William G. Bowen and Derek Bok
William G. Bowen and Derek Bok; drawing by David Levine

1.

For over thirty years America’s best universities and colleges have used race-sensitive admissions policies to increase the number of their black, Hispanic, Chicano, Native American, and other minority students.1 Conservative writers and politicians have attacked this policy of “affirmative action” from its inception, but the policy is now in the greatest danger it has yet faced—on two fronts, political and legal. In 1995, by a 14 to 10 vote, the regents of the University of California declared that race could no longer be taken into account in admissions decisions at any of the branches of that university. In 1996 California voters approved Proposition 209, which ratifies and broadens that prohibition by providing that no state institution may “discriminate against, or grant preferential treatment to, any individual or group on the basis of race, sex, color, ethnicity or national origin in the operation of public employment, public education, or public contracting.”2

The effect of the regents’ decision was immediate and, in the views of many of the university’s faculty, disastrous: the Boalt Hall law school at Berkeley—the state’s premier public law school—had enrolled an average of 24 black students each year for the last 28 years. In 1997 it enrolled only one, and he had been admitted the previous year but had deferred entering.3 The political campaign against affirmative action will continue, encouraged by the success of the California initiative, in other states. A similar prohibition will be presented to Washington voters next November, and others are certain to follow.

The second danger may be even more menacing. In 1978, in the famous Bakke case, the Supreme Court in effect ruled that race-sensitive admissions plans do not violate the Fourteenth Amendment of the United States Constitution, which declares that “no state may deny any person equal protection of the laws,” so long as such plans do not stipulate fixed quotas for any race or group, but take race into account only as one factor among others.4 In 1996, however, in the Hopwood case, the Fifth Circuit Court of Appeals declared the admissions program of the University of Texas Law School at Austin unconstitutional, and two of the three judges who made up the majority in that case declared that the Bakke rule had been overruled, even though not expressly, by more recent Supreme Court decisions.5

The immediate consequences of the Fifth Circuit decision were, once again, dramatic: though the Texas Law School had enrolled 31 black students in 1996, it could enroll only 4 in the following year. The Supreme Court declined to review the Fifth Circuit’s decision, which therefore stands as law in Texas and the other states of that circuit. Last October, the Washington, D.C.-based Center for Individual Rights, which had spawned the attack on the University of Texas in the Hopwood case, filed a similar suit in Michigan, arguing that the University of Michigan’s admissions program is also unconstitutional, and similar suits are expected in other states. Sooner or later the Supreme Court will be required to take some such case for review, and if the Court does overrule or substantially restrict Bakke, affirmative action henceforth will be crippled across the country. Without a constitutional amendment or another change of heart in the Court, not even a shift in the political climate could bring it back. 6

Much of the political and legal attack on affirmative action has centered on its consequences: critics say that it has lowered educational standards by admitting students who are unqualified to benefit from the education they receive, and that it has exacerbated rather than relieved racial tension. It is therefore opportune that the first comprehensive and statistically sophisticated examination of the actual effects of thirty years of affirmative action in American universities has just been published. The Shape of the River, by William G. Bowen, who was president of Princeton University, and Derek Bok, the former president of Harvard, analyzes an enormous database of records, called the College and Beyond (C&B) database, which was compiled by the Mellon Foundation, of which Bowen is the president, over four years.7

That database contains information about each of more than eighty thousand undergraduates who matriculated at twenty-eight selective colleges and universities in 1951, 1976, and 1989; these institutions are representative of the elite schools that have used affirmative action, and they range, in their selectivity in admitting students, from Bryn Mawr and Yale to Denison and North Carolina (Chapel Hill). 8 In the case of the 1976 and 1989 cohorts, the database records the undergraduates’ race, gender, high school grades, SAT scores, college majors and grades, extracurricular activities, any graduate or professional school record, and, for many, family economic and social background. It also presents information about the post-university experience of all those in the sample who answered detailed questionnaires sent out when the database was being compiled. An unusually high number of those surveyed did so—80 percent for the 1976 and 84 percent for the 1989 cohorts.

Advertisement

Bowen, Bok, and their colleagues have used advanced statistical techniques to analyze, so far as possible, the distinct impact of each of the great variety of variables the study isolates. They have done so in an attempt to chart the consequences affirmative action has actually had, over its now substantial career, for individual students and graduates, for their colleges and universities, and for race relations in the country as a whole. Their book is an extremely valuable sociological study quite apart from its specific findings about affirmative action, and it offers, in detailed appendices, a clear description of the complex statistical techniques it employs.

The River study has limitations, of course, which its authors are careful to acknowledge. A statistical survey, no matter how substantial its data or careful its techniques, is not a laboratory experiment, and though the authors show considerable ingenuity in finding and using controls and other checks on their conclusions, certain conclusions, as they point out, inevitably include some surmise. The study is confined to affirmative action in higher education, and its results may have little bearing on the effects of racial classifications for other purposes—in hiring, for example, or in awarding opportunities to minority-owned businesses. Most university affirmative action plans are designed to increase the enrollment of a variety of minority groups, but, except for some discussion of Hispanic students, the study presents and analyzes mainly data about black students and graduates. The institutions in the C&B list are representative of highly selective universities and colleges, moreover, and the study’s findings may not hold for less selective sectors.

