Affirmative action cannot solve the American dilemma of racial inequality.

Marie Gryphon is a senior fellow at the Manhattan Institute’s Center for Legal Policy and an adjunct scholar at the Cato Institute, specializing on issues related to education policy. She holds a J.D. from the University of Washington and maintains a private law practice, working in the areas of ERISA, securities, class action, and constitutional law among others.

In the wake of the Supreme Court’s 2003 decision upholding admissions preferences,1 affirmative action remains a deeply divisive issue. Ward Connerly has called it the civil rights struggle of our time. This tendency to frame the argument over preferences in terms of fundamental values is common to both sides of the debate. Because our nation’s history with respect to race is so painful, the resulting argument is heated, personal, and ultimately unproductive.

Overwhelmingly, such debates turn on considerations of “fairness” or “merit,” as if there were one best way to admit students to college. For those who favor little or no role for government in higher education, however, these are red herrings. There is no “fair” way to admit students to elite public institutions at the expense of taxi drivers and construction workers. Subsidies to particularly talented and capable students are especially difficult to justify. In the private sphere, on the other hand, institutions deserve broad latitude to create the educational environments they deem effective for their institutional mission.

The most broadly appealing argument against racial preferences in college admissions is that they are uniquely harmful, both legally and socially. In public universities, preferences have broken down constitutional protections against classification by race—protections that form a still insecure bulwark against habits of racial abuse and oppression that have festered for centuries. Erosion of the legal doctrine of racial neutrality is a high price to pay for a system of preferences that moves only a few thousand students a year from one college to another, but it is a price the Supreme Court has unwisely chosen to pay. Preferences are only permitted, not required, however, and policymakers should reassess whether the benefits of racial classification in schools outweigh the costs.

This Policy Analysis addresses support for racial preferences on the narrowest possible ground: the claim that they benefit formerly oppressed racial groups and promote racial healing. This study shows that this claim is untrue. Administrators and policymakers of all political persuasions should therefore oppose racial preferences in universities.

The Resurgence of Preferences

In the late 1990s racial preferences appeared to be on the decline. Critics of preferences persuaded voters in California and Washington that such policies were harmful and divisive, and the voters in those states approved initiatives banning racial preferences at public universities. A federal appeals court struck down affirmative action at the University of Texas, holding that preferences violated the U.S. Constitution’s guarantee of equal protection under the law.

But the tide has swiftly turned since the Supreme Court’s decision in Grutter v. Bollinger to uphold racial preferences at the University of Michigan School of Law. Whereas the Court struck down the university’s “mechanistic” approach to affirmative action in its undergraduate school in a related case, it upheld the law school’s non‐​quantified, “individualized” approach to preferences. As a result of these cases (collectively “the Michigan Cases”), racial preferences in public colleges and universities are unambiguously legal as long as they are implemented without numbers, weights, or stringent guideline.

Supporters of affirmative action seized this opportunity to reaffirm existing preferential programs and re‐​institute programs previously abandoned or struck down. The University of Texas system, which had dropped affirmative action under a now‐​obsolete court order, immediately announced a plan to resume consideration of race in its admissions process for the class of 2005. Virginia Tech, which briefly abandoned preferences due to legal concerns, reinstituted their program pending the Supreme Court’s decision in the Michigan Cases.

The California General Assembly passed a bill last summer to reintroduce preferences in the University of California system. Gov. Arnold Schwarzenegger vetoed the bill only because he believed a constitutional amendment would be necessary to override Proposition 209, which banned preferences in the state in 1996, and supporters of preferences are now seeking to pass an amendment there restoring affirmative action. In Washington State, Gov. Gary Locke has asked the legislature to pass a bill restoring preferences there as well, though that measure has not yet come to a vote.

Administrators at the University of Michigan quickly altered the school’s undergraduate admissions program to allow consideration of race in the same way that its law school does. Ohio State University also tweaked its affirmative action system to comply with the Court’s ruling, and University of Minnesota president Robert Bruininks expressed relief that his school’s affirmative action program already complied with the new ruling.

Because legal barriers to racial preferences in state universities have been eased, it is more important than ever for policymakers to consider whether these policies, even if legal, offer the benefits that supporters claim.

The Myth

The myth about preferences is perpetuated by some of America’s most influential academic and political leaders. It holds that racial preferences in selective universities benefit minority students in concrete ways, and that without preferences colleges would become “re‐​segregated,” depriving American students of the educational benefits of a diverse student body. It also holds that the social and psychological costs of preferences are modest—as University of Michigan Dean Earl Lewis writes—that affirmative action “is not about the weakening of standards or the fraying of interracial relations.”

William G. Bowen, the former president of Princeton, and Derek Bok, the former president of Harvard, became standard‐​bearers of “The Myth” with the publication of their book, The Shape of the River. Using a privately owned database assembled with the permission of a handful of the nation’s most selective colleges and universities, Bowen and Bok offer a “graphic and quantifiable” defense of The Myth: that the “net social benefits” of preferences at selective schools are “impressive,” and are achieved at “a tolerable cost.” Their work was cited at length in amicus curiae briefs filed with the Supreme Court in 2003 by dozens of elite colleges and universities in the Michigan Cases.

