Desegregation, Then and Now

R. Shep Melnick

Winter 2020

One of the oddest features of the 2019-20 Democratic primary season has been the return of the busing issue. Half a century ago, it nearly tore the party apart. Judicially mandated reassignment of students to achieve racial balance proved to be the most unpopular policy since Prohibition, opposed by overwhelming majorities of white voters. In city after city, Hispanic and Asian-American leaders and parents fought to preserve neighborhood schools. By the mid-1980s, support for busing had fallen below 50% among African-Americans. The parents of African-American students subject to busing were often the most vocal opponents of the policy. Yet here were Kamala Harris and other Democratic candidates attacking Joe Biden for his position on an issue that had long ago faded into political oblivion.

The candidates who pounced on the former vice president seemed to have little interest in actually bringing busing back. Senator Harris later offered a muddled statement explaining why she does not currently support the policies she slammed Biden for once opposing. From its origin in the 1960s to its slow demise in the 1990s, busing had only one reliable patron: the federal courts. They presented their racial-balance mandate as a remedial measure designed to undo prior discriminatory practices by state and local officials. After supervising school practices for decades, most judges determined that the lingering effects of such practices had been sufficiently overcome to return control of schools to the ordinary political process. The Supreme Court has gone even further, holding that once a school system has been released from court supervision, it can no longer assign students on the basis of race.

Given this, it would be easy to dismiss the return of the busing issue as little more than a debater's ploy to attack a leading candidate for the Democratic nomination. But Senator Harris's nostalgic defense of desegregation in Berkeley unleashed a series of articles calling for the return of busing to address what is misleadingly called the "re-segregation" of American schools. The leading example is the New York Times' long "news analysis" by staff writer and former MacArthur genius Nikole Hannah-Jones with the provocative title, "It Was Never About Busing. Court-ordered desegregation worked. But white racism made it hard to accept." An NPR piece on the Supreme Court's 1974 decision in Milliken v. Bradley bore the equally misleading title, "This Supreme Court Case Made School District Lines a Tool for Segregation." Other articles and op-eds offered similar arguments, based in large part on the assertion that the United States is so irredeemably racist that it rejects policies that have proved effective in providing equal educational opportunity to minority children. Meanwhile, New York City is considering a proposal to eliminate exam schools and gifted programs in order to "desegregate" a school system that is 85% non-white.

Not only is this line of argument a clear political loser for Democrats, but it badly distorts the history of school desegregation, preventing us from learning from that searing experience. Most important, by using the ambiguous term "desegregation" to cover vastly different policies, it keeps us from distinguishing between the features of desegregation that improved opportunities for minority children and those that did not. Yes, in many instances, court-ordered desegregation worked. In others, it did not. Yes, many who opposed busing were motivated by racial prejudice. But others — including many African-American and Hispanic leaders and parents — had stronger reasons for criticizing court mandates. Calling anyone who opposed any form of desegregation racist shuts down a badly needed discussion about the most effective ways to improve public education for minority children.


The central, unstated assumption of Hannah-Jones and other defenders of court-ordered busing is that the meaning of the crucial term "desegregation" remained constant from Brown v. Board of Education in 1954 to busing in Boston two decades later. To oppose busing in Detroit, Dayton, or Delaware (as Biden did), so the story goes, is to reject the most important Supreme Court decision of the 20th century. Moreover, if "desegregation" has a coherent, constant meaning, we should focus on the benefits of the entire project, from desegregation of the border states in the 1950s to the "reconstruction of Southern education" in the 1960s and 1970s, to the effort to create racially balanced schools outside the South during the 1970s and 1980s. There is no need to distinguish among the diverse projects lumped under the heading "desegregation."

This story is just plain wrong, as anyone who has studied the history of desegregation well knows. The starting point for any serious evaluation of desegregation must be acknowledging how much the meaning of that key term changed over time. We can argue over the merits of this transformation, but not over the nature and extent of the change.

In his opinion for a unanimous court in Brown v. Board, Chief Justice Earl Warren presented arguments he hoped would not only convince legal skeptics, but appeal to the better angels of citizens in the North and South. He never explained what school districts must do to achieve desegregation. For that omission the Court has been justly criticized, since its silence allowed the South to evade its constitutional responsibilities for a decade and a half. Nor did Warren provide an adequate explanation for why state-sponsored segregation is wrong. His reliance on dubious social-science evidence needlessly left him open to attack by segregationists.

