Civil Rights and Racial Preferences:
A Legal History of Affirmative Action
Last fall, the voters of California approved the California Civil Rights Initiative, which amends the state constitution to say that public officials shall not "discriminate against, or grant preferential treatment to, any individual or group on the basis of race, sex, color, ethnicity, or national origin." Before the election, both sides in the debate over CCRI made dramatic claims about what would happen if the initiative passed. It will be a while, however, before those predictions can be tested. At the end of the year a federal court issued an injunction staying implementation of CCRI while its opponents challenge its legality. The court held that the measure may well deprive minorities and women of their constitutional rights to equal protection.
Anyone who has ever read the text of the Civil Rights Act of 1964 might be baffled at this prospect, since it says, in rather plain language, much the same thing CCRI says. For example, Titles VI and VII of the Act tell school officials and employers that they "shall not discriminate" against any individual on account of race or gender; and Title VII further tells employers that in order to meet their legal duty under the Act, they are not required to give "preferential treatment" to anyone in order to achieve a racial or gender balance. Indeed, the original supporters of the Act insisted that Title VII actually prohibited preferential treatment. But, who knows? Perhaps the Civil Rights Act is unconstitutional, too!
How, in thirty years, have we come from the Civil Rights Act to CCRI? In this essay I want to set the current affirmative action debates -- typified by the controversy surrounding CCRI -- in some context. I want to distinguish very different programs with very different rationales, all subsumed under the phrase "affirmative action." I want to explain how programs of preference emerged out of antidiscrimination law. And I want to suggest the relevant framing of arguments about racial preferences in particular.
I. CONTEXT: EMPLOYMENT
In the Beginning
The Civil Rights Act of 1964 set American law and policy four-square against racial discrimination in workplaces, schoolrooms, and public accommodations. Along with companion legislation against discrimination in housing and voting, the Act swept aside widespread customary practices as old as the Republic itself. One year later, President Lyndon Johnson's Executive Order 11246 threw the weight of the massive federal contracting process behind the Title VII prohibition of employment discrimination. The executive order required that every federal contractor create and abide by an affirmative action plan as a condition of receiving federal money.
What constituted the racial discrimination that the Civil Rights Act meant to prohibit? To those who supported the law, the matter was simple. Everybody knew what discrimination was, they insisted. The sign in the factory window saying "No Colored Apply" was discrimination. The high school's refusal to put blacks on its sports teams was discrimination. The drug store's denying blacks a seat at the lunch counter was discrimination. What could be plainer?
Soon, however, as the federal courts began to enforce the Civil Rights Act, matters became far less plain. Suppose, for example, that to comply with the law the factory took the sign out of its window but had in place a rule requiring each applicant to supply a recommendation from a current employee. Suppose that another factory abolished its explicit segregation of blacks into maintenance jobs and whites into technical jobs but had in place a rule requiring workers transferring from one job category to another to give up their accumulated seniority. These facially neutral employment practices effectively reproduced the racially exclusionary practices abolished by the law. If all a factory's employees were white, the chance that a black applicant might be able to present the required recommendation was vanishingly small. If a factory's black maintenance workers had to yield years of seniority to transfer to a technical job, few if any could afford to make the sacrifice. In each instance, the factory's rules carried forward the effects of its own past discrimination.
Confronted with case after case like these, the federal courts had to decide how to conceive of practices that locked in place patterns of past exclusion, even though the practices were neutral on their face and not designed with the intent to discriminate. From early on, the courts began to construe such practices as being precluded by the Civil Rights Act in just the same way the sign in the window was precluded. In 1971, the Supreme Court, in Griggs v. Duke Power Company, ratified this line of interpretation, holding that a company rule disproportionately excluding blacks counted as discrimination under Title VII whether or not the company intended such a result, unless the company could show the practice was necessary to its doing business.
This ruling dramatically broadened the sweep of antidiscrimination law. Now all sorts of work rules and employment practices, from aptitude tests to physical requirements to educational qualifications, were up for grabs. Did black firefighters fail the fire department promotion test in greater proportions than white firefighters? Then the city had better show a demonstrable connection between scoring well on the test and good performance as a promoted firefighter. Did the high school diploma that the factory demanded as a condition of employment exclude more blacks than whites? Then the factory had better show why it needed high school graduates as janitors or laborers. The city and the factory had better show their practices to be necessary.
But what constituted "necessity" in such matters? What practices could a court reasonably require a firm or institution to abandon, should those practices exclude blacks disproportionately? In some instances, the answer to this question was easy, in others contentious. As with any reasonableness test, fully informed and well-meaning people could disagree about its application in a particular circumstance.
This evolution in the legal concept of discrimination grounded a further realization. Discrimination need not be the effect of this or that discrete, superficial rule or policy, which a court or government could simply require an employer to abandon. Discrimination could now be seen as built deeply into institutions. Webs of nominally innocent institutional habits and procedures could work to exclude blacks or burden them with special disadvantages. Likewise, they could work to exclude other minorities and women, who were also covered by the Civil Rights Act and beginning to assert their claims through litigation as well. An employer's exclusion of minorities or women might not be the upshot of any single, easily detectable rule or practice that could be isolated and changed; it might be the effect of many different practices working together. To comply with the law, institutions needed to rethink their operations from the ground up. They needed a mechanism that could shake them out of complacency and habit. That mechanism was affirmative action.
