Debating The Shape of the River: Long-Term Consequences of
Considering Race in College and University Admissions
by William G. Bowen and Derek Bok


BENEFITS AMERICA! Deposition of WILLIAM G. BOWEN in Gratz v. Bollinger summarizing conclusions from his book, The Shape of the River: Diversity is a benefit for all students -- minorities and nonminorities alike. The data overwhelmingly demonstrate that minority students admitted to selective schools had strong academic credentials, graduated in large numbers and did very well after leaving college. By every measure of success (graduation, attainment of professional degrees, employment, earnings, civic participation, and overall satisfaction), the more selective the school, the more blacks achieved (holding constant their initial test scores and grades).

HARMS AMERICA! Below, a critical review by CURTIS CRAWFORD of The Shape of the River: Of the benefits from race-sensitive admissions urged by Bowen and Bok, only one survived analysis as a clear, net gain: a reduction of the racial gap in incomes and status. Against this benefit are seven costs: minority achievement discredited, the racial nondiscrimination rule suspended, its moral authority undermined, the quality of work and the course of individual development impaired, racial preference injected into course content, academic policy made hostage to deception, and intellectual freedom abridged.

[The Shape of the River was greeted by the New York Review of Books as "the first comprehensive and statistically sophisticated examination of the actual effects of thirty years of affirmative action in American universities." The study's findings, summarized in Dr. Bowen's deposition, are challenged in Mr. Crawford's critique (below.)]


Curtis Crawford

A Review of The Shape of the River: Long-Term Consequences of Considering Race in College and University Admissions
by William G. Bowen and Derek Bok. Princeton, N.J.:
Princeton University Press. 1998. Pp xxxvi, 472. Cloth, $24.95.

[Published as a Review Essay in Society 37:4, pp 71-80, May/June 2000]

This book hopes to reassure the academy and persuade the courts that preferential college admission for African Americans is indeed the right national policy. Ex-presidents respectively of Princeton and Harvard, the authors have long been proponents and practitioners of race-based affirmative action. Their book was promptly hailed by Ronald Dworkin in the New York Review as containing "much more comprehensive statistics and much more sophisticated analysis than has been available before." The statistics and the analysis are supposed to demonstrate that the benefits from preferential admission far outweigh the costs.

Bowen and Bok acknowledge that, without racial preference, our top schools would lose most of their black students (p 51). Until recently this was not the kind of statement one heard in public from the friends of affirmative action. On the contrary, it was opponents (like myself) who were crying "preference," and (unsuccessfully) challenging the colleges to disclose how much. The people in charge cultivated the impression that little or no preference was involved. They denied that they were "discriminating." They were simply "leveling the playing field," "assuring equal opportunity" and "fostering racial diversity." However, this rhetoric suddenly shifted, when recent ballot initiatives and court decisions threatened to outlaw racial preference in the public colleges and universities of several states. Without preference, the supporters of affirmative action declared, their programs could not exist.

The transition to candor concerning the nature of affirmative action has been difficult and incomplete. Although Bowen and Bok admit that the policy they defend confers substantial preference, based on race, they never call it "preferential treatment" or "racial preference." Their label, instead, is "race-sensitive admissions." Preference for whites over blacks, wherever it occurs, they do not hesitate to call "discrimination," but this term is never allowed any contact with the kind of preference they favor.

Nevertheless, despite the euphemisms, preferential treatment based on race is now acknowledged and defended as standard procedure in selective college admissions. The high principle of equal treatment for persons of different race is openly challenged. Fruitful debate, exceedingly difficult so long as one side evaded the nature of the policy at issue, can finally take place on the merits.

In response to this new situation, it is not enough for critics of racial preference to insist that it is immoral, regardless of the circumstances. Bowen and Bok look at the long-term consequences of preferential college admission for blacks and find them overwhelmingly beneficial. Are they right? Is their conclusion sustained by the relevant data?


The book’s facts come primarily from an extensive database concerning 62,000 white and black students who entered 28 selective colleges and universities in 1976 and 1989 (pp xxvii-xxx). These schools are divided into three groups, according to their degree of selectivity, as measured by the SAT scores of their freshmen. Duke, Stanford, Yale and five other schools with average SAT scores above 1300 are ranked as SEL 1, the highest category. Fourteen schools with scores from 1150 to 1300, including Columbia, Oberlin, Vanderbilt and Wellesley, are SEL 2. Six, including the Universities of Michigan and North Carolina, with scores between 1000 and 1150, are SEL 3. (Freshmen at all these schools now average higher SAT scores, not because they are smarter, but because the SAT was re-centered in 1995.) Incorporated in the database are undergraduate academic records for all the 62,000 students, and responses from an overwhelming majority to a 1995 survey.

The first major finding of fact must have stunned those who thought that the advantage conferred by racial preference in college admissions is insignificant. Bowen and Bok estimate that, if selection became race-neutral, the top-tier schools would lose three-fourths of their black students, down from about 8% of their enrollment to 2%. Schools in the second tier would lose one-half, the percentage falling from 6 to 3. Third-tier schools would lose one-third of their blacks, down from roughly 6.5% to 4.5% (p 41).

These results are less drastic than they appear, since far more American colleges lie in the third category than in the upper two combined. Of all the students who attend the kind of school studied by Bowen and Bok, only 5% go to a college that the authors would rate as "SEL 1." The nation’s "SEL 2" schools enroll 13%; "SEL 3" schools, the remaining 82% (p 29). Thus, the decline in black enrollment at the overwhelming majority of selective campuses would average about two percentage points.

