The war over unconscious bias

Wal-Mart and others are facing class actions for job discrimination. But the biggest problem isn't their policies, it's their managers' unwitting preferences. Can any company be immune?

By Roger Parloff, Fortune senior editor

(Fortune Magazine) -- Last February a federal appeals court panel in San Francisco decided, 2-1, to allow the largest class action employment discrimination case ever convened to go forward against Wal-Mart Stores. The class includes the more than two million women who have worked at any of the company's more than 4,000 retail stores nationwide since Dec. 26, 1998.

The case, known as Dukes v. Wal-Mart, accuses the retailer of discouraging the promotion of women store employees to managerial positions and of paying them less than men across all job positions. The suit seeks changes in the company's internal procedures, more than $1 billion in back pay, and punitive damages.

theodore_boutros.03.jpg
Theodore Boutrous Jr. is representing Wal-Mart in Dukes v. Wal-Mart.
brad_seligman.03.jpg
Brad Seligman, who heads a nonprofit called the Impact Fund, brought the class action.
Interactive graphic
10 years of powerful women
In honor of the 10th edition of Fortune's Most Powerful Women list, we're looking back at who made the top 3 each year.
See them all
Video
A gathering of powerful women
Each year, women from the top level of business convene for Fortune's Most Powerful Women Summit. Pattie Sellers gets the lowdown on the event.
Video
The dark horse CEO
Angela Braly, No. 4 on Fortune's Most Powerful Women list, was an unexpected pick to head U.S. healthcare giant Wellpoint.

Wal-Mart (Charts, Fortune 500) denies any wrongdoing and asserts that it has put "enormous resources in seeking to have a diverse workplace and to make sure that women and minorities are having the best possible opportunities to succeed," as its lead lawyer, Theodore Boutrous Jr. of Gibson Dunn & Crutcher, puts it. (Boutrous has represented Time Inc., which publishes Fortune.)

Wal-Mart has asked the full appeals court to reconsider the panel's approval of the class. ("Class certification" means the court believes that the claims by various employees share enough common elements to proceed as one combined suit rather than endless individual suits.) If Wal-Mart fails to win that, it will probably seek U.S. Supreme Court review.

Gender discrimination is always an incendiary accusation, and this case carries added emotional freight because Wal-Mart is Wal-Mart. It's not just the world's biggest private employer and the most admired company in this magazine's 2003 and 2004 surveys but also the most abhorred in other circles, particularly for its anti-union, penny-pinching labor practices.

But while the Dukes case does make some charges of intentional wrongdoing, it focuses mainly on three generic, almost abstract accusations that have become fixtures of nearly every contemporary employment discrimination dispute. These one-size-fits-all charges are less criticisms of Wal-Mart than of our society as a whole. It will be the rare large company that feels confident it could repel any of them.

Whether the defendant is Wal-Mart, Costco, Home Depot, FedEx, General Electric, MetLife, Merrill Lynch, Smith Barney, Morgan Stanley, Deloitte & Touche, or American Express - to name just a few that have been hit with such suits - the cases all stem from statistical evidence showing that the percentage of women or minority employees in portions of the workforce is markedly lower than in the available labor pool.

The percentage of women or minorities in lower-level positions often far exceeds the percentage of those who are ever promoted to managerial positions, and their share keeps shrinking the higher up the corporate hierarchy one goes. (Sound like any companies you know?) Experts will often also present statistical analyses purporting to show that women or minorities are, on average, paid less than men or whites.

Once the statistical disparities have been established, the plaintiffs go on to make a second generic accusation: that "unconscious" bias is rife at the company.

Supported by expert psychological testimony, the plaintiffs argue that managers in charge of promotion and pay decisions are unwittingly engaging in "spontaneous" and "automatic" stereotyping and "in-group favoritism" that results in the most desirable jobs at the company being filled by people who look like the incumbents, who are usually white males. (The companies' only defense is to challenge the admissibility of the expert testimony, which almost never works.)

Finally, the plaintiffs present the testimony of an expert organizational sociologist who explains that the company's promotion and pay procedures provide too much discretion to managers, allowing their unconscious biases to run rampant. (Know any companies where pay and promotion decisions have significant discretionary components?)

In case after case, these themes are debated by a regular cast of experts. Each side relies on a handful of statisticians, economists, psychologists, and sociologists, and the two teams go to battle in courtrooms around the country, much as the New England Patriots and Miami Dolphins play their scheduled rematches at least twice a year, sometimes at home and sometimes away.

Usually the contests end with the class being certified and the defendant then settling for an eight-digit sum rather than risk facing a jury and the enormous exposure that a class action carries.

No scam is being perpetrated. What we are seeing is the clash of two sharply opposed philosophies of how active a company must become in the face of a phenomenon that is endemic and appropriately controversial in our country: workforce numbers that, when analyzed in certain plausible ways - though not in others - show discrepancies between how men and women, or blacks and whites, or the disabled and abled, are paid and promoted.

Such statistical disparities are not illegal per se, according to the Supreme Court, nor do they create any duty on the employer's part to ensure the numbers improve. But as we'll see, in practice employers are under precisely such pressure.

As maybe they should be. Whether you think forcing employers to monitor and increase their diversity is good or bad probably depends on your intuition as to whether existing statistical disparities generally reveal discrimination or whether they instead reflect a complex stew of social, historical, and cultural legacies that no company can or should be expected to correct.

It also depends on how you feel about racial or gender quotas and preferences. Though such mechanisms are illegal, they will obviously be tempting to employers who want to avoid being hit with class-action employment discrimination lawsuits.

For there is only one sure-fire way to inoculate oneself against such suits, and that is to have workforce numbers that look good even when analyzed by a plaintiffs' expert. And the cheapest and fastest way to get those is to use quotas or preferences.