. And even if a sufficiently compact majority-minority district can be drawn in a desegregated area, plaintiffs are unlikely to be able to show that voting is racially polarized. Homogeneous precinct analysis breaks down when most precincts are racially heterogeneous, and even regressions become unreliable when two (or more) racial groups coexist throughout a region.30

Last, consider school segregation, which the Court forbade in perhaps the most celebrated decision in its history.31 School enrollments are linked to residential patterns because of the American norm of neighborhood schools. Children tend to attend schools located near their homes, thus reproducing at the school level the racial makeup of local housing. However, the correlation between residential and school segregation is imperfect. The latter also is influenced by parents’ decisions to enroll their students in private schools, as well as by an array of school district policies. Some of these policies are integrative (and often adopted due to court order): busing, magnet schools, attendance zone adjustment, and the like. Other policies, such as vouchers and charter schools, usually are enacted for nonracial reasons.

Because school segregation is a function of residential segregation and other factors, its trajectory since Brown v Board of Education of Topeka32 has not been a steady descent. Instead, it plummeted in the late 1960s and 1970s, at a much faster rate than residential segregation, as courts ordered far-reaching integrative policies in hundreds of school districts.33 But since the late 1980s, it has remained roughly constant.34 The continuing decline in residential segregation has exerted a downward pressure on school segregation, but this effect has been offset by the release of many school districts from judicial supervision.35 At present, thanks to the removal of most court-mandated remedies, the connection between residential and school segregation is the strongest it has been in decades.36 Going forward, this means that trends in the two metrics should be similar.

Doctrinally, then, residential segregation plays a role in school desegregation litigation to the (substantial) extent that it determines school enrollments. At the liability stage, racially uneven enrollments caused by racially uneven residential patterns support an inference of segregative intent on the part of the school district. Uneven residential patterns also make it more likely that policies like attendance zone demarcation and new school construction will have a segregative impact, from which an illicit motive can be inferred as well. After liability has been imposed, courts presume that enrollment imbalances are “vestiges” of the original constitutional violation that make it improper for judicial supervision to be lifted.37 Since these imbalances often are the result of residential segregation, it often prevents school districts from attaining unitary status.

Once again, desegregation complicates the picture. At the liability stage, it is more difficult for plaintiffs to establish segregative intent if school enrollments, like the residential patterns that help drive them, are integrating. There still may be an improper motive in this scenario, but it is harder to discern if it does not manifest itself in racially skewed student bodies. Likewise, when a district requests to be released from court oversight, its claim is more likely to succeed if its schools are desegregating thanks to the ongoing residential trend. School enrollment statistics are vital evidence in any unitary status proceeding, and the better they look, the better the district’s odds of terminating the litigation.

So what might we conclude about the state of civil rights law in an America in which racial and spatial divisions are (gradually) mending? I would deliver a mixed verdict. On the one hand, some of the evils the law has long fought are fading, which is cause for celebration. One of the FHA’s aspirations, in particular, is the creation of “truly integrated and balanced living patterns,” as its chief Senate sponsor put it.38 We certainly are not there yet, but this goal’s achievement is no longer wholly fanciful. Similarly, even though it is invidious intent that Brown and its progeny proscribe, the cases still envision a future “without a ‘white’ school and a ‘Negro’ school, but just schools.”39 Progress toward school integration has stalled since the late 1980s, but it is likely to resume now that residential patterns and school enrollments are so tightly coupled.

On the other hand, segregation is not the only ill that civil rights law tries to cure, and its improvement does not mean that other problems have been solved. For instance, both the FHA and the cases from Brown onward are deeply concerned about discrimination too—the adverse treatment of real estate customers and schoolchildren because of their race, irrespective of the segregative consequences. True, discrimination is one of the most potent drivers of segregation. But discrimination also can occur in a more integrated society, and the law needs to remain wary of it even as segregation continues to decline.

