Discounting Discrimination

This article was written before the Supreme Court decided Wal-Mart v. Dukes.

Wal-Mart sends paychecks to more people than live in any one of eleven American states.Photo by Drew Allen.

It is no secret that anti-discrimination laws are, at best, a limited and controversial success. They have not done enough to stop discrimination: women still earn roughly eighty percent of what men earn for comparable work,[1] many workplaces and entire job categories remain racially segre­gated, unemployment among blacks—especially young black men without a high school education—is at a record high,[2] older job seekers face long odds in finding employment, and people with disabilities are shut out of work­places where they could thrive if given a chance and reasonable accommo­dations. This leads liberals to insist that we need more and stricter civil rights laws. But employment discrimination suits have multiplied at a far greater rate than other civil litigation: in the nine years between 1991 and 2000 they have almost tripled, from about 8,300 to over 22,000, and today they are the most numerous type of lawsuit on the federal civil docket.[3]

We don’t need more or tougher civil rights laws; we need better and smarter civil rights laws that can attack the subtler and less blatant bias of today, in jobs where decision making is decentralized and job criteria are fuzzy and subjective. We can get at today’s discrimination by using familiar and accepted evidentiary methods and long-standing theories of discrimina­tion. But we must be willing to admit that we often can’t pinpoint discrimi­nation accurately enough to blame a specific individual perpetrator, or to identify a specific individual victim, even though we can identify more gen­eral patterns of discrimination. This requires a new focus for civil rights law—one that abandons, or at least de-emphasizes, individual entitlement and individual culpability in favor of social justice. The recent case of Dukes v. Wal-Mart Stores, Inc.[4]—celebrated by civil rights advocates and maligned by free market conservatives—is a prime example of the new and more elusive type of discrimination that plagues the modern labor market and demonstrates the limits of conventional civil rights law.

Wal-Mart: Everyday Discrimination?

Wal-Mart, famous for its everyday low prices and notorious for its eve­ryday low wages, epitomizes many of the twenty-first century’s defining so­cial controversies. It is the predatory big-box chain store that drives small businesses under and exacerbates suburban sprawl; the mass retailer that en­courages a soulless and ecologically destructive habitual consumption of dis­posable products; the avaricious multinational corporation that exploits the land and the labor of developing nations; and the skinflint employer that pays its workforce a poverty wage. In this last respect, Wal-Mart’s influence has been profound: the largest employer in the United States with 1.4 million employees,[5] Wal-Mart sends paychecks—such as they are—to more people than live in any one of eleven American states.

Sam Walton was opening his first stores in Arkansas at just the time mechanization was replacing agricultural labor across the Ozarks. Walton stepped into the role of straw boss, offering unemployed former farmhands the jobs of the future at fifty cents an hour.[6] According to historian Nelson Lichtenstein, when Congress extended the minimum wage to retail employ­ers, Walton responded with the ingenuity that would make Wal-Mart a mar­ket leader: the minimum wage law applied only to businesses with $250,000 or more in annual revenues, so Walton divided his business into smaller corporations that would fall under the statutory threshold.[7] A federal court saw through the scheme and ordered Walton to compensate his employees retroactively.[8] Walton had no choice but to pay up, but he also dug in with the tenacity that employees and suppliers alike would come to know and fear. He delivered a warning along with the back pay: “I’ll fire anyone who cashes the check.”[9]

Wal-Mart’s Southern roots, scofflaw tendencies, and plantation-style la­bor practices seem a recipe for rampant race discrimination. Indeed, Wal-Mart has faced more than its share of complaints: the Equal Employment Opportunity Commission (EEOC) has filed about sixty employment-dis­crimination lawsuits against Wal-Mart.[10] In 2009 Wal-Mart settled a dis­crimination suit brought by black applicants for jobs as truck drivers for $17.5 million. That same year, the predominantly West African employees of a Colorado Wal-Mart complained that a manager remarked, “I don’t like some of the faces I see here. There are people in Eagle County who need jobs.”[11] But Wal-Mart’s most serious civil rights problem is sexism. Wal­Mart appears to have had little success in changing widespread patriarchal attitudes among its managers: according to American Prospect editor and Washington Post columnist Harold Meyerson, “[t]he Southern traditional­ism of Walton and his lieutenants dictated that the stores’ managers would be men and its salesclerks women.”[12] Because over ninety percent of Wal­Mart’s cashiers are women,[13] sexism would affect far more people than other forms of discrimination, even if it were no more pervasive or severe in de­gree. Not surprisingly then, sexism has become a serious financial liability for Wal-Mart as the company has faced numerous discrimination lawsuits. The EEOC sued Wal-Mart claiming that the company’s London, Kentucky, distribution center systematically excluded women from warehouse jobs; on March 1, 2010, Wal-Mart agreed to pay $11.7 million to settle the suit out of court.[14]

That’s small change compared to what Wal-Mart will have to pay if Betty Dukes has her way. Dukes worked for Wal-Mart in Pittsburg, Califor­nia, a working-class city about thirty miles east of San Francisco.[15] She had been on the job for seven years and sought promotions that, she claims, went to less qualified men.[16] Dukes claims that Wal-Mart routinely assigned wo­men to stereotypically feminine departments, such as baby clothing, and ex­cluded them from masculine departments, such as hardware. “I can mix a can of paint,” she complained. “I want a chance to do it.”[17] According to Dukes, openings for managerial positions were never announced and were filled by men before she could apply.[18] In 2001, Dukes, along with five other women, sued Wal-Mart in a class-action lawsuit filed on behalf of every woman who worked for Wal-Mart at any time since 1998—roughly 1.5 million women.[19] If allowed to proceed, this would be the largest class-action lawsuit in American history[20] and could cost Wal-Mart billions of dollars. Indeed, Wal-Mart’s 2009 annual report to shareholders deadpanned that “the Company cannot reasonably estimate the possible loss or range of loss that may arise from the litigation.”[21]