Perhaps the book’s most important findings concern the advantages of attending more selective schools (Chs 3-7). In the course of the book, these findings become important empirical premises of the authors’ case for race-sensitive admissions. Bowen and Bok conclude that students who attended an SEL 2 school were more likely (1) to graduate, (2) to feel very satisfied with their college experience, (3) to gain a postgraduate degree and (4) to earn a higher occupational income than students who attended an SEL 3 school. The same held for SEL 1 students as compared with SEL 2 students, and it was true for blacks as well as whites.

This pattern is partly ascribed to the fact that students in the more selective schools had greater cognitive ability, as indicated by their higher SAT scores. But the authors find, surprisingly, that students with the same SAT scores, if they attended more selective schools, were more likely to graduate, to feel very satisfied with their education, to attain advanced degrees, and to make more money. These advantages accrued even to students whose scores were several hundred points below the average of their fellows. All this was true for both blacks and whites.

The postgraduate success of the black students did not occur without additional boosts from racial preference. Although such information is not included in their database, the authors cite other sources for their belief that preference was substantial in graduate and professional school admissions, and widespread in corporate and professional employment.

These findings are not without significant uncertainties and exceptions. First, the 28 schools in the database are not shown nor asserted to be statistically representative of the other 286 American schools of comparable selectivity. One cannot be sure that what was true for the former concerning graduation rates, et al., would be true for the latter. Second, the database did not identify which black students received preference, either at college admission or later. We know which students were most successful with respect to graduate degrees and occupational income, but not which of them were preferential admittees. Third, for students of similar SAT scores there was an important disadvantage to attending a more selective school: they got lower grades, which may mean that they learned less. This was true for both blacks and whites.

Notwithstanding such objections, the Bowen/Bok findings concerning the advantages of attending a more selective school have a solid basis in the book’s data. Solid, but mistaken? They have been persuasively challenged in subsequent research: "Estimating the Payoff to Attending a More Selective College" (Stacy Berg Dale and Alan Krueger, 1999). This challenge was mounted from the same Mellon Foundation dataset on which River was based, and co-authored by one of River’s principal pilots.

Dale and Krueger hypothesize that career benefits attributed to the impact of a superior school might be due instead to the superior personal qualities that help students get into superior schools. The Bowen/Bok statistical analysis controls for the student attributes that are available in the database: race, gender, SAT score, class rank in high school, and parents’ social and economic status. Dale and Krueger do not dispute the Bowen/Bok calculations, when only these variables are included. But they argue that other applicant traits, such as ambition, discipline, imagination, perseverance, maturity, may be important in explaining individual success, both as college applicants and as working adults.

How does one test this hypothesis? In two ways. First, match students who fare alike in the selectivity of schools that accept and that reject them, but who attend different schools. Add this match to the other variables that are held equal: race, gender, SATs and so on. Then run the numbers on future earnings and see if the difference in the selectivity of the school attended changes the outcome. Or second, match students according to the average selectivity of the schools to which they apply, but who attend different schools. Add the other variables, run the numbers and see whether the difference in the selectivity of the school attended affects the result. The first approach may capture applicant qualities as perceived by admissions officers; the second, applicant qualities as perceived by the applicant himself. The Dale/Krueger results: when these controls are added to the statistical analysis, the selectivity of the school does not affect the likelihood of graduation or the size of future income, although it does increase somewhat the chance of acquiring a graduate degree.

These findings, if valid, erase or reduce much of the supposed career advantage from attending a more selective school. If such attendance does not predict more income, does it perhaps lead to greater job satisfaction? On this question, Bowen and Bok find that as school selectivity goes up, future job satisfaction actually goes down. (Would this school effect also disappear with the Dale/Krueger approach?) It may be that attending a more selective school is correlated with future attainment of greater power or prestige, but these possibilities are not examined in River. Nor the possibility that the chief advantage might be a better education.


As consequences of race-sensitive admissions, Bowen and Bok claim benefits for the students who receive preference, for the schools that give it, for other students, and for society as a whole.

  • The student recipients benefit from the advantages of attending a more selective school.
  • The schools do a better job of estimating student potential.
  • Other students benefit from the educational impact of racial diversity.
  • Society benefits from a reduction of the national gap in income and status between black and white Americans.

The Benefit to Recipients. Chapters Four through Six posit career advantages for the students who attend more selective schools. The text maintains that these advantages are generally shared by the portion of black students at these schools who are admitted via preference. The second proposition is in some doubt, since the database lacks statistics for preferential admittees as a separate category. The first proposition is in grave doubt, thanks to the Dale/Krueger research. Fortunately, these doubts do not have to be resolved here. This section of my critique reaches the same conclusion whether the advantage from attending a more selective school is large or small.

Against the benefits to those who enter via racial preference, one must weigh the corresponding losses to the applicants (mostly white, some Asian) who are excluded because of racial preference. In any school with fewer places than applicants, for every student admitted because of race another is rejected for the same reason. The greater the benefit from attending a more selective school, the greater the loss to the person deprived of this opportunity.

The authors may feel that the gain to the favored blacks is greater than the loss to the displaced whites. But such a finding would be hard to prove, and no attempt is made to do so. Some facts in the study suggest that black students were getting more than whites out of the educational opportunities available: blacks with SAT scores and college grades similar to whites’ were more likely to achieve graduate professional degrees. But other facts suggest that blacks were getting less: their college grades were lower than whites with similar SAT scores.

Taking into account the students who were rejected by race-sensitive admissions, as well as those who were accepted, no net benefit has been shown. To justify a preferential policy one must go beyond the position that it helps the persons it favors. Thus the authors, in their final chapter, argue that preferential admissions also benefit the schools, other students and society as a whole.

Estimating Student Potential. The book contends that race-sensitive admissions help colleges to identify applicants of high potential (pp 278-279). If true, this would be a notable benefit. Unfortunately, the argument as stated is brief and inconsequent, providing no support for the conclusion. The authors assert that "[a]n individual’s race may reveal ... what barriers were overcome, and what the individual’s prospects are for further growth." But they do not explain how this revelation occurs: how a school can reliably infer either past barriers or future prospects from an applicant’s racial identity.