Despite the academic establishment’s uncritical defense of preferences, recent research confirms what many academics, policymakers, and students have quietly suspected: this view of affirmative action is a myth. Preferences do not offer substantial benefits to preferred racial groups, and they do impose social, psychological, and practical costs on students of all backgrounds.

Preferences Do Not Send More Minority Students to College

Affirmative action defenders frequently and correctly tout the importance of college to the goal of improving life prospects. Bowen and Bok comment at length about the importance of a college education. They write, “The growing numbers of blacks graduating from colleges and professional schools, and the consequent increase in black managers and professionals, have led to the gradual emergence of a larger black middle class.” They are right. Few things foster professional success more reliably than a college education. College has helped many minority students achieve middle‐​class lives.

NAACP attorney William Taylor’s remarks are typical of efforts to connect racial preferences at elite schools to the issue of college access: “There can also be little question that affirmative action policies of colleges and universities [have] played a large role in the major increases in minority college enrollment that we saw during the 1970s and 1980s.” But preferences have not increased college access. In fact, Thomas Sowell observes that black college enrollment increased at least as quickly in the 1950s and early 1960s, prior to the establishment of affirmative action policies, as it did afterwards.

The reason that affirmative action does not affect college access is that most four‐​year colleges and universities in America are not selective; they take anyone with a standard high school education. Preferences are policy only at the 20–30 percent of American colleges that have substantially more applicants than places. Students attending those schools have many other college options.

The reason that minority students do not get college degrees as often as white students is not competitive admissions policies. Rather, the problem is that most minority students leave high school without the minimum credentials necessary to attend any four‐​year school, selective or not.

Freshmen must be “college ready” at almost all four‐​year colleges. That means that students must be literate, have a high school diploma, and have taken certain minimum coursework. Overwhelmingly, minority students are not college ready. Political scientist Jay Greene of the Manhattan Institute found that only 20 percent of black students and 16 percent of Hispanic students leave high school with these basic requirements.

Minority underrepresentation in college is caused by public schools’ failure to prepare minority students. It is a failure that affirmative action does not remedy. “College‐​ready” minorities are already slightly more likely to attend college than their white counterparts. Even if affirmative action were ended, every minority student affected by the policy change would have a college opportunity at some four‐​year school.

Preferences Are Not “Plus Factors”

Elite public and private universities claim that affirmative action is only a light “thumb on the scale”—a “plus factor” for deciding between candidates with virtually equal qualifications. University of Minnesota general counsel Mark Rotenberg says that the school uses race as “a plus factor together with many other factors in building a class that will meet the diversity objectives that [its] Regents have set.” Dean Herma Hill Ray of UC Berkeley’s Boalt Hall described affirmative action as a way of choosing “between two equally qualified persons.” But that is not true. Preferences for minority applicants to such flagship schools are enormous, and they generate painfully obvious gaps between racial groups on campus in terms of academic preparation.

Brookings Institution economist Thomas J. Kane estimated the size of preferences at selective schools and found that black applicants enjoyed an enormous advantage over white and Asian applicants to selective schools. The preference was, on average, equal to the combined effect of 200 points on the SAT and over one‐​third of a grade point (on a 4.0 scale), and was generally larger at the very most selective institutions.

Preferences this large inevitably produce large gaps in average academic preparedness between students of different races on college campuses. University of Pennsylvania sociologist Douglas Massey and his colleagues write, “While we are not privy to actual admissions processes, we do know that they operate to produce a freshman class composed of two very distinct subpopulations. On one hand are whites and Asians and on the other are Latinos and blacks.”

These differences in preparation cause minority students to receive low grades. African‐ American college students earn grade point averages about two‐​thirds of a letter grade below their non‐​minority peers. They are far more likely to drop out, and those who graduate finish, on average, in the bottom 25 percent of their college class.

University representatives often equate racial preferences with the preferences given to children of alumni and consideration of other “diversity” factors such as musical talent. Attorneys for several selective institutions write, “Admissions officials give special attention to, among others, applicants from economically and/​or culturally disadvantaged backgrounds, those with unusual athletic ability, those with special artistic talents, applicants who write exceptionally well, [those] who show a special dedication to public service, and those who demonstrate unusual promise in a wide variety of fields.”

But it is not true that racial preferences are comparable in size to the boost one gets from being a violinist or the child of an alumnus. For no group other than preferred racial minorities (and varsity athletes) are preferences so large as to leave that group visibly and consistently at the bottom of their college class.

The point system formerly used by the undergraduate program at the University of Michigan offers insight into the relative weights given to various nonacademic admissions factors. Special talent in music or other extracurricular activities were worth a maximum of 5 points in the system, whereas membership in a preferred racial group was worth 20 points, conferring an advantage equal to the difference between a “B+” grade point average and a “C+” average.

Preferences Do Not Increase Earning Power

No contention is more central to The Myth than that preferences are a catalyst for upward financial mobility. Moderate supporters of affirmative action tolerate the social costs of preferences because they hope that preferences will improve the concrete well‐​being of minority students after graduation.