Betty Dukes and her co-plaintiffs claim that Wal-Mart systematically discriminated against women in pay and promotions.[22] They point out that almost three-fourths of Wal-Mart’s hourly wage sales employees are women; by contrast, only about one-third of its managers are.[23] They argue that Wal­Mart’s executives in Bentonville, Arkansas, set the tone for the more than 3,000 stores in the United States[24]—the same tone of traditional Southern chauvinism and contempt for civil rights laws that Sam Walton exhibited in the 1960s. They’ve gathered declarations by more than 100 women who say that they were victims of sex discrimination at Wal-Mart.[25] And they’ve en­listed sociologist William Bielby, who claims that Wal-Mart’s employment practices are “vulnerable” to the influence of sex stereotypes because they leave personnel decisions to local store managers who use subjective criteria of evaluation such as “teamwork, ethics, integrity, [and] ability to get along with others.”[26]

Not surprisingly, Wal-Mart denies all of this. “We don’t have policies and practices in place that promote discrimination of any kind,” Wal-Mart spokesman Bill Wertz insisted.[27] Wal-Mart and its supporters argue that the Dukes plaintiffs can’t point to any specific company-wide policy or prac­tices; instead, they portray the plaintiffs as citing isolated and “widely diver­gent” anecdotes and a vague hypothesis of a monolithic corporate culture to conjure up the specter of a common pattern of discrimination.[28] According to one observer, the plaintiffs’ claims amount to little more than “accusations by women that supervisors, including female supervisors, made disparaging remarks about women workers—something entirely possible in a company of more than 1 million employees . . . . [T]here’s nothing in these anecdotes that adds up to a company-wide pattern of discrimination.”[29]

If Wal-Mart had exercised tight centralized control over personnel deci­sions, any gender imbalance could be attributed to systematic bias among the upper management in Bentonville, Arkansas. If it had promulgated de­tailed objective rules for personnel decisions, perhaps the rules themselves could be attacked as discriminatory in effect. Either of these scenarios might justify a class action attacking the practices or rules that affected all Wal-Mart employees nationwide. But Wal-Mart did not control personnel deci­sions centrally: it left them in the discretion of local managers, who, like many managers in the service industries, employed vague and subjective criteria such as “integrity” and “teamwork.”[30] Such decentralized manage­ment is less susceptible to a firm-wide class action because each store argua­bly operates under its own decision-making regime. Even if many Wal-Mart employees suffered sex discrimination, if each individual store has its own distinctive personnel practices, there are no questions of law and fact com­mon to all Wal-Mart employees.

The plaintiffs struggled to make the square peg of Wal-Mart’s decentral­ized and subjective personnel decision-making process fit the round hole of a class action with an injury common to all class members. They insisted that Wal-Mart had a centralized corporate culture—if not a centralized deci­sion-making process—that affected all 1.5 million women who had worked at Wal-Mart since 1998.[31] And how did Bentonville’s alleged ethos of sex­ism make its way to Wal-Marts from Maine to Monterey? Why, through decentralization of control over personnel decisions, which gave individual store managers almost absolute discretion, thus leaving their decisions “vul­nerable” to the influence of sex stereotypes. The plaintiffs wanted to have their centralization and eat it too: on the one hand, they argued, Wal-Mart was so centralized that a monolithic corporate culture affected decisions na­tionwide; on the other, it was so decentralized that individual managers could do whatever they pleased. Would it be better for Wal-Mart to be more centralized or less? One commentator complained of a “contradiction in describing a corporate system as both excessively subjective and at the same time centralized,” and worried that, under the plaintiffs’ legal theory, “de­centralization in management—one of the key productivity-boosting princi­ples of U.S. business in the last two decades—is something that firms must avoid or limit.”[32]

But the core of the plaintiffs’ discrimination case is statistical. Wal-Mart draws most of its managers from its hourly wage employees, of which 72% are women.[33] But only one-third of Wal-Mart’s managers are women, and “even this figure overstates the proportion of female managers [because it] . . . includes traditionally ‘female’ positions, such as assistant managers . . . and the lowest level of managers . . . . [W]omen comprise less than 10% of all Store Managers and approximately 4% of all District Managers. There are very few female Regional Managers. There is only one woman among the 20 executive officers of the company.”[34] By contrast, “among [Wal-Mart’s] 20 top competitors women comprise over 56% of management . . . . In fact, female representation among managers at Wal-Mart is at a substantially lower level today than [it was] among Wal-Mart’s competitors in 1975.”[35]

The Dukes class action would sever the question of Wal-Mart’s blame­worthiness from the question of any individual woman’s desert. According to the trial plan established by the District Court:

Plaintiffs would attempt to prove that Wal-Mart engaged in a pattern and practice of discrimination . . . . [Then] . . . a formula would be used to calculate the “lump sum” . . . that Wal-Mart owes to the class . . . . A separate procedure would then be used to distribute these lump sums to those class members entitled to share in them—a stage in which Wal-Mart would no longer have an interest.[36]

The District Court sensibly concluded that although it may be able to deter­mine whether or not Wal-Mart is guilty of discrimination, it will not be able to determine which women were victims of discrimination and which were not.[37] The class action will allow the court to order Wal-Mart to change discriminatory practices and to impose monetary damages in order to deter potential future discrimination by Wal-Mart and other employers.

But Wal-Mart complained that the class-action mechanism is unfair to individual women.[38] Under the proposed trial plan, if Wal-Mart is found liable, the jury will award injunctive relief and monetary damages to the class as whole.[39] The money will be apportioned to individual members in a separate proceeding, in which Wal-Mart would have no interest. Since Wal-Mart won’t have a chance to advance defenses to individual claims, this pro­cess “virtually guarantees [that] non-victims will share in a monetary dam­age award along with those truly harmed by Wal-Mart’s alleged discrimination.”[40] As Ninth Circuit Judge Kleinfeld—who voted to stop the lawsuit from going forward—argued:

Women employed by Wal-Mart who have suffered sex discrimina­tion stand to lose a lot if this sex discrimination class action goes forward. All the members of the class will be bound by the judg­ment . . . . What if the plaintiffs’ class loses? Worse, for many women in the class, what if the plaintiffs win? Women who have suffered great loss because of sex discrimination will have to share the punitive damages award with many women who did not . . . . Women who have left Wal-Mart will get injunctive and declara­tory relief of no value to them, while new female Wal-Mart em­ployees will benefit from the injustice done to other women.[41]