Although the Court never cited the famous words of Justice John Marshall Harlan's dissent in Plessy v. Ferguson — "Our Constitution is color-blind, and neither knows nor tolerates classes among citizens" — that understanding lay at the heart of the NAACP's legal argument and formed the foundation of the Court's discussion of remedies. In oral argument, the NAACP's Robert Carter explained that the "one fundamental contention which we will seek to develop" is that "no state has any authority under the equal protection clause of the Fourteenth Amendment to use race as a factor in affording educational opportunity among its citizens." Thurgood Marshall assured the Court that "the only thing that we are asking for is that the state-imposed racial segregation be taken off, and to leave the county school board, the county people, the district people, to work out their solutions of the problem to assign children on any reasonable basis they want to assign them on."

Both Marshall and the justices assumed that in most cases school districts would follow the standard practice of neighborhood schools. The simple job of drawing new district lines on a non-racial basis, Marshall maintained, "might take six months to do it one place and two months to do it another place." The chief justice explained that school officials have a constitutional responsibility "to achieve a system of determining admission to the public schools on a nonracial basis" (emphasis added). Neighborhood schools and prohibition of race-based assignments — in 1954 no one considered these commitments "racist."

The re-interpretation of "desegregation" to mean just the opposite — that is, to mandate use of racial assignments in order to replace neighborhood schools with racially balanced ones — came in two stages, the first directed by the Fifth Circuit Court of Appeals in the mid-1960s and the second by the Supreme Court from 1968 through 1973. After enactment of the 1964 Civil Rights Act, the Department of Justice and the Department of Health, Education, and Welfare joined forces with the Fifth Circuit finally to end Southern school segregation. By then, the main obstacles were the "freedom of choice" plans instituted by many school districts to maintain the status quo. These plans were almost always structured so as to inhibit black students from transferring to formerly white schools and to allow virtually all white students to remain where they were. Where administrative connivance proved inadequate, threats of violence and economic reprisal greeted African-American families audacious enough to switch schools. Whatever its merits in the abstract, in practice "freedom of choice" was a fraud.

Faced with thousands of school districts that remained nearly as segregated as a decade before, judges and administrators used numerical benchmarks to determine whether schools were making a good-faith effort to comply with Brown. These benchmarks did not require strict racial balance, only evidence that the old patterns had been dislodged. The resulting desegregation orders were remedial — tough measures designed to confront officials who had defied the courts for many years. Some advocates of Justice Harlan's "color-blind" understanding of the 14th Amendment have criticized this move, but it is hard to see how federal judges and administrators could have acted differently without giving up on the entire desegregation effort. At the time, such use of racial targets was considered a temporary measure to remedy egregious constitutional violations.

The second stage came in three decisions the Supreme Court issued when it re-entered the picture after a decade and a half of silence. In Green v. County School Board of New Kent County, a 1968 case from rural Virginia, the Court issued a short, rhetorically powerful but deeply ambiguous opinion that upheld the general approach of the Fifth Circuit, but then added the enigmatic command to eliminate all "racially identifiable" schools. According to Justice William Brennan's clever turn of phrase, previously segregated school districts have a constitutional duty "to convert promptly to a system without a 'white' school and a 'Negro' school, but just schools."

How does one know when such racial identification has been eliminated and a "unitary" system established? In Green, the Court implied that a "unitary" school system was one in which the racial composition of each school reflected the racial composition of the school system as a whole. That made sense for rural school districts like New Kent County, Virginia, which had few schools, little residential segregation, and a long history of racial discrimination. But what about large urban districts with substantial residential segregation?

In its 1971 Swann v. Charlotte-Mecklenburg Board of Education decision, the Court appeared to say that for urban school districts with a history of legally mandated segregation, desegregation requires racial balance in all its schools, even if that requires extensive busing of students (including those in elementary school) far beyond their neighborhood schools. I say "appeared" because Chief Justice Warren Burger's opinion was one of the most confused and internally contradictory opinions ever issued by the Supreme Court. As then-judge (and later attorney general) Griffin Bell put it, "It's almost as if there were two sets of views laid side by side." On the one hand, Burger wrote, "The constitutional command to desegregate schools does not mean that every school in every community must always reflect the racial composition of the school system as a whole" (emphasis added). Indeed, if the lower court had required "as a matter of substantive constitutional right, any particular degree of racial balancing or mixing, that approach would be disapproved and we would be obliged to reverse." On the other hand, the Court upheld in full a desegregation plan for the Charlotte-Mecklenburg school system in North Carolina that did precisely that.

