Justice Powell
Justice Powell; drawing by David Levine

The decision of the Supreme Court in Bakke was received by the press and much of the public with great relief, as an act of judicial statesmanship that gave to each party in the national debate what it seemed to want most. Such a sense of relief, however, hardly seems warranted, and it is important to explain why it does not.

Everyone knows something of the facts of the case. The University of California medical school at Davis administered a two-track admission procedure, in which sixteen of a hundred available places were in effect set aside for members of “minority” groups. Allan Bakke, a white applicant who had been rejected, sued. The California Supreme Court ordered the medical school to admit him, and forbade California universities to take race into account in their admissions decisions.

The United States Supreme Court’s decision affirmed the California court’s order that Bakke himself be admitted, but reversed that court’s prohibition against taking race into account in any way. So opponents of affirmative action plans could point to Bakke’s individual victory as vindication of their view that such plans often go too far; while proponents were relieved to find that the main goals of affirmative action could still be pursued, through plans more complex and subtle than the plan that Davis used and the Supreme Court rejected.

But it is far too early to conclude that the long-awaited Bakke decision will set even the main lines of a national compromise about affirmative action in higher education. The arithmetic of the opinions of various justices, and the narrow ground of the pivotal opinion of Mr. Justice Powell, mean that Bakke decided rather less than had been hoped, and left more, by way of general principle as well as detailed application, to later Supreme Court cases that are now inevitable.

Bakke’s lawyers raised two arguments against the Davis quota plan. They argued, first, that the plan was illegal under the words of the Civil Rights Act of 1964, which provides that no one shall “on the ground of race…be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any program” receiving federal aid. (Davis, like all medical schools, receives such aid.) They argued, second, that the plan was unconstitutional because it denied Bakke the equal protection guaranteed by the Fourteenth Amendment.

Five out of the nine justices—Justices Brennan, White, Marshall, Blackmun, and Powell—held that Bakke had no independent case on the first ground—the 1964 Civil Rights Act—and that the case therefore had to be decided on the second—the Constitution. They said that the language of the Civil Rights Act, properly interpreted, was meant to make illegal only practices that would be forbidden to the states by the Equal Protection Clause itself. They decided, that is, that it is impossible to decide a case like Bakke on statutory grounds without reaching the constitutional issue, because the statute does not condemn the Davis program unless the constitution does. The remaining four justices (Chief Justice Burger, and Justices Stewart, Rehnquist, and Stevens) thought that Bakke was right on the first ground, of the Civil Rights Act, and that they therefore did not have to consider the second, the Constitution itself, and they did not do so.

Of the five justices who considered the second, constitutional argument, four—Brennan, White, Marshall, and Blackmun—held that Bakke had no case under the Constitution either. Mr. Justice Powell held otherwise. He held that the Equal Protection Clause forbids explicit quotas or reserved places unless the school in question can show that these means are necessary to achieve goals of compelling importance, and he held that Davis had not met that burden of proof. But he also held that universities may take race into account explicitly as one among several factors affecting admission decisions in particular cases, in order to achieve racial diversity in classes. (He cited the Harvard undergraduate admissions program as an example.) He said that the Constitution permits this use of race and, since the California Supreme Court had held otherwise, he voted to reverse that court on this point. So a majority of those considering the matter voted against Bakke on both of his arguments; but Bakke nevertheless won, because five justices thought he should win on some ground even though they disagreed on which ground.

What does all this mean for the future? The Supreme Court has now decided, by a vote of five to four, that the Civil Rights Act does not in and of itself bar affirmative action programs, even those, like Davis’s, that use explicit quotas. It has decided, by a vote of five to none, that the Constitution permits affirmative action plans, like the Harvard undergraduate plan, that allow race to be taken into account, on an individual-by-individual basis, in order to achieve a reasonably diverse student body.