The authors have not been able to answer all the questions that their data raise. They concede that they are unable fully to explain, for example, the particularly worrying fact that black students as a group underperform in college grades compared to white students in the same institution who had the same SAT scores and other academic qualifications.9 None of these limitations compromise the force of the conclusions the study reaches, however, and many of these conclusions, as we shall see, flatly contradict premises and assertions that have become staples of the affirmative action debate in recent years.

In order to gauge the importance, and the limits, of the River study, we must take care to distinguish the two main strands of that debate.10 The first is an issue of principle: Is affirmative action for blacks unfair because it violates the right of every applicant to be judged on his or her individual merits? The second is a matter of policy or practical consequence: Does affirmative action do more harm than good, because it enrolls some blacks in studies beyond their capacities, or stigmatizes all blacks as inferior, or makes the community more rather than less conscious of race? These two questions are connected, because many people think that affirmative action is fair if it does substantial good, either for those it is intended to benefit or for the community as a whole, but unfair if it does not, because the damage it does to the admissions prospects of other applicants (who include not only whites but other minorities, like Asian-Americans, whose test scores as a group are relatively high) is then pointless. The questions are nevertheless independent, however, because race-sensitive admission policies may be unfair to rejected applicants or to blacks as a group even if they achieve exactly what they are designed to achieve.

The practical question has been the more sharply debated in recent years. Advocates of affirmative action often insist that race-sensitive policies of different sorts are essential, in the short run, if we are to have any genuine hope of eradicating or diminishing the impact of race in a longer term. The most prominent critics of such programs, both white and black, reply that affirmative action has been in every way counterproductive: that it has “sacrificed” rather than helped the blacks admitted to the programs, perpetuating a sense of black inferiority among both whites and blacks themselves, and promoting black separatism and a race-conscious society rather than black integration and a genuinely colorblind community.11

Both advocates and critics rely, however, on only sketchy factual evidence to support their large claims. They cite newspaper accounts of isolated incidents of interracial cooperation—or of racial disharmony—in universities. They rely on introspective or anecdotal reports of successful blacks who credit affirmative action with having given them a chance, or blame it for stigmatizing, insulting, or cheapening them. Most of all they appeal to supposedly common-sense assumptions about how whites and blacks “must” or “may” feel or react.

It would be wrong to blame proponents and critics for relying on such thin evidence for their large claims, however, because though some excellent studies have been produced on particular issues—Bowen and Bok refer to several of these—there have been few studies of the scope needed. That is why The Shape of the River is so important: it offers much more comprehensive statistics and much more sophisticated analysis than has been available before. It has already made a considerable impact: its findings have been widely reported and discussed in the press.

Advertisement

We must be careful, of course, not to accept even such an apparently imposing study uncritically. The statistical analysis it offers may later be shown to be flawed. Or even more comprehensive studies may later be published that refute some or all of its main conclusions. But it would be surprising and shaming if The Shape of the River did not sharply improve the character of the long political and legal debate. Its analysis has significantly raised the standard of argument. Impressionistic and anecdotal evidence will no longer suffice: any respectable discussion of the consequences of affirmative action in universities must now either acknowledge its findings or challenge them, and any challenge must match the standards of breadth and statistical professionalism that Bowen, Bok, and their colleagues have achieved.

2.

The two former university presidents are cautious and judicious scholars, and they are careful, as we have already noticed, to limit their claims to what the evidence justifies. Yet they have no doubt as to the most important result of their study.

If, at the end of the day, the question is whether the most selective colleges and universities have succeeded in educating sizable numbers of minority students who have already achieved considerable success and seem likely in time to occupy positions of leadership throughout society, we have no problem in answering the question. Absolutely….

Overall, we conclude that academically selective colleges and universities have been highly successful in using race-sensitive admissions policies to advance educational goals important to everyone.

We cannot evaluate that overall conclusion, however, without noticing the large variety of distinct findings on which it is based. I cannot, of course, adequately summarize all of these, or describe the often ingenious techniques used to obtain and defend them. I shall concentrate instead on those findings that seem most germane to the political and legal argument.

Does affirmative action accept unqualified blacks? In 1951, there were a total of 63 blacks—only an average of 0.8 percent per school—in the entering classes of the nineteen C&B institutions for which records are available. In 1989, blacks made up 6.7 percent of the entering class in all C&B schools, and 7.8 percent in the most selective group. Much of this increase must be attributed to race-sensitive admissions. Bowen and Bok estimate, by factoring out the impact of other variables, that a race-neutral admissions policy would have reduced the number of black entrants to between 2.1 and 3.6 percent for all schools in the study (depending on different assumptions about how many of the blacks admitted decided to attend). The decline would be greatest at the most selective schools.

It would be a serious mistake, however, to assume that these “retrospectively rejected” blacks were unqualified for the education they received.12 White applicants to the schools did have significantly higher test scores, as a group, than the black applicants. The difference narrows sharply, however, when we compare the scores of blacks who enrolled with the lowest decile of the scores of whites admitted: a study of law school admissions showed a difference, in LSAT scores, of only 10 percent. The difference in scores between white and black applicants is better explained, in any case, by the extraordinary improvement in recent decades in the academic qualifications of white applicants to selective schools—Bowen and Bok call these applicants “spectacularly” well qualified—than by any assumption that the black applicants were not qualified. Five of the C&B schools, which were otherwise representative of them all, retained full information about all their 1989 applicants, and more than 75 percent of the black applicants had higher math SAT scores, and more than 73 percent had higher verbal SAT scores, than the national average of white test-takers. The professional success of black graduates from C&B schools, discussed below, in itself rebuts any assumption that these blacks were, as a group, unqualified for their education.