The authors believe that black youngsters are more likely than whites to confront substantial obstacles. Agreed. But if a school wishes to know the disadvantages that applicants have faced, it need not speculate, based on racial assumptions. It can ask for specific information from the candidate or his references. Take, for example, the disadvantage of growing up in a family where income and parental education are low. The book’s 1989 figures for the 28 colleges indicate that 14% of the black freshmen came from such families, as compared to only 2% of the whites (p 341). But no one had to guess from an applicant’s race whether his family was in this category. The relevant information was in his application file.

To show a connection between racial identity and student potentiality, the authors might have referred to their figures concerning racial differences in the meaning of SATs. Are these tests, as is often charged, unfair to blacks? Many supporters of affirmative action have long believed that the SATs are culturally biased, or demand knowledge not acquired in inferior schools, or both. Are they right? Do the SAT scores of black high-school students understate how well they would do in a good college?

This question is easy to answer and the authors have answered it (pp 72-78). All they had to do was compare the SAT scores of entering students with their subsequent college grades (as measured by class rank). The book finds that the SAT scores of black students, instead of understating their academic potential, greatly overstated it. As compared to whites, black students’ SATs overpredicted their college grades by a wide margin. For example, the average class rank of blacks with SATs over 1300 was below the average rank of whites with SATs under 1000.

The benefit claimed for schools must be set aside as unclear and unsubstantiated. The authors assert that race-sensitive admissions help colleges to assess applicant potential, but they fail to say how. They do not specify the racial differences to which schools should pay attention, or the inferences concerning potential that should be made therefrom. They seem less interested in obtaining better potentiality estimates regardless of race than in boosting the acceptance rates of black applicants.

Racial Diversity. The third benefit urged is that, by increasing the number of blacks on campus, race-sensitive admissions enrich the educational opportunities of other students (pp 279-280). The authors contend that a significant black presence brings different "life-experiences and perspectives" to the campus dialogue, helping students learn "to understand how others think and function, to cope across racial divides, and to lead groups composed of diverse individuals . . ." The book’s data indicate that most of the white students at the 28 schools got to know well two or more black students. Also, that as the proportion of blacks increased, so did the extent of interracial acquaintance (pp 233-234).

Without racial preference, the authors warn, the number of black students at the more selective schools, and the educational opportunities that accompany them, would decline. This is true. But it should also be pointed out that the applicants thus turned away would be admitted by less selective schools, at which the number of blacks and the opportunity for black/white interaction would rise. The authors do not notice this side of the coin. They emphasize the educational benefits to white students of interacting with blacks, overlooking the fact that this argument cuts both ways, in favor of adding blacks to less selective schools by ending racial preference, as well as for retaining them in more selective schools by keeping it.

Bowen and Bok could reply that the marginal interracial educational benefit of additional black students is greater for schools that start with an enrollment of 3% rather than 7%. They might argue, also, that since the students at more selective schools are more likely to be national leaders, an increase in their interracial experience is more valuable to society. These are reasonable points. On the other hand, it may be that campus black/white interaction would lead more readily to friendship, if the schools were racially neutral and the student racial contingents more similar in academic ability. This is a plausible empirical hypothesis that the authors failed to test.

Whether there is any net educational benefit to other students, by the preferential shifting of black students from less to more selective schools, is not easy to assess. On this point, the book suffers from the methodological flaw of having considered and documented advantages, but not the concomitant disadvantages.

The Racial Gap in Attainment. The asserted benefit to society is that race-sensitive college admissions act to reduce the black/white gap in average earnings and societal leadership (pp 280-286). This claim follows logically from (a) the alleged career advantages of attending a more selective school, and (b) the operation of preferential admissions, which increase the number of blacks and reduce the number of whites who gain access to these advantages. Combine these black gains and white losses and ipso facto the racial gap is reduced.

As we have seen, the advantages of attending a more selective school in regard to professional preparation and future income are disputed by Dale and Krueger. They find that school selectivity has no impact on one’s future earnings, and very little on the chances of earning a graduate degree. Are they right? Their approach is ingenious, reasonable and conscientious, but their results feel counterintuitive. It is very hard to believe that, all else equal, the selectivity of one’s college contributes nothing to one’s future wealth, status or power.

Bowen and Bok stress the substantial proportion of black students in their 28 schools who go on to professional degrees, and within this group the remarkable percentage who receive these degrees from top institutions. These students seem credentialed for the higher ranks in the practice of medicine, law and business. Many of them presumably were brought into college and/or professional school via race-sensitive admissions. Citing the Dale/Krueger research, one could argue that as undergraduates these students would have done just as well, without racial preference, at less selective colleges, which have plenty of openings for students of their capability. But this argument would not hold for professional schools, where the total number of applicants far exceeds the total capacity of the schools. Bowen and Bok cite studies indicating that under race-neutral admissions the number of blacks studying to be lawyers would drop by over 50%, and the numbers preparing for medicine and business would also substantially decline (pp 44-46).

The question is whether and to what extent race-sensitive college admissions help to reduce the racial gap in income and influence. The answer is uncertain. For the purposes of this critique, let us stipulate that the help exists and is significant.

In the book’s prose, the impact of race-sensitive admissions sounds more dramatic than I have expressed. Indeed, the authors speak as if racial preference were the only factor that operates to reduce the racial gaps (p 285). But this they do not attempt to substantiate. They celebrate a large increase during the last three decades in the proportion of black professionals, executives, managers and government officials (p 10). But they have no data that show how much these changes were due to racial preference, and how much to the reduction of anti-black discrimination and to the achievements of blacks who did not receive preference.

