The year was 1965. The ink on the Civil Rights Act of 1964 was hardly dry. Duke Power Company executives at the Dan River Steam Station near Draper, North Carolina, needed a new personnel policy. The old one had been simple: segregation. African Americans worked as laborers; only whites could do other jobs. But Title VII of the new Act forbade employers to discriminate by “race, color, religion, sex, or national origin.” On July 2, 1965—the day the Act took effect—Duke Power announced a new policy. New hires had to have a high school education and pass two standardized tests—unless, that is, they wanted to work as … laborers. If laborers wanted to transfer to other jobs, they either had to have a high school diploma or had to pass the tests.

At that time, 34 percent of whites in North Carolina had finished high school; only 12 percent of blacks had done the same. Federal agencies ran experiments using the tests; 58 percent of whites passed, but just 6 percent of blacks.

The company argued that tests would produce an educated, skilled workforce. It would also produce a new post-segregation plant that looked a lot like the old segregated one, but there was no evidence that this was its aim.

Black employees sued under Title VII of the new Civil Rights Act. In 1971, the Supreme Court ruled in their favor, finding that, even without intentionally using race, Duke Power had violated the Act. The use of job testing was permissible; but when the tests operated unequally, the Court held, the company had to show that the tests were “significantly related” to the jobs they were being used to fill. “[G]ood intent or absence of discriminatory intent,” Chief Justice Warren Burger wrote for eight members of the Court, “does not redeem employment procedures or testing mechanisms that act as ‘built-in headwinds’ for minority groups and are unrelated to measuring job capacity.”

Thus was born the legal idea of “disparate-impact” claims. That idea—an important tool in civil rights contexts in areas from employment to education to housing—was under fire Wednesday at the Supreme Court. Disparate impact turns a civil rights lawsuit from a hunt for villains into a discussion of inclusion. If a policy hits minority groups, or women, or older employees, harder than others, the employer can’t just say “bummer.” Instead, it must show that the new policy furthers a real interest.

Disparate-impact claims can only be made under a few civil rights statutes. In 1976, the Court refused to apply “disparate-impact” analysis to claims under the 14th Amendment’s equal-protection clause. That would “raise serious questions about, and perhaps invalidate, a whole range of tax, welfare, public service, regulatory, and licensing statutes that may be more burdensome to the poor and to the average black than to the more affluent white,” Justice Byron White wrote for seven justices. To prove a violation of equal protection, a plaintiff must show intentional discrimination.

But disparate-impact claims are permitted under a number of civil rights statutes. And Section 2 of the Voting Rights Act was amended in 1982 to include an “effects test,” which frees plaintiffs from proving racist intent if a voting practice makes it harder for a minority group to elect candidates they favor.    

Disparate-impact analysis is very important in claims under the Fair Housing Act. That statute, passed just days after the assassination of Martin Luther King Jr., makes it unlawful to “make unavailable or deny a dwelling to any person because of race, color, religion, sex, familial status, or national origin.” It applies to a wide range of policies, from zoning decisions to the selection of sites and subsidies for low-income housing. Courts and commentators have reasoned that in housing, in particular, the needs of society—and the intent of Congress—go far beyond simply compensating individuals for intentional discrimination. Years of official racism by government—federal, state, and local—has left behind a corrosive pattern of racially identifiable neighborhoods, and indeed whole cities and towns. Even changing population dynamics do not break these blocs down. (Think, for example, of Ferguson, Missouri, which was transformed from a "white" suburb in 1970 to a "black" one today, barely pausing at "integration." Local governments usually end up entrenching separation by making choices about roads, schools, and new developments. And intentional racism is hard to show in the complex decisions about where roads and developments are built.

