More Ricci Fallout…

The estimable Mickey Kaus, citing the estimable Walter Olson, worries about employers who will get sued if they ignore the results of tests that have a disparate impact … and sued if they don’t, but he goes on to argue that

… Justice Anthony Kennedy’s solution to this problem in the Ricci case — that a city can’t throw out a job test that winds up promoting whites and no blacks unless there is a “strong basis in evidence” that it would lose a subsequent discrimination lawsuit — seems an unsatisfying solution to the litigation vise.

What if a city, after reading the Ricci decision, decides there’s just a wee a bit less than a “strong” basis for thinking it will lose a discrimination case, and doesn’t throw out the test (lest SCOTUS smack it down)–and then loses the discrimination case anyway?

Alas, Kaus does not offer a solution of his own, nor does he seem worried now (though I suspect he will be later) about the liberal solution, which is to remove uncertainty by regarding all tests — indeed, all employment policies or practices — that have a disparate impact as by definition discriminatory by virtue of their disparate impact alone. As I noted two days ago, here, quoting Stuart Taylor, Sotomayor and other disparate impact mavens believe that New Haven’s refusal to promote the white (and one Hispanic) successful test-takers should have been upheld

based solely on the fact that highly disproportionate numbers of blacks had done badly on the exam and might file a “disparate-impact” lawsuit — regardless of whether the exam was valid or the lawsuit could succeed.

Fortunately, the solution to this dilemma is conceptually (if not politically) easy: demote disparate impact to its proper role, which is suggestive evidence of the possibility of disparate treatment, a possibility that can be successfully refuted by an employer’s production of credible evidence that the challenged test, policy, or procedure bears a reasonable relationship to the organization’s activities. (Of course, credible evidence that an employer adopted even such a reasonable test for a discriminatory purpose would also be barred as disparate treatment.)


In Wards Cove Packing Co. v. Atonio, 490 U.S. 642 (1989) the Supreme Court provided a compelling version of what disparate impact should mean, i.e., a version remarkably similar to the one I favored above.

As I discussed here, quoting from an EEOC legislative history,

Wards Cove reinterpreted the disparate impact method of proof, and held that an employer can avoid liability merely by showing a business justification for the practice causing a disparate impact, and that the plaintiff has the burden of proving a lack of a business justification.

Indeed, the Court’s definition of disparate impact was so sensible and reasonable that the Democratic Congress went ballistic and reversed it in the Civil Rights Act of 1991.

Say What? (4)

  1. Mike Bertolone July 7, 2009 at 7:49 am | | Reply

    “Indeed, the Court’s definition of disparate impact was so sensible and reasonable that the Democratic Congress went ballistic and reversed it in the Civil Rights Act of 1991.”

    The misnamed “Civil Rights” Act of 1991.

  2. Marcus July 7, 2009 at 7:07 pm | | Reply

    the problem with your solution is that it’s no less vague than kennedy’s, and basically “demotes” disparate impact law to small hurdle that can be easily cleared by a company clever enough to offer “credible evidence” that its practices are related to its activities (or purposes). i suspect the realm of “credible evidence.” aside from adding another vague and confusing layer to a already confused area of law, you’d basically turn one of our core 14th amendment principles on its head.

  3. David Nieporent July 8, 2009 at 3:08 pm | | Reply

    Your test might slightly mitigate the problem by making it easier for the employer to prevail, but I would suggest two alternatives:

    1) The first is the Wards Cove approach you mention, putting the burden of proof on the employee rather than the employer to prove that the test isn’t job-related.

    2) The second is more radical, and has no chance of being adopted except by the Supreme Court circumventing Congress: require an intent element. The issue is discrimination, not merely unequal outcomes. Unless the policy was adopted for the purpose of filtering out minorities, it’s not the government’s business. It just doesn’t make sense for the government to sit there and say, “Well, yes, an educated workforce might be nice, but it isn’t a necessity, so you can’t require college degrees, because not enough blacks have them.”

    Marcus is correct that any of these approaches would drastically cut back on disparate impact law, but that’s the point. Disparate impact law is an idea whose time has passed. It’s one thing to continue to ban facially-neutral policies that are clearly pretextual or applied discriminatorily; it’s another to ban legitimate policies just because the outcomes are unequal and they’re not perfect policies.

  4. John Rosenberg July 12, 2009 at 9:16 am | | Reply

    I certainly didn’t mean for my recommendation on how to mitigate the damage of disparate impact to preclude better ideas, and David Nieporent’s ideas — both of them, but especially the second one — are much better.

    But note that that better idea — no intent, no discrimination — was also at the core of my recommendation. It would require that evidence of disparate impact (statistical disparities, etc.) be treated as just that, evidence of possible discriminatory intent. The burden of proof issue is indeed important, often crucial, and I’m open on that, but it does not strike me as unreasonable to require an employer to offer a plausible, reasonable justification for any policy that has a severe disparate impact. I would set that bar low (reasonable, business-related, etc.) and certainly would not require proof of “business necessity.”

Say What?