Indeed, research used to suggest that attending a more selective college was related to substantial, though not huge, financial gains. Generally, studies indicated that attending a school with an average SAT score 100 points higher would increase a student’s future earnings by 3–7 percent.

But those studies suffered from a serious methodological problem. They were unable to take into account many of the factors that colleges look at when deciding which students to admit. Academic researchers generally have only high school GPA and SAT scores at their disposal, so they must compare students with the same grades and scores and assume that the students are otherwise the same. Teacher recommendations, the difficulty of the high school attended, and student motivation as reflected in an admissions essay are all unavailable to researchers. As a result, researchers attributed wage premiums to “equally qualified” students who attended more prestigious schools, when in fact the students were not equally qualified at all.

But recent research has shown that this part of The Myth, like the others, is untrue. Attendance at a more selective school does not raise students’ future incomes, regardless of race.

Economists Stacy Dale and Alan Krueger developed an ingenious method to solve these problems and compare students who were truly alike. They “matched” several thousand students nationwide on the basis of selectivity of the schools that accepted and rejected them and compared members of the matched groups only to each other. This was possible because only 62 percent of students in the sample chose to attend the most selective school that accepted them.

Thus, Dale and Krueger were able to compare students who were accepted by a top tier school and actually attended that school to students who were accepted to that same top school but chose instead to attend a less selective school. Comparing students with identical acceptances takes into account (and “controls for”) all of the factors that colleges take into account when they accept students.

Dale and Krueger found that when genuinely equivalent students were compared, students attending less selective schools made just as much money as students who attended more selective schools. The idea that a selective university will make you rich is just another part of The Myth.

Affirmative action supporters frequently claim popular support for their cause among elite college students, graduates, and faculty. Bowen and Bok, for example, find that admissions preferences are popular on the basis of surveys showing that college alums thought their institutions ought to place even more emphasis on diversity.

But they got the “right” answer by asking the wrong question. Students and faculty do value diversity, but that does not mean that they support differential admissions standards in order to achieve racial balance. Most polls suggest that students and faculty are closely divided on the issue of preferences but that majorities of both groups do not support them.

A poll of Berkeley students taken at the time that Proposition 209, which banned preferences in state university admissions, was on the ballot in California showed that most students opposed affirmative action. New York Times columnist James Traub reported, “Berkeley students, it turns out, are like most Americans: they want diversity without the zero‐​sum calculus that inevitably accompanies affirmative action.”Similarly, a Roper poll found that UC faculty members were split on the issue, with 48 percent opposing admissions preferences and only 31 percent expressing support.

Aware of these polls, economists Harry Holzer and David Neumark, who support preferences, make a more cautious statement, that “public opinion polls still indicate public support for some forms of affirmative action.” They are right, but only those forms of “affirmative action” that do not involve preferences (such as outreach and remediation) command support. Prof. Stephen Cole reports, “Surveys suggest that a majority of both students and faculty are opposed to policies in which race trumps qualifications.”

Moreover, in highly charged university environments, faculty members are sometimes afraid to admit that they oppose preferences. Berkeley professor Martin Trow writes, “Very few academics wish to offend both the senior administrators who govern their careers and budgets and the well‐​organized affirmative action pressure groups that will quickly stereotype faculty members as ‘racists’ or, at very least, ‘right‐​wingers.’” Thomas Sowell recalls “bitter fights” that have erupted among faculty members about whether affirmative action policies should be decided by secret ballot, because whether the votes were public might affect the results.

The Harm

The foregoing suggests that many benefits attributed to preferences do not exist. But The Myth is worse than useless. It perpetuates a policy that is harmful to students of all backgrounds, especially minority students.

That is the argument against preferences that their supporters assail most energetically. Bowen and Bok optimistically asserted that their findings “have essentially disposed of the ‘harm‐​the‐​beneficiary’ line of argument. There is no empirical support for it.” This epitaph has proved premature. Recent research contradicts this claim on the basis of far more sophisticated methods than those used by the former university presidents.

Dropout Rates

Black students are less likely than white students to graduate from any institution of higher learning. Latino students also graduate at relatively low rates. That persistent problem depresses the wages of minority workers and is of concern to policymakers who seek to close the socioeconomic gaps between racial groups. Opponents of affirmative action have long contended that preferences increase minority dropout rates.

Bowen and Bok argued on the basis of SAT scores alone that equally qualified students are actually more likely to graduate if they attend more selective schools. However, their analysis assumes that the average minority student with an SAT score of 1250 at the University of Michigan is as academically prepared as the average minority student with the same SAT score at Yale. That is unlikely. The student accepted to Yale probably presented additional evidence, such as advanced placement work or an excellent essay, that made their application more attractive by reflecting skills likely to be useful in college. Bowen and Bok admit that SAT scores alone do not reflect differences between students as well as instruments that combine several measures of preparedness.

Moreover, like efforts to predict the effect of college selectivity on wages, predicting the effect of selectivity on dropout rates is made difficult by the presence of unobserved factors, such as motivation, that effect student outcomes. The techniques used by Bowen and Bok cannot take these differences into account.