As the backlash to civil rights laws built in the 1970s and 80s, however, so did the critique of disparate-impact cases. As a young Justice Department lawyer, John Roberts was passionately involved in trying to block the “effects test” of the Voting Rights Act, which he regarded as imposing quotas on politics. In a 2009 concurrence, Justice Antonin Scalia suggested that Title VII’s disparate-impact requirement is unconstitutional: “Title VII’s disparate-impact provisions place a racial thumb on the scales, often requiring employers to evaluate the racial outcomes of their policies, and to make decisions based on (because of) those racial outcomes,” he wrote. He added that “the war between disparate impact and equal protection will be waged sooner or later, and it behooves us to begin thinking about how—and on what terms—to make peace between them.”

At the Supreme Court Wednesday, the solicitor general of Texas tried to suggest that what Scalia called that “evil day” has arrived. The focus of Texas Department of Housing and Community Affairs v. The Inclusive Community Project, Inc. is a series of subsidized low-income housing developments in Dallas. The plaintiffs charged that the Department was concentrating them in predominantly minority areas—in effect, solidifying residential segregation of the races. Each application is judged by a complex mathematical formula, partially set by the federal government, without any overt racial content. But if the effect of the approvals is to “pack” minorities, then these decisions have a disparate impact, and the Department must justify its policies to a court. The project won below; when the Court granted review, plaintiffs were joined by the solicitor general's office, which is defending disparate impact as a tool of FHA enforcement.

Depending on your count, nine or eleven federal circuits have approved disparate-impact claims under the FHA; not one has rejected them. Remarkably, though, Texas insists that the statute is “unambiguous.” The unanimous contrary judgment of the appellate courts, in other words, is not a difference of interpretation but utter judicial incompetence. Texas even goes a step further, arguing that disparate-impact analysis “effectively compels state and private entities to engage in race-conscious decision-making to avoid legal liability.” Thus the Court should read the statute as not allowing disparate impact analysis in order to avoid violating the equal-protection clause.

That raises the long-time conservative insistence that disparate-impact analysis is unconstitutional in any context. At the oral argument Wednesday, Texas Solicitor General Scott Keller blew that dog whistle softly but clearly, saying that a victory for the project would be “the functional equivalent of a quota system.”

Chief Justice Roberts suggested that placing housing projects in affluent white neighborhoods was just another kind of “disparate impact,” and Justice Anthony Kennedy seemed troubled by that idea as well. Justice Samuel Alito seemed hostile to the government as well. Most interesting, however, was Justice Antonin Scalia. He noted that Congress had passed amendments in 1988 that seem to ratify the use of disparate-impact claims under the FHA. “[W]hy doesn't that kill your case?” he asked Keller. “I mean, when we look at a provision of law, we look at the entire provision of law, including later amendments. We try to make sense of the law as a whole.”

In this case, as in the Roberts Court’s previous school desegregation and Voting Rights Act decisions, two visions of the world collide. One view looks at contemporary America and sees a searing racial divide, as pervasive and dangerous, if not as overt, as the wall separating the races half a century ago. That view favors using government tools to dismantle that separation. That’s especially true in housing; residential segregation is the root of racial isolation in education, employment, and voting.

The other view believes that over the past half-century we have left the old racism behind, and the remaining incidents of racism are the work of a few bad apples. Chief Justice Roberts is the leading spokesman for the “past is past” camp. Any government notice of race—for inclusion or exclusion—makes him visibly furious. He breezily gutted a key portion of the Voting Rights Act with the assertion that “nearly 50 years later, things have changed dramatically"; in a case involving policies to diversify public schools, he asked, “How is that different from the separate but equal argument?” The central racial problem today, this view implies, is the government’s intervention to tilt the board toward minorities and against whites. (Witness Scalia’s sneer in remarks from the bench that the Voting Rights Act was not a landmark achievement but a “permanent racial entitlement.”)

The sunny, “look how far we’ve come” view seems particularly hard to justify in housing. Anyone who can look at American cities—their housing patterns, their employment figures, or their police policies—and see a new dawn of color-blindness is wearing glasses unavailable to me. A ruling cutting back the FHA would compare to the Court’s earlier mistakes in the context of schools and voting. An opinion destroying disparate impact altogether would be a disaster.