Over at the American Prospect, Richard Kahlenberg has a fascinating look at what the Louisville school district has done in response to last year's Supreme Court decision in Meredith v. Jefferson County Board of Education, which struck down a scheme that explicitly used race as a factor in assigning students to schools. Louisville has instead devised a plan whose stated goal is socioeconomic and geographic integration: Instead of giving a boost to minority students per se, it treats equally all students from poor, minority areas:
Under the new program, the school district is divided into six clusters, each of which was drawn to be racially and economically diverse. Families will choose schools within the cluster and their choices will be honored so long as there is room and the placement furthers the diversity requirements. To be diverse, each school is required to have between 15 percent and 50 percent of students from elementary school enrollment areas that are designated as "Geographic Area A"--those where the majority of students reside in census blocks that are below the district's average median household income, below the district average educational attainment, and above the district average for public school minority students. (Currently, 48 percent of the district's 98,000 students are from minority groups.) All students in a geographic area are designated by their neighborhood demographics, regardless of their individual race or economic status. All other areas--poor white areas, middle-class minority areas, and middle-class white areas--are designated as "Geographic Area B."
This scheme strikes me as quite clever: It's more narrowly targeted at the most serious problem Louisville faces--the effective segregation of entire neighborhoods of poor, minority students into bad schools--without explicitly discriminating against individual students on the basis of race. As Kahlenberg notes, it could still face legal challenges, since it's not entirely colorblind, although it would be in the (unlikely, but theoretically possible) event that Louisville's residential segregation were to disappear. From a political standpoint, of course, geographic and class-based affirmative action are somewhat easier to sell, even if they are designed in such a way as to benefit poor minority students disproportionately.
It's also worth noting, I think, that the relative ease with with Louisville appears to have devised a reasonably suitable alternative to the overtly discriminatory scheme suggests that perhaps the five conservative justices had it right in striking down the original plan, since one of the questions asked in a strict-scrutiny analysis is whether the state-sanctioned discrimination is the least restrictive means possible for achieving the interest it serves.