Both of these decisions are important. The Civil Rights Act issue was, in my opinion, not a difficult issue, but it is useful that it is now removed from the argument. The argument of the California Supreme Court—that racially conscious admissions programs are always unconstitutional—would have been disastrous for affirmative action had it prevailed in the United States Supreme Court, and it is therefore of great importance that it was rejected there. It is also important that at least five justices are agreed that at least a program like the Harvard undergraduate plan is constitutional. The Harvard model provides a standard; if the admissions officers of other universities are satisfied that their plan is like the Harvard plan in all pertinent respects, they can proceed in confidence.

It is equally important to emphasize, however, that the Supreme Court has not decided that only a program such as Harvard’s is constitutional. It has not even decided that a program with a rigid quota such as the one Davis used is unconstitutional. Mr. Justice Powell drew the line that way in his opinion: he said that a quota-type program is unconstitutional, and his arguments suggest that only something very like the Harvard program is constitutional. But his opinion is only one opinion; no other justice agreed, and four other justices expressly disagreed with him on both points. So Powell’s line will become the Supreme Court’s line only if not a single one of the four justices who remained silent on the constitutional issue takes a position less restrictive than Powell’s on that issue. In these circumstances it seems premature to treat Powell’s opinion, and the distinction he drew, as the foundation of the constitutional settlement that will eventually emerge.

There seems little doubt that the four justices who remained silent on the constitutional issue will have to break that silence reasonably soon. For there are a variety of affirmative action cases likely to confront the Court soon in which no statute can provide a reason for avoiding the constitutional issue. The Court has now remanded a case, for example, which challenges the provision of the Public Works Employment Act of 1977 that at least 10 percent of funds disbursed under the act be applied to “minority” businesses. Since Congress enacted this statute, there can be no argument that its provisions violate congressional will, and the four justices will have to face the question of whether such quota provisions are unconstitutional when this case (or, if it is moot, when some similar case) finally arrives before them. Of course these will not be education cases, and Powell’s opinion is carefully tailored to education cases. But the arguments of principle on which he relied, in taking a more restrictive view of what the Equal Protection Clause permits than did the other justices who spoke to that issue, are equally applicable to employment and other cases.

Indeed, it is arguable that, in strict theory, the four justices who remained silent would have to speak to the constitutional issue even if another education case, like the Bakke case, were for some reason to come before the court. Suppose (though this is incredible) that some university that administers a quota-type system like the Davis system were to refuse to dismantle it in favor of a more flexible system, and the Supreme Court were to review the inevitable challenge. Since Bakke decides that the Civil Rights Act is no more restrictive than the Constitution, the four justices might well consider this point foreclosed by that decision in any future case, in which event they would have to face the constitutional issue they avoided there.

(Anthony Lewis, in The New York Times, said that it was surprising that these justices did not give their opinion on the constitution even in Bakke, since they knew that the Court as a whole rejected their argument that the case could be decided under the statute. Lewis speculates that one of the five justices who rejected the statutory argument might have held the contrary view until just before the decision was released, and so left the four little time to address the constitutional issue.

He may be right, but there is at least another possibility. Suppose at least one of the four believed that even the Davis quota-type plan was constitutional. If he had said so, but nevertheless voted in favor of Bakke on the statutory issue, then the Court would have ordered Bakke admitted even though a majority of the full court, and not simply a majority of those speaking to each issue, was against Bakke on both grounds, and even though the Court would have been constrained, by precedent, to approve a quota-type program in the future. That would have been even more bizarre and confusing than the present decision. But all this is simply guess-work squared.)


Is there any point (other than as an academic exercise) in these speculations about the position that members of the silent four would take on the constitutional issue? Some practical lawyers have already said that the main goals of affirmative action programs, at least in university and professional school education, can be served by programs that fall comfortably within what Mr.Justice Powell expressly permitted. If that is so, then it might be wise to proceed as if the Powell opinion, even though the opinion of only one justice, states constitutional law for university educational programs, and then try to work out a similar settlement for other areas, like employment, in other cases. But I am not so sure that it is so, because the Powell opinion, at least until clarified by later decisions, is less coherent and may well be less permissive than it has widely been taken to be.

