"Detailed discussions of 'disparate impact'... are not big ratings-grabbers," Howard Kurtz of the Washington Post observed during the Senate Judiciary hearings on Sonia Sotomayor.

Indeed not. That's why most of the media understandably took a minimalist approach to explaining the issue at the core of Judge Sotomayor's most controversial decision, in the New Haven firefighter case, and the Supreme Court's 5-4 ruling on June 29 reversing it.
Disparate-impact lawsuits -- typically brought by blacks and/or Hispanics who challenge as discriminatory employers' use of objective tests on which those minorities do poorly -- are unavoidably complex.

But perhaps the educated public might find the complexities worth understanding if the media explained how such lawsuits can -- depending on how the extraordinarily vague rules codified by Congress in 1991 are interpreted -- either open opportunities to qualified minorities, or foster discrimination against better-qualified whites, or some of both.

I explored some of this terrain in my May 31 column. But it's worth revisiting because the firefighter decision, Ricci v. DeStefano, both clarified the law and pointed to the need for further clarification.

The majority made it somewhat easier for employers to defend against disparate-impact suits -- and harder to justify discriminating against whites based on fear of such suits.

But that will change, the four liberal dissenters made clear, if they can get one more vote to impose broader disparate-impact liability -- which, as the majority explained, would be (and sometimes has been) an engine of quotas and discrimination.

This is not to deny that disparate-impact lawsuits can play a valuable role in protecting minorities against tests and other selection criteria that disproportionately exclude minorities for lacking skills of little relevance to the positions they seek. One example is using IQ scores to screen applicants for jobs that don't involve difficult cognitive tasks. But the tests in use by most employers now are carefully designed and reasonably related to the requirements of the job.

The firefighter case was not itself a disparate-impact suit. But it dramatized how even the mere threat of such suits can lead to overt racial discrimination against whites, especially by employers who -- as in New Haven -- may already have political motives to discriminate.

The justices held that the city had intentionally and illegally discriminated against the 18 plaintiffs (17 white, one Hispanic) when it denied them the promotions that they had earned by doing well on qualifying tests designed to measure the skills and leadership qualities needed to be a good fire lieutenant or captain.

Justice Anthony Kennedy's majority opinion also held that the city's purported justification for dumping the test results -- fear that promoting the high-scoring whites would bring a disparate-impact suit by blacks who did poorly -- was not good enough, because any disparate-impact suit by blacks would (or at least should) have been rejected.

The difference between the facts of the original disparate-impact case, Griggs v. Duke Power Co., in 1971, and those of the firefighter case illustrate how disparate-impact liability can either "remove artificial, arbitrary and unnecessary barriers to employment when the barriers operate invidiously to discriminate" against minorities, as Griggs held, or lead to "a de facto quota system" discriminating against whites, as Ricci warned.

Griggs created the disparate-impact rules as an interpretation of the 1964 Civil Rights Act. As codified by Congress in 1991, those rules are that any qualifying test with a disparate impact on blacks or Hispanics -- leading to scores substantially lower on average than those of whites -- is presumed invalid unless the employer can prove it to be "job related" and "consistent with business necessity." And even if the employer can prove that, the plaintiffs still win if they can demonstrate the availability of an equally valid test on which minorities would do better. Lack of intent to discriminate is no defense.

It was a bit of a stretch to derive these rules from the 1964 Civil Rights Act. It prohibits intentional discrimination against whites as well as against blacks; specifies that employers may use "professionally developed ability tests"; and says that they have no obligation to seek racial balance or grant minorities "preferential treatment."

But the facts of Griggs show why the Court stretched, by an 8-0 vote. The company had overtly excluded blacks from all but the lowliest jobs until the 1964 Act had made such discrimination illegal. And even after the 1964 Act, the company continued to exclude almost all blacks from the better jobs in practice by requiring both a high school diploma and a passing score on a general intelligence test.

Only 12 percent of black males in the state (North Carolina) had finished high school. And the pass rate for blacks on the intelligence test was only 6 percent. The Court attributed this to "inferior education in segregated schools." It also held that the company's own past experience suggested that neither a high-school diploma nor a passing score was "a reasonable measure" of ability to do the specific jobs in question.

In a 1975 decision, the Court held that employers must prove a "manifest relationship" between the test and the job to justify a disparate impact.

But Griggs had made clear that "testing or measuring procedures... obviously... are useful" -- as long as reasonably job-related -- and that "the less qualified [need not] be preferred over the better-qualified."

At the time of Griggs and for some years thereafter, it was hoped that the large white-black gaps in high-school attendance, academic achievement and test scores would diminish over time, which would in turn lead to more proportionate hiring and promotions of blacks. But the academic and test score gaps -- including on tests of ability to learn job-related skills -- stopped closing more than 20 years ago. They remain very large today.

The result has been that objective tests of job-related skills very often produce racially disparate impacts, making it hard for many blacks to compete for good jobs and leading to racially unbalanced workforces.

Some judges have responded by placing a virtually insuperable burden of proof on employers to prove that their tests -- even the best of which are inherently imperfect and imprecise -- are not just reasonably job-related, but flawless. This in turn has spurred many employers to drop objective tests and use subjective evaluations that facilitate reaching whatever racial numbers seem likely to shield them from liability.

In the Ricci case, Sotomayor and her colleagues disregarded the fact that the city spent $100,000 to buy professionally designed written and oral tests for promotion to fire lieutenant and captain that were carefully designed to be race-neutral and to measure knowledge, skills and leadership qualities that fire officers should have. Justice Kennedy detailed all this on pp. 28-33 of the majority opinion.

In addition, as Justice Samuel Alito explained in a concurrence, there was strong evidence that "the city's purported reason for scrapping its test -- concern about disparate-impact liability -- was a pretext and that the city's real reason was illegitimate, namely, the desire to placate a politically important racial constituency."

For the four liberals' arguments to the contrary, see pp. 26-39 of Justice Ruth Bader Ginsburg's dissent.

Meanwhile, in another concurrence, Justice Antonin Scalia suggested that even as clarified by Ricci, the disparate-impact rules "place a racial thumb on the scales" that can lead to discrimination against whites.

And that, asserted Scalia, puts disparate impact law at "war" with precedents interpreting the 14th Amendment's equal protection clause as barring governmental discrimination on the basis of race. Richard Primus, of the University of Michigan Law School, explored this tension in depth in a 2003 Harvard Law Review article, "Equal Protection and Disparate Impact: Round Three."

Scalia hinted at one point that the disparate-impact provision may be unconstitutional unless narrowed to be "simply an evidentiary tool used to identify genuine, intentional discrimination." That would go too far, in my view, especially given the lack of consensus even among conservatives that the Constitution bans all racial discrimination against whites.

But Scalia also hinted that he might be satisfied with a disparate-impact regime in which an employer could avoid liability by "proving that its motives were pure and its actions reasonable." That sounds about right.

This disparate-impact regime has not much affected the high-income job market because employers of white-collar workers require good grades at good colleges instead of their own objective tests.

Of course, highly disproportionate numbers of blacks don't have good grades from good colleges. But the nation's judges, who got started in their own careers by passing bar exams with a gigantic disparate impact on minorities, have not extended disparate-impact liability to people like themselves. Nor have many judges racially diversified their own overwhelmingly white and Asian law-clerk hires by looking to minority candidates with less than outstanding academic credentials.

The implicit assumption seems to be that, unlike lawyers, firefighters -- whose learned skills can mean the difference between life and death -- and other working-class types don't need to know much.