Let me offer a very specific example to answer these questions: metropolitan Detroit, Michigan. (For the record, I could just as easily make the same case using metropolitan Cleveland, St. Louis, Chicago, Boston, Philadelphia, New York, Newark, Cincinnati, Atlanta, etc.)

In metropolitan Detroit, there are 83 separate school districts. About 80 percent of African American students attend schools in just 3 of those districts. Why? Because zoning laws, racial discrimination in the real estate market, the lack of affordable housing in many suburban communities, the long-term effects of federal policies that disadvantaged blacks, the enormous black-white wealth gap (see Orlando Patterson on this), and persistent racial tension limits blacks' choices in the housing market.

The three Detroit-area school districts that are home to most black students are overwhelmingly poor and working class. Students coming from disadvantaged backgrounds are concentrated together in classrooms. There is a broad consensus among education researchers that putting a bunch of poor students together in the same classroom hurts their educational outcomes (Rick Kahlenberg elaborates on this point in his post). There is high teacher turnover. Classroom order is more difficult to maintain.

The fact that Detroit (and most other metros have dozens, even hundreds of local school districts) is the result of a long-standing American tradition of privileging local control of education.

One of the biggest reasons for educational inequality is the reliance on local taxes. This hurts poorer school districts that have suffered decades of disinvestment and the resulting tax loss.

To be very specific, it is perverse to concentrate the neediest students together in districts with the oldest buildings, the highest teacher turnover, and weakest local tax bases.

And it is perverse to cluster together the richest students in the districts with the newest facilities, the most committed teachers, and the strongest local tax bases.

We have long dreamed of education as the ticket up and out--the path to upward mobility. But in our balkanized metropolitan areas, the opposite is the case. Our educational system creates and reinforces inequality--it helps the rich stay rich and the poor stay poor.

The fragmentation of school districts has had another perverse consequence. When African Americans (even middle or upper middle class folks) move into formerly all-white communities in any sizeable numbers, whites pick up and flee, most of them moving to nearby majority-white school districts. The result is resegregation.

Southfield, an upper-middle-class suburb of Detroit, offers a good example of resegregation (other examples might be Mt. Vernon or Hempstead, NY, or Yeadon, PA, or Decatur, GA). Southfield went from being nearly all-white to mostly black in a period of less than 20 years. The well-documented negative effects of segregation kicked in. School test scores fell. And the district has grown poorer.

Americans have long touted the virtues of local control in education. Yes, local control can be empowering (at least to the relatively small segment of the local population who bothers to vote in school board elections and the smaller number who attend meetings).

But when it comes to race and education, there is an unfortunately long history of communities using local control to benefit themselves to the exclusion of others. That's what happens in most big metropolitan areas (Detroit, Philadelphia, New York, Boston, Chicago, etc., where minorities remain concentrated in the center city and a few surrounding school districts).

Localism often fosters a very narrow sense of responsibility: residents of privileged communities can look out onto nearby poor school districts and claim that "it's not my problem--it's their problem" or "I don't want my hard-earned tax dollars helping someone else's children." This is what I mean by a narrow, cramped version of community.

Thinking about education in the metropolitan context means rethinking community--drawing the boundaries more broadly, more inclusively.

On this topic, I strongly recommend the work of the two Orfields. Gary Orfield at UCLA has written extensively about the resegregation of American public education in the last twenty years. Myron Orfield of the University of Minnesota has written compellingly about commonalities of cities and suburbs, the irrationality of purely local governance, and the need for regional solutions to problems that can't be solved by the even the most well-meaning local officials.