Burger's muddled opinion led federal judges throughout the South to mandate extensive busing to create racially balanced schools. Swann has since become the foundational opinion for those who believe that desegregation requires rather than prohibits the use of race in assigning students to particular schools. For example, Justice Stephen Breyer's defense of race-based assignments to achieve racial balance in his 2007 Parents Involved in Community Schools v. Seattle School District dissent repeatedly invokes Swann, mentioning Brown just a handful of times.

The third Supreme Court decision on busing, issued in Keyes v. School District No. 1, Denver, Colorado in 1973, in effect applied Swann's expansive remedies to cities outside the South. In a deceptively bland opinion, Justice Brennan held that if a judge can find evidence that a school district has drawn attendance lines or sited school buildings so as to keep some schools predominantly minority, then he must issue an injunction eliminating "racially identifiable" schools throughout the city. The Court later lowered the evidentiary threshold so that a school district's failure to maximize racial balance constituted evidence of discriminatory intent. That meant that few cities could prove their innocence if targeted by lawyers from the Department of Justice or the NAACP. Soon, scores of cities outside the South were required to abandon neighborhood schools in order to create racial "balance" — a term that remained ambiguous despite the weight it was expected to carry.


In the Green-Swann-Keyes trilogy of cases, the Supreme Court departed so far from the original understanding of Brown — as well as from the Civil Rights Act, which stipulated that "'desegregation' shall not mean the assignment of students to public schools in order to overcome racial imbalance" (emphasis added) — that one must ask what lay behind this fateful transformation. The answer is apparent in the many lower-court decisions applying the Court's contradictory, maddeningly ambiguous, and sometimes disingenuous opinions. The underlying problem, they explained, is not state-sponsored racial segregation, but racial isolation, whatever the cause. The most direct statement of this argument appeared in an influential 1967 report by the United States Commission on Civil Rights:

The central truth which emerges from this report and from all of the Commission's investigations is simply this: Negro children suffer serious harm when their education takes place in public schools which are racially segregated, whatever the source of such segregation may be.

Negro children who attend predominantly Negro schools do not achieve as well as other children, Negro and white. Their aspirations are more restricted than those of other children and they do not have as much confidence that they can influence their own futures. When they become adults, they are less likely to participate in the mainstream of American society, and more likely to fear, dislike, and avoid white Americans. [Emphasis added.]

This message was transmitted to federal judges by respected expert witnesses. They confidently testified that the academic performance of minority children could be improved substantially if placed in schools that are 70-80% white. (Over 80% will leave minority children feeling isolated; less than 70% approaches the "tipping point" for white flight.) For example, in Swann, district-court judge James McMillan wrote,

The uncontradicted evidence before the court is that segregation in Mecklenburg County has produced its inevitable results in the retarded educational achievement and capacity of segregated school children....The experts and administrators all agreed that transferring underprivileged black children from black schools into schools with 70% or more white students produced a dramatic improvement in the rate of progress and an increase in the absolute performance of the less advanced students, without material detriment to the whites. There was no contrary evidence. (In this system 71% of the students are white and 29% are black.)

Judges throughout the country heard similar "uncontradicted" testimony and came to similar conclusions. Judicially mandated reassignment of students to achieve racial balance was thus transformed from an extraordinary judicial remedy for egregious racial discrimination to an education policy designed to improve the academic performance of minority students even in districts praised by judges for their desegregation efforts.

This ambitious education-reform plan faced three challenges. First, the evidence on which it was based simply did not justify the confidence with which it was propounded. The basic argument was that when African-American students from lower socioeconomic backgrounds go to school with white students from higher socioeconomic backgrounds, the schools' culture and the attitudes of minority children will change fundamentally. Although expert witnesses frequently relied on the Office of Education's famous Coleman Report to justify this position, the findings of that report in fact made this conclusion highly dubious.

The second problem was white flight: If desegregation orders lead white families to flee urban areas for the suburbs or private schools, then the presumed benefits of the project will be substantially reduced. During the 1960s and 1970s, families both black and white had many reasons for leaving the cities, and busing orders accelerated the process. Especially if the few remaining white students are also poor, it is hard to see how the educational opportunities of minority children will improve.

