Is Affirmative Action Doomed?

On September 7 Deval Patrick, the assistant attorney general for civil rights, filed a brief in a New Jersey case arguing that it is legal to fire a white teacher over a black teacher purely because of her race. And on August 19 a federal district judge in Austin, Texas, held that aspects of the affirmative action program at the University of Texas law school are unconstitutional. One or both of the cases may reach the Supreme Court before long. Each on its own could revive the debate about racial preferences and ventilate their more troubling assumptions. Together, they suggest that the current legal foundations of affirmative action are highly unstable. If some kind of argument for racial preferences is to be sustained at all, it has to come from new--and more candid--premises.

For nearly a decade the Supreme Court has said that affirmative action in college and law school admissions is generally permissible; but firing someone on account of race crosses a crucial line, for it imposes all the burdens of racial preferences on a single white victim. By challenging this distinction, which is widely accepted by liberals and conservatives, Deval Patrick has committed the Clinton administration to a vision of racial preference that fulfills the most extravagant fantasies of a conservative attack ad. ("You lost that job because you were white....") Rather than honestly confronting the costs of affirmative action, Patrick has blithely endorsed the most extreme form of racialism. And principled liberals in the administration are concerned. Assistant Attorney General Walter Dellinger, for example, argued strenuously against Patrick's position.

The facts of the New Jersey case, Taxman v. Piscataway, make it a particularly unfortunate platform for Patrick's crusade. A white teacher, Sharon Taxman, and a black teacher, Debra Williams, were hired on the same day in 1980 to teach typing and secretarial studies at Piscataway High School. In 1989 the Piscataway school board decided, in the face of budget cuts, that it couldn't afford both Taxman and Williams. Ordinarily, when two teachers are equally senior, the board flips a coin. But in this case, it decided that since Williams was the only black teacher in the business education department, she should be retained and Taxman fired. What makes the decision so peculiar is that there is no scarcity of black teachers at Piscataway High School. On the contrary, 10 percent of the teaching staff is black, nearly twice the percentage of blacks eligible to teach in the county. In firing Taxman and retaining Williams, the school board had decided that it was important to maintain racial diversity within the department of secretarial studies itself. This is the radical end-point of the diversity argument: the notion that all races should be proportionately represented, not on the faculty as a whole, but in each department of every high school in the district.

The most disappointing aspect of Patrick's brief is the coyness with which it calls into question the basic principles of the Supreme Court's affirmative action doctrine. Patrick's first innovation is his argument that " faculty diversity" is a sufficiently compelling interest to justify firing people on the basis of race. In a letter to the editors of TNR, Patrick argues that "many Supreme Court justices" have described faculty diversity as a "laudable" goal. (See Correspondence, page 6.) He neglects to mention, however, that no Supreme Court majority has ever endorsed this view.

Patrick is also less than frank when he claims that "nothing in the controlling case law" clashes with his position. In fact, Justice Lewis Powell's plurality opinion in the 1986 Wygant case explicitly rejected the argument that white teachers could be fired over black teachers "to provide role models for minority children." Carried to its logical conclusion, Powell said, the idea that black students are better off with black teachers could lead to the very system the Court rejected in Brown v. Board of Education. More pointedly still, Powell objected that the role model theory allows the board to use racial preferences "long past the point required by any legitimate remedial purpose."

Patrick dismisses Justice Powell's concerns. Citing the dissenting opinion in Wygant by John Paul Stevens, he argues that there are a number of reasons why schools that have never discriminated in the past might seek racially diverse faculties. "It is one thing for a white child to be taught by a white teacher that color, like beauty, is only skin-deep,'" wrote Stevens. "It is far more convincing to experience that truth on a day-to-day basis." The argument may be intelligible for social studies teachers; but Stevens never suggested that students would benefit from being exposed to racial diversity in typing and accounting classes.

In any event, the "faculty diversity" argument is a red herring, since Justice Powell said explicitly that it is not compelling enough to justify firing people on the basis of race. Patrick argues that the Piscataway school board did not "unnecessarily trammel" Taxman's rights. But in fact, the board violated Taxman's right to have the same chance of keeping her job as anyone with equal seniority. The coincidence that Taxman and Williams were hired on the same day doesn't change the constitutional principle: in a random, colorblind system, each would have had a 50 percent chance of keeping her job. "We're not taking a position on whether this is constitutional," Patrick told me on the telephone; but his brief takes a very strong position. It explicitly rejects Justice Powell's central insight in Wygant: "While hiring goals impose a diffuse burden, often foreclosing only one of several opportunities, layoffs impose the entire burden of achieving racial equality on particular individuals, often resulting in serious disruption of their lives. That burden is too intrusive."