Nonpreferential Affirmative Action
The basic idea of affirmative action was simple: motivate firms to carry on continuous, conscious appraisal of their procedures and rules to detect and eliminate those that excluded minorities and women without appropriate justification. The mechanism to embody this idea was the ubiquitous affirmative action plan, imposed on all federal contractors by Executive Order 11246. Make a plan, the government told firms in 1972, that includes these steps:
Step 1. Take action to make sure your selection pool is expansive.
Step 2. Given the racial composition of the expanded pool, predict the results over time of your selecting nondiscriminatorily from it. Your prediction constitutes your affirmative action "goals." They give you a benchmark against which to compare actual outcomes.
Step 3. At intervals, compare your actual selections with your "goals." If you are not meeting your goals, then reexamine your rules and procedures to see what is causing the problem.
If your selections match your "goals," the government went on to say, we are not going to look your way. But if your selections fall short of your "goals," we will come to inquire why. If the inquiry shows that your firm has made "good faith" efforts to carry out your affirmative action plan, you will suffer no penalties. After all, your requirement under the law is not to select the predicted number of blacks, but to select without discrimination. Since the predictions are often based on crude assumptions, there can be many legitimate reasons why the actual outcome wasn't the predicted outcome. But still, we will ask some hard questions: you had better have some very good answers.
This was nonpreferential affirmative action. Nonpreferential affirmative action was a color-conscious, self-monitoring device to aid firms and institutions in achieving nondiscrimination. And it was also a device enforcers of the law could use to measure compliance.
The controversy surrounding this version of affirmative action centered on the outcomes that were to be expected in the absence of discrimination. "This predictive aspect of Affirmative Action could be called any number of things," Stanley Pottinger, director of the Office of Civil Rights in the Department of Health, Education, and Welfare, observed in 1975. "They happen to be called goals." Would that they had been called other things! "Goals" is a misleading description, given the model of affirmative action set out above. In plain English, goals are things you aim at. But in a nonpreferential affirmative action plan, "goals" aren't what you aim at. What you aim at is nondiscriminatory selection. As a byproduct of achieving your aim (nondiscriminatory selection), you expect a certain number of blacks or women to be selected. If that expectation is disappointed, you are prompted to wonder why and seek an answer.
Had this conception of affirmative action prevailed as the dominant public understanding, much less controversy would have attached to it. But the model had to compete with quite different interpretations. Since institutions were required to adopt something called "goals," most people, whether sympathetic or unsympathetic, naturally took the word at face value and construed goals as aims incumbent on institutions. Government did not help matters here by prescribing rules that seemed to make elimination of gender and minority "underutilization" the first duty of institutions while at the same time disavowing any intent to require institutions to give preferential treatment.
Ordinary citizens could hardly be faulted for failing to see that the government meant, when it said "goals," not-goals. Nor could critics of affirmative action be greatly faulted for disbelieving the disclaimer of preferential intent. The disclaimer seemed to rest on a morally specious distinction that the Labor Department and other agencies drew between "goals" and "quotas." The government maintained that "goals" (good) are flexible aims whereas "quotas" (bad) are rigid ones. But for those critics concerned that affirmative action amounted to a policy of coerced proportional representation by race and gender, it did not matter morally whether mandated racial and gender preferences were rigid or flexible. Preferences were still preferences. After all, the problem with that sign in the factory window, "No Colored Apply," was not its rigidity. A more flexible sign, "No Colored Apply, More or Less," or "Almost No Colored Apply," would hardly have constituted a moral or legal improvement.
Even on the Pottinger model, affirmative action raised questions that needed forthright and persuasive answers. If government did not look closely at a contractor who was making the numbers but posed some searching and tough questions to the contractor who wasn't, wouldn't this create an incentive for contractors to make sure they were making the numbers, whatever it took? The prudent contractor, concerned about the bottom line, would hardly scruple at employing covert preferences, would he, if that was the price of legal peace? Unfortunately, the answers to these questions were frequently neither forthright nor satisfying to the critics.
Preferential Affirmative Action
The waters of controversy were further muddied by the parallel existence of genuinely preferential affirmative action. By the early 1970s, courts had begun ordering some employers to select by the numbers. Moreover, as the decade went along, the government used the threat of litigation to impose genuine hiring goals on several large, high-profile employers. The AT&T Corporation was a case in point. It agreed to a "Model Affirmative Action Plan" requiring it to "recruit without discrimination," but at the same time to achieve, "within a reasonable time, an employee profile, with respect to race and gender in each major job classification, at a pace beyond that which would occur normally." The Model Plan made explicit provision for the use of racial and gender preferences. Whenever the company failed to meet its quarterly "targets" through its ordinary procedures, the Plan required that "selections be made from among any at least basically qualified candidates for promotion and hiring of the group or groups for which the target is not being met," even if there were more senior and more qualified candidates first in line.