The third problem was that in many cities the percentage of minority students exceeded the "optimal" 30% even before judicial intervention. This was a more serious problem in the Northeast and Midwest than in the South because of the smaller geographic size of their school districts. The Charlotte-Mecklenburg school district now covers 525 square miles; Oklahoma City and Houston are even larger. Boston's school district, in contrast, covers only 48 square miles, Cleveland's 82, and Cincinnati's 91. Southern school districts typically included both urban and suburban areas. Outside the South, cities and suburbs are generally distinct political bodies. There, addressing "racial isolation" by making schools majority white required the creation of mega-school districts, fundamentally changing the way schools are governed.


This became most apparent in Detroit, where the number of white students was small and rapidly dwindling. In the pivotal case of Milliken v. Bradley, district-court judge Stephen Roth praised the school board for its efforts to integrate the system, but nevertheless found it guilty of unconstitutional segregation. Recognizing the futility of rearranging student assignments within this educational Titanic, he mandated the creation of a new school district comprised of the city and 52 suburbs, and ordered extensive busing between them. Roth offered scant evidence that the suburbs had engaged in discriminatory behavior; indeed, he never gave them an opportunity to present their legal arguments during the liability phase of the litigation. His ruling flowed from the extensive testimony he heard on the effects of racial isolation: Regardless of the cause of the separation, minority students cannot receive a constitutionally adequate education if they are not in majority-white schools.

The Detroit busing order produced furious opposition. The political effects were evident in the results of the 1972 presidential primary in Michigan. Although the state's UAW-led Democratic Party had long been among the country's most liberal, arch-segregationist George Wallace won the Democratic primary by a substantial margin. Politicians who supported the court's desegregation plan faced near-certain defeat. Nor did the plan receive much support from African-American leaders in Detroit or civil-rights organizations. Many black leaders (including future mayor Coleman Young) preferred a decentralization plan that would give neighborhoods within the city more control over their schools.

Before the Detroit plan could go into effect, the lower-court decision was reversed by the Supreme Court. By a narrow 5-4 vote, the Court held that the suburbs could be included in mandatory desegregation plans only if there was evidence that they had engaged in "segregative acts." Implicitly rejecting the "racial isolation" argument, the majority required evidence of intentional government action. Since the boundaries of Detroit had been drawn long before it had a significant black population, the lower court would need to identify other "segregative acts" by government officials before issuing such a far-reaching injunction. Only rarely after Milliken did federal courts mandate busing that crossed political boundaries. One of those court orders came in Joe Biden's Delaware, the product of an inscrutable circuit-court opinion that the Supreme Court refused to review.

For years, the Court's Milliken opinion has been roundly criticized by advocates of the racial-isolation argument. They rightly point out that it often led to the worst of both worlds: massive disruption of school systems that still left most of them overwhelmingly black and Hispanic. But those who attack Milliken fail to acknowledge the likely political consequences of a contrary ruling. Milliken took the steam out of the effort to amend the Constitution to prohibit race-based school assignments and to mandate neighborhood schools. If courts throughout the North and West had ordered similar urban-suburban busing, that amendment may well have passed — and the few Democrats who opposed it would likely have been driven from office. Representing the sole district subject to two-way interdistrict busing, Senator Biden recognized this danger earlier than most of his fellow Democrats.

The Supreme Court's tacit embrace of the racial-isolation argument in the Green-Swann-Keyes trilogy combined with the rigid constraint imposed by Milliken placed the lower courts in a terrible bind. In the 1970s, most district-court judges required desegregation by the numbers, though some judges allowed more variation than others. Gradually, they shifted away from a preoccupation with racial ratios toward experimentation with more extensive educational reforms. Some of these worked — most notably magnet schools. As Joshua Dunn's detailed history of the Kansas City case shows, others did not. Meanwhile school demographics were rapidly shifting, with the number of white students continuing to decline and the number of Hispanic students escalating. In contrast to the 1960s, desegregation issues were no longer black and white.


Today, it is hard to find a newspaper article on race and education that does not confidently assert that our schools are becoming "re-segregated." Yet a 2019 analysis conducted by the Washington Post found that "[t]he number of children attending U.S. public schools with students of other races has nearly doubled over the past quarter-century, a little-noticed surge that reflects the nation's shifting demographics."

Readers have good reason to be confused about what the terms "segregation," "desegregation," and "re-segregation" now mean. Clearly, those who claim our schools are being re-segregated do not mean that states are enacting laws to mandate racial segregation or even that school officials are locating school buildings or drawing attendance zones to keep the races apart. Most large urban school districts have made substantial efforts to reduce the impact of housing segregation through magnet schools, "controlled choice," majority-to-minority transfer options, and careful siting of new schools. The original meaning of the term "segregation" and the pivotal role de jure segregation played in the abominable racial caste system of the South are distant memories.