Patrick tries to minimize Taxman's injury with a series of evasions. Race, he says, "played no greater weight in the board's decision than any other qualification." This is ludicrous: as the Justice Department acknowledged before it switched sides in the case--it had originally supported Taxman rather than Williams--race was the only factor that tipped the scales. "Let me assure you once again, Sharon, that this board action is not related to any assessment of your professional performance," wrote Gordon Moore, the director of staff personnel, when he fired her. Patrick also argues that Taxman's injury was only "temporary," because she was rehired in 1990. (She was fired again in 1991, and then rehired the following year.) In fact, the judge calculated her lost income and benefits as $134,000--hardly a " temporary" loss.

There's something a little glib, finally, about Patrick's characterization of Supreme Court precedents. He invokes William Brennan's opinion in the Metro Broadcasting case, which said that Congress could set aside fcc licenses for minority broadcasters. But Brennan went out of his way to stress that the program didn't disturb anyone's vested rights or legitimate expectations. Patrick also skims over the Johnson and Weber cases, which say that racial preferences are permissible only when they're temporary, and intended to remedy the effects of past discrimination, or to eliminate a " manifest imbalance" in the racial makeup of a workplace as a whole. The Taxman case meets none of these criteria.

Patrick's attempt to rewrite the law of affirmative action, rather than prudently withdrawing the government's brief in this explosive case, must be ranked among the most peculiar decisions of the Clinton Justice Department. " We do not support quotas," Patrick said when he filed the Taxman brief, but " affirmative action is a different animal." But when whites are fired because of their race, the euphemistic distinction between affirmative action and reverse discrimination becomes impossible to sustain. As Samuel Issacharoff of the University of Texas argues, if there is any point at which liberals should balk at racial preferences, the Taxman case presents it squarely.

The University of Texas case is the most important challenge to racial preferences in university admissions since the Bakke decision in 1978, which struck down a rigid set-aside program at the u.c.-Davis medical school; and it may give the Court another chance to re-examine the boundaries of its affirmative action doctrine. While Justice Powell in Bakke coyly refused to look behind the curtains of the Harvard College affirmative action program, which he praised as a model, Judge Sam Sparks of the U.S. district court in Austin unblinkingly revealed the raw data--grade point averages and law school admission test scores--that separate white, black and Mexican-American applicants to the University of Texas. The numerical gap between the groups is so stark that it undermines the central premise of Bakke: that race may be used as a tie-breaker in admitting similarly qualified candidates, but that using race as a decisive consideration is "discrimination for its own sake."

With Hopwood v. Texas, history has made an awkward circle. In 1950 Thurgood Marshall represented Heman Sweatt, who sued the University of Texas law school because it barred blacks; four decades later four white plaintiffs are suing the same law school because it is too welcoming to blacks. In the 1940s Texas tried to resist integration by setting up a makeshift "separate but equal" law school in the basement of the state capitol; in the 1990s Texas tried to guarantee integration by imposing separate admissions standards for blacks and whites. And there are plenty of ironies from the decades in between. As late as 1980 the Department of Health, Education and Welfare found that Texas had failed to eliminate the vestiges of segregation. When the university pledged to increase the numbers of black and Hispanic graduate students in 1982, Ronald Reagan's assistant secretary of education, Clarence Thomas, rejected the plan. The numerical goals, Thomas said, were simply too low to satisfy Title vi of the Civil Rights Act of 1964. Texas, in short, finds itself trapped between what federal law requires and what the Constitution forbids.

The plan that Thomas eventually endorsed is the genesis of the plan that Cheryl Hopwood, a white, working-class graduate of a Texas junior college, challenged. Under the Texas plan, which was developed by moderate and conservative faculty members, the law school pledged to admit blacks and Mexican-Americans in proportions approaching their graduation rates from Texas colleges. Each class, the admissions committee decided, should be approximately 5 percent black and 10 percent Mexican-American.

