George Stephanopoulos turned up at the Supreme Court last week, sitting next to Joel Klein, the deputy White House counsel. Their joint appearance seemed to illustrate the administration's anxiety about the case, Adarand v. Pena, in which the Court is being asked to strike down racial preferences in the construction industry that have been endorsed by every president since Nixon. But Klein assured me afterward that Stephanopoulos, who had never seen a Supreme Court argument before, had come along purely out of curiosity. He picked a good day. Drew Days, the solicitor general, was defending the constitutionality of federal set-asides for minorities and women; and he was in a very delicate spot.
In 1980, as Jimmy Carter's assistant attorney general for civil rights, Days had stood before the justices in the Fullilove case and defended a similar program, which set aside 10 percent of federal contracts under the Public Works Employment Act for "Negroes, Spanish-speaking, Orientals, Indians, Eskimos and Aleuts." But in 1987 Days dramatically recanted, and he published an article in the Yale Law Journal confessing his doubts. "It may appear to some either that I have experienced a profound conversion or that I have now resolved to acknowledge my own hypocrisy," Days wrote, joking that the headlines about his article might read: "respected civil rights attorney repudiates minority business enterprise set-asides."
Now Days has repudiated his repudiation. In his brief for the United States in Adarand, he resurrects some of the arguments his article rejected, and defends some of the policies his article attacked. Of course, Days was wearing different hats as an advocate and a scholar; and his hands were tied as solicitor general; and the racial preferences in the Fullilove and Adarand cases are not identical. Nevertheless, a comparison of Days's dutiful brief with his more candid article shows how the apparatus of congressional set- asides continues to rest on fiction.
In his Adarand brief, Days declares that "Congress's findings that disadvantage is highly correlated with race were based on extensive evidence of racial discrimination affecting government contracting." But the central accomplishment of his 1987 article had been to prove the opposite: "One can only marvel at the fact that the minority set-aside provision was enacted into law without hearings or committee reports, and with only token opposition." Days concluded that Congress had failed to establish, and the Supreme Court had failed to demand, that the 10 percent set-aside in Fullilove was "responsive to findings of racial discrimination" and designed to end after achieving its remedial objective. "This is an indefensible state of affairs," Days insisted. "It ought to stop."
Days's article also highlighted the interest group politics that prompted Congress to single out certain minority groups for special treatment. The current version of the congressional preference declares that the following groups should be presumed to be "socially disadvantaged": "Black Americans; Hispanic Americans; Native Americans (American Indians, Eskimos, Aleuts or Native Hawaiians); Asian Pacific Americans (persons with origins from Japan, China, the Philippines, Vietnam, Korea, Samoa, Guam, U.S. Trust Territory of the Pacific Islands, Northern Mariana Islands, Laos, Cambodia or Taiwan)." Women were added to the list in 1987. And the Small Business Administration, on authority delegated by Congress, has extended the preference to businessmen from Burma, Singapore, Laos, Republic of the Marshall Islands, Federated States of Micronesia, Fiji, Kiribati, Sri Lanka and Bhutan. "What does it mean," Days asked in his article, "to say that there is evidence of discrimination against minorities'? One of the major flaws in the set-aside upheld in Fullilove, one that subsequent federal, state and local programs have replicated, is that the record did not explain why six racial groups were selected to receive the 10 percent preference." The remedial objectives that convinced Congress to favor Eskimos and Aleuts may be even more of a fiction when applied in a state with a relatively small minority population. " It requires far more," Days concluded, "to justify the inclusion of all or most of these groups in any state or local government set-aside program."
At the argument last week, Justice Stephen Breyer declared: "I assume that, as a matter of fact, it is rational to assume that black people have suffered from prejudice." But is it equally rational, as a matter of fact, to assume that all Micronesians, Northern Mariana Islanders and women have suffered from prejudice? Perhaps the presumption of social disadvantage for groups other than African Americans is not really rational after all; for Congress, as Days illustrated, lazily enshrined the presumption in the Surface Transportation Act of 1982 without considering evidence of discrimination against many of the favored subgroups.
Is there a meaningful difference between the 10 percent set-aside that Days criticized in Fullilove and the more subtle preference that he is now defending in Adarand? The beneficiaries of the current program are called " socially and economically disadvantaged business enterprises" rather than " minority business enterprises," and the preference is called a "rebuttable presumption" rather than a set-aside. To trigger the preference, a subcontractor has to show that at least 51 percent of its stock is owned by individuals who have experienced "racial or ethnic prejudice or cultural bias. " But members of the preferred gender racial groups are "presumed" to be socially disadvantaged, and they get the benefit of the preference without having to prove that they have actually suffered from prejudice. In theory, an Appalachian farmer or a Hasidic Jew can try to prove that he has been victimized as well; but, unlike the favored minorities, white men have to offer real evidence.