The two most commonly used metrics for monitoring racial trends in our schools reflect the redefinition of "desegregation" described above. One is the "black-white dissimilarity" or "evenness" index. It measures the extent to which white and black students are evenly distributed across the schools within a district. How many students would need to move to ensure that each school has exactly the same racial balance? Given the Supreme Court's insistence that previously segregated districts must eliminate all "racially identifiable" schools, this would seem to be the most appropriate way to judge whether a school system has achieved "unitary status."

Complicating the matter is the rapidly growing number of Hispanic and Asian students in most urban school systems. One can construct "dissimilarity indexes" not only to measure the distribution of black and white students, but also to measure Anglo / Hispanic segregation, white / non-white segregation, and segregation among non-white groups. The multiplicity of these "dissimilarity indexes" indicates how complex the "segregation" question has become. Should we consider a school with a substantial number of African-American, Hispanic, and Asian students but few Anglos "diverse" or "segregated"? Behind each index lies an understanding of what type of racial or ethnic balance we consider most important.

Despite the fact that many school districts have been released from judicial supervision since 2000, "dissimilarity" indexes have changed very little. In 2014 Sean Reardon of Stanford and Ann Owens of the University of Southern California concluded, "it seems fair to say that the last 25 years have been characterized by largely stable patterns of sorting of students among schools (unevenness)." Using a measure of evenness that considers the distribution of white, black, Hispanic, and Asian students, Brian An and Adam Gamoran "found little evidence that school segregation has increased over time." "If anything," they wrote, "there was evidence that school segregation decreased over time, especially between Latino and white students."

A shortcoming of all "dissimilarity" indexes is that they do not measure the extent to which one group actually comes in contact with another. A school system in which each school is exactly 60% black, 30% Hispanic, 5% Asian, and 5% white would have a perfect score on the "dissimilarity index," yet white and black students would rarely see one another. To address this problem, social scientists have developed "exposure" indexes. The most commonly used exposure index measures the percentage of white students in the typical black student's school. As with the dissimilarity index, though, there are several variants. Should we focus on white / black exposure, or Anglo / minority, or black / non-black? Should Asian students be lumped together with whites or with African-Americans? The use of exposure indexes reflects the concerns over racial isolation that drove desegregation policy in the early 1970s. But demographic change complicates the picture: isolated from whom?

The white / black exposure index improved substantially in the late 1960s and 1970s, stabilized in the 1980s, then began to drop. The chief culprit here was not government policy, but demographic change. Exposure indexes are highly sensitive to the relative size of racial and ethnic groups living within the school district. Within the first two decades of the 21st century, the percentage of white students had declined to less than 15% in Los Angeles, Baltimore, Memphis, Chicago, and Miami-Dade County, Florida; it was down to 5% in Dallas, 2% in Detroit, and 15% in New York City. Today, white students constitute less than half the public-school population. By 2023, the percentage of Anglos is predicted to drop to 45% while Hispanics climb to nearly 30%. As a result, measures of white / black and white / minority exposure will continue to fall regardless of what government does.

The drop in the white / black exposure index does not mean that our schools are becoming increasingly separated by race. In fact, schools attended by the typical white student are less white today than in years past. Our schools are simply becoming more diverse, with more Hispanic and Asian students in schools attended by both the typical black student and the typical white student.

Another key difficulty with the "exposure" index is that it does not really measure contact between students of different races and ethnicities. It assumes, quite unrealistically, that there is no segregation within school buildings as a result of tracking and peer-group norms. What we most want to know — are black and white students sitting together in math class and in the cafeteria? — is extremely hard to measure.

In short, because "desegregation" has so many different meanings, there are many ways to measure its progress. As Reardon and Owens note, "depending on the definition used, segregation has either increased substantially or changed little" over the past several decades. Given how loaded the term "segregation" is, we should take a close look at the meaning of the statistics used by advocates who throw that term around.


Nikole Hannah-Jones, Senator Harris, and many other self-described progressives claim that, even with the limitation imposed by Milliken, busing to create racially balanced schools "worked." Hannah-Jones relies almost entirely on the work of Rucker Johnson discussed below. Harris painted a rosy picture of busing in Berkeley in the 1970s based on her experience as a young student. In his detailed study of desegregation in that city, Berkeley professor David Kirp concluded that it "failed to move beyond racial balance" to address "broader concerns of educational purpose." The school system "became fragmented in purpose, seemingly rudderless." Racial tension increased; many black students and teachers demanded separate black schools. The achievement gap between white and black students widened. Senator Harris did not experience this downside because she moved to Montreal at age 12.