Powell expressly ruled out admissions programs, like the Davis program, that reserve certain places for minority members only. He approved programs, like the Harvard undergraduate program he cited, that do not even set target numbers for minority acceptance. Such programs are aimed at diversity in the student body. They recognize that racial diversity is as important as geographical diversity or diversity in extracurricular talents and career ambitions, and so take race into account in such a way that the fact an applicant is black may tip the balance in his favor just as the fact that another applicant is an accomplished flutist may tip the balance in his.

But a great many affirmative action admissions programs fall between these two extremes. They do not expressly reserve a set number of places for which only minority applicants compete, but they nevertheless do set rough “target” figures representing a general decision about the proportion of the class that should on general principles be filled by minority applicants. The number of such applicants accepted will vary from year to year but will hover within a range that will be less varying than the proportion of accomplished musicians, for example, or of applicants from a particular section of the country. In most cases, the admissions committee will report the number of minority applicants selected to the faculty at large, as a separate statistic, and will attempt to explain a particularly low percentage in any particular year. Minority applications will in this way be treated very differently from applications of musicians or West Coast residents. Do such rough “targets,” used in this way, make a program unconstitutional under the analysis Powell proposed?

The answer may depend on the goal or purpose of the “target.” Powell considered a number of goals that affirmative action programs in a medical school might be expected to achieve, and he said that some goals were constitutionally permitted, while others were not. He rejected, in particular, “the purpose of helping certain groups whom the faculty…perceived as victims of ‘societal discrimination.”‘ (He said that this goal must not be pursued by any classification that imposes disadvantages on others who had no responsibility for the earlier discrimination.) He accepted as permissible the goal of supplying more professional people for under-served communities, but denied that Davis had shown that a program “must prefer members of particular ethnic groups” in order to achieve that goal. He also accepted the goal of educational diversity, which in his opinion justified the flexible Harvard plan though not the Davis quota plan.

The constitutionality of an affirmative action plan therefore depends, according to Powell, on its purpose as well as its structure. It is not altogether plain how courts are to decide what the purpose of a racially conscious admissions program is. Perhaps they should not look behind an official institutional statement that the plan seeks educational diversity, if such a statement seems plausible.

But in the case of some professional schools it may not be plausible, and Powell says, in this connection, that “good faith” should be presumed only “absent a showing to the contrary….” Perhaps the motives of individual members of the admissions committee or of the faculty as a whole are not relevant. It is nevertheless true that many faculty members, particularly of professional schools, support racially conscious admissions programs because they do believe that such programs are necessary to provide more professional people for the ghettos. Even more support them because they are anxious that their school help groups that have been disadvantaged by discrimination, by providing models of successful professional men and women from these groups, for example.

The leaders of many institutions are now on record, in fact, that these are their goals. (They may or may not also believe that the level of diversity in their classes that would be reached without racially conscious programs is unsatisfactory for purely educational reasons.) May disappointed applicants to such institutions now sue, placing in evidence statements faculty members have made about the purposes of racially conscious plans, or subpoenaing officers of admission to examine their motives under oath?

Powell’s opinion raises these questions, but it does little to help answer them, even in principle, because the argumentative base of his opinion is weak. It does not supply a sound intellectual foundation for the compromise the public found so attractive. The compromise is appealing politically, but it does not follow that it reflects any important difference in principle, which is what a constitutional, as distinct from a political, settlement requires.