It is a striking fact, moreover, that the average SAT scores of black entrants to the most selective schools in 1989 was higher than the average of all matriculants in the same institutions in 1951. As the authors observe, middle-age and elderly graduates should reflect on that fact before insisting that blacks accepted through affirmative action programs are unfit for their universities. It is also striking that the test scores of the retrospectively rejected blacks—those who the study predicts would not have been admitted had race-neutral tests been used—were not much different from the scores of the blacks who would have been accepted anyway. In the five schools just mentioned, the average SAT score of the former was 1145 and of the latter 1181. So while abolishing affirmative action would very greatly decrease the number of blacks who attended selective schools, it would not much improve the average scores of those who did.

Do blacks waste the opportunity they are offered? Would they be better off in less demanding institutions where they would “fit” better? In their recent book, America in Black and White, Stephan and Abigail Thernstrom report that at some three hundred “major” colleges and universities between 1984 and 1987 the drop-out rate was 43 percent for white students and 66 percent for black students. They quote an article in The Journal of Blacks in Higher Education describing this fact as “disastrous,” and they cite the same fact as justifying their own conclusion that

Affirmative action admissions policies…did work to increase enrollments, but if the larger aim was to increase the number of African Americans who would successfully complete college, preferential policies had disappointing, even counterproductive, results.

But (as would have been evident had they quoted the two sentences immediately following the Journal sentence from which they did quote13 ) their argument is highly misleading. The figures they cite are drawn from records of 301 schools in the National Collegiate Athletic Association (NCAA)’s Division I, and though these schools do include some institutions that are plainly “major,” they include a great many that are equally plainly not.14 Though the drop-out figure for blacks in the 301 schools taken together is indeed much greater than that for whites, even the figure for whites is so alarmingly high that it suggests that affirmative action cannot be the main problem. The Thernstroms provide no evidence, in any case, as to how many of those 301 institutions practice affirmative action. (Bowen and Bok refer to a study that suggests that only the top 20 percent of all four-year institutions do.) If many do not, the difference in graduation rates would be unaffected even if that practice were eliminated elsewhere.15

The River study is much more discriminating and useful than the study cited by the Thernstroms. It shows that the black dropout rate in the C&B schools is small by national standards: 75 percent of the 1989 black C&B cohort graduated from the school they entered within six years compared with 59 percent of white students at the 301 schools that belong to Division I of the NCAA. Even so, the black graduation rate is lower than the white rate even at the C&B schools—by 11 percent in the 1989 group.

Some of the gap can be explained by obvious factors—blacks come, on average, from worse-off households, and are more likely to be forced to leave college for financial reasons, for example—but not all of it can be explained in that way. However, the black graduation rate, for blacks at each level of SAT scores, is progressively higher at more selective schools within the C&B group. Even blacks with the lowest SAT scores (1000 and less) graduated at higher rates when they attended more selective and demanding schools within the C&B group, where the difference between their scores and the average of other classmates was greater.16

Bowen and Bok consider a variety of explanations for these results. The more selective schools are also richer schools with more resources available for scholarships and other forms of student aid. Since the economic value of a college degree increases with the prestige of the school, moreover, all students have a greater financial incentive to remain in a more selective school. Such schools also have the resources to set up “mentoring” and other programs designed to help blacks with less adequate prior training in study and research skills to cope, and the study demonstrates the value of that help in other ways. In any case, these findings are of great importance because they seem to refute the “fit” hypothesis that the Thernstroms and others defend: that blacks would graduate in higher numbers if affirmative action were abolished and they attended less selective and competitive schools as a result.

That dismal hypothesis is also contradicted by much of the other data and analyses in the River study. Blacks as a group do not suffer, financially or otherwise, when they have attended a more selective school. At each SAT level, blacks earn more after having attended a more selective school, and report themselves as more satisfied with their careers. Nor do most blacks who attended more selective schools report any discomfort or regret when they reflect on their undergraduate experience, or otherwise suggest that they were “sacrificed” by affirmative action programs. Black graduates of C&B schools report satisfaction with their university experience at the same very high rate—in the 1989 cohort, 91 percent declared themselves either “very” or “somewhat” satisfied—as that of all other students. At every SAT level, moreover, blacks who attended the more selective C&B schools, where the gap between their SAT scores and the norm for the school was greater, reported a higher degree of satisfaction, which is the contrary of what the “fit” hypothesis claims. That hypothesis has played a very prominent part in the affirmative action debate in recent years: the River study—at least until it is challenged by evidence rather than shoddy scholarship or anecdote—should put an end to its role.

3.

Has affirmative action produced, as hoped, more successful black businessmen, professionals, and community leaders? If we measure success by income, it certainly has. Black male graduates from the 28 C&B schools in the 1976 cohort found less well paid jobs than their white classmates who had parallel test scores and college or professional school grades,17 and that sad fact is alone enough to refute any suggestion that racism has disappeared from our economy. But black C&B graduates earn considerably more than the average black with a BA degree: black women in the 1976 cohort earn on average 73 percent, or $27,200, more than the average of all black women BAs, and black men in the cohort earn 82 percent, or $38,200, more than the average of all black male BAs.