Extensive research on the topic presents a more complicated picture. One challenge that researchers face is untangling the many factors that affect students' performance and their long-term prospects. School desegregation was just one part of a huge change in the economic, cultural, social, and political status of African-Americans after the civil-rights revolution of the 1960s. Another challenge is the wide variety of forms desegregation took over the years. Desegregation was never a coherent policy applied uniformly across the country. In effect, lower-court judges custom-built plans for each city, then constantly revised them.

In his book-length review of research on the topic, Roger Levesque notes, "only after research indicated that desegregation pervasively did not lead to the intended positive effects," did researchers try to isolate those "contextual and process factors" that might make the difference. In other words, instead of asking the cosmic and essentially unanswerable question "does desegregation work?" researchers began to study the consequences of a variety of desegregation techniques.

In the early years of the desegregation effort, social scientists had made expansive claims on the basis of very little evidence. In Brown, Chief Justice Warren relied on social-science studies to suggest that segregation lowers minority students' self-esteem and aspirations. To nearly everyone's surprise, subsequent research has repeatedly shown that African-American students have higher self-esteem than do white children, and that those in predominantly minority schools often have higher self-esteem than those in majority white schools. Other studies have shown that the aspirations of black students in predominantly black schools are at least as high as those of white students in predominantly white schools.

Another pervasive but not persuasive argument for desegregation is based on "contact theory." It seems like common sense: More contact between members of different racial and ethnic groups would reduce prejudice and hostility. As Justice Thurgood Marshall put it, "unless our children begin to learn together, there is little hope that our people will ever learn to live together." Unfortunately, a number of studies have indicated that desegregation is as likely to increase as to decrease racial animosity and stereotyping. It all depends on how it is done. Desegregation also seems to reduce civic-mindedness among students. This distressing finding is consistent with the conclusion Robert Putnam reluctantly reached in a 2006 lecture: Racial and ethnic heterogeneity tends to decrease social capital.

Perhaps the most important "what works" questions focus on the black / white academic-achievement gap. Here, the 30,000-foot picture is pretty clear. From 1970, when desegregation finally took hold in the South, through the mid-1980s, the racial achievement gap measured by the National Assessment of Educational Progress test narrowed significantly, declining in both math and reading for almost all age groups. For some age groups, the gap was cut nearly in half. Then progress stalled. Although there have been some minor ups and downs since then, the gap has been depressingly persistent for several decades now. Other indicators tell a similar story. The white-black high-school-graduation gap narrowed from the late 1960s to the mid-1980s, when the rate for white and black women and white men stabilized, but the rate for black men declined. The racial difference in years of education also narrowed through 1990, then grew slightly.

To what extent has desegregation policy contributed to these trends? Large numbers of studies and meta-analyses have examined this question, yet answers remain elusive. In 1984, the National Institute for Education commissioned an analysis of many of the most reliable studies of desegregation conducted over the preceding two decades. The lead investigator concluded that desegregation produced no increase in minority students' achievement in math and a small increase in reading. But he added, "I have little confidence that we know much about how desegregation affects reading 'on the average' and, across the few studies examined, I find the variability in effect sizes more striking and less well understood than any measure of central tendency."

Two decades later, another leading desegregation scholar, Charles Clotfelter of Duke, concluded, "Studies that have sought to determine the effect of desegregation on the achievement of blacks have come up with a decidedly mixed set of results. In general, the research suggests no effect on mathematics achievement for blacks and some modest positive effect on reading for blacks." In 2014, Reardon and Owens noted that, although this research has been motivated by concern about the achievement gap, we still don't understand "the mechanisms through which segregation affects student outcomes." This they attributed to the fact that researchers "lack a detailed and comprehensive theoretical model (or models) of exactly how segregation might affect educational and social outcomes." Levesque's 2018 review of the literature reached a similar conclusion: Many years of study "reveal weak and inconsistent effects and generally view desegregated schooling neither as a demonstrated success nor as a demonstrated failure." These "mixed effects," he argues, reflect the variety of ways that desegregation has been carried out.