Wal-Mart also argued that a class action deprives Wal-Mart of many of its statutory defenses under federal civil rights law.[42] For instance, in an individual case, Wal-Mart could avoid paying money damages by showing that a decision that was influenced by sexism would have been the same even without the sexism. In the sex harassment context, Wal-Mart could defend itself by showing that it had an internal process in place to deal with sex harassment complaints and the plaintiff unreasonably failed to use it. The class action plan doesn’t allow for such individual defenses.[43] But such individualized defenses are irrelevant in a case argued largely on the basis of statistics. The Dukes plaintiffs offered anecdotal evidence of discrimination, a theory of a centralized and sexist corporate culture, and sociological evi­dence that Wal-Mart’s organization structure is “vulnerable” to influence by stereotypes. None of this evidence, however, comes close to proving that Wal-Mart discriminated against anyone: ultimately, the case stands or falls on the statistical evidence. The law doesn’t require an employer to have a workforce that reflects the available labor pool, but it’s fair to infer discrimi­nation from an unusually lopsided workforce. Of course, if Wal-Mart can offer convincing evidence of an alternative explanation, it should avoid lia­bility; if it can offer evidence that partially explains the disparity, its liability should be reduced accordingly.

Can The Civil Rights Of Yesterday Fix The Problems Of Today?

In a sense, Dukes v. Wal-Mart is a back-to-the-future civil rights case, using a disfavored type of litigation and an older conception of what civil rights litigation should accomplish in order to confront a very contemporary social problem. Shortly after the Civil Rights Act was passed in 1964, the class action became one of the most important types of discrimination litiga­tion.[44] Legal challenges to a pattern and practice of discrimination—a col­lective injury—have been at the center of civil rights litigation since the early 1970s. The law prohibiting polices with a discriminatory or disparate effect on minorities and women—widely considered the most important part of employment discrimination law for years after it was established in 1971—addresses a group-based injury.[45] As late as 1978, legal commenta­tors considered the individual civil rights case a relatively minor part of civil rights enforcement.[46] Early civil rights cases saw individual rights as a means to the end of equal opportunity for everyone: the individual plaintiff was considered a “private attorney general” who basically reinforced the efforts of the EEOC.[47]

The prestige of other collective civil rights claims suffered in the 1980s. Legal changes made class actions more difficult to bring, and by 1988 the official reporter of the Federal Rules of Civil Procedure could opine that “class actions had their day in the sun and kind of petered out.”[48] Mean­while, the Burger Supreme Court refused to extend Title VII’s disparate im­pact theory of discrimination to constitutional litigation in 1976’s Washington v. Davis.[49] Throughout the 1980s, the Rehnquist Supreme Court limited disparate impact under Title VII of the Civil Rights Act, insisting that individual discriminatory treatment was the law’s defining civil rights violation.[50]

But in today’s job market, collective claims make more sense than ever before. Wal-Mart has gained notoriety because it exhibits all of the defining features of the contemporary service-sector employer in exaggerated form: it has a large low-wage workforce, high turnover, a decentralized management structure, and it evaluates its employees based on highly subjective criteria. Wal-Mart is larger than any other employer, and many of its practices may be more extreme than those of its competitors, but all of the things Wal­Mart’s critics hate about it are typical of service-sector employers generally.[51]

These features—especially high turnover, decentralization, and subjec­tive evaluation—make it hard to apply the individual civil rights model that has developed since the 1970s. The main method of proving individual dis­crimination was established in 1973, in the context of manufacturing jobs that required objective skills and formal certifications. Percy Green, a black employee of the St. Louis McDonnell Douglas factory who was laid off from his job, sued when McDonnell Douglas rejected his application a year later, after it started hiring again.[52] Green insisted that McDonnell Douglas rejected him because of his race; McDonnell Douglas claimed it rejected him because an earlier protest he organized disrupted the company’s fac­tory.[53] The Supreme Court in McDonnell Douglas v. Green held that if a plaintiff shows that (1) he’s a member of a racial minority; (2) he applied for and was qualified for the job; (3) the employer rejected him; and (4) the employer continued to seek similarly qualified applicants for the position, he has made out a prima facie case of racial discrimination.[54] At that point the employer must offer a non-discriminatory reason for rejecting the plaintiff.[55] Finally, the plaintiff can try to convince the judge or jury that the defendant’s reason is really a pretext for discrimination—if he does, he wins his lawsuit; if not, he loses.[56]

This method worked pretty well in a manufacturing context, like Mc-Donnell Douglas. Employees needed objective skills, such as formal train­ing or certification to use specific types of equipment, and employers typically were indifferent between employees with the requisite skills. When McDonnell Douglas refused to hire Green, it raised the suspicion that its reasons were unusual. But notice how anachronistic the McDonnell Douglas approach is in today’s service-sector labor market. Consider factor number two of the prima facie case: the plaintiff applied for and was quali­fied for the job. Today, many jobs are filled by word of mouth or through informal internal promotion processes. The Wal-Mart plaintiffs complained that women were not mentored and groomed for upper-management posi­tions, so they weren’t prepared when the better jobs opened up, and some­times openings weren’t even announced. What does it mean, then, to “apply” for a management position at Wal-Mart? And the qualifications for many service-sector jobs are hard to define. The Wal-Mart promotions pro­cess was left almost entirely up to local managers, who said they considered factors such as “teamwork, ” “ethics,” “integrity,” and “the ability to get along with others” in their decisions. Getting a promotion at Wal-Mart—as in many service-sector jobs—isn’t a simple matter of being “qualified,” be­cause it requires being the best of a large group of aspirants, based on “soft skills” such as demeanor, poise, and personality. It’s easy for bias and ste­reotypes to hide in the fog of subjectivity such evaluations entail: a manager who thinks women shouldn’t be in charge might claim a female candidate lacks “team spirit,” or his opinion of a female candidate’s people skills might be influenced by his stereotyped views of appropriate feminine demeanor.