Sociologist Robert Lerner, now commissioner of the National Center for Education Statistics, observed, “Despite its size, The Shape of the River includes largely cursory statistical analysis of applicant data.” Bowen and Bok are aware that their work is rudimentary. “In due course,” they write, “we expect others, using more sophisticated econometric techniques, to expand the analysis presented here.”

Economists did subsequently analyze the question of dropout rates in more detail and got very different results than Bowen and Bok. Economists Audrey Light and Wayne Strayer were able to better predict university completion patterns among students of different abilities. They did this by using methods that took into account unmeasured student qualities, as Dale and Krueger did in their study about wage rates.

When student differences were held equal, Light and Strayer found that the likelihood of graduating from college depended on how close the “fit” was between a given student and his or her classmates in terms of academic preparedness. They write: “Our estimates reveal that the ‘match’ between student ability and college quality does have a causal effect on college completion.”

Light and Strayer divided both students and schools into four categories based on standardized test scores, and predicted the probability that students in each score category would graduate from colleges in each selectivity category.

Light and Strayer found that the least prepared students were most likely to graduate if they attended the least selective schools. Their graduation rates are lower at more selective institutions. The most prepared students exhibited the opposite pattern: their chances of graduating were highest at the most selective schools. Students with middling levels of preparedness did best at colleges of middling selectivity, with their graduation rates tailing off slightly both at nonselective schools and at highly selective schools.

Although minority college attendance has increased rapidly in recent decades, minority graduation rates have not kept pace. Research that suggests how graduation rates may be maximized is thus important to policymakers who seek to close racial gaps in educational attainment and earnings. Light and Strayer’s findings suggest that students are most likely to graduate at colleges attended by peers of roughly equal academic strength.

Affirmative action may increase minority dropout rates by mismatching students and schools. Massey and his colleagues also find that a student’s sense of being a poor fit at his or her school makes the student more likely to drop out. Academically, students feel like a poor fit at college if their classes are either too easy or too difficult for them.

Status over Substance

Too often today, Americans view college as a zero‐​sum status competition rather than a learning opportunity. Status‐​conscious parents have so personalized this process that one selective school has banned them from student campus tours so that students feel free to ask their own questions.

Washington Post education reporter Jay Mathews observes, “We are talking about colleges the same way we talk about wine or left‐​handed pitchers or American Idol contestants. This is fashion and marketing and branding, not real value being added to our lives, or to our children’s lives.”

Parents love to talk about their children’s accomplishments, and admission to a selective school is an accomplishment to the extent that only those who excel academically can achieve that goal. But a myopic parental focus on bragging rights ill‐​serves students by transforming what should be a learning opportunity, a chance to build new skills and better oneself at no cost to any one else, into a winner‐​take‐​all competition.

Partly as a result, discussions of affirmative action tend to focus on notions of “fairness” and “merit” rather than concrete evidence, as if college admission were a cash prize or a commendation for good behavior. Sowell writes, “Discussions of college admissions opportunities often proceed as if the issue is the distribution of benefits to various applicants, when in fact the issue is selecting those applicants who can best master the kind and level of academic work at the particular institution.”

Affirmative action exacerbates our cultural tendency to look at college selection in terms of prestige because preferences only promote equity if selective colleges are objectively “better” than others, rather than merely better fits for some students. Having promoted for decades the notion that prestige matters, selective schools now generate resentment by apportioning this prestige according to race.

Even from the perspective of status, affirmative action harms minority students. In our stratified system, the college a student attends says quite a lot about her level of academic preparedness. But at elite schools, admission now signals two different levels of achievement— one for white and Asian students, and another for black and Latino students—which diminishes the cachet of admission for the latter group. Berkeley linguistics professor John McWhorter writes, “I was never able to be as proud of getting into Stanford as my classmates could be. After all, growing up [middle class], how much of an achievement can I truly say it was to have been a good enough black person to be admitted, while my colleagues had been considered good enough people to be admitted?”

One of the self‐​defeating effects of affirmative action is that, in a university culture that attaches inordinate social value to credentials, preferences dilute those credentials for minority students who would be admitted to selective schools without them. To the extent that an acceptance letter from a “top school” is a trophy signifying an extraordinary accomplishment, America’s highest‐​achieving minority students are being robbed of the recognition they deserve.

Stereotype Threat and Under‐​performance

Most critically, recent research shows that affirmative action impedes academic achievement by undermining minority students’ confidence. This hypothesis is one of many that researchers have generated to explain the mysterious phenomenon of minority under‐​performance in college.

The term “under‐​performance” does not refer to differences in minority college grades and graduation rates that can be explained by available measures of preparedness, such as high school grades and SAT scores. Rather, “under‐​performance” is what researchers call the tendency of African‐​American and Latino students to obtain lower college grades and graduation rates than white and Asian students with identical previous grades and test scores.

Nuanced, difficult‐​to‐​measure aspects of academic preparedness (the same ones that confounded economists before Dale and Krueger) play some role in the phenomenon of under‐​performance, but they cannot explain it entirely.

Critics of preferences have long argued that double standards in admissions are harmful to preferred students’ self‐​esteem in competitive situations, and thus contribute to under‐​performance. Shelby Steele observed, “The effect of preferential treatment—the lowering of normal standards to increase black representation— puts blacks at war with an expanding realm of debilitating doubt, so that the doubt itself becomes an unrecognized preoccupation that undermines their ability to perform, especially in integrated situations.”