In the AT&T case, then, the "goals" were quite undeniably goals. They constituted aims the company had to achieve, even by extraordinary measures if necessary. Nonetheless, the Plan offered the usual disclaimer, insisting that its "goals, intermediate targets and time frames" were "neither rigid nor inflexible quotas, but objectives to be pursued by mobilization of available company resources for a 'good faith effort.'" Once again, the effect of this rhetorical strategy was to confirm the critics' view that the whole apparatus of affirmative action was bent to the same aim. If in the AT&T case the standard disclaimer of nondiscriminatory intent was plainly disingenuous, wasn't that proof positive that the government's own disclaimers were equally mendacious?
As it was, the government had good reason for the plan it imposed on AT&T, whatever the plan's numerical components were called. Consider the especially acute problem that gender discrimination posed at AT&T. Half of the company's 700,000 employees were female, but they were all operators and secretaries. What was notable about AT&T on close inspection was how deeply the concept of "woman's job" and "man's job" was built into everything the company did. The government concluded that any approach short of force-feeding large numbers of women into "men's" occupations would founder on the profound inertial force of AT&T's inherited ways of doing things. So, with its "targets," "goals," and "objectives," the government sought to break open and destroy an entire corporate culture premised on men's and women's separate spheres.
Similar reasoning lay behind the government's strategy in other cases. For example, after nearly a decade of litigation involving the Mississippi and Alabama highway patrols, the federal courts concluded that discrimination was so deeply built into the culture and operations of the two institutions that the only effective remedy was to require them to hire one black trooper for each white trooper hired until the patrols reached a substantial, specified level of racial integration. The effective way to approach the patrols was not to change their cultures in order to get more blacks into them, but to put more blacks into them in order to change their cultures.
Reverse Discrimination
By the middle of the 1970s, employers were caught in a dilemma. If they "underutilized" blacks and women on their workforces or in particular job categories, then even their most "innocent" -- i.e., facially neutral -- practices could ground a legal charge of discrimination that might be difficult to defend against. Litigation was chancy and settling with the government meant meeting its possibly onerous terms. Look at what happened to AT&T -- and to many other giants in the airlines, banking, and steel-making industries, to name a few. An employer, concerned to avoid such a fate, could tinker with this or that work practice and still fail substantially to reduce its "underutilization," and thus fail to reduce the jeopardy of government action.
Yet if the employer embraced programs effectively assuring increased utilization of blacks and women, it faced jeopardy on another front: from reverse discrimination lawsuits. When a Kaiser Chemical Corporation plant in Louisiana decided to remedy the complete absence of blacks in its craft jobs by creating its own on-the-job training program and admitting blacks and whites to it on a one-to-one ratio until 35 percent of craft jobs were filled by blacks, the company was hauled into court by an aggrieved white worker.
In a pivotal decision in 1979, the Supreme Court upheld Kaiser's program; and the Equal Employment Opportunity Commission immediately issued rules effectively immunizing employers from reverse discrimination lawsuits where they were acting on the basis of an approved affirmative action plan. These developments were crucial in stabilizing affirmative action in the 1980s and 1990s among America's private employers. Firms could set about changing work practices to bring them in line with the expanded antidiscrimination mandate, and do so with a conscious eye to the numbers. They could adopt the devices best for them in remaking their total operations so as to accommodate blacks and women without fear of litigation whiplash.
By the mid-1980s, then, the basic patterns of affirmative action in employment settled into stability. All federal contractors that underutilized minorities or women had to put in place affirmative action plans containing goals and timetables. The asymmetry of enforcement pressure -- threat of government scrutiny for employers who weren't making the numbers and relative immunity from reverse discrimination lawsuits for those who were -- undoubtedly sometimes led to the use of racial preferences. More typically, however, the stable legal environment let institutions remake practices and alter routines that had formerly perpetuated discrimination. In special cases, as we have seen, employers used overt racial or gender preferences under court order or consent decree; but the order or decree was always limited in duration. Indeed, AT&T's Model Affirmative Action Plan expired after six years. Even so, the company voluntarily retained most of its features.
II. CONTEXT: UNIVERSITIES AND GOVERNMENT SET-ASIDES
Preferential Admissions
Out of the prohibition against racial and gender discrimination mandated by the Civil Rights Act grew -- paradoxically -- occasional uses of gender and racial preferences. Because discrimination was not a simple thing, courts and federal agencies sometimes concluded that short of imposing such drastic measures as preferences, institutions would continue wrongfully excluding minorities and women. In areas outside employment, however, preferences were being put in place without reference to an institution's own discrimination. Universities, for example, began adjusting their admissions policies to assure a certain percentage of minorities in undergraduate and graduate entering classes.