Even more worryingly, the VRA seeks to secure representation for minorities, but this aim is directly threatened by desegregation. To win districts in which they can elect their preferred candidates, minorities need to prove geographic compactness and voting polarization—both daunting tasks if they are residentially integrated. Fortunately, these hurdles are the product of the Court’s case law rather than the statute itself, and so could be lifted without legislative intervention. The Court could drop the compactness requirement that it conjured out of thin air. It could allow nonelectoral evidence, survey results in particular, to be used to establish polarization. And most promisingly, it could embrace remedies other than single-member districts, thus enabling even dispersed minorities to be represented.

The Article unfolds as follows. First, in Part I, I discuss the sociological literature on racial segregation. I cover definitions of segregation, its trends for various racial groups, and the factors that cause it. Then, in Parts II through IV, I analyze the implications of declining segregation for the three bodies of civil rights law to which it is most relevant: the Fair Housing Act, the Voting Rights Act, and school desegregation doctrine. For each area, I show how it historically has relied on the existence of segregation, how it is challenged by greater residential integration, and how it might be rethought in a less racially separated environment.

One last point before beginning: While the lessening of black-white segregation is striking, not all the news here is good. For one thing, black-white segregation has not fallen at the same rate throughout the country. In numerous metropolitan areas, especially in the Midwest and Northeast, it remains stubbornly high.40 In addition, segregation scores for other minorities, namely Hispanics and Asian Americans, have not declined in recent years. Instead, they mostly have held steady, albeit at lower levels and despite these groups’ rising populations.41 And even as racial segregation wanes, income segregation is worsening. Mixed-income neighborhoods are becoming rarer, and the poor and the rich are increasingly isolated from each other.42 None of these developments refutes the optimistic premise of this project. But it is important to remember the clouds and not just the silver lining.