In his article, Days emphasizes the distinction between a "rebuttable presumption" and a rigid set-aside. But the difference is hard to detect. A racial presumption is a form of racial discrimination, as Judge Richard Posner argued in a recent opinion. Clearly, the state couldn't have a rebuttable presumption that black subcontractors should be barred from highway projects. Congress and the states, Posner notes, are conferring a valuable economic benefit--access to the presumption of disadvantage--by pronouncing entire groups to be victims of discrimination. This neatly defies the Supreme Court's warning that generalizations about racial groups are almost never enough to justify racial classifications, that only empirical evidence of discrimination will do.
At the oral argument, Days undercut his own distinction. He revealed that the presumption of economic disadvantage is rebuttable, but the presumption of social disadvantage is not. A disgruntled white competitor who has lost out to a Hispanic firm can challenge the decision by proving that the Hispanic firm isn't, in fact, small or poor; but he can't offer evidence that the owner of the firm was actually head of the Porcellian club at Harvard. So Congress now presumes, as a matter of law, that all Hispanic, Samoan and Aleutian businessmen, and all businesswomen, have been victimized by prejudice. And it refuses to entertain any evidence to confirm or deny the presumption--which is really an ethnic and gender stereotype--in individual cases.
Days's empirical arguments as a scholar and an advocate are hard to reconcile; but his legal arguments are ultimately consistent. Despite his criticisms of Congress and the Supreme Court, Days stopped short, in his article, of claiming that federal set-asides were unconstitutional. And, on the constitutional point, he may be right. For all the tendentious claims of the conservative justices, it's hard to find in the text and history of the Fourteenth Amendment a principle authorizing judges to strike down racial preferences that are endorsed by Congress.
The Adarand case exposes the Achilles' heel of the conservative justices' affirmative action jurisprudence. They have never begun to explain why the Fourteenth Amendment, which says that no state shall deny any person the equal protection of the laws, should restrict the powers of Congress in any way. In 1954, without any analysis, Chief Justice Earl Warren discovered an " equal protection component" in the Fifth Amendment's due process clause; but this is the sort of result-oriented judging that conservatives are supposed to reject. And, while there is a more convincing way of deriving constitutional limits on Congress's power to draw racial classification, it doesn't quite justify striking down federal set-asides in the construction industry.
In his dissents in the civil rights cases of the 1880s, Justice John Harlan made clear that the central purpose of the Civil War amendments was to declare that the rights of national citizenship should not be withheld on the basis of race. But Harlan did not claim--as Justices Antonin Scalia and Clarence Thomas like to suggest--that all racial classifications are automatically unconstitutional. He claimed instead that the privileges and immunities of national citizenship can't be granted to one race and denied to another. Harlan's argument wouldn't justify striking down racial preferences for minority contractors, because the privileges and immunities of national citizenship are limited to public benefits that are extended to all citizens on equal terms. They don't include being hired by the government to build highways.
The constitutional question turns, in the end, on whether Congress should be able to legislate on the basis of dubious generalizations and stereotypes. Maybe so. Congress has broad authority, under the Fourteenth Amendment, to choose among competing notions of racial justice; and judges don't ordinarily require Congress to compile a convincing empirical record before passing laws. Still, there's something unsettling about Congress's increasingly casual resort to racial presumptions. When the Public Works Act was passed in 1977, it was the first federal law of general application containing an explicit racial classification. (Even the Freedman's Bureau Act of 1865, which set aside land and benefits for freed slaves, was amended to include white refugees, because the Reconstruction Republicans were committed to the principle of "no distinction according to color.") In the past two decades, however, federal race and gender preferences have proliferated, from television licenses to foreign aid provisions, culminating in Senator Howard Metzenbaum's Multiethnic Placement Act of 1994, which implicitly encourages adoption agencies to engage in racial discrimination when choosing prospective parents.
Judges are not very well suited to second-guess Congress's generalizations about race and gender, as appellate judge Clarence Thomas showed when he struck down the FCC's gender preferences for broadcast licenses. But, even as one recites the mantra of judicial restraint, the truth is that race and gender preferences for federal contracts are getting harder and harder to defend with anything less than dismay. After almost twenty years, any notion that the preferences are remedial, or compensatory, or even arguably linked to discrimination, is increasingly strained. In their current form, federal set-asides are not, perhaps, unconstitutional. They're merely indefensible. As Professor Days once argued, they ought to stop.