The medical school at the University of California at Davis instituted one of these policies, reserving 16 of its 100 entering places for minority students. Allan Bakke, a white applicant rejected by the medical school, sued. The Supreme Court's ruling in 1978 struck down the medical school's program, but the lead opinion written by Justice Louis Powell seemed to give universities permission to take race into account in their admissions process, if they were not too blatant about it. On the basis of this permission, minority preferential programs proliferated in academic institutions, ranging from informal preference-granting to formally race-normed admissions lists to specially reserved scholarships. Because this whole edifice rests on the fragile support of Regents v. Bakke -- and because no form of affirmative action has stirred greater controversy, in California and elsewhere -- it is worth looking in some detail at this quite unusual legal decision.
The medical school defended its preferential policy not as a necessary tool to remedy its own discrimination but as a useful mechanism for increasing the number of minorities in the medical profession. Such an increase, argued the school, was beneficial in a number of ways: it would supply medical practitioners to urban and rural minority populations, it would speed the integration of medicine as a whole, and it would stimulate future cohorts of minority youth to pursue medical training.
The Supreme Court split three ways in its decision. Four justices thought that the medical school's admissions policy violated the prohibitions of Title VI of the Civil Rights Act, whose language, after all, seems pretty plain. Five did not. So this five moved to a second question: given that the medical school was a public institution, did its admissions policy violate the constitutional guarantee of equal protection of the laws? Four of the five, led by Justices William Brennan and Thurgood Marshall, argued that discrimination in favor of minorities ought to be judged by a permissive constitutional standard rather than the quite strict standard used to assess discrimination against minorities. Moreover, mindful of a decade of jurisprudence tying legitimate preferences to remedies for discrimination, they argued that the medical school was justifiably acting to remedy not its own but societal discrimination.
Thus, the Court stood four to four, one group of justices voting to invalidate the medical school's preferential program, the other group voting to sustain it. The odd man out was Justice Powell, whose vote tipped the decision against the medical school. He argued that all preferential programs, whether favoring or disfavoring minorities, must meet the same strict constitutional test and that the medical school's program failed that test. However, he also conjectured that less rigid university policies might meet the test. After all, academic freedom was protected by the Constitution itself, and an aspect of academic freedom was the right of academic institutions to manipulate their admissions policies to produce educationally interesting mixes of students. Universities that took race or ethnicity into account as one factor among many might be acting within constitutional limits, argued Powell. And on the basis of this singular opinion, the academy proceeded to institutionalize many sorts of preferential programs.
Since the Bakke decision, two jurisprudential developments have further undermined the Brennan-Marshall rationales for racial preferences in university admissions. First, the effort to expand the legal permission, "preferences to cure your own discrimination," into the broader privilege, "preferences to cure societal discrimination," has been decisively rejected by the Court. Second, the effort to install a two-tier constitutional test for racial preferences, with an undemanding standard of review for minority-favoring preferences and a very demanding standard of review for minority-disfavoring preferences, has also been decisively rejected. The Court has on several occasions in recent years insisted that racial classification in particular must meet the most demanding review, and that only the most compelling reasons can justify a public agency in giving racial preferences of any sort.
Already since 1994 two federal circuits have struck down university preferential policies -- racially reserved scholarships at the University of Maryland and two-track admissions at the University of Texas Law School. Although Maryland tried belatedly to show that its scholarship program was intended to remedy the effects of its own past discrimination, the court couldn't find any current discrimination to remedy. In the Texas case, the court rejected both the law school's claim that it needed to cure its own discrimination and its assertion that it could promote social and educational aims by using preferences. Thus, common campus practices face jeopardy not only from political initiatives like CCRI but from the courts as well.
Set-Asides
These common university practices are called "affirmative action," though they differ in their aim and operation from the employer affirmative action I described above. Indeed, they are probably what most people think of as affirmative action. In addition, many agencies of the federal, state, and municipal governments parcel out contracts by race and gender, reserving a certain percentage of their contracts to firms operated by minority or female entrepreneurs.
Congress's first set-aside legislation -- affecting the federal government's public works contracting -- arguably fell within the ambit of the antidiscrimination approach embodied in employment affirmative action. This is because the "preferences" granted through the original set-aside programs could be seen as attempts to counteract the effects of the government's own discriminatory policies. For example, the federal government maintains a public works market structured by the Davis-Bacon Act, requiring federal contractors to pay "prevailing wages" (i.e., wages pegged to prevailing union wage scales). This structuring adversely affects minority contractors by negating their main competitive advantage, the ability to offer low-wage labor; yet the Davis-Bacon Act is certainly not "necessary" to the government's doing business. By extending the Griggs test for employment discrimination, we could easily think of the government as a discriminator in its public works contracting, and think of its minority set-aside program as a compensation for its own discrimination, given that Congress is unwilling to scrap the Davis-Bacon Act.
Eventually, however, state and municipal governments, as well as many federal agencies, created their own set-aside policies, generally without any reference to their own discrimination. The Federal Communica-tions Commission, for instance, set aside some broadcasting licenses for minority firms in the name of increasing "diversity" in broadcasting. In recent years, many of these set-aside programs have failed legal challenge in the courts. Thus, many different policies may be described as "affirmative action." And many different reasons are at work in the programs that go by that name.