The problem, put starkly, is that Texas cannot begin to achieve its goals without adjusting its admissions standards. The admissions committee relies heavily on a number called the Texas Index, which is a weighted average of every applicant's LSAT score and undergraduate gpa, and which has proved-- both sides concede this--to be a reliable predictor of first-year grades. The school, with its first-rate faculty and low tuition, is highly selective: in 1991 white students admitted to Texas had a median lsat in the ninety-third percentile and an average GPA above 3.5. In 1991 white Texas applicants with Texas Index scores above 199 were presumptively admitted, while those with scores below 192 were presumptively rejected.

The university's lawyers began the trial by frankly conceding that, without racial preferences, only one out of the 280 black applicants to Texas in 1992 had a score high enough to be presumptively admitted. Out of all the minority applicants to all the law schools in the country, only 289 blacks, and ninety- six Mexican-Americans, had scores high enough to put them in the discretionary zone for white applicants to the University of Texas; and only eighty-eight blacks, and fifty-two Mexican-Americans, in the country had scores higher than the median for white students at the University of Texas. If the law school had based its 1992 admissions on a strictly colorblind standard, the entering class of 500 students would have included, at most, nine black and eighteen Mexican-American students, all of whom were being courted by the most prestigious law schools in America.

Faced with the prospect of returning to a nearly all-white class (which would have been illegal under Title VI), the University of Texas balked. It established a separate admissions committee to consider applications from blacks and Mexican-Americans; and it imposed lower scores for admission and rejection. The presumptive admission score for the preferred minorities dropped to 189, three points lower than the presumptive rejection score for whites. Consequently, a score in the range of 189 to 192 was a presumptive rejection for a white applicant but a presumptive admission for a black or Mexican-American applicant.

The gap between the scores of black and white applicants is worrying on several levels. And it exposes the three central fallacies underlying the Bakke opinion: that race can be a plus factor rather than a decisive factor in admissions decisions; that the need for racial preferences will soon disappear; and, most implausibly of all, that only institutions with a history of discrimination are permitted to avail themselves of racial preferences today.

To be consistent with Powell's opinion in Bakke, Judge Sparks said, the law school had to compare black applicants to white applicants at every stage, rather than considering both sets of applications on racially segregated tracks. The procedural requirement comes from Powell's opaque notion that racial preferences in universities "must be limited to seeking the educational benefits that flow from having a diverse student body and to addressing the present effects of past discriminatory practices." But what Powell meant by the educational benefits of a diverse student body has never been clear. He seemed to be saying that a university could use race as a rebuttable presumption that a black applicant would contribute to intellectual diversity; but that admissions officers would have to decide, on a case-by-case basis, whether a given white applicant would bring more diversity to the mix. For example, Powell seemed to suggest, an Italian-American farmboy who had overcome great hardships might have a legitimate constitutional objection if he were rejected in favor of a son of the black middle class.

Accurately applying the vague standards of Bakke, Judge Sparks held that the lack of individual comparison between white applicants and minority applicants makes the Texas procedure unconstitutional. After Bakke was decided in 1978 the law school initially had concluded that its use of two separate tracks for minority and white applicants was unconstitutional, and decided to merge the separate committees into a single unit. But during the 1980s, Texas, like many other top law schools, such as Stanford, got sloppy and decided that it would be more efficient to revive the separate minority admissions committee once again. The persistence of separate minority committees, despite Powell's explicit prohibition, shows how little influence Bakke has had on affirmative action as it is actually practiced.

After the Hopwood suit was filed, however, Texas administrators became more careful about appearances. "When one gets sued in federal court, it catches one's attention," the admissions director testified. For the 1995 entering class, the law school has eliminated the minority subcommittee and the presumptive admission and rejection scores. All admissions decisions will be made by a single committee of three faculty members. The new, unified admissions procedure, Judge Sparks emphasized, is perfectly constitutional under Bakke, which is why the law school is properly treating the decision as a victory.

But the victory is incoherent. The only practical effect of Bakke, as correctly applied by Judge Sparks, is to punish schools that have the courage to be honest. The court is saying, in effect, that Texas can rely on the same strong racial preferences that it did before, as long as it doesn't overtly adopt separate admissions standards. But it's hard to see why it should be constitutional to do covertly what is unconstitutional when done openly. Furthermore, by making its separate admissions standards explicit, rather than inviting three faculty members to apply racial preferences in secret, Texas was trying to minimize the possibility for abuse.