The Wal-Mart plaintiffs come close to insisting that decentralization and subjective job criteria are inherently discriminatory.[57] This is the weak­est part of their case: decentralized and subjective decision making are legiti­mate and effective management styles; a company shouldn’t be punished for adopting these very common practices, nor should we presume that they are a cloak for bias. In a stereotypical manufacturing job, it might make sense to be suspicious of subjectivity: the main thing the employer should care about is how well and how quickly an employee performs his job on the production line—something that can be objectively measured. But in the service-sector jobs that are the fastest growing part of the modern labor mar­ket, objective performance is only part of the job. A cashier who efficiently completes a transaction but is surly to customers is a bad cashier. A man­ager who knows how to organize his subordinates but can’t manage conflicts among them or inspire them to work diligently is a bad manager. In fact, even in manufacturing, the most successful businesses today know that teamwork, cooperation, and the ability to innovate are important skills for everyone from upper management to the assembly-line worker. State-of­the-art management science suggests that giving local managers, who are closest to the specific challenges, the latitude to respond to them often leads to dramatic improvements in productivity and efficiency. These innovations in management may be more “vulnerable” to bias than formal and objective job criteria, but they are becoming more and more prevalent for good reasons.

Decentralized management and subjective job criteria may be vulnera­ble to bias, but the real problem for civil rights law is that whatever bias there is will be harder to detect, making it very near impossible to prove specific instances of discrimination, even when discrimination is rampant. Judge Kleinfeld insisted in his dissent to the Dukes class certification that the main justification for a class action is to “solve an attorneys’ fee problem . . . . [S]mall recoveries do not provide the incentive for any individual to bring a solo action prosecuting his or her rights.”[58] If so, the Dukes class action is inappropriate because “[m]uch of the bar now earns a living by litigating sex discrimination claims. . . . [W]omen discriminated against by Wal-Mart do not need a class action. They can, with contingent fee agree­ments, afford to hire their own lawyers . . .”[59] But if the Wal-Mart plaintiffs are right about Wal-Mart’s pervasive culture of sexism, individual lawsuits can’t capture the magnitude of the social injustice, nor can they provide the appropriate incentives for Wal-Mart to change its ways. The evidence of statistical disparity can’t tell us whether any given woman was a harmed by sexism, but it might tell us a lot about the pervasiveness of sexism at Wal-Mart. Since the nature of Wal-Mart’s personnel management makes it hard to detect individual discriminatory intent, most of the Wal-Mart plaintiffs would lose individual suits, even if they were victims of discrimination. Judge Easterbrook of the Seventh Circuit explained this problem in a race discrimination case:

Suppose 1,000 employees apply for 100 promotions; 150 of the workers are black and 850 are white. If all are equally qualified and the employer ignores race, then 85 white workers and 15 black workers will be promoted, plus or minus some variation that can be chalked up to chance. Suppose only 10 black workers are pro­moted. Is that the result of discrimination or chance? Econometric analysis . . . may suggest the answer . . . [but] it cannot reveal with certainty whether any given person suffered . . . . [Suppose that] [b]ut for discrimination, 15 [black employ­ees] would have been promoted . . . . Which of the 140 non-promoted employees would have received the other 5 promotions? The statistical analysis does not tell us—and in civil litigation, where the plaintiff’s burden is to show more likely than not that he was harmed by a legal wrong, data of this kind will not get a worker over that threshold.[60]

Even when we can be pretty sure a given employer discriminated, it’s possi­ble that no individual will be able to prove that she was discriminated against. Because civil rights law requires such proof, much of the modern job market has left civil rights law behind. The Dukes suit, with its heavy reliance on statistics, is an attempt to bring employment discrimination law into the twenty-first century.

What’s new is that today it is often very hard to tell when an individual injustice has occurred, even when we know from looking at society as a whole that a lot of injustices must be occurring somewhere. Even if the statistics cited by the Wal-Mart plaintiffs prove that Wal-Mart must have discriminated against women, we have no way of knowing which women were victims of discrimination and which suffered due to their own poor performance of lack of initiative.

Sex discrimination is pervasive in the American labor market. Despite decades of effort by women’s groups, women still earn on average roughly eighty percent of what men do for comparable work, and the glass ceiling prevents qualified women from getting many of the most lucrative posi­tions.[61] Wal-Mart may well be worse than average, but it’s a prime example of a widespread problem. Wal-Mart makes a good target, not because it is unique, but because it is conspicuous. The Wal-Mart plaintiffs called atten­tion to a social injustice that they as individuals may or may not have suf­fered by zeroing in on a target that may be no worse than most. An individual claim is a potentially misdirected way to address a real injustice. But what is the alternative? It is hard to oppose these efforts, as poor a fit as they may be, if there is no other way to confront the injustices that civil rights protests and litigation try to correct.

As Judge Kleinfeld complained in opposing class certification, the trial plan aspires to only rough justice for the millions of putative victims of Wal­Mart—it will almost certainly give an undeserved windfall to some women who were not victims of discrimination and under-compensate some who were.[62] This is a damning objection only if individual compensation is the overriding goal. Because we have defined social justice in terms of rights, it is natural to assume that some sort of individual restitution is the ultimate goal. But perhaps individual compensation should not be the main objec­tive: maybe the means has become confused with the ends. Consider the idea of the civil rights plaintiff as a “private attorney general”—a sort of desegregation bounty hunter who is deputized to help enforce the law and offered a reward, in the form of monetary damages, for so doing. Here, private damages are not a moral entitlement, but instead an incentive offered to entice people to bring civil rights violations to court. From the defen­dant’s perspective, damages are designed to deter discrimination. For pur­poses of deterrence, it does not matter who gets the money as long as the defendant has to pay for violating the law. In this sense, civil rights laws are designed to encourage more responsible behavior so that a lawsuit is not necessary. Effective regulation can bring about similar changes in behavior directly, making individual entitlements unnecessary.