Until recently, little research was available to support or refute this view. But two separate studies, one by sociologist Stephen Cole of the State University of New York at Stony Brook and Elinor Barber and another by Massey’s Pennsylvania group, confirm what seems to be intuitively true about preferences. Preferences harm students’ self‐​images, and this harm has practical costs in terms of grades and graduation rates.

Both studies build on earlier work by Stanford University sociologist Claude Steele, who coined the term “stereotype threat” to refer to the decline in performance suffered by members of groups who become afraid of confirming negative group stereotypes. Steele tested his theory by giving standardized exams to groups of white and African‐​American undergraduates at Stanford University.

Testers told some groups that the exam evaluated psychological factors related to testing, and that it was not a measure of ability. They told other groups that the exam measured their intellectual abilities, and in some instances had them indicate their race on the exam. The African‐​American students who had been implicitly “threatened” with the stereotype of minority academic inferiority did markedly worse on the exam than black students in the other groups.

Steele and colleague Joshua Aronson conclude, “Making African‐​Americans more conscious of negative stereotypes about their intellectual ability as a group can depress their test performance relative to that of whites. They also find that stereotype threat can be triggered by “quite subtle changes of environment” and that reducing stereotype threat “can dramatically improve blacks’ performance.”

Cole and Barber established a connection between stereotype threat and racial preferences in a book published in 2003. Titled Increasing Faculty Diversity: The Occupational Choices of High‐​Achieving Minority Students, their book sought to determine why there are so few minority college professors and how their numbers might be increased.

Cole and Barber found that high levels of academic self‐​confidence were critical to a student’s decision to follow up on an interest in a career as a professor. They also found that minority students at highly selective universities suffered from lower academic self‐​confidence than their counterparts at less selective schools. This diminished confidence caused minority Ivy Leaguers to abandon their academic aspirations at twice the rate of comparable non‐​minority students in state universities.

Cole and Barber concluded that stereotype threat is activated among high‐​achieving minorities by racial preferences at selective schools. Preferences ensure that minority students as a group will be less prepared than their peers. Even minority students who do not need preferences respond to an environment characterized by the relative academic weakness of minorities by worrying about confirming a negative stereotype.

Stereotype threat is not merely a personal problem affecting feelings of satisfaction or school friendships. As Steele’s early work suggested, it has concrete effects on minority achievement in academic settings.

To try to understand the mystery of minority under‐​performance, Massey and his colleagues at the University of Pennsylvania researched the histories of students attending elite universities. The group found that difficult‐ to‐​measure socioeconomic factors and finer‐​grained measures of academic preparedness played some role in the performance gap. They also determined that vulnerability to Claude Steele’s stereotype threat is related to the lower grades earned by minority students.

Massey and his group found that those black and Latino freshman particularly susceptible to stereotype threat received grades that were on average .122 points lower on a 4.0 scale than minority students who felt less threatened.97 This is not as small a difference as it may appear to be: It is one‐​third of the entire black‐​white GPA gap of .36 at the sampled schools and more than half the gap of .22 that persists after background and academic preparation are held equal.

Even if minority students who were not particularly vulnerable to stereotype threat were wholly unaffected by it, the Pennsylvania group has shown that stereotype threat explains at least half of the mystery of minority under‐​performance at elite colleges. If, as seems likely, even minority students who are not especially vulnerable feel threatened to some extent, stereotype threat becomes the primary explanation for under‐​performance.

This finding is consistent with the fact that African and Caribbean immigrants do not under‐​perform in American colleges. Because they do not carry the heavy psychological baggage of slavery and segregation with them to school, they are far less vulnerable to stereotype threat than African Americans.

Stereotype threat may do even more harm than lowering grade‐​point averages. Massey and colleagues found that susceptibility to stereotype threat increases the likelihood of dropping or failing a first semester class, events that are related to the likelihood of dropping out of school.

Feelings of insecurity worsened by double standards in university admissions are no small cost to be paid for the cause of practical benefits. Stereotype threat, always present to some extent in academic settings, is exacerbated by affirmative action. It has measurable costs to minority students in the form of lower levels of academic achievement and the abandonment of at least one academically ambitious career goal: that of college professor.

Isolation and Stigma

Nearly as bad as the problem of under‐​performance is the harm that preferences do to race relations among America’s highest‐​achieving young people. Thomas Sowell writes, “Even in the absence of overt hostility, black students at M.I.T. complained that other students there did not regard them as being desirable partners on group projects or as people to study with for tough exams.” Law professor Eugene Volokh relates the story of a law student who claimed that he and his friends chose classes with high minority enrollments because they believed that competition for good grades would be less severe. Such preconceptions can contribute to feelings of social distance between peers of different races.