The procedural distinction on which Justice Powell and Judge Sparks place so much weight--the distinction between separate admissions committees and a single admissions committee--collapses in light of the admissions data. The gap between the scores of black and Mexican-American candidates and other candidates is so persistent that it is simply not possible to compare minority candidates with white candidates on a case-by-case basis, and to weigh race and ethnicity in "competitive consideration" with other, nonracial qualities. In order to admit meaningful numbers of black and Mexican-American students, race must be used not as a "plus factor" but as the decisive factor, and a black professor's son will have to be preferred to a less privileged white working-class applicant--like Cheryl Hopwood--in case after case. The Hopwood data suggest, therefore, that the consequence of a purely class-based affirmative action policy would be virtually all-white colleges and law schools.

The Hopwood case highlights yet another fallacy of Bakke: Powell's notion that racial preferences would be short-lived. In a new biography of Powell, John C. Jeffries Jr. reveals that during the Court's private conference about Bakke, John Paul Stevens and Powell agreed that racial preferences might be acceptable as a temporary measure, but not as a permanent solution. Perhaps, Stevens suggested, blacks would not need preferences for much longer. At this point Thurgood Marshall broke in. Racial preferences would be necessary, he said, for another 100 years. Powell was speechless. Afterward, Powell's clerk speculated that if Marshall had predicted ten years rather than 100, Powell might have provided a fifth vote for the liberal's unqualified approval of racial preferences. Powell simply recoiled at the prospect of generation after generation of racial quotas.

But the evidence presented at the Hopwood trial suggests Marshall was right and Powell was engaging in wishful thinking. Sixteen years after Bakke, the gap between the scores of white and black applicants remains so persistent that it is hard to believe things will improve anytime soon. In an amicus brief submitted to the Bakke Court, the Association of American Law Schools pointed out that in 1976, 20 percent of white law school applicants, but only 1 percent of black and 4 percent of Chicano applicants, had LSAT scores of 600 or more and GPAs of 3.25 or higher. Without strong racial preferences, therefore, the number of blacks students admitted to law schools in 1976 would have been about 1 percent of the entering class. This is roughly the same as the percentage actually admitted to law schools in 1964, and it is similar to the University of Texas's estimate of the percentage that would have been produced by a colorblind admissions process in 1994. "Certainly, an indefinite program would violate the equal protection clause," Judge Sparks declared confidently; but in light of the lack of progress after thirty years, it's hard to share his optimism that racial preferences will end anytime soon.

The Hopwood case exposes one last fallacy of Bakke: the impossibility of measuring and addressing, with any precision, the present effects of past discrimination. Faced with a choice between two coherent theories of affirmative action, Powell chose neither. On the one hand, William Rehnquist, now joined by Antonin Scalia and Clarence Thomas, argues that the Constitution is colorblind and that all racial classifications are unconstitutional. Rehnquist, Thomas and Scalia would allow racial preferences only in the narrowest circumstances, as compensation from an institution that is guilty of discriminating to an individual who has been discriminated against. Although this theory is rhetorically appealing, it is flamboyantly inconsistent with the original intentions of the framers of the Fourteenth Amendment, who considered, and rejected, language that would have prohibited racial classifications in all circumstances.

The most liberal position is also unsatisfying. According to William Brennan, Thurgood Marshall and Harry Blackmun, the Constitution forbids racial classifications only when they reflect prejudice or contempt for a particular group. Virtually all racial preferences designed to help blacks are permissible, according to this view, because they're not based on stigmatizing prejudice for whites, but instead on a desire to remedy the effects of discrimination in society as a whole. The weakness of this theory is obvious enough: not every form of discrimination is based on contempt or loathing. On the contrary, as Ruth Bader Ginsburg has argued, discrimination against women has long been rooted in the paternalistic idea that women need special and benevolent protection. And plenty of economic data suggest that most racial discrimination today is motivated not by hatred but by the administrative efficiency of generalizing from the average performance of a group to the likely performance of one of its members.

Instead of siding with either the liberal or conservative theories of affirmative action, Powell tried to mix and match them in cases after Bakke. Borrowing the individual responsibility position from the conservatives, Powell said institutions could engage in affirmative action only if they themselves had been guilty of discrimination. And borrowing the group oppression position from the liberals, Powell said the minorities could benefit from affirmative action even if they themselves had not been victims of discrimination.