Berkeley professor Rucker Johnson's 2019 book with Alexander Nazaryan, Children of the Dream: Why School Integration Works, offers a more optimistic analysis. Unlike most studies, his examines the long-term consequences of attending a desegregated school. To control for the many other factors that could influence his measures of well-being, Johnson took advantage of the fact that the year in which school districts were subject to desegregation orders was "driven largely by a number of chance factors, not systematic differences in families or neighborhood conditions that may have independently affected children's outcomes." Combining extensive data on school resources and quality with longitudinal data on the educational attainment and income history of thousands of students, he compared students who attended desegregated schools with those who attended schools that were similar in other respects but not yet desegregated. Johnson found a strong relationship between the number of years students spent in a desegregated school and five key outcome measures: total years of education (up 30% for those attending a desegregated school for 12 years), adult annual earnings (also up 30%), adult health status, incarceration rates (down 22% for adults), and poverty rates (down 25%).

Johnson's data did not allow him to distinguish among the many phases and forms of desegregation, so his analysis can't tell us how much of this improvement was the product of the initial rounds of desegregation in the border states and the South, and how much of the later rounds in the North. Moreover, it examines children born between 1945 and 1970, which means that it begins with those who entered school between 1950 and 1975 and stops with those who left school between 1963 and 1988. In other words, it focuses on years in which the achievement gap and the high-school completion gap were narrowing, stopping soon after those gaps plateaued.

Hannah-Jones relies on Johnson's analysis to support her claim that busing "worked," but ignores what he says about which features of desegregation had the most beneficial effects. Two proved particularly important: "sharp increases in per-pupil spending (by an average of 22.5 percent) and significant reductions in the average class sizes experienced by black children." These changes were especially significant in the South, where previously black schools had been egregiously underfunded.

Johnson found that money mattered much more than black-white student exposure. Where resources increased significantly but exposure did not, students did well. Conversely, "in court-ordered desegregation districts in which school spending for black children did not appreciably change, however, although the children experienced greater classroom exposure to their white peers, they did not make a comparable improvement in their educational and socioeconomic trajectories." In other words, what matters most are the tangible resources devoted to schools serving minority children, not the number of white students in the school building. Johnson's sophisticated analysis provides substantial support for some forms of desegregation, but not for those that simply bused children around to achieve racial balance.


Because the Supreme Court's rulings on school desegregation were so ambiguous and contradictory, district-court judges were often left on their own to devise plans for the complex local circumstances they faced. Some demanded racial balance in schools throughout the district with little variance. Others placed greater emphasis on magnet schools and what became known as "controlled choice." In some cities, school officials worked diligently to make desegregation plans succeed. Others dragged their feet and did whatever they could to subvert these plans.

It is easy to identify cities in which court-ordered desegregation plans failed. The most discouraging example is Kansas City, Missouri. There a judge ordered the state and school district to spend an additional $2 billion to increase the "desegregative attractiveness" of city schools with the hope of drawing in more white suburban children. Meanwhile, test scores flat-lined, and the percentage of white students in Kansas City schools dropped. Joshua Dunn's Complex Justice: The Case of Missouri v. Jenkins is a chilling reminder that noble intentions are not enough to improve public schools. In 2002, 26 years after the Dayton school system had been placed under court supervision, the state of Ohio labeled the poor performance of its students an "academic emergency." The district negotiated a settlement that would reinstate neighborhood schools and bring in more state money. The NAACP agreed. In nearby Cleveland, the school district experienced "total fiscal and administrative collapse" in 1995. This led the district-court judge to ditch the student-assignment plan put in place years before and to endorse the school district's reform package. "Twenty years of transportation to achieve racial balance," he wrote, "has not resulted in improved academic performance profiles." In Cleveland, too, African-American leaders and parents approved the change.

What about the cities deemed to be success stories by advocates of busing plans? Here the leading example is Charlotte-Mecklenburg school district (CMS) in North Carolina, the site of the Swann case. The district initially offered the optimal white / black student ratio, 70-30. School officials and local business leaders were determined to demonstrate that their city could made desegregation work. There was no violence. Although some white students left for private schools, there was less white flight here than elsewhere. As three busing advocates have put it, the school system was "widely considered the epitome of a successfully desegregated school district — an example for the nation of how desegregation could be accomplished in ways that fostered student achievement, interracial civility, political stability, and economic growth."

