This week, as the U.S. Supreme Court took up the constitutionality of voluntary racial school integration, all eyes were on Justice Anthony Kennedy. In two cases, which involve the question of whether the Seattle and Louisville school districts can use race as a factor in deciding where students attend elementary and secondary school, Kennedy is seen as the swing vote on a court whose other justices are likely split on the issue, 4–4.

Kennedy, like most Americans, wants to see schools integrated by race but is uncomfortable with policies which deny individual students seats on the basis of skin color. Kennedy told the lawyer representing Seattle schools, “You’re characterizing each student by reason of the color of his or her skin. . . . And it seems to me that that should only be, if ever allowed, allowed as a last resort.”

To many liberals, like Justices David Souter and Ruth Bader Ginsburg, the notion that one wants a certain racial result, but is uneasy about using race to get there, is nonsensical. But there is another alternative—used in about forty school districts across the country—that produces significant amounts of racial integration without relying on race: socioeconomic school integration. Given the overlap in American society between class and race, policies that seek to avoid concentrations of poverty also often produce substantial racial integration.

In Cambridge, Massachusetts, for example, the school district achieves racial integration with a policy that seeks a rough balance of students by income, measured by eligibility for free or reduced price lunch (185 percent of the poverty line). The Cambridge policy holds that all schools should be within plus or minus ten percentage points of the district average for subsidized lunch. In the four years since it was adopted, the program has always produced both economic and racial diversity. But Cambridge allows race to be a determinative factor in student assignment should socioeconomic integration fail, in the future, to produce sufficient racial integration.

Souter and Ginsburg attacked the idea of using “proxies” for race, but Cambridge and other communities can point to a broad body of evidence going back forty years to suggest that socioeconomic integration is crucial to promoting academic achievement, whatever its impact on racial integration. Indeed, the research suggests that what makes for a difficult learning environment is not having “too many minorities,” but having concentrations of poverty.

Most importantly, the Cambridge model may be able meet Justice Kennedy’s standard: that race be used not as a first resort, but as a “last resort.” It zeroes in on the prime driver of academic achievement—the socioeconomic mix in a school—and uses race only when necessary to garner the important benefits of racial integration: increased tolerance and social cohesion.

Richard D. Kahlenberg is a Senior Fellow at The Century Foundation.