Is racial inequality in earnings and influence bad for the country? Bowen and Bok consider it a great evil, whose correction should be a top national priority (pp 283-286). They believe "that present disparities in racial outcomes are dismayingly disproportionate." Pointing to the rapid growth in our minority population, they argue that government and business will make wiser decisions and enjoy greater public confidence if more people of color occupy top positions. They add that minority communities need "well-trained, articulate leaders to represent them in the political arena;" minority-owned businesses, which are "more likely than white-owned companies to hire minority employees"; and minority physicians, who are more likely "to include minorities and poor people among their patients." (pp 11-13) The authors are passionate about the need to reduce the racial gap in wealth, status and power. Indeed, this need is the central policy assumption of the book.

The benefit to minority communities of having more people able and willing to serve their needs is surely incontestable. Whether the need to remove or reduce the "present disparities in racial outcomes" is as clear or compelling as the authors contend is another question, to be addressed later.

In summary, a reduction in the black/white gap in income and status is the only asserted benefit of race-sensitive admissions that remains on the scales. The first declared benefit, career advantages for the black students preferred, was offset by the corresponding disadvantages for the students they displace. The second benefit, enhanced institutional ability to estimate student potential, was set aside as unsubstantiated. The third benefit, a net educational advantage to campus life from shifting black students via preference to more selective schools, was possible but unsubstantiated.


What costs of race-sensitive admissions should be weighed against this acknowledged benefit? I suggest at least the following:

  • Blacks have been stamped as less able and their achievements discredited.
  • In an important area of American life, the rule forbidding racial discrimination has been suspended.
  • The moral authority of this rule has been undermined.
  • The quality of academic and professional work has suffered.
  • Racial preference has spread to the selection of academic course content.
  • A culture of deception concerning racial policy has flourished.
  • The intellectual freedom of students and faculty has been abridged.

Black Achievement Discredited. It has been argued that preference for some members of a group lowers the status and devalues the achievements of all members. According to this view, the use of a double standard signals that the favored group is inferior. Moreover, since no one knows which members of the group have received special treatment, no one knows whose attainments were fully earned. As the authors put it, "The very existence of a process that gives explicit consideration to race can raise questions about the true abilities of even the most talented minority students ..." (p 264)

Bowen and Bok consider this cost real, but minor (p 265). It is not severe enough, they conclude, to prevent black students from strongly supporting race-sensitive admissions. Unfortunately, the data on which this judgment is based are indirect and ambiguous. Black alumni were asked, not about race-sensitive admissions or racial preference, but about their school’s emphasis on "a racially/ethnically diverse student body." (p 322) The overwhelming majority answered that the emphasis should be very great, more than had actually occurred (p 444). The authors interpret this as enthusiasm for preference.

The data did not have to be ambiguous. Racial diversity and racial preference are two different things: the former, a result that can take many forms; the latter, one means of achieving some form of the result. It is poor social science to infer, from a respondent’s approval of an end, his approval of a means. Moreover, it is not at all clear that students attending selective colleges in the late seventies or early nineties were aware that affirmative action admissions relied on racial preference or to what extent. The survey should have asked alumni what they had known about the nature and degree of racial preference in their school, and how they felt about its consequences.

My own hunch has been that many talented blacks resent the stain of preference on their abilities and achievements. But few publicly complain. Concerning the severity of their resentment, I lack a solid basis for judgment. And so do the authors.

Does preference cause whites to think less of blacks? Black academic accomplishments, no doubt, are often discounted. But is the existence of preference the primary cause of this response? Preference is given to alumni children and state residents, but their academic achievements are not usually discounted. Athletes (derisively known in my day as "jocks") are often thought to be academically inferior, but this generalization seems due more to their academic performance than to the fact of preference. I suspect that black academic accomplishments are questioned by others primarily from negative experience of, or prejudices concerning, black academic ability.

It can be argued that racial preference, by admitting comparatively underqualified black applicants, only makes the situation worse: the more such admissions, the more opportunity for other students to experience black academic performance as substandard. For this problem, however, the abolition of preference is a doubtful remedy. The black students remaining at elite schools would be perceived as more capable, but scarce. It is not clear which fact is more apt to breed negative inferences concerning a group’s ability: the comparative absence of its members under preference-free admissions, or the presence of less able members admitted via preference.

The Bowen/Bok survey might have asked white students whether their opinions concerning blacks had changed during their college experience and what had caused the change. No such questions were used. The authors could reasonably have thought that the answers would not have been fully candid. Whites these days are under considerable social pressure not to entertain negative opinions about black ability, character or history.

The Antidiscrimination Rule Suspended. For the sake of race-sensitive admissions, the rule prohibiting racial discrimination has been suspended in an important area of national life. By "racial discrimination" I mean a difference in treatment that favors or disfavors people, based on their race or ethnicity. The policy defended by Bowen and Bok permits discrimination always against whites, sometimes against Asians, but never against blacks or Hispanics (except when historically black colleges give preference to non-blaks?). The right not to be racially discriminated against is thus denied to the vast majority of college applicants.

Is this a substantial cost? I believe that the rule against racial discrimination is society’s best defense against racial or ethnic favoritism and its historical effects: racial partisanship, prejudice, hatred, injustice, oppression and war. The frequency and grievousness of these consequences, in every century, on every continent, are indisputable. The obligation of racial nondiscrimination, by all toward all, seems the only rule with both the power to check favoritism, and the fairness to be voluntarily accepted on all sides. Legislation prohibiting racial discrimination in the distribution of important goods, services and opportunities has been adopted in Austria, Australia, Brazil, Bulgaria, Canada, China, Colombia, Cuba, Cyprus, Finland, France, Germany, Iran, Norway, Pakistan, Nigeria, New Zealand and Great Britain. The widespread realization that racial nondiscrimination is a social necessity seems to me one of the great moral triumphs of the twentieth century.