There are indeed important differences between the “quota” kind of affirmative action program—with places reserved for “minorities” only—and more flexible plans that make race a factor, but only one factor, in the competition for all places. But these differences are administrative and symbolic. A flexible program is likely to be more efficient, in the long run, because it will allow the institution to take less than the rough target number of minority applicants when the total group of such applicants is weaker, and more when it is stronger. It is certainly better symbolically, for a number of reasons. Reserving a special program for minority applicants—providing a separate door through which they and only they may enter—preserves the structure, though of course not the purpose, of classical forms of caste and apartheid systems, and seems to denigrate minority applicants while helping them. Flexible programs emphasize, on the other hand, that successful minority candidates have been judged overall more valuable, as students, than white applicants with whom they directly competed.

But the administrative and symbolic superiority of the flexible programs, however plain, cannot justify a constitutional distinction of the sort Powell makes. There should be no constitutional distinction unless a quota program violates or threatens the constitutional rights of white applicants as individuals in some way that the more flexible programs do not.

Powell does not show any such difference, and it is hard to imagine how he could. If race counts in a flexible program then there will be some individual white applicant who loses a place but who would have gained one if race did not count. However that injury is described it is exactly the same injury—neither more nor less—that Bakke suffered. We cannot say that in a flexible system fewer whites lose places because race figures in the decision; that will depend on details of the flexible and quota programs being compared, on the nature of the applicants, and on other circumstances. But even if it could be shown that fewer whites would lose in a flexible plan it would not follow that the rights of those individuals who did lose were different or differently treated.

Powell argues that in a flexible plan a marginal white applicant is at least in a position to try to show that, in spite of his race, he ought to be taken in preference to a black applicant because he has some special contribution that the black applicant does not. His race does not rule him out of even part of the competition automatically.

This argument may be based on an unrealistic picture of how admissions committees must deal with a vast volume of applications even under a flexible plan. An individual admissions officer will use informal cut-off lines, no matter how flexible the program is in principle, and a majority applicant with a low test score may be cut off from the entire competition with no further look to discover whether he is a good musician, for example, though he would have been rescued for a further look if he were black.

But even if Powell’s sense of how a flexible plan works is realistic, his argument is still weak. An individual applicant has, at the start of the competition for places, a particular grade record, test score average, personality, talents, geographical background, and race. What matters, for a white applicant, is the chance these give him in the competition, and it does not make any difference to him in principle whether his race is a constant small handicap in the competition for all the places, or no handicap at all in the competition for a slightly smaller number of places. His fate depends on how much either the handicap or the exclusion reduces his overall chances of success; and there is no reason to assume, a priori, that the one will have a greater or lesser impact than the other. That will depend on the details of the plan—the degree of handicap or the proportion of exclusion—not which type of plan it is.

Powell sees an important difference between a handicap and a partial exclusion. He says that in the former case, but not the latter, an applicant is treated “as an individual” and his qualifications are “weighed fairly and competitively.” (He chides Justices Brennan, white, Marshall, and Blackmun for not speaking to the importance of this “right to individualized consideration.”) But this seems wrong. Whether an applicant competes for all or only part of the places, the privilege of calling attention to other qualifications does not in any degree lessen the burden of his handicap, or the unfairness of that handicap, if it is unfair at all. If the handicap does not violate his rights in a flexible plan, a partial exclusion does not violate his rights under a quota. The handicap and the partial exclusion are only different means of enforcing the same fundamental classifications. In principle, they affect a white applicant in exactly the same way—by reducing his overall chances—and neither is, in any important sense, more “individualized” than the other. The point is not (as Powell once suggests it is) that faculty administering a flexible system may covertly transform it into a quota plan. The point is rather that there is no difference, from the standpoint of individual rights, between the two systems at all.

There is a second serious problem in Powell’s opinion which is more technical, but in the end more important. Both Powell and the other four justices who reached the constitutional issue discussed the question of whether racial classifications used in affirmative action programs for the benefit of minorities are “suspect” classifications which the Supreme Court should subject to “strict scrutiny.” These are terms of art, and I must briefly state the doctrinal background.