Several factors help to explain those differences: blacks in the C&B cohort had on average better test scores, and came on average from higher socio-economic backgrounds, than black graduates generally. But the selectivity of the school they attended is nevertheless an important part of the story: the more selective a black graduate’s school, the higher his or her anticipated income, even holding all else equal. As the authors point out, “While graduation from a selective college hardly guarantees a successful career, it may open doors, help black matriculants overcome any negative stereotypes that may still be held by some employers, and create opportunities not otherwise available.”

The income advantage for blacks of attending a more selective school is revealing, not only because it shows that these admissions policies have helped, not harmed, their intended beneficiaries, but because it provides further evidence, as I said earlier, that the beneficiaries were fully qualified to profit from the education they received. Black graduates of the C&B institutions would not get or retain their well-paid positions in business, law, and medicine if their ability and education did not enable them to earn those salaries, in competition with others, by genuine contribution to their firm or profession.

We cannot measure the success of affirmative action by concentrating only on salaries, however, or only on the increased numbers of black executives, lawyers, doctors, and professors the study shows that the policy has helped to produce. Affirmative action, Bowen and Bok write, “was also inspired by a recognition that the country had a pressing need for well-educated black and Hispanic men and women who could assume leadership roles in their communities and in every facet of national life.” In this respect, too, the study reports success. Nationwide, black and white college graduates are equally likely to participate in various kinds of civic and professional groups. But among C&B graduates, black men are strikingly more likely to do so, especially in those activities that seem most important to black communities, including social service, youth clubs, and elementary and secondary school organizations. Almost twice as many blacks as whites from the 1976 cohort have participated in community service organizations, for example. In every type of activity cited, moreover, black males were more likely than white ones to hold leadership positions.

These findings are particularly interesting in view of the widespread fear, voiced by Henry Louis Gates and Orlando Patterson,18 among others, that educated middle-class blacks will take up new lives at a distance from the concerns of the larger black community. That fear remains, but the study’s statistics offer hope.19 “The fact that this group is consistently providing more civic leadership than its white peers indicates that social commitment and community concerns have not been thrown aside at the first sign of personal success.”

Does racial diversity in a university’s student body help to break down stereotype and hostility among the students, and, if so, does the benefit endure in post-university life? Or does racial preference generate animosity on campus, and a backlash that increases rather than decreases racial tension in the community generally? Critics cite well-publicized incidents of racial hostility on campus, and practices like “black tables” in university dining halls, to suggest that racial diversity has done nothing to reduce racial isolation and hostility, and may even have exacerbated it.

It is difficult to test attitudes and emotions, but the Bowen-Bok study has nevertheless produced impressive statistics about them. The River’s questionnaire asked graduates in the cohorts it studied how important they thought race relations are, whether they thought their undergraduate education contributed to improving their own relations with other races, what interactions they had with members of other races as undergraduates, and whether they thought their university’s admissions policies had emphasized racial diversity too much, too little, or about in the right degree. Answers to multiple-choice questions can only partly capture the complexity of personal experience and opinion, but the results are nevertheless telling.

Predictably, more blacks than whites thought knowing people of other races particularly important. In the 1976 cohort, 45 percent of whites thought it was “very important” to get to know people of “different beliefs” and only 43 percent to know people of different races, while 74 percent of blacks in that cohort thought the latter very important and only 42 percent the former. The number of both whites and blacks who thought race relations very important increased in the 1989 cohort, however—by a modest 2 percent for blacks but by a dramatic 13 percent for whites. (For those white graduates who occupy leadership positions in civic organizations, the increase was even greater—to 59 percent.)

When asked to rank the value of their college experience in improving their ability to “get along with” people of other races, 46 percent of white and 57 percent of black respondents in the 1976 cohort rated that value at either 4 or 5 (5 indicated “very important”), and 18 percent of white and 30 percent of black respondents rated it at 5. In the 1989 cohort, these figures had jumped: 63 percent of the white respondents ranked the value at either 4 or 5, and 34 percent at 5; 70 percent of blacks ranked it either 4 or 5, and 46 percent at 5. These are significant differences between the two cohorts: the authors speculate that students might have become more aware of the importance of racial interaction by 1989, or that universities might have become more adept at creating an environment that facilitates that interaction, or, most likely, both.

It is important to try to confirm these subjective judgments of the importance of diversity by seeing how far they were actually reflected in behavior, particularly in view of the widespread belief that student groups are often racially isolated from one another. The study asked its respondents whether they had “known well” two or more students different from themselves in each of several categories, including geographical and economic background, general political orientation, and race. Even though black students made up less than 10 percent of the student body (except in one school, where they were 12 percent) 56 percent of the white respondents in the 1989 cohort said they knew two or more black students well. (Eighty-eight percent of the black students said that they knew two or more white students well.) The authors conclude that even though there was undoubtedly some self-segregation of races on these campuses, in clubs and dining arrangements, for example, “the walls between subgroups were highly porous.”20

The further question the study put to respondents, about whether they approve of the level of concern for racial diversity that they believe their institution has shown, is doubly important, because it helps us to assess not only the value graduates place on that diversity in their own lives, but also the degree to which they, as members of the general public, resent racial preferences. Most of the whites in the 1976 cohort think that their institution places too much emphasis on alumni concerns, intercollegiate athletics, and faculty research. But only 22 percent of them think it places too much emphasis on racial and ethnic diversity, compared to 39 percent who think it places too little. Blacks in that cohort agree with their white classmates about alumni, athletics, and research, but, understandably, many more of them think their institution places too little emphasis on race. (The opinions of the 1989 cohort are surprisingly similar to those of the 1976 one, except that the later graduates think that their university places a greater emphasis on racial diversity than the earlier ones do.)