Bowen and Bok dismiss as "largely conjectural" the thesis that treating people "differently because of their race" tends "to increase racial animosities." (pp 268-269) They ignore the historical experience that produced the thesis. As evidence that race-sensitive admissions are not "poisoning race relations" at the 28 schools, they cite two findings from their study: (1) Approval of racial diversity was as strong among whites who had been rejected (possibly due to racial preference) by their school of first choice as among whites who had not (pp 251-252). (2) Approval of racial diversity, already strong among whites who entered in 1976, was even stronger among those who entered in 1989 (p 250).

These findings are not to the point. Here again the authors infer views concerning racial preference from views concerning racial diversity. To find out what these students assumed concerning preference and whether they approved, why not ask them? The purpose of the book is supposedly to provide a long-needed, first-ever, comprehensive assessment of the consequences of race-sensitive college admissions. Why does the alumni survey, on which so many findings depend, omit any questions that invite student beliefs about or reactions to the policy being assessed?

Do Bowen and Bok really see no risk in abandoning the nondiscrimination rule in college admissions? Do they consider America immune from the dangers of racial or ethnic favoritism? How could they - how could anyone who is familiar with our history? Would they say that racial discrimination may be dangerous when one group inflicts it on another (whites vs blacks, Serbs vs Croats or Albanians, Hutu vs Tutsi), but not when it occurs within a group (white administrators vs white applicants)? One could argue that race-sensitive admissions are not going to upset blacks and Hispanics, and that predominantly white schools will not carry the policy so far that whites are massively antagonized.

But causes are in motion that we ignore at our peril. Blacks and Hispanics are learning to claim that a certain number of places at elite universities are rightfully theirs. These claims are always based on demographic statistics (the percentage of state residents or high-school graduates or college applicants who belong to these groups), never on academic prowess. If whites resolutely objected to being discriminated against, would blacks and Hispanics obligingly back down? More likely they would continue to feel entitled to proportional representation, and fight its national repudiation as "unwelcoming," "unjust" and "racist."

When the nondiscrimination rule is generally accepted, the moral force of the whole society is available against any group that seeks preference. But when the rule is set aside, the society easily falls into partisan racial or ethnic camps, with no common standard of what is fair.

The rule against racial discrimination would not exist without ample human experience of the disastrous effects of its absence. Any claim that suspending this rule in college admissions does not harm race relations should be supported by solid evidence and analysis, which the authors have failed to provide.

The Moral Authority of Racial Nondiscrimination. The moral authority of racial nondiscrimination, as a duty of each person to all others, has been undermined. The racial conduct of our respected colleges and universities is a powerful teacher. What it has taught, during the last three decades, is that racial discrimination is permissible, even desirable, when it favors certain people. Blacks and Hispanics have been glad to learn that racial preference for themselves and their fellows is praiseworthy, and increasingly feel free to demand and to practice it. And how many whites have absorbed and put to their own use the lesson that nondiscrimination is passé?

The principle that unequal treatment based on race is morally wrong was for many generations in America the lodestar of racial justice. Its comparative eclipse at present is symbolized by the fact that the effort to restore it in American law is routinely and vehemently attacked as "racist."

Do I overstate? One could argue that race-sensitive admissions are an exception, suspending the nondiscrimination rule in only one area of life. In housing, health care, voting, travel, entertainment, shopping, business, finance and criminal justice, racial nondiscrimination is still the law. However, this "exception" is not unique. It gains leverage by its kinship with widespread, robust, preferential policies in employment and public contracting. More important, race-sensitive admission has never been urged or practiced as an exception to the rule against discrimination. For the policy to be thought of as an exception, its proponents would have to admit that it discriminates. No such acknowledgment occurs.

Good rules are often improved and strengthened by exceptions that deal with special circumstances. (As, for example, the rule against killing is made more just and powerful by the exception for cases of self-defense.) But if an exception is to strengthen a rule, the fact of its being an exception, its reason, its limits, and beyond them the sovereignty of the rule must be clear. In this way exception and rule are both affirmed.

Race-sensitive admissions, however, have not been advocated as a carefully defined exception to the rule of racial nondiscrimination toward all. Instead, they have been implemented as the prototype of a contrary rule: racial preference for "underrepresented" minorities. The "exception" thus metastasizes, crowding out the rule to which it was purportedly subordinate. Race-sensitive admission undermines the moral authority of the rule against racial nondiscrimination, partly by the prestige of the institutions that practice it, but mostly by the failure to subordinate it to the general nondiscrimination rule. This cost, in my view substantial, is ignored by the authors.

The Quality of Work. As one would expect, the quality of student work has suffered. It stands to reason that people with less ability will do work of lower quality. A procedure that selects the less able over the more able is bound to vitiate the calibre of performance. Bowen and Bok agree that under race-sensitive admissions the less able (as measured by SAT scores) displaced the more able (pp 31-32). They grant also that the black students admitted via preference did poorer academic work (as measured by class standing) than the displaced students would have done (pp 72-78).

Such, of course, would be the result of any admissions preference based on something other than fitness for academic work. Private colleges give preference to children of alumni; public colleges, to residents of the state. In both cases the less able displace the more able, and learn less than the displaced students would have. The degree of harm to the quality of student work does not depend on whether the basis of preferential admission is race, alumni parentage or local residence, but on the size of the advantage conferred and on the number who receive it.

The society always suffers when its work is done less well, from the first year of school on up to the top ranks of the professions, business and government. Human problems and needs require the best hands we can find, with the best training we can give. But there are additional costs, to individual development and happiness, when important doors are shut or opened for reasons other than personal fitness. Many individuals are blocked thereby from doing what they can; others are pulled into doing what they cannot.

