On Tuesday the Supreme Court will hear oral arguments in Dukes v. Wal-Mart, the largest employment discrimination class-action suit in American history. In 2001, Betty Dukes sued Wal-Mart for sex discrimination in a lawsuit filed on behalf of every woman who worked for the company since 1998—roughly 1.5 million women. The only question now before the court is "class certification": whether it makes sense for all 1.5 million women to sue together as a group. But even though the legal issue is a narrow one, Dukes v. Wal-Mart may be the most important case the court will decide this term. At stake is the continuing viability of one of the most important means of enforcing laws against discrimination.
The law allows a group of people to join together and sue when their claims involve "common questions of law and fact"—basically, when they have all suffered more or less the same injury. You may have been a party to such a collective lawsuit. If you bought one of the first iPods, you were probably a plaintiff in a suit against Apple for using a plastic that scratched easily, making the screens unsightly and hard to read. Because everyone who bought one of those iPods suffered the same injury, all of their claims could be lumped together in one big lawsuit: If several representative iPods were sufficiently scratch-prone to violate the (deep breath) implied warranty of merchantability and fitness for intended purpose, then they all were.
Class actions serve many important purposes. They save the courts from hearing basically the same case over and over. By pooling a lot of small injuries together, they make it worthwhile to sue for injuries that aren't worth the cost of litigation individually. But class actions don't make sense if the claims are significantly different from each other. Imagine a lawsuit where some plaintiffs have scratched iPod screens, some have batteries that won't hold a charge, some have faulty hard drives, and some are claiming that Apple encouraged iPod buyers to violate copyright laws. Here there are too many different issues to justify lumping all of the claims together. The claims involve different facts and legal theories and therefore should be litigated in different lawsuits.
The question before the court in Dukes is whether the women suing Wal-Mart have enough in common to justify a collective lawsuit. According to the plaintiffs, a common culture of sexism led to a pattern and practice of discrimination against women working at Wal-Mart stores nationwide. In their complaint they claim that, at the time the suit was filed, almost three-fourths of Wal-Mart's hourly wage sales employees were women; by contrast, only about one-third of its managers were. The plaintiffs have gathered declarations by 120 women who say that they were victims of sex discrimination at Wal-Mart. And they've enlisted sociologist William Bielby, who claims that because Wal-Mart gives local store managers discretion to use subjective criteria such as "teamwork, ethics, integrity, and the ability to get along with others" in evaluating employees for raises and promotions, its decisions are especially vulnerable to the influence of sexism and sex stereotypes.
But it's just this discretion that Wal-Mart says undermines the plaintiffs' case. Because Wal-Mart gives managers at the store level almost complete discretion to make personnel decisions, there's nothing that connects the decisions of one Wal-Mart manager to those of another. Even if a lot of Wal-Mart employees suffered sex discrimination, if each individual store has its own distinctive practices, there are no questions of law and fact common to all Wal-Mart employees that would justify joining the claims as a class action. Judge Alex Kozinski made the point with characteristic aplomb in his dissent to the 9th Circuit's affirmation of class certification: "The half-million members of the majority's approved class held a multitude of jobs … in 3,400 stores, sprinkled across 50 states, with a kaleidoscope of supervisors (male and female), subject to a variety of regional policies that all differed depending on each class member's job, location and period of employment. … They have little in common but their sex and this lawsuit."
Wal-Mart doesn't deny that some of these things might have actually happened. But the company does deny that they were related to each other. And if any individual Wal-Mart manager discriminated, the individual victim can sue in an individual lawsuit.
The core of the plaintiffs' discrimination case is statistical. Wal-Mart draws most of its managers from its hourly wage employees, of which 72 percent are women. But at the time the lawsuit was filed only one-third of Wal-Mart's managers were women, and according to the plaintiffs, "even this figure overstates the proportion of female managers [because it] … includes traditionally 'female' positions, such as assistant managers … the lowest level of managers. … Women comprise less than 10% of all Store Managers and approximately 4% of all District Managers." 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."
Are these statistics enough to establish a firm-wide pattern of discrimination? They might be. A plaintiff can prove a "pattern and practice" of employment discrimination by showing an unexplained statistical disparity between the composition of the employer's workforce and the composition of the qualified labor pool from which the employer draws its workforce. A nondiscriminating employer would wind up with a workforce that reflects the demographics of the relevant labor market, absent some unusual circumstances. Of course, that's just where the analysis starts: If fewer women are qualified for or interested in the jobs in question, the imbalance wouldn't suggest discrimination. Sophisticated statistical methods can take account of the other factors that might contribute to disparities, like differences in qualifications, interest, or availability.
Once such legitimate factors have been taken into account, it's reasonable to conclude that any remaining disparity is the result of discrimination. And here Wal-Mart's size works in the plaintiffs' favor. Unexplained disparities in small workforces may be attributable to chance. But in larger workforces like Wal-Mart's, random accidents will cancel each other out, making statistical evidence more reliable.
In the classic pattern-and-practice case, the plaintiff doesn't have to point to a corporate policy that joins all of the individual cases of discrimination—it's enough to show, using statistics, that such a pattern exists. The employer as an entity is responsible for the decisions of its managers: Even if each manager exercises his or her own independent judgment, he still acts on behalf of the employer whenever he hires, fires, promotes, or disciplines an employee. So, if a plaintiff can show—using a combination of statistics and a sample of representative cases—that an employer discriminated repeatedly, it is considered sufficient to establish liability. That's what the Dukes plaintiffs hope to show at trial.
But while statistics might tell us that a lot of women have been discriminated against, they can't tell us which ones. A lot of unexplained decisions that harm women suggests a pattern of discrimination. But any single unexplained decision that goes against an individual woman might be due to chance. And it's not practical for the court to hear evidence about 1.5 million cases. As a result, if the case goes ahead and the plaintiffs win, the trial plan calls for any damages to be determined for the entire class and then apportioned to the individual women according to a mathematical formula. As some of the judges who opposed class certification at earlier stages of the litigation complained, this guarantees that women with strong claims will effectively have to share their damages with those who have weak claims. For individuals, the class action offers only rough justice.
Still, rough justice is better than no justice at all. Most of the women in the Dukes class action would never bring an individual lawsuit. Proving individual discrimination is hard. Even if the statistics prove that Wal-Mart discriminated against a lot of women, very few would be able to prove that they were one of them. And even for those who could prove it, the damages they would be entitled to often aren't worth suing over.
More important, civil rights have always been as much about social justice as individual justice. And social justice requires that employers who discriminate be held accountable so they won't keep discriminating. The deterrent effect of a large class-action lawsuit may be the only thing that will encourage employers to root out discrimination among their managers: Tellingly, Wal-Mart has made an impressive effort to treat women more equitably in the 10 years since the case was filed. That's why large pattern-and-practice civil rights cases have been one of the most important types of civil rights litigation since the passage of the Civil Rights Act of 1964. If large civil rights class actions like the one in Dukes can't get off the ground, then for the most part the government will be the only entity able to sue for firm-wide patterns of discrimination.
Ultimately what's at stake in Dukes v. Wal-Mart is whether class-action lawsuits will continue to be a way to address pervasive discrimination, or whether America's battle against prejudice will have to be fought on a case-by-case basis. This has been the central question in many of the most important civil rights disputes of the last 30 years. And in almost every case since the early 1980s, the Supreme Court has come down on the side of individualism against social justice. That statistic doesn't look so good for the women suing Wal-Mart—or for equal opportunity generally.