III. GETTING THE DEBATERS TO ENGAGE
Discrimination: What and How Much
There are so many ways that the public debate about affirmative action misfires. To start with, people often treat affirmative action as a unitary policy, rather than as a range of programs with different aims and effects. The temptation is to embrace or condemn affirmative action as a whole, by reference to certain putative rationales or putative abuses, without attending to nuances and distinctions. Even worse, critics and defenders of affirmative action are often on completely different pages, conceptually and factually, about what discrimination is.
As we have seen, the legal concept of discrimination evolved from referring to that paradigmatic sign in the factory window to encompassing all sorts of facially neutral practices that adversely excluded blacks and women and couldn't be defended as "necessary" to doing business. An institution could be found guilty of discrimination because of effects, not intent.
Many critics of affirmative action have never reconciled themselves to this definition of discrimination. In the early years of the Reagan administration, Senator Orrin Hatch pushed very strongly to overturn the effects test and make discriminatory intent an essential element of any legal standard of discrimination. The effects test was "pernicious," he claimed, "one that undermines everything that the civil rights movement has traditionally stood for -- by undermining some of the most basic principles of fairness and due process." The "basic principles" Hatch had in mind require that a person intend a wrong before he be judged guilty of it.
Such a requirement does underlie some parts of the law, especially those that are "judgmental," i.e., imply a moral condemnation of the guilty. Yet the equation that Hatch supposed between legal charges of discrimination and moral condemnation is inappropriate, given the evolution of the law. The authors of the Civil Rights Act of 1964 undoubtedly thought in terms of the everyday notion then current, in which discrimination involved deliberate exclusion based on racial hostility. That everyday notion was a "moralized" concept, meaning that it implicitly condemned the acts it described. However, as it developed through court cases as a legal term of art, "discrimination" became a partly or wholly de-moralized concept in some circumstances. No longer could a finding of discrimination directly imply a morally bad character in the discriminator. Some discrimination still involved direct intention wrongfully to exclude; but most discrimination fell on different parts of the moral spectrum.
Consider what happened when institutions were told to eliminate practices that "unnecessarily" excluded blacks and women. Some of those found guilty under this standard simply failed to be very diligent about reconsidering and revising their practices. They didn't take very seriously the new legal standard. These discriminators certainly merited some moral disapprobation, just as intentional discriminators did. But an institution that took the new legal standard seriously and tried to adhere to its spirit as well as its letter still could find itself judged guilty of discrimination by a court that disagreed with the institution's assessment of necessity. Someone who acted as conscientiously and as responsibly as we could want still risked legal liability under the law.
However, if the legal understanding has evolved in this way, ordinary understanding of discrimination generally has not. This can be seen on both sides in the affirmative action debate. Many who in the 1970s and 1980s supported affirmative action and the most extensive reach of antidiscrimination law quite freely imputed moral fault to all those found guilty of discrimination. They used the term "discrimination" as a rhetorical weapon that would have lost a good deal of its power had they acknowledged the partial de-moralization of the concept. So too, critics like Senator Hatch automatically equated a legal with a moral finding, and then demanded a change in the law because they thought the moral finding too often inappropriate.
But the standard that Hatch wanted changed isn't just the holding of some unelected judges who don't represent the American people; it is now embedded in the Civil Rights Act of 1991, passed by Congress in its final form with bipartisan support. Those who oppose affirmative action need to take heed of this standard as they argue about how much discrimination there is and whether affirmative action in weaker or stronger forms is needed to combat it.
Even taking heed, however, won't generate much agreement about the extent of discrimination. For one thing, the legal concept doesn't refer to a stable set of practices or activities. The Civil Rights Act of 1991 defines "business necessity" as a practice "significantly related" to job performance or to a "significant employer objective." It takes little acumen to see that this vague idea has to be fleshed out case by case in litigation. What a decade ago didn't count as discrimination does today; and what today doesn't count as discrimination quite possibly will a decade from now. If what counts as discrimination keeps changing, we will have a hard time tracking discrimination over time. Moreover, the sum of what counts as discrimination will exhibit a quite lumpy contour, not very well captured by the gross statistical data on employment gaps, wage gaps, etc., utilized by economists. Thus, even if those on both sides of the affirmative action debate take seriously the legal concept of discrimination (in both its extension and its partly de-moralized nature), they are not likely to have eliminated their differences. At least they will be on the same page, however, able to address the threshold issue: If there remains discrimination resistant to non-aggressive measures, what do we do about it?
Winners and Losers
The original affirmative action that grew out of Title VII of the Civil Rights Act and Lyndon Johnson's executive order program was primarily a means of changing institutions -- of overcoming and altering entrenched practices that perpetuate discrimination. Both nonpreferential and preferential affirmative action were premised on the same proposition: institutions must end discrimination. All institutions had to embrace nonpreferential affirmative action and some, for temporary periods, had to utilize preferential affirmative action. "Preferences to cure your own discrimination" -- so spoke the law, as interpreted by the courts.