The institutional or social benefits from preferences based on alumni parentage or in-state residence may outweigh the costs to the quality of work and to individual development. But these costs certainly deserve to be counted. Likewise, the same costs should count when due to preference based on race. The authors partly admit the consequences described in this section, but they do not count the costs.

Affirmative Action in Course Content. Installed for the admission of students (and the employment of faculty and staff), racial preference has invaded course content. In literature and history courses, for example, considerations of racial or ethnic balance influence which authors are assigned and which people and problems are studied.

At first there seems to have been a genuine effort to base such decisions on merit. In literature courses, outstanding minority writers, previously neglected because of their race, were assigned because of their excellence. But this attitude soon gave way to racial bias: the worth and importance of minority authors were deemed greater because of their race. Reasons sprang up for not judging these writers according to the same standards by which others were held worthy of a place in the curriculum. Thus, the affirmative action process, with its racial double standard in the selection of students and teachers, was copied in the selection of what is studied and taught. This sacrificed what should be the educational opportunity of all competent students - to experience the very best poets, playwrights and novelists regardless of their race, religion, culture, politics, nationality or century.

The same kind of thing has occurred in history courses. Initially, one made sure that outstanding persons and movements, previously overlooked because of race, were included because of their importance. But race soon became a factor in deciding which eras, persons, actions and events were important. One now sought to assure blacks, Latinos, Asians and American Indians their share of pages in the textbook, whether or not their achievements or influence met impartial standards of historical significance. My objection here is not to ample coverage of racial conditions and relations, but to racial preference in determining what shall be studied and what shall be praised and condemned.

But is affirmative action in course content, however deplorable, a consequence of affirmative action in student admissions? It can be argued that the national solicitude for blacks after centuries of maltreatment would have produced some racial preference in faculty hiring and course content, irrespective of student admissions policy. Nevertheless, the impact of preferential admissions should not be discounted. This was the first area of academic life in which the racial nondiscrimination rule broadly gave way to affirmative action. Suspended in one area, the rule was easier to override in others. Larger numbers of minority students (admitted via affirmative action) led the demand for more minority faculty (hired via affirmative action), who joined the demand for course revisions (based on affirmative action).

Nothing should be more precious to a university than its standards concerning what is taught. To debase these standards is a greater evil than the compromises on competency induced by the preferential selection of faculty and students. Concerning the scope of the evil, I lack comprehensive data. This consequence of race-sensitive admissions is not mentioned by Bowen and Bok.

A Culture of Deception. The pretense that no substantial discrimination by race or sacrifice of quality occurs under affirmative action has spawned decades of evasion, equivocation and duplicity by its academic sponsors. Insistent on the justice of their conduct, they nonetheless conceal its nature: from black students, who might be humiliated; from white students, who might be resentful; from the public, who might disapprove; from the courts, who might forbid.

The determination to conceal is illustrated by a large hole in the authors’ database. For none of the 28 schools do we actually know how many students owed their admission to racial preference and how large a boost they required. The database does not contain the rules that the schools followed in granting preference, or enough data concerning the credentials of accepted and rejected applicants for an investigator to make a direct estimate of the race advantage. Either this information was refused or it was not sought. This, despite the importance of such information to the book’s initial premise - that without racial preference our top schools would lose most of their black students. (The authors had to rely instead on a plausible but circuitous calculation, pp 39-42, based largely on College Board statistics concerning the national SAT pool.)

For thirty years, evasion and deceit have largely succeeded in concealing the nature of race-sensitive admissions, partly because journalists and scholars have not been very inquisitive. The favorite label for racial preference in admissions (and elsewhere) is "affirmative action," a term that epitomizes evasion. It promises that something will be done, implies that it will be good, but carefully avoids saying what it is.

Schools with race-based admissions officially deny that they practice "racial discrimination." This denial is either a falsehood or an equivocation. The usual meaning of "racial discrimination" is "a difference in treatment or favor based on race." If this is what the school means, the denial is simply false. But officials may have in mind another definition, such as "a difference in treatment unfavorable to racial minorities," or "a race-based difference in treatment that we consider unjust." By either of these meanings the denial is literally true, but equivocal because intended to deceive. The college counts on the public to apply the usual definition of "racial discrimination," and thus to take the denial as a repudiation of racial preference.

Another example is the equivocal use of the term, "equality of opportunity." The original meaning of the phrase, in the Civil Rights Act of 1964, was "the same chance of selection for people with the same qualifications, regardless of race." This continues to be the common understanding. When an institution promises "equality of opportunity," people applaud, assuming that the policy bars unequal treatment. But the school may mean something else: e.g., "the same acceptance rate for black and white applicants, despite differences in qualifications." In this case the promise is literally true, but intended to deceive, unless the school clearly explains that it has reversed the generally accepted meaning of "equal opportunity."

An important practical cost of concealing and misrepresenting a policy is the prevention of its public consideration, based on fact and reason. A more profound cost, when the concealment and deception are practiced by an institution whose reason for being is the acquisition of truth, is corruption of soul. These costs of race-sensitive admissions are not mentioned in the book.

Intellectual Freedom Abridged. The intellectual freedom of students and faculty has been abridged. A climate of orthodox ("politically correct") opinion has developed concerning racial matters. Its primary function is to deny, obfuscate or attribute to white "racism" a painful fact of campus experience: the comparatively poor academic fitness of students brought into the community by race-based affirmative action. Without preferential admissions, this problem and therefore this remedy would not have arisen in anything like its present scope.