  • 1. See generally Report of the National Advisory Commission on Civil Disorders (1968) (“Kerner Commission Report”). See also generally John Charles Boger, Race and the American City: The Kerner Commission in Retrospect—an Introduction, 71 NC L Rev 1289 (1993) (commenting at length on the report’s significance).
  • 2. Kerner Commission Report at 10 (cited in note 1). See also id at 1 (“This is our basic conclusion: Our Nation is moving toward two societies, one black, one white—separate and unequal.”).
  • 3. value="3">See generally Douglas S. Massey and Nancy A. Denton, American Apartheid: Segregation and the Making of the Underclass (Harvard 1993). See also Patrick Sharkey, Stuck in Place: Urban Neighborhoods and the End of Progress toward Racial Equality 25 (Chicago 2013) (calling American Apartheid one of two “major work[s] on urban poverty” published in the last several decades).
  • 4. Massey and Denton, American Apartheid at 17–78 (cited in note 3). See also Douglas S. Massey and Nancy A. Denton, Hypersegregation in U.S. Metropolitan Areas: Black and Hispanic Segregation along Five Dimensions, 26 Demography 373, 373–74 (1989).
  • 5. Massey and Denton, American Apartheid at 9 (cited in note 3).
  • 6. Robert G. Schwemm, Housing Discrimination: Law and Litigation § 2:1 at 2-2 (Thomson Reuters 2014).
  • 7. Stacy E. Seicshnaydre, The Fair Housing Choice Myth, 33 Cardozo L Rev 967, 970 (2012).
  • 8. Brief of Housing Scholars as Amici Curiae Supporting Respondent, Texas Department of Housing and Community Affairs v Inclusive Communities Project, Inc, Civil Action No 13-1371, *4 (US filed Dec 23, 2014) (available on Westlaw at 2014 WL 7405732). For other examples of legal scholars characterizing segregation levels as high and stable, see Rigel C. Oliveri, Beyond Disparate Impact: How the Fair Housing Movement Can Move On, 54 Washburn L J 625, 642 (2015) (noting “the persistent and pervasive nature of residential racial segregation across the nation”); Daria Roithmayr, Locked in Segregation, 12 Va J Soc Pol & L 197, 198 (2004) (pointing out “[t]he persistence of residential segregation” and observing that it appears “to be a [ ] stable feature of the American socio-economic landscape”); Abraham Bell and Gideon Parchomovsky, The Integration Game, 100 Colum L Rev 1965, 1979 (2000) (commenting on the “prevalence of segregation as a social phenomenon”).  
  • 9. Michael T. Maly, Beyond Segregation: Multiracial and Multiethnic Neighborhoods in the United States 2 (Temple 2005).
  • 10. value="10">See generally Reynolds Farley, The Waning of American Apartheid?, 10 Contexts 36 (Summer 2011).
  • 11. See generally Edward Glaeser and Jacob Vigdor, The End of the Segregated Century: Racial Separation in America’s Neighborhoods, 1890–2010 (Manhattan Institute, Jan 2012), archived at For other optimistically titled works, see generally William H. Frey, Diversity Explosion: How New Racial Demographics Are Remaking America (Brookings 2015); Maly, Beyond Segregation (cited in note 9); John Iceland, Gregory Sharp, and Jeffrey M. Timberlake, Sun Belt Rising: Regional Population Change and the Decline in Black Residential Segregation, 1970–2009, 50 Demography 97 (2013).
  • 12. See Glaeser and Vigdor, The End of the Segregated Century at *3–4 (cited in note 11).
  • 13. See id.
  • 14. See, for example, Jacob S. Rugh and Douglas S. Massey, Segregation in Post-Civil Rights America: Stalled Integration or the End of the Segregated Century?, 11 Du Bois Rev 205, 206 (2014) (“Public policies . . . appear largely to have ended overt racial discrimination in real estate and lending markets.”).
  • 15. See Margery Austin Turner, et al, Housing Discrimination against Racial and Ethnic Minorities 2012 *xxiii (HUD Office of Policy Development and Research, June 2013), archived at (“Long-term trends in patterns of discrimination suggest that the attitudes and actions of rental and sales agents have changed over time.”).
  • 16. See Farley, 10 Contexts at 40 (cited in note 10).
  • 17. See id.
  • 18. See John R. Logan and Charles Zhang, Global Neighborhoods: New Pathways to Diversity and Separation, 115 Am J Sociology 1069, 1070–72 (2010) (explaining the “buffer” phenomenon and concluding that “stable diversity is possible . . . if black entry is preceded by a substantial presence of both Hispanic and Asian residents”).
  • 19. Pub L No 90-284, 82 Stat 81 (1968), codified as amended at 42 USC § 3601 et seq.
  • 20. See Havens Realty Corp v Coleman, 455 US 363, 376–78 (1982); Gladstone, Realtors v Village of Bellwood, 441 US 91, 109–11, 114–15 (1979); Trafficante v Metropolitan Life Insurance Co, 409 US 205, 208–12 (1972).
  • 21. Gladstone, 441 US at 111, 115.
  • 22. See Texas Department of Housing and Community Affairs v Inclusive Communities Project, Inc, 135 S Ct 2507, 2522 (2015) (stating that the FHA targets practices “creating discriminatory effects or perpetuating segregation”) (emphasis added).
  • 23. Schwemm, Housing Discrimination § 10:7 at 10-52 (cited in note 6).