In the early years of desegregation in CMS, test scores for both black and white students rose significantly. Yet the racial achievement gap did not close. By the mid-1980s, CMS children had spent their entire elementary and secondary education in the desegregated system. The gap between white and black students not only remained large — exceeding one standard deviation in 1986 — but somewhat greater than the national average. Starting in the 1990s, the district allowed greater variation in the racial composition of individual schools. Surprisingly, the racial composition of students' schools had no apparent effect on their test results: Scores for black students in schools that were less than 25% minority were virtually identical to those in schools that were more than 75% minority. Perhaps the most distressing statistic about the educational system in Charlotte is one cited by Rucker Johnson. Despite many years under an ambitious desegregation plan and despite a booming regional economy, the pioneering study of intergenerational mobility produced by Raj Chetty and his associates "found that Charlotte ranked last among the fifty largest cities in the United States in terms of a child's prospects for escaping poverty."

In 2001, the federal courts held that Charlotte had achieved "unitary status," and could no longer use race to assign students to schools. By then the student body was majority non-white. CMS retained its magnet schools, but eliminated the racial quotas it had previously imposed on them. Students who did not chose to go to magnets would attend neighborhood schools. The district's "Equity Plus" program provided more funding to schools with high percentages of poor children. CMS's white / black "exposure" index dropped significantly. The percentage of black students attending a school where two-thirds of students or more were black increased from 12% to 21%.

In the years that followed, there was a slight increase in the black / white achievement gap for math, but not for reading. But as Jacob Vigdor has shown, the gap remained smaller in Charlotte than in nearby Wake County, which continued its extensive busing program. Test scores for black children in middle school showed little change — perhaps as a result of "Equity Plus" spending increases. In contrast, the scores of high-school students (both black and white) who attended predominantly minority schools were lower than those attending majority-white schools.

The most serious negative consequences of the post-2002 CMS policy were "large and statistically significant increases in crime" by poor minority males assigned to predominantly minority schools. This was the principal finding of a 2013 study by Stephen Billings, David Deming, and Jonah Rockoff that compared similar groups of students who had, largely by chance, been assigned to schools with different racial compositions. Peer effects, their research suggests, are particularly powerful in middle school and high school, leaving a deeper imprint on non-academic behavior than on test scores. They also found that poor white students (but not black students) are more likely to drop out if they attend a predominantly minority school. This indicates that under optimal conditions, busing can improve the lot of some students living in poor neighborhoods, especially boys. But this is a far cry from the expansive claims made by busing advocates.


Over the past seven decades, "desegregation" has acquired many meanings, and a variety of policies have flown under its banner. Some of these have markedly improved the educational opportunities we provide to minority students. Others have not. All other things being equal, there are significant advantages in reducing the number of predominantly minority schools — not because their students are black or Hispanic, but rather because they are likely to be poor. Yet "other things" are rarely equal. Long bus rides can tire students and reduce the amount of time they spend in class or on school work. Eliminating neighborhood schools can dampen parents' involvement in their children's schools. Regular rejiggering of school assignments to maintain racial balance threatens the stability and continuity that promote learning. White flight can drain schools of both wealthier students and political support. Not surprisingly, the parents of African-American students have often become frustrated with these features of desegregation plans, and have argued for a return to neighborhood schools over which they have more control.

Given demographic trends, ending "racial isolation" by making urban schools majority white is a pipe dream. In most cities, that would require far more busing now than in 1974. Today, the major legal question is not whether the federal courts will mandate greater use of race-based assignments in order to achieve racial balance, but whether they will allow school officials to do so on their own.

Treating "desegregation" as an undifferentiated whole that we must either accept — busing and all — or reject — and thus be tarred as racist — is not only politically imprudent, but prevents us from appreciating which forms of desegregation did work. A more equitable distribution of resources, smaller classes, magnet schools, more experienced teachers, early learning opportunities — these are some of the things that seem to have made a difference.

In the decade following the civil-rights revolution, education reformers could be forgiven for viewing judicially imposed racial balance as a magic bullet that would produce a more equitable education system throughout the nation. In the years that followed, we learned that education reform is never so simple. Today, education experiments and innovations abound. Many of them will fail; a few might succeed. Far better to attend to these incremental reforms than to sit around waiting for the reappearance of the magic school bus.

R. Shep Melnick is the Tip O’Neill Professor of American Politics at Boston College and the author, most recently, of The Transformation of Title IX: Regulating Gender Equality in Education.


from the


A weekly newsletter with free essays from past issues of National Affairs and The Public Interest that shed light on the week's pressing issues.


to your National Affairs subscriber account.

Already a subscriber? Activate your account.


Unlimited access to intelligent essays on the nation’s affairs.

Subscribe to National Affairs.