Massey and his colleagues surveyed thousands of students attending selective schools to find out how they felt about members of other racial groups in general, and affirmative action beneficiaries in particular. They found that all students generally had positive feelings about members of other racial groups. However, white and Asian students had notably cooler feelings towards “affirmative action beneficiaries” than others of any race. The researchers conclude:

Such perceptions of distance from “affirmative action beneficiaries” carry important implications for the general tone of race relations on campus because [many students believe] that without affirmative action most black and Latino students would not be admitted. To the extent that such beliefs are widespread among white students at elite institutions, they will not only increase tensions between whites and minorities on campus; they will also increase the risk of stereotype threat by raising anxiety among minority students about confirming these negative suspicions.

Preferences generate distrust between racial groups that works against the mission of diversity in education: promoting mutual respect and understanding between students of different backgrounds.

The Ratchet Effect

Although only 20–30 percent of colleges and universities use racial preferences, they enlarge gaps in academic preparedness between white and minority students at other colleges because of what researchers call the “ratchet effect.” The ratchet effect ensures that the policies of a handful of elite public and private schools have harmful effects at colleges all along the selectivity continuum.

The ratchet effect begins at Harvard College. Harvard has long been able to attract an extremely high percentage of the tiny number of black and Latino students who graduate from high school each year with truly Ivy League credentials. As a result, the academic gap between white and preferred minority students at Harvard is among the smallest anywhere. African‐​American freshmen at Harvard have average SAT scores that are only 95 points below those of their non‐​minority peers. Not surprisingly, the African‐​American graduation rate at Harvard is the highest in the country.

But Harvard’s gain is a loss for the rest of the Ivy League. To remain as racially diverse as Harvard, Princeton must employ preferences large enough to produce a freshman class with a 150‐​point black‐​white SAT gap. Columbia tolerates a 182‐​point gap. Because every Ivy League school other than Harvard has attracted and admitted those minority students who would, under race‐​neutral standards, be well‐​qualified to attend schools like Wellesley and NYU, these schools must in turn admit minority students whose grades and scores more nearly match those of white and Asian students at schools such as the University of Virginia or the University of Texas. Those flagship state university systems then come under tremendous political pressure to employ preferences also, since if they do not, they will lose their successful minority applicants to even more selective institutions.

The result, pictured in Figure 2, is what Thomas Sowell has called the “mismatching” of minority students and colleges. The ratchet effect ensures that even colleges that do not have preferences struggle with large gaps in academic preparedness, because their white and Asian applicants are far stronger than their African‐​American and Latino counterparts. Sowell describes the downstream effect of preferences in the University of California system: “Thus, San Jose State University had 70 percent of its black students fail to graduate [during the 1980s], just like Berkeley, though it is doubtful that the minority students at Berkeley would have failed at San Jose State. That is the domino effect of mismatching.”

With selective schools educating only a few thousand of the approximately 100,000 black and Latino students who receive BA degrees each year, affirmative action is the tail that wags the dog. Preferences at elite private schools exacerbate the political pressure on much larger flagship state institutions to use racial preferences to avoid becoming racially homogeneous. These public institutions often choose to respond to this pressure by adopting preferences, which contribute to painfully large academic gaps between racial groups at many nonselective public and private institutions.

Why The Myth?

Given the falsity of The Myth, it is natural to wonder why the educational establishment vigorously embraces it. Many scholars who pride themselves on the fearless pursuit of truth are mute about problems with affirmative action. The answer may be that the academic establishment wants to free itself from the taint of historical racial prejudice while retaining its exclusive status in American society.

Moral Redemption of Schools

Affirmative action programs are the primary way that college administrators offer an institutional apology for the exclusionary policies of decades past. Affirmative action is thus an expressive act as much as a policy decision.

Institutions that have discriminated in the past should acknowledge and remedy those wrongs. But racial preferences are a poor vehicle for doing that. The academic establishment’s desire to redeem its institutions from past sins does not justify such a harmful policy. Instead, selective schools should focus on outreach designed to build real academic skills and confidence among students of all backgrounds, and should work hard to ensure that the students they do admit have the support they need to succeed in demanding academic programs.

Preservation of Academic Elitism

There was a time when, as one author wrote, “selective colleges would rather be selective than integrated.” Much contemporary debate centers on whether these priorities are now reversed—whether these same schools would rather be integrated than academically selective. But one thing is certain: affirmative action has been their way of avoiding this uncomfortable choice.

One reason that elite schools defend racial preferences so heatedly is that alternative methods for producing diversity, such as Texas’s guarantee of admission to the top 10 percent of students from every state high school or lotteries among qualified students, would make the non‐​minority students at those schools a less elite group. This is so because those alternative policies admit many non‐​minority students with lower grades and scores as well as minority students. Attorneys for several selective colleges contend that ending affirmative action “would compel them to trade selectivity to obtain diversity.”

Selective schools enjoy their exclusive cachet and don’t want to admit a larger cross section of white and Asian students in order to achieve racial diversity. Because they value their status so highly, they instead subject their students and the larger society to harmful policies that mix far less qualified pools of minority students into student bodies otherwise composed of very highly qualified white and Asian students. Then, they dissemble about the size of academic disparities that are nonetheless obvious to students and teachers.

Legal theorist Charles R. Lawrence III notes that affirmative action is a conservative policy in the sense that, by maintaining separate admissions standards, it allows for more racial mixing while protecting the exclusivity of selective schools. Instead of creating educational environments that embrace a greater variety of students of all races, preferences “do not challenge … conventional selection processes or standards of merit.” Rather, they bolster popular support for flagship state universities and other top schools whose mission is “the education and legitimization of an intellectual and professional elite.”