The study also reports interesting figures for the 1951 cohort, who are now in their mid-sixties. That group might be expected to be more conservative about affirmative action, which did not exist in their college years, than later cohorts. But 41 percent of the 1951 cohort (as compared to 37 percent and 48 percent of the 1976 and 1989 cohorts) believe that a great deal of emphasis should be placed on seeking racial diversity. Though roughly a third of its members think that their institution now places too much emphasis on it, a half think the present emphasis right, and 17 percent would prefer more.

It might seem plausible to assume that white students with relatively low SAT scores, who might have worried more about the impact of race-sensitive admissions standards on them before they were admitted, would disapprove of them more. But there is no significant difference in approval or disapproval of such policies across the full range of SAT levels. Perhaps even more surprisingly, those white students who were not admitted to their first-choice school, and might well be expected to blame their failure on racial preferences, do not disapprove of seeking racial diversity any more than their initially more successful classmates do. The figures for the once-rejected white graduates in both the 1976 and 1989 C&B cohorts are nearly identical to those of all white graduates.

These statistics seem important in any attempt to assess the degree of general backlash against affirmative action in the USgenerally. The political wars against affirmative action that I described earlier have concentrated on racial preferences in hiring, which many working-class voters believe they have personal cause to resent, and some commentators have doubted whether there is a genuinely deep national resentment against affirmative action even over hiring. Louis Harris, for example, has argued that the success of the California proposition banning all affirmative action was determined by a misleading presentation; his own polls suggest, he reports, that a fairer presentation of the proposition would have led to its defeat. 21 In any case, the River study gives some reason to doubt whether there is any general and deep-seated antagonism to affirmative action specifically in university admissions. Of course, many rejected university applicants (including, presumably, the plaintiffs in the lawsuits I mentioned) are indeed resentful. But the study estimates the number of once-rejected students who are resentful as relatively low.

Does affirmative action damage blacks by insulting or mortifying them, or destroying their self-respect, or poisoning the black image? The most moving arguments against affirmative action are made by those blacks who feel insulted or damaged by the assumption that blacks need special favors. Anyone, whether a black graduate or a successful child of rich or prominent parents of any race, will resent any suspicion that undervalues his personal achievements, and the fact that many prominent blacks fear that affirmative action has encouraged such suspicion is an undoubted and regrettable cost of the policy.

It is obviously important, however, in estimating the extent of that cost, to discover how many blacks hold this view. If many do, then the cost is great. But if the view is firmly rejected by most black graduates of elite institutions, who are especially likely to suffer professionally and personally from any assumption that their credentials or achievements are tainted, or from any damaged black image in the nation—if most such graduates believe, on the contrary, that the pursuit of racial diversity through race-sensitive admissions policies has been good for them and for their race—then the pain suffered by the small minority who disagree, while genuine, cannot be thought significant enough to outweigh the advantages the majority believe the race has gained. In fact, the overwhelming majority of blacks canvassed in the River study applaud the race-sensitive policies of their university. They think that their universities should now place more, not less, emphasis on racial diversity, and they accept what the study confirms: that affirmative action has been good for them, both in raising their income and in other, less material, ways.

Could the proportion of blacks in prestigious institutions be maintained if affirmative action was abandoned and race-neutral standards used instead? The River study calculates, on the basis of plausible assumptions, that a strict race-neutral admissions policy would have reduced the number of blacks in the C&B schools by between 50 percent and 75 percent.22 The impact of race-neutral policies on the professions would be particularly dramatic and damaging: blacks would have made up only 1.6 to 3.4 percent of the total number of students accepted to the 173 law schools approved by the American Bar Association if those schools had relied only on college grades and test scores, and less than 1 percent in the most selective law schools.23

Some scholars, including many who are anxious not to reduce the number of blacks in elite schools, have suggested that roughly the same number would be admitted if schools gave preference to low-income applicants instead of black applicants, because so many black applicants are poor. The study shows that this suggestion is based on a fallacy: though black applicants are disproportionately poor, poor applicants are still dominantly white, and even race-neutral tests that aimed at economic diversity would result in greatly decreased numbers of blacks.24

That calculation assumes, it is true, that institutions like those in the C&B list would be content to accept so dramatic a reduction in black presence in their classrooms, and would not try to escape whatever political or legal decisions had forced them to use race-neutral standards. That assumption is not necessarily valid. Boalt Hall and other branches of the University of California are studying changes in admissions procedures, including treating top grades at much less selective colleges as just as important as top grades at Harvard, and relying less on test scores altogether.25

The Texas state legislature, responding to the Fifth Circuit’s Hopwood decision, enacted a new program requiring the public universities of the state to accept all graduates of state high schools in the top 10 percent of their classes: since some high schools are almost exclusively black, this change can be expected to increase black enrollment at the formerly most selective Texas universities. Even if such adjustments succeed in their goal, they may well do so by substituting less qualified black matriculants for those that an open and acknowledged affirmative action program would have selected. Some former opponents of affirmative action, worried by that possibility, are having second thoughts. Professor John Yoo of Boalt Hall, who campaigned for Proposition 209, now says he realizes that conventional affirmative action is a useful way of maintaining racial diversity while still, as he puts it, “limiting the damage” to academic standards generally.26

