Epstein Says Overrule Teal
For a long time now, I have made no bones about my strong opposition to the employment discrimination laws. In my view, this limitation of freedom of contract has imposed heavy compliance costs on industry when the needed protection for workers comes through competition, not government regulation, over a wide range of employment practices that the Equal Employment Opportunity Commission does not understand and cannot oversee. The basic indictment contains these two key elements.
First, the 1971 judicial adoption from left field of the disparate-impact theory of liability in Griggs v. Duke Power has, in particular, placed huge limitations on the ability of employers to use various psychological tests that have a useful, if limited, role in making hiring and promotion decisions. Courts have taken a tough line because they are deeply suspicious that firms harbor latent racial bias or operate on the strength of "unconscious stereotypes" that prejudice minority workers. Ironically, this vigilance is counterproductive. Employers now do have a rational incentive to engage in "statistical discrimination" on grounds of race, precisely because they are cut off from tests that supply better and more individuated information. How do we know? Because whenever disparate impact is not an issue, the tests are in widespread use.
Second, the clear color-blind imperative of the original 1964 civil rights act, which made it illegal to fail or refuse to hire, or to discriminate against "any individual" (exclamation point), put a serious crimp in sensible programs of affirmative action that could do much to ease tensions in this important market. Remember, without employment discrimination law, private decisions on affirmative action would be a legitimate application of the principle of freedom of association.
As a political matter, no working politician will touch the employment discrimination laws in their current configuration. But without any political ado whatsoever, Congress or the Supreme Court could vastly improve their operation by making a small technical adjustment to one of the worst decisions under the 1964 act, Teal v. Connecticut, a Justice William Brennan special from 1982.
The Connecticut Office of Income Maintenance relied on a written test to determine which of its current employees were eligible for promotion. As is so often the case with these tests, the black workers scored on average significantly lower than the whites. Winnie Teal was one of those workers who fell beneath the cutoff score and claimed that the test denied her an opportunity to advance.
The state of Connecticut did not seek to defend the test as such, but instead asserted a "bottom line" defense by showing that the percentage of black workers who received promotions was higher than that of white applicants, notwithstanding the overall difference in test scores between the two groups. Brennan nixed this defense on the ground that it denied Teal "the opportunityto compete equally with white workers on the basis of job-related criteria" (his italics).
There was no showing, of course, that Teal had been excluded from the internal employee pool allowed to compete for promotions. Nor was there any indication that the overall results by race would have been different if the state had gone through the expensive procedure to use a validated test. The decision was a classic illustration of how the race-neutral language of the law was twisted into a one-directional test.
Bottom line is what counts
For these purposes, however, I want to assume that all affirmative action programs are legitimate. The question is: Why worry about the intermediate steps used in the selection process if that bottom line comes out right? Even if the employer had been taken over by overt racists, their unconscious biases could not have played any role in a decision process that on average was consciously tilted in favor of black applicants.
In Teal, we cannot justify using Griggs to ferret out latent discrimination. The affirmative action program that promoted a disproportionate number of black employees wholly dispels any such inference. Instead, a misdirected search for overall racial justice-a group conception-compromises the reliability of the selection process for all applicants, regardless of race. Allowing in the bottom-line defense thus achieves two goals. It leaves intact the affirmative action norm of such concern to the left, and it opens up the return to a rational testing system to any efficiency-minded firm willing to pay the price of proportionate representation, a "price" many are eager to pay.
The overruling of the Teal rule should not be hard to achieve. I am hard-pressed to see why there should be any legislative opposition to the proposal. Nor does it seem that there is any reason why the present Supreme Court should not revisit the precedent, which is in no way compelled by the original 1964 act, which did not seem to leave place for either disparate impact or affirmative action, or blocked by the 1991 Civil Rights Act, which was comfortable with both.
These observations should have some indirect bearing on the selection of a Supreme Court justice. Only on rare occasions, if ever, can sitting justices take it upon themselves to do a U-turn with the dominant trajectory of legal precedent. Proposing those shifts is a job for law professors and commentators, like myself. But a strong technical lawyer with solid institutional instincts can do much to clean up the interstices of the law to make it simpler, more efficient and, yes, more just. John Roberts fits within that mold, as someone who has worked with such distinction as an appellate advocate. Brennan's devotion to the ideals of the civil rights laws led him badly astray in this case. Now it's time for the entire court to rectify one of his worst mistakes.
Richard A. Epstein, an NLJ columnist, is a professor of law at the University of Chicago Law School and a senior fellow at the Hoover Institution.
Copyright 2005 ALM Properties, Inc.