  • 24. See, for example, Personnel Administrator of Massachusetts v Feeney, 442 US 256, 278 (1979) (reciting this common-law presumption). Of course, this presumption alone is insufficient to establish discriminatory intent, at least under the Equal Protection Clause. See id at 278–80.  
  • 25. Pub L No 89-110, 79 Stat 437 (1965), codified as amended at 52 USC § 10101 et seq. See also 52 USC § 10301(b) (banning practices that result in minority members having “less opportunity . . . to elect representatives of their choice”).
  • 26. Thornburg v Gingles, 478 US 30, 50 (1986).
  • 27. See id at 51. Gingles’s second prong requires minority political cohesion, and its third prong requires white bloc voting. See id. In combination, these two prongs amount to a racial-polarization criterion.
  • 28. See id at 52, 53 n 20 (quoting the district court as referring to “extreme case analysis” carried out by the plaintiffs’ expert as “standard in the literature”). See also D. James Greiner, Ecological Inference in Voting Rights Act Disputes: Where Are We Now, and Where Do We Want to Be?, 47 Jurimetrics 115, 155–57 (2007) (listing dozens of VRA cases employing homogeneous precinct analysis).
  • 29. This ban originated in the landmark case of Shaw v Reno, 509 US 630 (1993), which recognized the “analytically distinct claim” that a district was drawn predominantly for racial reasons. Id at 652.
  • 30. See generally D. James Greiner, Re-solidifying Racial Bloc Voting: Empirics and Legal Doctrine in the Melting Pot, 86 Ind L J 447 (2011) (discussing difficulties caused by desegregation for racial-polarization analysis).
  • 31. See generally Brown v Board of Education of Topeka, 347 US 483 (1954). More specifically, the Court forbade de jure but not de facto school segregation. See id at 487–88, 495. I refer to school segregation as “de jure” or “intentional” when I wish to call attention to its constitutionality. When I refer to school segregation without any qualifiers, I mean de facto segregation: schools’ actual level of racial separation. Consistent with this usage, I treat “integration” and “desegregation” as synonymous, both referring to de facto rather than de jure conditions.
  • 32. value="32">347 US 483 (1954).
  • 33. See, for example, John R. Logan and Deirdre Oakley, The Continuing Legacy of the Brown Decision: Court Action and School Segregation, 1960-2000 *15 (Lewis Mumford Center for Comparative Urban and Regional Research, Jan 28, 2004), archived at (showing a decline in the dissimilarity index of the average school district from 79.4 in 1968 to 45.4 in 1990).
  • 34. See, for example, Sean F. Reardon and Ann Owens, 60 Years after Brown: Trends and Consequences of School Segregation, 40 Ann Rev Sociology 199, 204 (2014) (“[T]he last 25 years have been characterized by largely stable patterns of sorting of students among schools.”).
  • 35. See, for example, Charles T. Clotfelter, Jacob L. Vigdor, and Helen F. Ladd, Federal Oversight, Local Control, and the Specter of “Resegregation” in Southern Schools, 8 Am L & Econ Rev 347, 350 (2006) (noting that “were it not for judicial rulings of unitary status, racial segregation across schools might have declined” due to “[t]he decline in residential segregation”).
  • 36. See Erica Frankenberg, The Role of Residential Segregation in Contemporary School Segregation, 45 Educ & Urban Society 548, 557–58 (2013) (showing an increase in the correlation between residential and school segregation to 0.91 in 2010).
  • 37. Swann v Charlotte-Mecklenburg Board of Education, 402 US 1, 15–18 (1971). See also, for example, Freeman v Pitts, 503 US 467, 505 (1992) (Scalia concurring) (observing that once a violation has been proved, “there arises a presumption, effectively irrebuttable . . . that any current racial imbalance is the product of that violation”).
  • 38. Trafficante, 409 US at 211, quoting 90th Cong, 2d Sess, in 114 Cong Rec 3422 (Feb 20, 1968) (“1968 Civil Rights Act Senate Debate”) (statement of Sen Mondale).
  • 39. Green v County School Board of New Kent County, 391 US 430, 442 (1968).
  • 40. See John R. Logan and Brian J. Stults, The Persistence of Segregation in the Met­ropolis: New Findings from the 2010 Census *9 (US2010 Project, Mar 24, 2011), archived at (labeling the “persistence of very high black-white segregation in a few major Northeastern and Midwestern metropolitan areas” a “striking feature” of recent decades).
  • 41. See John Iceland and Gregory Sharp, White Residential Segregation in U.S. Metropolitan Areas: Conceptual Issues, Patterns, and Trends from the U.S. Census, 1980 to 2010, 32 Population Rsrch & Pol Rev 663, 665 (2013) (“Hispanic and Asian segregation has not declined markedly over the past three decades.”).
  • 42. See Sean F. Reardon and Kendra Bischoff, Income Inequality and Income Segregation, 116 Am J Sociology 1092, 1116 (2011) (“Average metropolitan area income segregation . . . [grew] from 1970 to 2000, with the fastest increase occurring in the 1980s.”).