Is the United States better off, judged strictly by the outcome, because its most selective universities and colleges have practiced affirmative action over the past thirty years? Most of the seven hundred “retrospectively rejected” black students from the 1976 cohort, who would not have attended a C&B school if race-neutral standards had been used, would have attended other, less selective, universities. But the high correlation the study establishes between the selectivity of the school attended and later success, for every level of SAT score, high school grades, and socioeconomic background, suggests that many fewer of them would then have become prominent professors, doctors, or lawyers, or high-salaried and powerful business executives, or political or community service leaders, than the actual graduates have become. The one thousand “retrospectively rejected” black students of the 1989 cohort give promise, already, of even greater success. So we can rephrase the question: Would America be better off if many fewer such important positions were held by blacks now and over the next generation? It seems incredible to suppose that it would. In all the dimensions in which our society is stratified—income, wealth, power, prestige, and authority—blacks are greatly underrepresented in the top levels, and the resulting de facto racial stratification is an enduring shame, waste, and danger. How could we think ourselves better off if that racial stratification were even more absolute than it is, and if we saw no or fewer signs of its lessening?

4.

Still, if affirmative action is unfair, because it violates the rights of white and other candidates who are refused places or of the few blacks who feel insulted, then it would be improper even if it does make the nation better off. We should notice, before we begin to examine that possibility, that the damage affirmative action inflicts on any particular non-preferred candidate is very small: the River study suggests that if race-neutral standards had been used in one set of schools it studied, and fewer blacks therefore admitted, the antecedent probability of admission of any particular white applicant who was in fact rejected would have risen only from about 25 percent to about 26.5 percent, because there were so many rejected white candidates at approximately the same level of test scores and other qualifications that adding a few more places would not much have improved the chances of any of them. When the Fifth Circuit declared the Texas Law School’s admission scheme unconstitutional, and remitted the case to a lower court to award damages to the rejected white plaintiffs who had brought the lawsuit, the lower court awarded each only $1 because it was so unlikely that any of them would have been admitted even under race-neutral standards.

Does affirmative action violate the right of candidates to be judged only on the basis of their individual qualifications? What counts as a qualification in this context? In some competitions, like a beauty contest or a quiz show, qualification is a matter only of some physical or intellectual quality: the winner should be the most beautiful or knowledgeable candidate. In others, like a book prize or awarding a medal for bravery, qualification is a matter of prior achievement: the winner should be the candidate who has produced the best work or product, or shown special character in some way, in the past. In still other competitions, however, qualification is a matter of forward-looking promise rather than backward-looking achievement or natural property: a rational person does not choose a doctor as a tribute to her skill or to reward her for past cures: he chooses the doctor that he expects will do best for him in the future, and he takes the doctor’s innate talent or past achievements into account only because, and so far as, these are good indicators of the doctor’s value to him in the future.

Competitions for university places are, of course, competitions of the last sort. Admissions officers should not award places as prizes for past achievements or effort, or as medals for inherent talents or virtues: their duty is to try to choose a student body that, as a whole, will make the greatest future contribution to the legitimate goals their institution has defined. Elite higher education is a valuable and scarce resource, and though it is available only to very few students, it is paid for by the community generally, even in the case of “private” universities that are partly financed by public grants and whose “private” donors benefit from tax deductions. Universities and colleges therefore have public responsibilities: they must choose goals to benefit a much wider community than their own faculty and students. These need not be economic or social or political goals in any narrow sense: on the contrary, we expect all our educational institutions, and particularly the best financed and most prestigious ones, to contribute to science, art and philosophy, whose advancement we take to be part of our collective public responsibility, and to select students and faculty very much with that goal in mind.

But the advancement of knowledge is not the only goal that we allow or expect educational institutions to pursue. We expect them all, particularly the best financed, to help both their students and the larger community in other, including more practical, ways as well—a great university may properly decide to study the treatment of AIDS or Alzheimer’s disease even when it knows that different, more basic, research would be theoretically more rewarding. Nor do we expect all schools to adopt the same goals, or to attach the same relative importance to the goals they do select. The great research universities in the C&B database have different priorities from the smaller liberal arts colleges in that base, and both of these have different goals from small agricultural colleges, community colleges, and other institutions of a kind not represented in the base at all. The academic freedom we prize means, among much else, that each institution is free, within broad limits, to set goals for itself, and to define the academic strategies, including admissions strategies, that it believes most appropriate to those goals.

The C&B schools have all traditionally regarded impressive high school or college grades and high SAT and other test scores as important qualifications for university and professional education. But none of them has treated these distinctly academic qualifications as exclusive: they have all from time to time rejected candidates with top SAT scores and grades—even black candidates—in favor of other students with lower grades and scores. The list of other qualifications is long: it includes motivation for public service, athletic ability, unusual geographical background and, in the case of some of these schools, “legacy status,” which means having parents who are graduates of that university. Admissions officers regard each of these attributes, and many more, as indicators (though far from perfect ones) that a particular applicant will contribute to one or more of the institution’s traditional goals. The River study shows, and the great American universities and colleges have recognized for thirty years, that at least two of these traditional goals are well served by including, as one among the myriad factors that admissions officers take into account, a candidate’s race.

First, as I said, American schools have aimed at student classes that are diverse in several ways. They have plausibly assumed that students are better equipped for commercial and professional life, and better prepared to act as good citizens in a pluralistic democracy, if they have worked and played with classmates of very different geographical background, economic class, religion, culture and—above all, now—race. Critics argue that selecting on the basis of race is an inappropriate means of pursuing diversity because it wrongly, and insultingly, assumes that all black students do, and only black students can, provide desirable diversity in class, political attitudes, or culture: it would be better, according to these critics, to accept students of whatever race whose parents are poor, or who appreciate soul music, rather than seek black students, some of whom have rich parents, or prefer Bach.