Legislatures and other institutions that make political decisions must use general classifications in the rules they adopt. Whatever general classifications they use, certain individuals will suffer a disadvantage they would not have suffered if lines had been differently drawn, sometimes because the classifications treat them as having or lacking qualities they do not. State motor codes provide, for example, that no one under the age of sixteen is eligible to drive an automobile, even though some people under that age are just as competent as most over it. Ordinarily the Supreme Court will not hold such a general classification unconstitutional even though it believes that a different classification, which would place different people at a disadvantage, would be more reasonable or more efficient. It is enough if the classification the legislature makes is not irrational; that is, if it could conceivably serve a useful and proper social goal. That is a very easy test to meet, but if the Court used a more stringent test to judge all legislation, then it would be substituting its judgment on inherently controversial matters for the judgment reached by a democratic political process.

There is, however, an important exception to this rule. Certain classifications are said to be “suspect” and when a state legislature employs these classifications in legislation, the Supreme Court will hold the legislation unconstitutional unless it meets a much more demanding test which has come to be called the test of “strict scrutiny.” It must be shown, not simply that the use of this classification is not irrational, but that it is “necessary” to achieve what the Court has called a “compelling” governmental interest. Obviously, it is a crucial issue, in constitutional litigation, whether a particular classification is an ordinary classification, and so attracts only the relaxed ordinary scrutiny, or is a suspect classification which must endure strict scrutiny (or, as some justices have sometimes suggested, falls somewhere between these two standards of review).

Racial classifications that disadvantage a “minority” race are paradigm cases of suspect classifications. In the famous Korematsu case the Supreme Court said that “[All] legal restrictions which curtail the rights of a single racial group are immediately suspect. That is not to say that all such restrictions are unconstitutional. It is to say that courts must subject them to the most rigid scrutiny.” But what about racial classifications that figure in a program designed to benefit a group of disadvantaged minorities? It had never been decided, prior to Bakke, whether such “benign” classifications are suspect.

The four justices who voted to uphold the Davis plan did not argue that “begin” racial classifications should be held only to the weak ordinary standard—i.e., that it could conceivably serve a useful social goal. But neither did they think it appropriate to use the same high standard of strict scrutiny used to judge racial classifications that work against minorities. They suggested an intermediate standard, which is that remedial racial classifications “must serve important governmental objectives and must be substantially related to achievement of those objectives.” They held that the Davis medical school’s purpose of “remedying the effects of past societal discrimination” was sufficiently important, and that the racial classification Davis used was “substantially related” to that objective.

But Mr. Justice Powell disagreed. He held that “benign” racial classifications should be held to the same extremely strict scrutiny that is applied to racial classifications that disadvantage a minority. He therefore required that the Davis classification be “necessary” to a “compelling” purpose, and he held that it was not. He argued that no distinction should be made between the test applied to racial classifications that benefit and those that disadvantage an established minority for two reasons. First, because any such distinction would rest on judgments, like judgments about what groups are, in the relevant sense, minorities, and which classifications carry a “stigma,” that Powell called “subjective” and “standardless.” Second, because constitutionally important categories would then be constantly changing as social or economic conditions (or the perception of Supreme Court justices of such conditions) changed, so that yesterday’s disadvantaged minority became a member of today’s powerful majority, or yesterday’s helping hand became today’s stigma.

There is plainly some force in this argument. All else being equal, it is better when constitutional principles are such that reasonable lawyers will not disagree about their application. But often the political and moral rights of individuals do depend on considerations that different people will assess differently, and in that case the law would purchase certainty only at the price of crudeness and inevitable injustice. American law—particularly constitutional law—has refused to pay that price, and it has become in consequence the envy of more formalistic legal systems.