The orthodoxy holds that any deficiencies in the academic fitness of American blacks and Hispanics are the legacy of past or present oppression. The idea that these deficiencies may be due partly to the groups’ cultures or genes is anathematized as "racist." Opposition to affirmative action is suspect, but may be tolerated if it is not vigorous.

"Correctness" is enforced and dissent curbed by campus speech codes, indoctrination, accusations of "racism," administrative punishment, public ostracism, protest rallies, letter-writing campaigns, the disruption of classroom instruction and public lectures, and the destruction of student newspapers. In the administrative response to such conflict, a double standard commonly prevails. Mere speech by a "heretic" is more likely to elicit administrative punishment than coercive actions by the "correct."

The speech codes typically bar "verbal conduct" that "harasses" an individual, based on his race or ethnicity, by creating an "offensive environment." Oral or written statements concerning minority persons or groups, which bear negatively on their ability, conduct or history, readily evoke charges of harassment. The codes are not mere exhortations to civility, but rules whose alleged violation may lead to severe punishment.

Indoctrination in various tenets of the orthodoxy takes place during freshman orientation, and in workshops on "diversity," "multicultural sensitivity" and "oppression." These activities are commonly sponsored by administrative staff for student unions and residence halls. In addition to the doctrines noted above, students are told that "racism" requires both racial prejudice and the power to oppress, and thus that only whites are "racist," since they alone have such power. Also, that the United States and this college are profoundly "racist" in history and present character. Also, that students can shake off their upbringing as "oppressors" and learn to side with the "oppressed." (According to standard doctrine, cast with whites as "oppressors" are males, heterosexuals and the prosperous; along with people of color, the "oppressed" include women, homosexuals and the poor.) These ideas are not presented as possibly true, open to impartial investigation and debate, but as articles of faith.

A laudable, initial purpose of helping minority students to feel welcome in predominantly white institutions has been overwhelmed by the disposition to favor the newcomers as victims and wards. Opinions that conflict with this disposition or its supporting ideology have been regularly suppressed. The abridgment of intellectual freedom as a consequence of race-based admissions is not addressed by Bowen and Bok, although it also reaches the soul of higher education.


Of the benefits from race-sensitive admissions urged by Bowen and Bok, only one survived analysis as a clear, net gain: a reduction of the racial gap in incomes and status. Against this benefit I have named seven costs: minority achievement discredited, the racial nondiscrimination rule suspended, its moral authority undermined, the quality of work and the course of individual development impaired, racial preference injected into course content, academic policy made hostage to deception, and intellectual freedom abridged.

These costs are formidable, in my judgment more than sufficient to outweigh a benefit much larger than the one asserted. How the authors would weigh the costs we cannot tell, since most were not addressed in the book. But might they argue that the value of the acknowledged benefit had not been fully appreciated?

The Census Bureau projects that, in three decades, the US population will grow to about 13% black and almost 20% Hispanic (Statistical Abstract of the United States 1998, p 19). Bowen and Bok fear a dangerous increase in racial tensions if by that time "the vast majority of the top jobs in government, business and the professions continue to be held by whites, while one-third of the population is composed of blacks and Hispanics who are largely relegated to less remunerative, less influential positions." (p 268)

This is what the authors foresee, if racial preference is abandoned. Their prediction of persistent racial gaps in attainment is based squarely on their judgment that the racial gaps in academic skills as measured by grades and test scores "are substantial and show no signs of disappearing in the near future." (p 51) The escalation of tensions would presumably be due primarily to Hispanic and black, not to white and Asian, disaffection.

Bowen and Bok believe that eliminating or greatly reducing the racial gaps in earnings and influence should be a primary objective of national policy. They apparently assume that this goal is achievable via racial preference, and that its pursuit is imperative for racial peace and racial justice. Unfortunately, no data or analysis is presented to support these assumptions.

The authors are emphatic that, absent preference, the attainment gap would remain wide. This may be so, but are the current depth and scope of preference great enough to remove or radically to reduce the gap? The evidence from recent experience is negative.

By "top jobs" Bowen and Bok mean what elsewhere they call the "most lucrative and influential occupations," the category designated "managerial and professional" in government data. One can measure occupational participation by dividing a group's share in an occupational category by its share of the labor force. A result of 100% indicates equal participation; a lower number signifies underrepresentation. We have figures (SAUS 1998, p 417) for the last decade and a half, a period of racial preference in college and graduate school admissions, and in private and public employment. In 1983, the black proportion of managers and professionals, compared to the black share of all workers, was 60%; in 1997, 68%. The parallel Hispanic figures were 49% and 51%. The black share has risen substantially; the Hispanic share started lower and rose much less.

Bowen and Bok also emphasize the racial gaps in income. These can be measured by comparing the average incomes of blacks and Hispanics with the average income of whites. In 1980 black income per capita was 59% of white income; in 1996, 62%. The corresponding Hispanic figures were 59% and 52% (SAUS 1998, p 476). Thus, under the status quo, the black/white income gap has narrowed slightly, while the Hispanic/white gap has widened.

Judging by these indicators, at current levels of racial preference substantial racial gaps in income and influence will persist far into the next century. If the authors believe that the removal or radical reduction of these gaps is necessary for racial peace and racial justice, shouldn't they advocate a massive increase in the depth and scope of racial preference? There are many steps, for example, that government could take. It could provide the following advantages, specifically for blacks and Hispanics: full tuition scholarships for college and graduate school, interest-free loans for buying a home or establishing a business, partial exemption from income taxes, larger benefits from Social Security and Medicare, campaign expenses for political candidates, greater preference in selection for government employment and public contracts. The civil rights laws could be amended to permit for blacks and Hispanics the kind of discounts now given to senior citizens in travel, entertainment, shopping, etc.