Social Justice Is Not An Individual Right

Suppose we were to suspend—if not abandon—the idea that every vic­tim of discrimination is entitled to an individual remedy. Suppose we shifted our focus from individual entitlements to public policies designed to reduce the prevalence of invidious discrimination and unjust social inequal­ity overall. This may seem unfair if you think that individuals have a right to equal treatment that no collective goal can outweigh. But the American le­gal system has never recognized such a broad and sweeping guarantee of individual fairness. In the United States, most employment is “at will”: employers can choose whom to hire, fire, and promote for any reason that isn’t explicitly forbidden by law.[63] A lot of bad motives—nepotism, favorit­ism, idiosyncratic prejudice (imagine an employer who hates redheads or people with bad skin)—are legally acceptable.

Employment at will is controversial, and sometimes it is superceded by other arrangements, such as civil service rules, union contracts, academic tenure, and state wrongful termination laws. Government could try to guar­antee that all employment decisions are made for “good cause.” But that would require detailed micro-managing of innumerable business decisions: government would need to determine what counts as good cause in a host of different situations and to review every controversial employment decision. Such a reform would be costly, both in terms of direct expenses on regula­tory enforcement and litigation, but also in terms of inconvenience and new inefficiencies for employers when the government makes a mistake and pro­hibits a useful practice or reverses a sound decision. Employment at will reflects valid concerns about the capacity of government to identify and pre­vent or remedy unfairness: it can be very hard to tell the difference between unfair treatment and the kinds of difficult judgment calls that every em­ployer must make. Would a more aggressive regulation of the employment market be worth the expense? If the labor market is healthy, a good em­ployee who is treated unfairly will be able to find employment elsewhere— in fact, many people who could sue their employers for discrimination prefer to cut their losses and find a new job. Perhaps it would be wiser to devote scarce resources to unemployment insurance and a robust social safety net to help cushion the blow of all unavoidable employment dislocations—whether due to capricious employers or the unpredictable shifts of a volatile economy.

Civil rights law prohibits only the most severe and socially destructive forms of discrimination—typically those that have become widespread be­cause of misguided laws or social customs in the past. For instance, race discrimination was part of an all-encompassing, legally structured system of racial hierarchy and many of the habits, attitudes, and effects of that system are still with us today. Similarly, sex discrimination was legally and socially enforced for much of American history, and the bad habits of male chauvin­ism are not yet extinct. The social injustice and individual harm caused by race and sex discrimination have been much more severe than those caused by other types of favoritism, nepotism, or idiosyncratic biases. The racial minority passed over because of his race or the woman limited in her ad­vancement because of her sex may not find better opportunities elsewhere; instead they may well face the same type of discrimination over and over again. But if these kinds of discrimination were less widespread, they wouldn’t be as damaging. Public policy that reduced segregation and irra­tional prejudice would make race and sex discrimination less severe and more like the kinds of unfairness we can get beyond without individual compensation.

Moreover, practically speaking, individual rights don’t provide every victim of discrimination with a remedy. Lots of people are victims of dis­crimination but do not know it or cannot prove it. Plenty of others find that the costs of pursuing their rights—time and expense spent in litigation and a possible tarnished reputation in one’s profession—outweigh the benefits. It is expensive and time consuming to sue for discrimination; as a result, very few people do so, even when they have valid claims. And the people who do sue are not always those with the strongest claims; instead, they are often those with the least to lose in terms of foregone opportunities and reputation by filing suit. Imagine a competent, conscientious, amiable employee with a good reputation in her field who is fired by a racist employer. She could spend the next several years suing her employer for foregone pay and the dubious privilege of returning to work for a bigot, or she could just find another job. Now imagine an incompetent, lazy, and belligerent employee who is fired for her poor performance and disruptive behavior. She is un­likely to find work elsewhere and has no good reputation to put at risk. With nothing to lose, such an employee has every incentive to file a trumped-up discrimination claim.

A focus on individual civil rights can inadvertently encourage weaker claims over stronger ones in other ways. For instance, people are much more likely to sue when they lose a job or are denied an expected promotion than when they are not hired in the first place. This is probably because most people value things that they currently have much more highly than they would value the same thing if they did not have it. This phenomenon is so well-known among psychologists and economists that they have a techni­cal name for it: the endowment effect.[64] It’s a sort of twist on the idea that a bird in the hand is worth two in the bush. Numerous studies have shown that people will demand much more—sometimes orders of magnitude more—to sell an item that they possess than they would offer to buy the same item if they did not have it. My colleague Mark Kelman has offered this simple example[65] of the phenomenon: suppose I were to inherit a rare and valuable bottle of wine, worth $1,000. An acquaintance, who is an ob­sessive oenophile, would happily pay me $1,000 for the wine. I, by contrast, am a martini drinker—a relatively plebian tipple by contrast to vintage Bor­deaux—and would never spend more than $100 on a bottle of wine even for the most special occasion. Still, I’d kind of like to know what a $1,000 bottle of wine tastes like. If I inherited one, it is quite plausible that I would decide to drink the wine myself rather than sell it. This is an example of ask/ offer asymmetry: I would not relinquish a bottle in my possession even for ten times the price I would be willing to pay for it if I didn’t already have it.

The same phenomenon seems to apply to jobs. Immediately after the Civil Rights Act was first passed, many employers had blatantly discrimina­tory hiring practices and deliberately segregated workforces—these offered easy targets for litigation. But civil rights plaintiffs quickly picked off the sitting ducks, and employers began to change their practices. Discrimination in hiring was harder to prove, and few individuals had the information or the incentive to sue over a job that got away. But many people, having landed a job, would fight to keep it. As a result, complaints of termination now out­number failure to hire complaints by almost seven to one.[66]

Devoting such an overwhelming share of resources investigating and adjudicating claims of discriminatory termination makes little sense. By so heavily favoring discriminatory termination claims, individual civil rights litigation looks for prejudice where we are least likely to find it while neglecting more promising hunting grounds. Only an employer who has in fact hired members of an underrepresented group can become a defendant in a discriminatory termination lawsuit brought by a member of that group. Such an employer is less likely to be bigoted than one who has never hired any minority employees in the first place, yet civil rights enforcement spends most of its resources policing the actions of just these employers while paying relatively little attention to employers who discriminate in hiring.