Supporters of preferences decry the possibility of making schools less selective by admitting more students of all types. They warn that alternative admissions plans such as Texas’s “top 10 percent” strategy produce “a spurious form of equality that is likely to damage the academic profile of the overall class … far more than would anything accomplished through race‐​sensitive admissions policies.” Racial preferences are popular among schools that (for better or worse) want to preserve their exclusive cachet.

Cover for Companies

University administrators often cite strong support for admissions preferences by industry as evidence that preferences are beneficial. It is true that much of corporate America has leapt to the defense of college admissions preferences in recent years. Indeed, 65 multinational companies including Nike, Microsoft, and American Express filed an amicus brief with the U.S. Supreme Court urging the court to uphold the University of Michigan’s affirmative action programs. General Motors stated in a press release that the “elimination of affirmative action in leading educational institutions would deprive businesses of the well‐​trained minority candidates who are essential to our nation’s economic success.”

But as we have seen, preferences do not improve the skills or wages of minority graduates. Rather, corporations support admissions preferences because they are trying to avoid civil liability for both “discrimination” and “quotas” at the same time. Affirmative action by colleges helps corporations disguise the fact that they, too, must employ preferences to achieve diverse workforces. By giving less prepared minorities the same alma maters as more prepared peers, affirmative action at selective schools makes workplace preferences less obvious.

Corporations thus encourage affirmative action at schools where they recruit graduates, and schools that hope to place students with these companies have an incentive to oblige them. Companies could recruit high‐​achieving minority students without admissions preferences, but some of them would be attending schools that are currently below the radar of top management training programs and investment banks.

In the absence of preferences, companies would have to do one of two things. They would either have to admit that they are willing to consider minority graduates from less selective schools even if they only recruit white and Asian applicants from the Ivy League, or they would have to consider applicants of all races from a wider variety of schools. This last option may be the wisest in light of Dale and Krueger’s finding that a student at a less selective school will be just as successful in time as her counterpart from a “top school.”

The Way Forward

Racial preferences in college admissions cannot offer the benefits their boosters have promised, and they harm American students of all races by impeding learning and generating unnecessary suspicion and distrust between groups. Whereas private universities have a right to pursue unwise admissions policies if they wish, policymakers should not allow selective state institutions to follow their example. Although the Supreme Court has ruled that preferences are constitutional, they can and should be banned at public institutions because they are bad public policy.

Ending preferences does not amount to abandoning the dream of real racial equality and healing in America. Affirmative action supporters sometimes deride opponents by saying effectively, “Well, we’re doing something about this terrible problem of inequality in American society. What do you want to do?” Although good intentions cannot excuse a harmful policy like racial preferences, the question is a good one. What follows are suggestions for improving the educational opportunities and achievements of minority students.

Acknowledge History

All policymakers, particularly those who oppose the use of racial preferences in admissions, should acknowledge the role that America’s shameful history of slavery and segregation have played in producing current disparities between white and minority students, particularly African Americans.

Prof. John McWhorter observed a sense of cultural distance on the part of African‐ American students toward academic endeavor that results from “whites having denied education to blacks for centuries.” He writes, “It is not the fault of black Americans that they have inherited anti‐​intellectualism from centuries of disenfranchisement, followed by their abrupt inclusion in American society before they had time to shed the internalization of their oppressor’s debased view of them.”

Only in the past few decades have minorities, and particularly African Americans, been offered a genuine chance to excel. But the opportunity to succeed carries with it the possibility of failure, and minority students are afflicted by a crisis of confidence due to centuries of oppression and negative stereotyping. McWhorter writes, “Black America today is analogous to a wonderful person prevented by insecurity from seeing the good in themselves.”

Acknowledging our history is an important prerequisite to taking the next overdue step in our relationships with each other as Americans: acknowledging that lack of skills, not present‐​day racism, explains the vast majority of current income and education gaps between whites and minorities. To narrow these gaps, we must acknowledge the historical role of racism, and then move on to address the current problem.

Focus on Skills, Not Credentials

Dale and Krueger have shown that affirmative action cannot close the earnings gap between white and minority workers because graduates of selective schools don’t make more money than their counterparts elsewhere. What can narrow that gap, however, are solid academic skills as measured by standardized tests.

Traditional labor market discrimination research starts with a random sample of workers of different races and controls for all of the nondiscriminatory variables that might account for wage differences, such as years of education, years of experience, hours worked, prevailing wage rates in the city where the worker is located, and so on. Any residual wage gaps between racial groups remain a mystery. They may result in part from discriminatory practices in the labor market and in part from unmeasured differences in workers.

Studies show that no other factor explains this residual gap as much as academic skills measured by tests. Sociologists George Farkas and Keven Vicknair reanalyzed existing study data by controlling results for performance on standardized tests and found that the test results explained the entire remaining wage gap between black and white workers.

In a separate study, economists Derek Neal and William Johnson used scores on the Armed Forces Qualification Test taken prior to college or workforce entry to measure skill. They found that performance on the AFQT explained the entire wage gap between black and white women and almost three quarters of the gap between white and black men, even without controlling for educational attainment.