Most Americans believe that racial justice requires racial equality before the law. But the policies that seem necessary if the racial gaps in attainment are to be radically reduced in "the foreseeable future" all require racial inequality before the law. They would provide unequal treatment of persons, based on their race, by statutes, regulations, courts and constitution. Racial justice cannot require a contradiction. It may command equality before the law, or progress toward equality of attainment through inequality before the law, but it cannot command both. If racial justice truly demands racial equality of attainment (or substantial progress thereto), as a higher principle than equality before the law, the latter can be sacrificed. But whence comes this demand? What is its authority or derivation? I can find no principle in the Declaration of Independence, the Constitution, the civil rights statutes or the American moral compact that supports any obligation to arrange an equal or indeed any particular level of attainment for a racial or ethnic group.

If racial groups are entitled to equal fortunes, why not religious groups? And groups defined by vocation, residence, hobby, ideology? Groups are composed of individuals, who lack any inherent right to equal attainment: how, then, can their groups possess such a right? If there are grounds for affirming the doctrine that racial justice requires racial equality of income and influence, they do not appear in the book.

But what about the dangers to racial peace from inequalities of attainment? If such gaps continue, will racial tensions escalate, exploding into riots, terrorism, interracial warfare?

Bowen and Bok simply assume - they present no evidence - that substantial racial inequalities of income and influence are enough to cause dangerous racial tensions. But the historical record suggests that racial inequality in fortune is neither a necessary nor a sufficient cause of dangerous racial tension, whereas racial discrimination is always necessary and often sufficient. Where racial or ethnic favoritism rules, severe racial antagonism, even racial warfare or subjugation, often results, whether or not the racial groups are equal in wealth or power. On the other hand, where racial nondiscrimination is the rule, these consequences do not result, whatever the racial distribution of wealth and power.

Racial discrimination and racial antagonism are closely connected, in a vicious causal circle. Racial discrimination breeds antagonism, in both the victim and the beneficiary; and racial antagonism breeds discrimination, both offensive and defensive. But the connection between racial inequality of attainment and racial antagonism is less strong. People do not move from the former to the latter without (1) strong feelings of racial identity, (2) a strong emphasis on racial difference and (3) a strong disposition to racial favoritism. These are the steps that take racial or ethnic groups from outcome inequality to antagonism. If these steps are not taken, group inequality of wealth and power is not dangerous. If they are taken, the society is in trouble, gaps or no gaps.

If these generalizations are unfounded, it should be easy to produce counterexamples. But Bowen and Bok do not mention any cases in which a wide racial or ethnic attainment gap has led to severe antagonism, in countries where nondiscrimination was the general rule. Nor do I know of any. Examples supporting my view appear in the historical relations of American ethnic groups. Despite great differences in wealth and influence between established whites and the more recently arrived Irish, Poles, Italians, et al., ethnic discrimination though frequent never predominated, and ethnic antagonism ranged from moderate to minimal. The close connection between antagonism and discrimination is illustrated by American racial relations. In the 19th century, severe antagonism between whites on the one hand and blacks, Indians and Orientals on the other, was always accompanied by strong racial discrimination; in this century, the antagonism and the discrimination have moderated in tandem, most successfully between whites and Asians.

The villain is racial or ethnic favoritism: inequality of treatment, not inequality of attainment. The latter is dangerous only if it results from discrimination or tempts people to discriminate. So long as the group gaps in grades and test scores remain, we cannot remove the gaps in income and influence without plunging deep into group favoritism. Fleeing a pseudo danger, this remedy runs straight into real and perhaps fatal evils.

America can live with these gaps justly and safely, if we resist the temptation they create to play racial or ethnic favorites. The focus needs to shift from group identities and comparisons to individual needs and abilities. It is the business of schools, from kindergarten through graduate study, to offer individual students the best education of which they are capable, regardless of the group to which they happen to belong. Membership in a racial or ethnic group whose performance is below average does not justify discrimination, whether favorable or adverse. It is wrong when teachers allow such membership to lower their assumptions and expectations concerning a student's ability and performance. And it is wrong when colleges make such membership a reason for lowering standards of admission.

Our multiethnic, multiracial country can boast an amazing record of successes in reducing first ethnic and then racial favoritism. But we are by no means immune to the forces that would increase such favoritism, to the damage it causes, and to the dangers it creates.


Other detailed reviews of The Shape of The River:

  • Ronald Dworkin, "Affirming Affirmative Action," The New York Review of Books, XLV:16, October 23, 1998, pp 91-102. [Enthusiastic]
  • Terrance Sandalow, "Minority Preferences Reconsidered," Michigan Law Review, Vol 97, No 6 (May 1999) pp 1874-1927. [Mixed]
  • Stephan and Abigail Thernstrom, "Reflections on The Shape of the River," UCLA Law Review, Vol 46, No 5, June 1999, pp 1583-1631. [Critical]

An earlier study, using a national database, of the impact of preferential college admissions on the graduation rates and later earnings of black and Hispanic students:

  • Thomas Kane, "Racial and Ethnic Preferences in College Admissions," in The Black-White Test Score Gap, edited by Christopher Jencks and Meredith Phillips, Brookings Institution Press, Washington, D.C., 1998, pp 431-456.

A recent study, using Bowen/Bok’s dataset but contradicting their findings concerning the impact of attending a more selective school on graduation rates and later earnings:

  • Stacy Berg Dale and Alan B. Krueger, "Estimating the Payoff to Attending a More Selective College": Mellon Foundation Working Paper #409, Princeton, N.J., July 1999. [Dale was thanked in River as one of the authors’ "principal collaborators."] Online at: http://papers.nber.org/papers/W7322

*My thanks to Roger Clegg, Carl Cohen, William Dickens, Mary Alice Fisher, Russell Nieli, Champe Ransom, Gary Rosen and Jo Smith for reading and criticizing the manuscript.