It is easy, moreover, to exaggerate the “subjectivity” of the distinctions in play here. Once the distinction is made between racial classifications that disadvantage an “insular”minority, like the detention of Japanese-Americans in the Korematsu case, and those that are designed to benefit such a minority, then reasonable men cannot sincerely differ about where the racial classification of the Davis medical school falls. Nor is the social pattern of prejudice and discrimination the Davis program attacked either recent or transient. It is as old as the country, tragically, and will not disappear very soon.

My present point, however, is a different one. Powell’s argument in favor of strict scrutiny of all racial classifications, which is that the putative distinction between benign and malignant classifications relies on “subjective” and “standardless” judgments, is not and cannot be consistent with the rest of his judgment, because his approval of flexible admissions programs, like the Harvard undergraduate program, presupposes exactly the same judgments. Powell begins his defense of flexible but racially conscious admissions programs with the following exceptionally broad statement of a constitutionally protected right of universities to choose their own educational strategies:

Academic freedom, though not a specifically enumerated constitutional right, long has been viewed as a special concern of the First Amendment. The freedom of a university to make its own judgments as to education includes the selection of its student body. Mr. Justice Frankfurter summarized the “four essential freedoms” that comprise academic freedom: “It is an atmosphere in which there prevail the four essential freedoms of a university—to determine for itself on academic grounds who may teach, what may be taught, how it shall be taught, and who may be admitted to study.”

Diversity is the “compelling” goal that Powell believes universities may seek through flexible (but not crude) racially conscious policies. But what if a law school faculty, in the exercise of its right to “determine for itself…who shall be admitted to study,” decided to count the fact that an applicant is Jewish as a negative consideration, though not an absolute exclusion, in the competition for all its places? It might decide that it is injurious to “diversity” or to the “robust exchange of ideas” that Jews should form so large and disproportionate a part of law school classes as they now do. Or what if a Southern medical school one day found that a disproportionately large number of black applicants was being admitted on racially neutral tests, which threatened the diversity of its student body, to the detriment, as it determined, of its educational process? It might then count being white as a factor beneficial to admission, like being a musician or having an intention to practice medicine in a rural area.

The four justices who voted to uphold the Davis program as constitutional would have no trouble distinguishing these flexible programs that count being Jewish as a handicap or being white as a beneficial factor. Neither of these programs could be defended as helping to remedy “the effects of past societal discrimination.” They could argue that, on the contrary, since these programs put at a disadvantage members of races that have been and remain the victims of systematic prejudice, the programs must for that reason be subject to “strict scrutiny” and disallowed unless positively shown to be both “necessary” and compelling.

Mr. Justice Powell did not, of course, have any such programs in mind when he wrote his opinion. He surely could not accept them as constitutional. But he, unlike the four justices, could not consistently distinguish such programs on their grounds, since the judgments I just described involve precisely the “subjective” and “standardless” judgments about “stigma” that he rejected as inappropriate to constitutional principles.

The point is, I think, a simple one. The difference between a general racial classification that causes further disadvantage to those who have suffered from prejudice, and a classification framed to help them, is morally significant, and cannot be consistently denied by a constitutional law that does not exclude the use of race altogether. If the nominal standard for testing racial classifications denies the difference, the difference will nevertheless reappear when the standard is applied because (as these unlikely hypothetical examples show) our sense of justice will insist on a distinction. If that is so, then the standard, however it is drafted, is not the same, and will not long be thought to be.

I raise these objections to Powell’s opinion, not simply because I disagree with his arguments, but to indicate why I believe that the compromise he fashioned, though immediately popular, may not be sufficiently strong in principle to furnish the basis for a coherent and lasting constitutional law of affirmative action. Later cases will, of course, try to absorb his opinion into a more general settlement, because it was the closest thing to an opinion of the Court in the famous Bakke case, and because it is the creditable practice of the Court to try to accommodate rather than to disown the early history of its own doctrine. But Powell’s opinion suffers from fundamental weaknesses and , if the Court is to arrive at a coherent position, far more judicial work remains to be done than a relieved public yet realizes.

This Issue

August 17, 1978