Yet many advocates of affirmative action took up a different rationale to defend such policies. It was perfectly fair, they argued, after generations of "affirmative action for white males," that blacks and women should get preferences in employment and elsewhere. Affirmative action preferences were a form of compensation, insisted some, making up for past wrongs against blacks and women. Others saw the preferences as a form of distributive justice, offsetting "white skin privilege" or other unjust advantages that white males took into the competition for jobs, university admissions, and other social positions. If affirmative action favored blacks and women, the favoritism was deserved.
Such defenses invited some pointed questions. Christopher Edley, writing in his recent book, Not All Black and White: Affirmative Action and American Values, puts one of the standard challenges this way: "Imagine a college admissions committee trying to decide between the white daughter of an Appalachian coal miner's family and the African-American son of a successful Pittsburgh neurosurgeon. Why should the black applicant get preference over the white applicant?"
Why, indeed? If we see affirmative action as a necessary mechanism to change institutions to make them more racially accommodating, we will give an answer that refers to the skin color of the two applicants, and not to their personal deservingness. Neither the privilege of the neurosurgeon's son nor the hardship of the coal miner's daughter has any bearing. After all, putting a black into a predominantly white institution changes it in the desired direction, even if only minimally; putting in a white does not.
If, however, we see affirmative action as a scheme for compensating people or fairly redistributing advantages and disadvantages, we have to give an answer to Edley's question that refers to the personal deservingness of the winners and losers. It has to be fair that the white applicant loses out, fair that the black applicant wins.
So, the privileged son of the black neurosurgeon really did deserve the preference, argued many defenders of affirmative action in the 1970s. They insisted that every black, no matter how apparently privileged, had been harmed by racial discrimination. But how much harm, offset by what other advantages, in comparison to the overall circumstances of the white loser? These are awkward questions that were never answered satisfactorily. Nevertheless, defenders of affirmative action found it difficult to give up the proposition that the black applicant's preference was deserved.
More importantly, they couldn't tolerate the idea that the white applicant's loss was undeserved. To admit as much called into question affirmative action's distributive fairness. So when the white male worker or applicant complained, "Why must I bear this loss? I am innocent of the wrongs you are trying to compensate; I am not responsible for the disadvantages you are trying to even out," the retort was, "You are not innocent." Some still insist on this point. Could we object that racial preferences burden innocent white males? No, responds one recent defender of affirmative action. White males are not innocent.
This line of argument, more prevalent in the 1970s and 1980s than now, embraces a toxic confusion. Intellectuals as different as Andrew Hacker and Alasdair MacIntyre have joined in ridiculing the protest by whites "that they bear neither responsibility nor blame for the conditions blacks face" (to use Hacker's words). While Hacker sees the protest as merely a manifestation of white racism, MacIntyre construes it as a sign of civic irresponsibility. But why does MacIntyre assume that the protest was a denial of civicrather than personal responsibility? After all, it was personal responsibility, or at least personal liability, that was being imputed to white males by many defenders of racial preferences. And why does Hacker assume that only a white racist would repudiate such responsibility or liability? Why shouldn't white males -- and everyone else -- reject it? If you think Allan Bakke is not innocent, but all you know about Allan Bakke is that he is white, and that's enough for you to know, then you must believe that guilt travels through skin color -- an obnoxious notion that ought to be rejected. Even if "angry white males" are nothing more than morons and racists, some common defenses of affirmative action gave them genuine reasons to reject it.
The confused focus on individual desert and personal liability distorted defenses of affirmative action and gave its critics reason to complain. The proper focus must lie elsewhere. Our nation's system of racial segregation and exclusion that affirmative action meant to overcome was not the result of myriad individual discriminations and personal choices coalescing into one uniform effect. It was official policy, national, state, and local -- policy of our people as a corporate body, executed through corporate agencies. The obligation to remedy that system and its effects is correspondingly a corporate obligation. Individuals have direct obligations to support and bear the costs of that corporate effort, but obligations in their capacity as citizens, not in their capacity as white persons. The white applicant or worker who loses out in preferential affirmative action owes no more or less as a citizen than any other citizen. That the burden of affirmative action falls unevenly on him is unfortunate if unavoidable.
Has it been unavoidable? This question points to an important area of neglect in affirmative action policy. When Congress passed the Civil Rights Act in 1964, it gave little thought to how to distribute the costs of the social change it was commanding. After all, the costs did not appear significant. What could it cost the factory to take the sign out of its window, or the drug store to serve blacks at its lunch counter? But in 1972, when Congress acknowledged that discrimination was more complex than it had thought, that it was a deep and extensive effect of "systems" operating in nominally neutral ways, it did not further acknowledge that the necessary social change required to reform these "systems" would involve substantial costs, both monetary and human. There was no public debate about how these costs should be shared. As the courts and the government imposed various forms of nonpreferential and preferential affirmative action on firms and institutions, they were content to let the costs fall where they may.