College graduation is important. Most studies do find a “sheepskin effect” associated with holding a college degree, although selectivity does not seem important. But actual skills, not credentials, matter most when it comes to raising the wages of minority workers.

Neal and Johnson emphasize that the academic skills measured by exams like the AFQT can be taught. Outcomes on these tests are powerfully related to parenting styles, education of parents, books in the home, and quality of education. In fact, the skills gap can be measured in academic terms: African‐​American students are, on average, the equivalent of four academic years behind white students. That gap can be narrowed, but only if policymakers recognize that it is the primary culprit producing current inequalities.

Effort and high expectations are critical. Thomas Sowell notes that Asian‐​American students routinely outperform whites with the same standardized test scores, both in college and in their later careers. We can narrow socioeconomic disparities by having high expectations of all students, not just some of them. African‐​American students and those who want to help them succeed should overcome a tendency to focus only on credentials—“getting that piece of paper”— and concentrate on building the skills that lead to labor market success.

Rethink College Tracking

Even if admissions preferences were ended tomorrow, America’s most selective schools would retain between one‐​third and two‐​thirds of their black and Latino students. This does not amount to re‐​segregation, particularly since the remaining minority students would be academically competitive with their peers. But if this amount of racial diversity is not enough, it may make sense to reevaluate the current practice of tracking students very narrowly into different colleges based on academic preparedness.

Affirmative action defenders present a choice between racial preferences on one hand and academic exclusivity on the other. But this is a false dichotomy, and opponents of preferences should not fall victim to this straw man argument. In fact, colleges can enjoy racial diversity without double standards if they are willing to maintain less exclusive admissions policies for students of all races.

Simply educating students of differing abilities at the same college is not the cause of most of the problems generated by racial preferences and catalogued in this study. Rather, problems such as stigma and stereotype threat result from the creation of isolated communities of minority students in selective schools that are substantially and visibly less prepared than their classmates. State flagship universities and other elite schools can have very diverse student bodies while avoiding affirmative action’s negative consequences if they are willing to admit more students of differing abilities and talents from all racial backgrounds.

Some opponents of affirmative action are appalled by the idea of less academic stratification between colleges. Law professor Jeffrey Rosen, for example, supports affirmative action only because he believes that if it were ended, colleges would “lower academic standards across the board” in order to maintain racial diversity. Stephen and Abigail Thernstrom worry that elite schools will give up on “maintaining intellectual standards” to avoid reducing minority enrollments. Members of the American academic elite are attached to their schools and to what they represent, but present little evidence that less exclusive admissions overall would be more harmful to student learning than racial preferences are, or than less diverse student bodies would be.

The current pigeon‐​holing of students on the basis of academic merit came about only in the years following World War II, when the GI bill and rising incomes increased college access for the middle class, and standardized testing was popularized. As recently as the 1950s, admission to college was not academically competitive. Harvard accepted three out of four applicants during that period, and its students had credentials virtually indistinguishable from the top half of many state universities.

Academic elitism is not a 300‐​year tradition in American colleges. It is a 50‐​year experiment. Originally conceived to break down an old‐​boy network based on inherited wealth and social connections, it is worth rethinking whether—in an era that regards college as a coming‐​of‐​age social experience— such rigid sorting of students along academic lines remains a good idea.

Careful study may show that tracking in college is academically beneficial enough to preserve at the cost of relatively homogeneous student bodies at highly selective schools. But that is not self‐​evident. Administrators and policymakers balancing the harms of racial preferences against the benefits of diversity should reassess selectivity’s costs and benefits.

Fix the Pipeline

Nothing is more important to the project of racial equality in America than increasing the numbers of black and Latino students who leave high school prepared for success. The NAACP agrees, for example, that racial disparities in Virginia’s state universities “stem directly from continuing inequalities in Virginia’s public schools.”

School choice can help by rescuing minority children from failing public schools that do not prepare them for college. Studies show that, while all students benefit from school choice, African‐​American students benefit the most, for reasons that are not well understood. One thing is clear: we can narrow the critical skills gap by empowering parents to choose their schools.

Universities that want to assist in this effort can sponsor programs that help minority high school students prepare for college. Economist Bruce Wydick found that intensive college preparation programs are the only way to increase minority representation in selective schools without harming minority graduates in the entry‐​level labor market.

Since the passage of Proposition 209 in California, which banned preferences at state schools, the University of California system has instituted programs that provide tutoring and counseling to local students who might not otherwise get the assistance they need. John Briggs, head of UC Riverside’s writing program, says about the university’s effort, “What affirmative action is supposed to be about is making a concentrated effort to increase the pool of available students, and that means better preparation and better counseling.”


Affirmative action cannot solve the American dilemma of racial inequality. Preferences are designed to harness what their boosters thought would be the formidable power of prestige in getting ahead. But those who hope to ride credentials into the sunset of racial equality have saddled the wrong horse. Not only do preferences fail to narrow racial disparities in income and educational attainment, they harm students of all backgrounds. Only no‐​fuss integration and a focus on building real skills will lead to success.