The Texas case shows the illogic of this compromise. It's beyond dispute that the University of Texas discriminated notoriously against blacks in the past. But 1994 is not 1946; and the law school that rejected Cheryl Hopwood is a very different place than the law school that rejected Heman Sweatt. And despite the historical record, Judge Sparks's efforts to link the beneficiaries of Texas's current racial preferences with the victims of Texas's past racial discrimination seem particularly strained. He emphasizes, for example, that the parents of black applicants to the University of Texas today were discriminated against by the state of Texas a generation ago. This local discrimination, he suggests, may account for the persistent gaps between the scores of white and black applicants to the university. But the argument falters on an inconvenient fact: Texas admits most of its black law students from the pool of out-of-state applicants, who can hardly claim to have been handicapped by the legacy of discrimination in Texas. The existence of a national applicant pool tends to undermine any claims that the university is engaging in geographically precise compensation for its sins.

The Hopwood case, in short, exposes the implausible and increasingly esoteric foundations on which the law of affirmative action now rests. To accept the intellectual framework of Bakke, you have to accept a series of premises that are hard to reconcile with the cold data: that race can be a plus factor rather than a decisive factor in admissions decisions; that the need for racial preferences will soon disappear; and, most exotically of all, that institutions that have discriminated in the past can atone for their sins by projecting the effects of their discrimination into the future. Whatever its value as an act of public relations, Justice Powell's opinion in Bakke clearly has failed as an act of constitutional interpretation.

If the Court chooses to use the Hopwood case as an opportunity to re- examine Bakke in the next few years, there is a more convincing theory. As Andrew Kull argues in The Colorblind Constitution, the framers of the Fourteenth Amendment did not intend to forbid all racial preferences, but only those that were unreasonable or unrelated to legitimate public purposes. They were especially concerned about laws like the Black Codes that extended unequal punishments, and unequal benefits, to blacks and whites. For better or for worse, therefore, fidelity to constitutional history supports, rather than undermines, the notion that the Supreme Court should pick and choose among racial classifications, upholding the reasonable ones and striking down the unreasonable ones. And in trying to refine the vague standard of reasonableness, the Court might indeed focus on whether the burdens of a racial classification are focused unfairly on particular individuals.

This weak colorblind principle would probably permit racial preferences along the Texas model, in college and graduate school admissions, because their burdens fall on a wide and diffuse group. Between the ninety-fourth percentile, which is the median for white applicants, and the eighty-ninth percentile, which is the median for black applicants, the University of Texas has to pass over 1,000 whites for every black applicant that it admits. The disappointment of the rejected whites should not be minimized; but law schools, like undergraduate colleges, have never pretended to cultivate academic distinction alone: they seek to train not only legal scholars but also politicians and local worthies.

Moreover, unlike the highly qualified Alan Bakke, who clearly would have been admitted to the University of California in a colorblind world, the University of Texas refused to concede that the four plaintiffs in the Hopwood case would have been admitted on the merits. All four had adequate but unexceptional academic records that put them squarely in the broad middle range of the Texas applicant pool; and most were admitted to less prestigious law schools. Since Texas has to reject hundreds of applicants like Cheryl Hopwood, it's hard to argue that she is being singled out unfairly to bear the full weight of the racial preferences. Perhaps a more qualified plaintiff could have argued more convincingly that her expectations were frustrated.

Once racial preferences are extended beyond college and graduate school admissions, however, the burdens become far less diffuse. Consider the case of faculty hiring. The number of positions are far more limited; and when a certain number of them are set aside only for blacks, there are especially heavy costs on the rejected whites, who have invested an extraordinary amount of time in specialized preparation and are likely to have many fewer alternatives than students applying to graduate schools. There also seems far less justification for using group membership as a stand-in for intellectual diversity than in student admissions, where the actual views of applicants are hard to discern. And while even mediocre law students may go on to be successful lawyers or politicians, the standards for scholarly success are far more precise and objective. All this suggests that the "faculty diversity" argument that Deval Patrick is pressing in the New Jersey case is constitutionally as well as educationally questionable.

As the over-reaching of Patrick demonstrates, the Supreme Court's careful distinctions between permissible and impermissible affirmative action are being eroded or ignored in the wake of the political impulse to achieve proportionate racial representation in every sub-department of society. This impulse, which comes perilously close to discrimination for its own sake, refuses to engage the real question in the affirmative action debate--namely, what are the justifications for favoring members of one race over another in very different spheres and contexts? The only hope of avoiding a descent into pure racialism is to admit candidly that racial preferences have costs, and to set precise and legally enforceable boundaries. That project, thankfully, has now begun.