The monetary costs of affirmative action plans fell upon the companies and institutions that had to implement them. Given that they were incurred by nearly all sizable firms and institutions, these costs could be passed along to consumers through higher charges for goods and services; they were a burden all of us shared, as we should have. However, the human costs -- promotions and employment opportunities denied, legitimate expectations frustrated, and the like -- fell upon individuals unable to transfer them to anyone else. These costs weren't large in the aggregate, but they were significant individually. They were borne by the white workers or applicants who happened to be at the wrong place at the wrong time.
This outcome may have been unavoidable; no scheme of amelioration may have been feasible. But no one tried to find out. The courts, the Congress, and the defenders of affirmative action turned a blind eye to the question of costs and their distribution. Indeed, as we have already seen, many of the defenders of affirmative action denied there were any human costs at all. When you have undeserved opportunities or advantages snatched away, you have no basis for complaint; you haven't suffered any real harm.
Racial Preferences: Always Impermissible?
When universities set up special scholarships for minorities or adopt preferential programs to speed the entry of women into the sciences, say, they act for perfectly worthwhile ends. When municipalities create set-aside programs in order to help support a growing stratum of minority or female business entrepreneurs, they are promoting a socially valuable goal. If we were to reach a broad consensus in favor of these goals, and if we adopted policies that took account of the trade-offs and human costs involved in pursuing them, would that be enough to end the controversy?
Unfortunately, no. The reason is that, for many of those supporting CCRI and similar proposals, there is something fundamentally unjust about distributing benefits by gender and race, but especially by race. "You point to the moral evil of past racial discrimination," the critics of preferences say. "That makes the case for our side, since discriminating against whites is just as wrong as discriminating against blacks." This asserted moral symmetry is consistent with the Supreme Court's position that, as a constitutional matter, all racial classifications are equally suspect and therefore subject to the same level of scrutiny.
Yet why should this symmetry be assumed, unless the purpose and consequences of a practice have absolutely no bearing at all on its moral status? As a plain matter of our history, racial discrimination after emancipation from slavery relegated African-Americans to the economic and cultural margins of this country. It denied them protection of the law and access to the voting booth. Its aggregate impact on blacks was enormously destructive. And achieving that impact was among the very purposes of discrimination, every aspect of which was meant to remind blacks tangibly and symbolically that they were fit only for a low and subservient place in our society. By contrast, the highly circumscribed and limited preferences attaching to some parts of affirmative action imply no official judgment of contempt toward those discriminated against and are aimed at overcoming a legacy of racial division and exclusion.
It is certainly true that discriminating against whites in order to humiliate and subjugate them would be just as wrong as discriminating against blacks for the same purpose. Further, it is plausible to urge that discriminating against whites is wrong when it serves trivial rather than significant social interests. But it is not plausible to equate preferences under affirmative action with the malign and crippling discrimination under our former regimes of racial oppression. There is nothing intrinsically morally objectionable in using racial preferences to ameliorate and overcome the legacy of racial oppression in this country. Why not, then, treat race on a par with the countless other bases of distinction we make in law and policy?
There is a two-part answer to this question which makes CCRI more plausible. The answer is this. Race is different; but not because of something intrinsic to race. It is different because of history -- our particular racial past. Race has obscenely deformed our politics and corrupted our sensibilities for our entire national history. Of course we can imagine good reasons in the abstract for using racial preferences; the real issue, however, is trusting real people and real institutions in this country properly to act on those reasons. We have a history that should make us wary of ever letting anybody allocate public benefits for racial reasons. The risks of abuse are too great.
Consider a parallel case. Just as in the abstract, it seems a perfectly reasonable idea to let doctors act on the requests of terminally ill patients to speed their deaths, in real life the picture is not so clear. We can imagine a host of subtle ways that inappropriate influences could bear on and distort the decision-making process once doctors are given such dangerous power. Perhaps we had better withhold the power.
Thus, the first part of the two-part answer says that we must not only conceive of reasons in the abstract for using (or not using) racial preferences; we must also make a historically based prudential judgment about how these reasons would play out in the real world, where they might easily be corrupted and perverted. Whereas we might be able to offer decisive, knock-down arguments for some of the abstract reasons, no such easy course is available when the issue is a summative judgment about what will happen in the future, given the evidence of our past.
The second part of the two-part answer points not to a judgment we must make but to one we have already made. The fact of the matter is that the text of the Civil Rights Act of 1964 is not hospitable to racial preferences, however many good reasons there may be for using them. Over the years, courts have squeezed out some space for preferences, but preferences always tied to furthering the purposes of the Act itself -- preferences to enable an institution to cure its own discrimination. When the Act was written, Congress treated gender, religion, and other factors more flexibly, acknowledging, for example, that sex might sometimes be a legitimate qualification for a task and granting that some institutions might legitimately take religion into account in their operations. It made no such hedges in the case of race. Under the terms of the Act, employers and school officials may not bestow or withhold benefits on account of race, period.