Such lopsided enforcement not only misdirects enforcement resources, it also creates perverse incentives. A well-crafted civil rights policy would reward employers who hire members of underrepresented groups and punish those who refuse to. But current civil rights litigation does just the opposite: it makes it relatively safe to refuse to hire members of underrepresented groups (because there is little risk of being sued for failure to hire) and risky to hire them (because doing so opens one up to the much more likely lawsuit for discriminatory termination). As my Stanford colleague, law professor and economist John Donohue notes,

A worker who is not hired in the first place is obviously in no position to bring a future firing suit. . . . With the enormous in­crease in discharge cases, the probability that a worker will bring a discriminatory firing suit is now substantially higher than the probability that a worker will bring a failure to hire suit. Conse­quently, anti-discrimination laws may actually provide a (small) net disincentive to hire women and minorities.[67]

Supplementing individual entitlements with more assertive administrative regulation in the public interest and eliminating the most questionable enti­tlements would change these perverse incentives and lopsided emphases.

The current civil rights model—premised on a struggle between indi­vidual victims and callous or belligerent perpetrators—pits business against social justice. But many of today’s businesses need not be made the enemies of equal opportunity—they could be powerful allies. Unlike individual plaintiffs suing in the adversarial legal system, administrative agencies could work with businesses to develop reforms that get results without undermin­ing productivity. Today hundreds of businesses try to recruit minority job applicants, improve job conditions for working women, and provide access to people with disabilities. Most do all of this without judicial coercion or the threat of litigation. Large corporations such as 3M, Coca-Cola, General Electric, Hewlett-Packard, Intel, Johnson & Johnson, Kraft Foods, Microsoft, Nike, Pepsi, Procter & Gamble, Reebok, and Xerox urged the Supreme Court to affirm the validity of affirmative action in 2003’s Grutter v. Bollin­ger.[68] They insisted that “today’s global marketplace and the increasing di­versity in the American population demand the cross-cultural experience and understanding gained from [a diverse] education[al environment].”[69]

But the antagonistic relationship between businesses and the civil rights community leads some employers to jealously guard the race and gender composition of their workforces. For instance, several Silicon Valley high tech companies, including Google, Apple, Yahoo, Oracle and Applied Materials, wanted to keep this information secret so badly that they fought to prevent San Jose Mercury News reporter Mike Swift from obtaining this routine workforce data from the EEOC.[70] The companies claimed that the information was a “trade secret.”[71]

It is hard to imagine how the raw data could reveal a firm’s recruitment strategy, and we will have to imagine because when Swift asked for elabora­tion he was told that the explanation was a trade secret, too. But it’s not hard to imagine why they would prefer to keep the race and gender of their workforces a secret: in 2008, only 1.5% of computer workers living in Silicon Valley were black, and 4.7% were Hispanic; by contrast nearly one-quarter of the working-age population in the region is Hispanic.[72] Similarly, men and women in technical careers in Silicon Valley are equally likely to hold mid-level jobs, but men are 2.7 times more likely than women to be promoted to high-ranking tech jobs.[73] Of course this doesn’t mean that the firms are discriminating: there are many possible explanations for these figures. Qualification for high-tech jobs varies by race and gender, as does interest and proximity. For instance, although more than twelve percent of the national working age population is black, only 2.9% of working age Silicon Valley residents are.[74] In this light, the 1.5% figure doesn’t look quite so bad. Perhaps the statistics simply reflect the qualifications and in­terest of the available labor force.

Still, it’s telling that the businesses of the future are willing to go to such lengths to keep their workplace diversity—or lack thereof—a secret. A more open and cooperative relationship between business and civil rights advocates might yield new ideas for improving diversity. Or, of course, it might demonstrate that the firms are as diverse as one could expect given the available and qualified labor pool.

When the Civil Rights Act of 1964 was debated in Congress, many of its sponsors proposed an administrative approach, modeled on the National Labor Relations Board. They saw private enforcement through lawsuits as a second-best compromise. Civil rights leaders worried that private enforce­ment was a way to weaken the new law, because “many of those discrimi­nated against would be poor and legally unsophisticated.”[75] As it has turned out, private enforcement has not led to too little enforcement, as liberal civil rights groups feared, nor has it led to too much, as many conservative com­plaints would suggest. Instead, it has given us some of both: many serious offenses go undetected and uncorrected, while minor, innocent, and ambigu­ous transgressions often receive inordinate attention. Many people with strong claims involving widespread injustices do not seek relief or can’t prove their cases, while people with weak or dubious claims can extract undeserved settlements from risk-averse businesses. Worst of all, individual civil rights claims are less and less likely to correct the nation’s worst injustices.

Can we do better? Yes, we can. But only if both liberals and conserva­tives are willing to put aside dogma and pet obsessions and focus on practi­cal solutions and concrete results. Liberals should abandon the fantasy that civil rights litigation can guarantee workplace meritocracy. Some people will lose their jobs, be denied job opportunities or promotions, or suffer poor treatment at the hands of mean-spirited, petty, or biased employers and co­workers. Most of these abuses will go undetected or will be impossible to prove. The best the law can do is to insure that inevitable labor market injustices don’t typically fall on the same types of people. A reasonable anti­discrimination law should require employers to make an honest effort to root out discriminatory practices, but it is probably impossible for any employer to eliminate discrimination completely. Statistical evidence can detect pat­terns of discriminatory conduct, even when individual cases are hard to prove, and sensitive reforms can reduce unjustified imbalances. Laws that encouraged a holistic approach, providing incentives for employers to reex­amine their hiring, promotion, and retention practices, would do more good than the threat of civil rights litigation. Perhaps employers who meet realis­tic diversity benchmarks could enjoy immunity from some types of discrimi­nation lawsuits. Or public policy could encourage progress through some combination of fines and tax credits.