But this objection misses the aspect of diversity that is in question, which is not what race may or may not indi-cate, but race itself. Unfortunately the worst of the stereotypes, suspicions, fears, and hatreds that still poison America are color-coded, not coded by class or culture. It is crucial that blacks and whites come to know and appreciate each other better, and if some of the blacks turn out not to have the class or cultural or other characteristics that are stereotypically associated with them, that obviously enhances rather than undermines the benefits of racial diversity.

Second, our schools have traditionally aimed to help improve the collective life of the community, not just by protecting and enhancing its culture and science, or improving its medicine, commerce, and agriculture, but by helping to make that collective life more just and harmonious—those are, after all, among the main ambitions of our law schools and schools of politics and public administration, and they should form part of the goals of the rest of the academy as well. Our universities and colleges are surely entitled to think that the continuing and debilitating segregation of the United States by race, class, occupation, and status is an enemy of both justice and harmony, and it is one of the most dramatic conclusions of the River study that affirmative action has begun to erode that segregation in ways no other program or policy probably could. We expect educational institutions to contribute to our physical and economic health, and we should expect them to do what they can for our social and moral health as well.

So affirmative action, in pursuit of either or both of the twin goals of student diversity or social justice, in no way compromises the principle that student places should be awarded only on the basis of legitimate and appropriate qualifications. No student has a right to a university place in virtue of past achievements or innate virtues, talents, or other qualities: students must be judged only by asking how likely each is, in combination with others selected on the same standards, to contribute to the various goals that the institution has legitimately chosen. I do not mean (as some critics have accused defenders of affirmative action of supposing) that black color is in itself a virtue or an aspect of merit. But it is nevertheless a qualification in the sense I have been describing. We do not count a person’s height as a virtue or a merit. But someone who is tall may just for that reason be better able to contribute, on a basketball court, to one of a university’s traditional goals, and in the same way, though for sadder reasons, someone who is black may for that reason be better able to contribute to its other goals, in the classroom and dormitory and in the course of his or her later career.

Why, then, is affirmative action so widely thought unfair? Why do even many of its supporters concede that it is a distasteful remedy, even if, in their view, a necessary one? We must take care to distinguish and consider a variety of answers to those questions, because each has had an important though sometimes inarticulate part in the public’s response. It is often said, first, that race-sensitive admissions policies do not judge applicants as individuals, but only as members of large groups. That objection was strongly pressed against early and relatively crude forms of affirmative action, like the quota system declared unconstitutional in the Bakke case, because, as Justice Powell said, once the white quota had been filled, no further white candidate could be compared, even on an all-things-considered basis, to a black who was accepted instead.27 Under contemporary versions of affirmative action in university admissions, however, no quotas are used: these plans are in that respect like the Harvard plan that Powell expressly approved. Admissions officers now do make case-by-case, all-things-considered judgments, and, as I said, they sometimes accept a white student with lower SAT scores than a rejected black applicant. No one is accepted or excluded simply by virtue of race.

Many people do feel very strongly that even if universities should view a wide variety of properties as among the qualifications for admission they recognize, race, for special reasons, should not be among these. It is crucial, however, to distinguish different ways in which race might be thought special, and to consider the implications of each. We have already discussed one: many people believe that race-sensitive admissions standards exacerbate rather than help relieve racial tension. But we may set that claim aside, in the light of the River study, unless that study is somehow impeached. Many people also believe, however, that racial classifications are always wrong in principle, even when their results are in themselves desirable. They point out that we would not accept a law school’s argument that it rejects all black applicants because it aims to help the community’s economy by producing graduates who will function effectively in local law firms that do not welcome blacks. They insist that we cannot in principle distinguish that invidious use of race to achieve results in themselves creditable from a so-called “benign” use. And even if we could do so in principle, we could not realistically do so in practice, because invidious uses might easily masquerade as benign ones.

The first of these arguments is the easier to answer: we can make the distinction between affirmative action and malign uses of race, at least in principle, in two ways. First, we can define an individual right that the malign forms of discrimination violate but that properly conceived affirmative action programs do not: this is the fundamental right of each citizen to be treated by his government, and by institutions acting with the support of his government, as equally worthy of concern and respect. A black citizen is denied that right when schools justify discriminating against him by appealing to the fact that others are prejudiced against members of his race.

But the case for affirmative action does not reflect, either directly or indirectly, prejudice against white citizens; seeking racial diversity no more reflects a prejudice against whites than seeking geographical diversity expresses prejudice against people from large urban centers. Second, though, as I said, it is important to allow universities ample latitude in designing their own purposes and goals, we can nevertheless insist that some goals a university might conceivably adopt are illegitimate and unacceptable, and we can dismiss, as such, a goal that panders to and reinforces the racial stratification of our society.

It is true, however, that these distinctions of principle may be hard to enforce in practice, particularly because they rely on judgments about institutional motives that are often hard to identify. How could we be sure, for example, that a program that gives preference to some minorities, like blacks and Hispanics, is not motivated by hostility to other groups of citizens—Asian-Americans or Jews, for example—who score well on tests and who would be admitted in greater numbers if admission policies were race-neutral? Or by an even cruder desire on the part of admissions officials in some institutions, who might themselves be black, to favor their own people at the expense of others? Would it not be better to guard against this possible corruption by flatly forbidding all use of race in university admissions?