Twice Congress has revisited the Civil Rights Act, in 1972 and 1991. Though it has amended the Act in various important ways, it has never modified the Act's blanket prohibition against the use of race in hiring or admissions. Congress could easily have revised the Act to read: "you may not discriminate against African-Americans and women, but you may discriminate against whites and males for very good reasons, such as achieving a racial and gender balance in your institution, or promoting integration in the professions, or creating role models for young blacks and women, or offsetting societal discrimination, or distributing the benefits and burdens of life more justly, etc." But it did not.
The two-part answer, then, says in effect: it would not be unreasonable as a society to bar putatively good uses of race as well as bad; and, in fact, we have. So the various good reasons we can imagine for extending preferences by race are beside the point. Legally, they are off the table.
IV. CONCLUSION
One reason I have described the early development of affirmative action is to jar us out of the habit of equating affirmative action with racial and gender preferences. There is a quite distinct and important nonpreferential variety of affirmative action. Most federal contractors still have to create and act on such affirmative action plans; and as the recent events at Texaco should impress upon us, the kind of self-monitoring and self-discipline demanded by such plans is very much still needed. Moreover, the failure to distinguish nonpreferential from preferential affirmative action can produce unfortunate confusion. The week after the passage of CCRI, the government in Loudon County, Va., abolished its entire affirmative action program, all of its nonpreferential elements along with anything possibly preferential, thinking it was following in the path blazed by CCRI. (Fortunately, the county government reinstated the program within weeks.)
The second reason I have described the early development of affirmative action is to show how preferential policies might arise directly out of the antidiscrimination mandate. Preferential policies do not all stand on the same ground, and it is important to be clear about this. The argument that an opponent of preferences must make against a program like that imposed on AT&T is very different from the argument he might make against a university's racially reserved scholarship program.
Any argument against AT&T-like preferences must engage the claim that an institution's discrimination can't be overcome without such a program; and engaging this claim means having an appropriate conception of discrimination. This is often the sinkhole into which all arguments about affirmative action and preferential treatment disappear, since people's conceptions of discrimination are so variable and, indeed, often so unarticulated even to themselves. The legal conception of discrimination that evolved under the Civil Rights Act of 1964 and that now stands embodied in the Civil Rights Act of 1991 ought to be the starting place of argument, even though that conception is itself full of difficulties. Getting some common ground on the idea of discrimination is vital, because not only do preferences like those in the AT&T case purport directly to be antidiscrimination measures, but those employed by universities and contracting agencies may also link up indirectly to antidiscrimination aims. If we cannot agree on what discrimination is, we will not be able to agree on what we need to do to end it.
Finally, there are many good arguments in the abstract for using race as a tool to promote worthy social purposes. What we must decide, however, is the likelihood those reasons would actually play out well if we allowed real people and real institutions to act on them. The right answer to this challenge is far from obvious. A society might reasonably choose any number of courses, from conceding limited permission to use racial preferences to barring them altogether. Arguably, our basic law has pretty much done the last.
Sources: The Civil Rights Act of 1964, as amended by the Civil Rights Act of 1991 (42 U.S.C. 2000 etc.), is the basic statutory starting point for thinking about discrimination and preferences. Revised Order #4 (41 C.F.R. 60-2), issued by the Labor Department in 1972, implemented Lyndon Johnson's Executive Order 11246 requiring "affirmative action" of all federal contractors. The seminal court cases mentioned or referred to in this essay are: Griggs v. Duke Power Company, 401 U.S. 424 (1971); Regents of University of California v. Bakke, 98 S. Ct. 2733 (1978); United Steelworkers of America v. Weber, 443 U.S. 193 (1979); Richmond v. J. A. Croson Co., 488 U.S. 469 (1989); Podberesky v. Kirwan, 46 F 3rd 147 (1994); Adarand v. Pena, 115 S. Ct. 896 (1995); and Hopwood v. Texas, 78 F 3rd 932 (1996). On "goals" and "quotas," see chapter 11, Robert K. Fullinwider, The Reverse Discrimination Controversy (Rowman & Littlefield, 1980); on the emergence of preferential attacks against discrimination, see Robert K. Fullinwider, "Achieving Equal Opportunity," in The Moral Foundations of Civil Rights, edited by Robert K. Fullinwider and Claudia Mills (Rowman & Littlefield, 1986). For recent versions of the theme "whites are not innocent," see Sharon Rush, "Understanding Affirmative Action: One Feminist's Perspective," in An Ethical Education: Community and Morality in the Multicultural University, edited by M.N.S. Sellers (Berg, 1994) and Thomas Ross, "The Rhetorical Tapestry of Race: White Innocence and Black Abstraction," William and Mary Law Review, vol. 32 (Fall 1990). The remarks by Hacker and MacIntyre can be found in Andrew Hacker, Two Nations: Black and White, Separate, Hostile, Unequal (Ballantine Books, 1992), p. 50, and Alasdair MacIntyre, After Virtue (University of Notre Dame Press, 1981), p. 205. The quotation about the neurosurgeon's son versus the coal miner's daughter comes from Christopher Edley, Jr., Not All Black and White: Affirmative Action and American Values (Hill & Wang, 1996), p. 138.