At this point many conservatives will begin to bleat about their pet obsession: quotas. Won’t any reference to numerical benchmarks just en­courage employers to ignore merit and simply hire to meet the benchmark? That depends. If the benchmarks were realistic and fair, an employer would be able to meet them simply by eliminating discrimination. Of course, in theory some employers may prefer to allow discrimination to continue and offset it with unwarranted preferences for members of the disfavored groups. But for most this would be a silly decision: working to rid the workplace of discrimination would yield a better workforce, whereas offsetting its effects with preferences would compound the damage to workforce quality done by the initial discrimination. That said, we should leave the choice to the em­ployer. An employer who is unable or unwilling to reduce discrimination should be free to offset the social damage done through other policies. This may sound perverse, but there may be sound reasons for an employer to make this decision. The workplace policies most likely to eliminate bias may be bad for other reasons. For instance, rigid personnel rules enforced by a centralized bureaucracy might reduce discrimination at Wal-Mart, but they would also eliminate hands-on flexibility and sensitivity to qualities that can’t be captured by objective qualifications. Monitoring of hiring and promotion decisions may be too costly or cumbersome. Employers are likely to resist and attempt to circumvent such requirements, and if they are forced to adopt inefficient personnel policies, the entire business will suffer. Allowing employers to keep policies that are unavoidably vulnerable to dis­crimination or have inadvertent discriminatory effects and to offset those effects with other policies may be the best solution for all concerned. I sus­pect businesses would prefer such a compromise to the omnipresent threat of litigation and the prospect of an injunction prohibiting policies that generally work well.

In a sense, my stereotypical liberal and conservative would object to this type of solution for the same reason—it sacrifices individual justice for social justice. The liberal thinks that if we detect, punish, and reverse every instance of individual discrimination, social justice will necessarily follow; hence, she develops ever more elaborate theories of discriminatory con­duct—unconscious bias is the latest in a long line—and presses for relaxed evidentiary burdens for plaintiffs. But decades of civil rights jurisprudence suggest that litigation will never root out all of the discriminatory conduct that distorts our labor market, and libertarians and conservatives make a per­suasive case that the costs of ever more intensive and intrusive regulation of workplaces will outweigh the benefits.

The conservative typically cares only about the individual injustices that can be definitively proven. But she opposes the detailed scrutiny and micromanaging that would be required to guarantee comprehensive individ­ual justice. Given the unavoidable spottiness of civil rights enforcement, this position amounts to acquiescence: for the most part the conservative is willing to allow discrimination to continue and let the chips lie where they fall. Libertarians like Richard Epstein, who would allow discrimination in all of its forms[76]—invidious bias as well as affirmative action—are at least more consistent and more honest on this point than are the many conserva­tives who are willing to tolerate discrimination as long as it doesn’t take the form of affirmative action.

Anti-discrimination rights serve a vital but limited purpose. The limita­tions of the anti-discrimination approach to social justice become more and more pronounced as the labor market becomes more volatile and job qualifi­cations become more subjective. This leads conservatives prematurely to celebrate the end of invidious discrimination and leads liberals to look for hidden bias and incognito bigots with all the more fervor and anxiety. Be­cause both of these reactions share a common obsession with individual big­otry, neither offers a viable way forward. Ultimately, racism and sexism are social evils whose effects can be seen in collective disadvantage. To con­front such collective injuries, we will have to suspend a once-appropriate emphasis on individual injury in favor of a renewed focus on collective in­justice. In order to fulfill the practical promise of civil rights, we must ac­knowledge their limitations; in order to honor the achievement of the civil rights movement, we must move beyond the tactics and ideas of the past.