That argument has been pressed in non-university situations, as an argument, for example, against allowing city councils, which might well be dominated by black members or dependent on black support, to set aside a quota of construction contracts for black-owned firms, or allowing a state legislature, which might be influenced by racial politics in these and other ways, to design electoral districts so as to elect more black officials.28 But however forceful this argument for a flat prohibition on racial classifications might or might not be in these other situations, it seems fanciful and misplaced when applied to higher education. The faculty and academic administrators who stipulate and use race-sensitive admissions standards are in no way beholden for power or financial support to any of the communities these standards benefit. They act in pursuit of traditional goals that the River study shows are most efficiently served in that way.

Moreover, any suspicion of hidden hostility to another group that has also been the target of prejudice could easily be tested by seeing whether that group is disproportionately represented among those who would probably have been admitted under race-neutral standards. True, these considerations do not wholly eliminate any conceivable possibility that illegitimate motives have played a role. But denying all universities the power to do what they can to improve diversity and social justice and stability, on the remote chance that some one or two institutions would abuse that power and escape undetected, would be like denying any use of public funds for medical research on the ground that a few researchers might be plagiarists or embezzlers.

So we may set aside these mistaken arguments of principle and policy. We must nevertheless recognize the important psychological fact that many people do think that being rejected by a university because they are not of the “right” race is far worse—more outrageous and more insulting—than being rejected because they lack some other quality, like a skill or physical ability, or even because their parents did not graduate from the school. That special outrage is understandable, however, not because race has some special importance in the metaphysics of personal identity—one’s skin color is no less a matter of choice, and no more genetically grounded, than the raw abilities that insure that some adolescents will never be able to score above 1400 on an SAT test no matter how hard they prepare. It is understandable because we are all so familiar with the character and consequence of invidious racial classification.

Racial discrimination expresses contempt, and it is deeply unjust and wounding to be condemned for one’s natural properties; racial discrimination is, moreover, wholly destructive of its victims’ lives—it does not merely close off to them one or another opportunity open to others, but injures them in almost all the prospects and hopes they might conceive. In a racist society people are indeed rejected absolutely and for who they are, and it is therefore natural that racial classifications should be seen as capable of inflicting a special form of injury. But it would nevertheless be perverse to disallow the use of such classifications to help combat the racism that is the true and continuing cause of that injury. The special psychological character of race is not a fixed fact to which policy must always defer. It is a product and sign of racism, and it must not be permitted to protect the racism that has generated it.

We should consider one final reason why race might be thought special, which lies at the intersection of moral and legal concerns. It is often argued that America’s social and constitutional history has committed us, as a people, to a society that is colorblind as a matter not just of our ultimate goals but also of the means that we are entitled to use toward any goal. According to this argument, the constitutional amendments adopted after the Civil War, which include the Fourteenth Amendment’s guarantee of “equal protection of the laws,” were a national commitment—moral as well as legal—to refuse race any official role in our affairs whatever. If so, then university affirmative action programs are wrong in principle, whether or not they violate anyone’s rights as an individual, because they cheat on that important national commitment.

But this argument, though popular, is unpersuasive. Some critics of affirmative action do argue, as we have seen, that a colorblind commitment would be a wise strategic decision: that we would do a better job of confronting and eliminating racism in the long run if we always avoided any racial classification, even those that might seem, in the short run, effective against racism. But they have offered no argument for that strategic hypothesis, and the River study seems to prove it false. Nor is there any ground for supposing that the Constitution, or anything else, has committed the nation to that strategy. The Fourteenth Amendment does not mention race, and no plausible interpretation of that amendment shows it automatically to rule out all racial classifications as means to greater justice.29 Nor have the American people, by any longstanding or sustained consensus, ruled out all such classifications for that purpose. The supposed national commitment is an illusion.

So, according to by far the best evidence yet available, affirmative action is not counterproductive. On the contrary it seems impressively successful. Nor is affirmative action unfair: it violates no individual rights and compromises no moral principle. Is it nevertheless unconstitutional, as the Fifth Circuit judges I mentioned ruled in the Hopwood case?

In a second article, to be published in the next issue of the Review, I consider the arguments of constitutional principle and the analysis of recent Supreme Court decisions on which those judges relied; I argue that both principle and precedent continue to support the Bakke principle that properly drafted race-sensitive admissions standards are constitutional. Of course affirmative action has had its costs—both to disappointed white applicants and to those successful black ones who resent any suspicion that they needed special preference to succeed—and the policy has undoubtedly provoked more general resentment, even if the scale of that resentment remains unclear.

But the moral and practical costs of forbidding it would be far greater. The systemic racial discrimination of the past has created a nation in which positions of power and prestige have been largely reserved for one race. It was not irresponsible for critics to oppose affirmative action, on the ground that it would do more harm than good, when the consequences of the policy were still uncertain. But it would be wrong for the nation to prohibit that policy now, when comprehensive statistics and analysis have apparently demonstrated its value. The River study, as I said, might be impeached by a better—larger or more sophisticated—study. But unless and until it has been impeached, we have no reason to forbid university affirmative action as a weapon against our deplorable racial stratification, except our indifference to that problem, or our petulant anger that it has not gone away on its own.30

—This is the first of two articles.

This Issue

October 22, 1998