---

[asterisk]George E. Osborne Professor of Law, Stanford Law School. This article is derived from excerpts from the upcoming book RIGHTS GONE WRONG, to be published in 2011 by Farrar, Straus and Giroux.[/asterisk]
[1] U.S. Bureau Of Labor Statistics, U.S. Dep’tof Labor, Highlights Of Women’s Earnings In 2008, at 1 (2009), available at www.bls.gov/cps/cpswom2008.pdf.
[2] See Erik Eckholm, Plight Deepens for Black Men, Studies Warn, N.Y. Times, Mar. 20, 2006, at A1.
[3] Kevin M. Clermont & Stewart J. Schwab, How Employment Discrimination Plaintiffs Fare in Federal Court, 1 J. Empirical Legal Stud. 429, 429, 433 (2004).
[4] 603 F.3d 571 (9th Cir. 2010).
[5] Walmart, Corporate Facts: Walmart by the Numbers 1 (2010), available at http://walmart.stores.com/download/2230.pdf.
[6] Nelson Lichtenstein, The Retail Revolution 117 (2010).
[7] See id. at 118.
[8] See West v. Wal-Mart, Inc., 264 F. Supp. 158, 164 (W.D. Ark. 1967).
[9] Harold Meyerson, In Wal-Mart’s Image, AM. Prospect, Sept. 11, 2009, at 33, available at http://www.prospect.org/cs/articles?article=in_walmarts_image.
[10] Dan Frosch, Immigrants Claim Wal-Mart Fired Them to Provide Jobs for Local Re­sidents, N.Y. Times, Feb. 9, 2010, at A1.
[11] Id.
[12] Meyerson, supra note 9.
[13] Wal-Mart Watch, Low Wages, Always 2 (2009), available at http://action.walmartwatch.com/page//Fact Sheets/Low Wages FACT SHEET _Feb_09.pdf.
[14] Press Release, U.S. Equal Empl’t Opportunity Comm’n, Wal-Mart to Pay More Than $11.7 Million to Settle EEOC Sex Discrimination Suit (Mar. 1, 2010), available at http://www.eeoc.gov/eeoc/newsroom/release/3-1-10.cfm.
[15] Plaintiffs’ Third Amended Complaint at 9, Dukes v. Wal-Mart Stores, Inc., 222 F.R.D. 137 (N.D. Cal. 2004) (No. C-01-2252 MJJ), 2002 WL 33645690.
[16] See id. at 10.
[17] See Reed Abelson, 6 Women Sue Wal-Mart, Charging Job and Promotion Bias, N.Y. Times, June 20, 2001, at C1.
[18] Plaintiffs’ Third Amended Complaint at 10, Dukes, 222 F.R.D. 137 (No. C-01-2252 MJJ).
[19] Dukes, 222 F.R.D. at 142 (order granting in part and denying in part motion for class certification).
[20] See id.
[21] Wal-Mart Stores, Inc., 2009 Annual Report 47 (2009).
[22] Plaintiffs’ Third Amended Complaint at 7–9, Dukes, 222 F.R.D. 137 (No. C-01-2252 MJJ).
[23] Id. at 6.
[24] See id. at 5–7.
[25] Dukes, 222 F.R.D. at 150.
[26] Declaration of William T. Bielby, Ph.D., in Support of Plaintiffs’ Motion for Class Cer­tification at 22, 25, Dukes, 222 F.R.D. 137 (No. C-01-2252 MJJ).
[27] Abelson, supra note 17.
[28] See Dukes v. Wal-Mart Stores, Inc., 603 F.3d 571, 610 (9th Cir. 2010) (en banc).
[29] Steven Malanga, The Tort Plague Hits Wal-Mart: A Federal Judge Dignifies a Flimsy Claim, 14 CITY J. 14 (2004).
[30] See Dukes v. Wal-Mart Stores, Inc., 222 F.R.D. 137, 146–49 (N.D. Cal. 2004); see, e.g., Declaration of William T. Bielby, supra note 26, at 22.
[31] See Dukes v. Wal-Mart, Inc., 509 F.3d 1168, 1175, 1178 (9th Cir. 2007).
[32] Malanga, supra note 29.
[33] Plaintiffs’ Third Amended Complaint at 2, Dukes, 222 F.R.D 137 (No. C-01-2252 MJJ).
[34] Id. at 6–7.
[35] Id. at 7.
[36] Dukes, 509 F.3d at 1190 n.16.
[37] See Dukes, 222 F.R.D. at 179–83 (No. C-01-2252 MJJ).
[38] See Petition for Rehearing En Banc at 13–17, Dukes v. Wal-Mart Stores, Inc., 603 F.3d 571 (9th Cir. 2010) (Nos. 04-16688, 04-16720), 2007 WL 1420550.
[39] Dukes, 603 F.3d at 624 n.49.
[40] Mark Fischer, Dukes v. Wal-Mart 18 (2007), available at http://works.bepress.com/cgi/viewcontent.cgi?article=1000&context=mark_fischer (unpublished manuscript, on file with the Harvard Law School Library).
[41] Dukes, 509 F.3d at 111 (Kleinfeld, J., dissenting).
[42] Opening Brief for Appellee-Respondents and Cross-Appellees at 9–17, Dukes, 603 F.3d 571 (No. 04-16688).
[43] Dukes v. Wal-Mart Stores, Inc., 222 F.R.D. 137, 174 (N.D. Cal. 2004).
[44] See Douglas Martin, The Law; The Rise and Fall of the Class-Action Lawsuit, N.Y. Times, Jan. 8, 1988, at B7.
[45] See John J. Donohue III, Employment Discrimination Law in Perspective: Three Con­cepts of Equality, 92 Mich. L. Rev. 2583, 2584 (1994).
[46] See Robert Belton, A Comparative Review of Public and Private Enforcement of Title VII of the Civil Rights Act of 1964, 31 Vand. L. Rev. 905, 914–16 (1978).
[47] See id. at 907.
[48] Martin, supra note 44.
[49] 426 U.S. 229 (1976).
[50] See, e.g., Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 657 (1989).
[51] Lichtenstein, supra note 6, at 131.
[52] McDonnell Douglas Corp. v. Green, 411 U.S. 792, 794–96 (1973).
[53] Id. at 794, 796.
[54] Id. at 802.
[55] See id. at 802–03.
[56] Id. at 804.
[57] See Plaintiffs’ Third Amended Complaint at 22, Dukes v. Wal-Mart Stores, Inc., 222 F.R.D. 137 (N.D. Cal. 2004) (No. C-01-2252 MJJ), 2002 WL 33645690.
[58] Dukes v. Wal-Mart, 509 F.3d 1168, 1199 (9th Cir. 2007) (Kleinfeld, J., dissenting).
[59] Id.
[60] Baylie v. Fed. Reserve Bank, 476 F.3d 522, 524 (7th Cir. 2007).
[61] U.S. Bureau Of Labor Statistics, supra note 1, at 1.
[62] Dukes, 509 F.3d at 1199 (Kleinfeld, J., dissenting).
[63] See generally Charles J. Muhl, The Employment-At-Will Doctrine: Three Major Excep­tions, Monthly Lab. Rev., Jan. 2001, at 3.
[64] Richard Thaler, Toward a Positive Theory of Consumer Choice, 1 J. Econ. Behav. & ORG. 39, 44 (1980).
[65] Mark Kelman, Consumption Theory, Production Theory, and Ideology in the Coase Theorem, 52 S. Cal. L. Rev. 669, 678–79 (1979).
[66] Laura Beth Nielsen Et Al., Contesting Workplace Discrimination In Court: Characteristics And Outcomes Of Federal Employment Discrimination Litigation 1987–2003, at 6 (2008).
[67] John J. Donohue III, The Changing Nature of Employment Discrimination Litigation, 43 Stan. L. Rev. 983, 1024, 1027 (1991).
[68] Brief for 65 Leading American Businesses as Amici Curiae, Grutter v. Bollinger, 539 U.S. 306 (2003) (Nos. 02-241, 02-516), 2003 WL 399056.
[69] Id. at *2.
[70] See Mike Swift, Five Silicon Valley Companies Fought Release of Employment Data, and Won, San Jose Mercury News (Feb. 14, 2010), http://www.siliconvalley.com/news/ ci_14382477 (on file with the Harvard Law School Library).
[71] Id.
[72] Mike Swift, Blacks, Latinos and Women Lose Ground at Silicon Valley Tech Compa­nies, San Jose Mercury News, (Feb. 13, 2010), http://www.siliconvalley.com/news/ci_14383730?nclick_check=1 (on file with the Harvard Law School Library).
[73] Id.
[74] Id.
[75] Paul Burstein, Discrimination, Jobs, And Politics: The Struggle For Equal Em­ployment Opportunity In The United States Since The New Deal 28 (1985).
[76] See generally Richard A. Epstein, Forbidden Grounds: The Case Against Em­ployment Discrimination Laws (2d ed. 1995).