Fifty-eight years ago, Justice Felix Frankfurter told his brethren to stay out of the business of drawing political maps. "Courts ought not to enter this political thicket," Frankfurter warned in Colegrove v. Green. "The fulfillment of this duty cannot be judicially enforced." Initially, they listened to him. The U.S. Supreme Court dismissed Colegrove's assertion that Illinois congressional districts were grossly unequal in population, were based on obsolete census data and were therefore unconstitutional. Your argument is reasonable, the court in effect said, but we don't go there.
Before long, however, on this issue as on many others, Frankfurter came to be seen as a nay-saying old sourpuss. New cases came up in which a congressman or state legislator represented three times as many people as his colleague in the adjoining district--so that voters in the smaller district had three times as much representation. A consensus developed that if legislatures would not fix this inequity, the Supreme Court would have to.
And so, beginning with Baker v. Carr in 1962, the justices defied Frankfurter's Curse and thrust themselves into the territory of redistricting and reapportionment. After all, it was argued, unequal districts were blatant violations of equal protection under the law. How could judicial intervention make the system any worse?
Forty years later, we have an answer to that last question. Not only could the courts make things worse--they did.
Decades of litigation and judicial activism have created a system in which bizarrely shaped districts exist to serve undisguised racial purposes. Partisan gamesmanship has brought us legislatures stacked with safe seats that preclude competition at election time. All of this has taken place under the auspices of a Supreme Court doctrine that virtually any political outrage is permissible as long as the census populations of the districts are mathematically the same--even though no such mathematical precision even exists.
Recognizing that they have made a mess, state and federal judges keep tinkering with the rules to try to bring some order to the chaos. But every ruling they hand down makes the whole process more confusing, more irrational and more unfair. Perhaps the most sensible thing the courts could do would be to start over from scratch. But of course, courts don't act that way. They go to illogical lengths to make decisions look consistent with earlier ones. In doing so, they create even more confusion.
The best example of this is the ever-changing constitutional doctrine on race as a factor in drawing political boundaries.
The federal Voting Rights Act of 1965 provides that districts cannot be drawn, for Congress or state legislatures, in a way that discriminates against the political rights of minorities. Since intentional discrimination is difficult to prove, Congress amended the law in 1982 to establish the concept of "unintentional discrimination," which required a distortion of the English language. Common sense tells us that if something isn't done on purpose, it can't be discrimination.
Nevertheless, the law was rewritten, and in 1986, in Gingles v. Thornburgh, the Supreme Court tried to explain what it meant. A district map would be unconstitutional if it caused a large and cohesive minority group to have "less opportunity than other members of the electorate to participate in the electoral process."
That was an impossibly vague instruction for any mapmaker, but legislatures all over the country assumed it to mean that whenever it was possible to create a district that would elect a minority-group candidate, they were required to do so. This was why, in the 1990s, states created districts with incomprehensibly bizarre shapes that bore no semblance of a political identity but contained enough minority voters to satisfy what it thought was the court's position.
In 1996, however, in Shaw v. Reno, the Supreme Court began trying again. It has held in a series of cases since then that race can be a factor in drawing districts--it just can't be the predominant factor.
That is an oversimplified account of what has happened, but it is a reasonably accurate summary of where the law on race and redistricting currently stands. A new district map must take race into account, but not too much--otherwise it may be unconstitutional. But where that line is to be drawn no one can say. And, in accordance with the Voting Rights Act, minorities can't be denied equal participation in politics--but what that means nobody knows, either.
And that is the situation confronting the U.S. District Court in Boston right now as it considers a package of lawsuits that challenge the redistricting of the Massachusetts legislature following the 2000 census.
In redrawing the maps to fit census numbers, the Democrats who run the legislature did what majority parties have always done: They jiggled the lines here and there to make incumbents happy. The House Speaker, Thomas M. Finneran, saw the demographic makeup of his Boston district go from 74 percent minority down to 61 percent. Most of the voters Finneran lost went to an African-American Democrat, Shirley Owens-Hicks, whose constituency became 92 percent black.
Some of Finneran's critics thought this was pretty shocking. Investigative reporters for a local TV station claimed credit for a big scoop: "The I-team has learned," WBZ announced last May, "that the Speaker changed his district to make it easier for him to win reelection." These intrepid sleuths might qualify for a reporting prize had they uncovered this practice 200 years ago, when Boston politicians first started doing it.
In fairness, however, there are several ways to look at the facts in Massachusetts. One can see these redistricting games, if one wishes, as an exercise in simple log-rolling that, while less than noble in its goals, is an inevitable part of the political process and wholly inappropriate to judicial intervention. That's what Frankfurter would have said.
Or one can see them as discrimination. In its final form, the legislative district map gave the city of Boston 17 state House seats, 12 of them mostly white and the rest "majority minority." There's little doubt that, had they wanted to, the legislature's mapmakers could have grouped black and Hispanic voters together to establish a sixth Boston district dominated by black and Hispanic voters. In failing to do that, they might be said to have diluted minority voting strength and violated the principles of Gingles v. Thornburgh. But had they done it too enthusiastically, they would have disobeyed Shaw v. Reno.
You get the idea. After more than three decades, there is essentially is no usable law on race and redistricting, or at least, no law that helps much in making decisions. Judges in a redistricting case can do just about anything their values and political prejudices urge them to do.
Before long, the rules will change again, not just on racial line- drawing but on the whole matter of what's permitted and forbidden in redistricting. Last month, the U.S. Supreme Court heard oral arguments in Vieth v. Jubelirer, a Pennsylvania case dealing with the issue that even judicial activists haven't wanted to touch: gerrymandering for purely partisan purposes.
The plaintiffs allege that the GOP leadership in Pennsylvania redrew the state's congressional maps with the explicit intention of electing more Republicans. They argue that Democrats were essentially deprived of a role in the process, thereby denying them equal protection. The mapmakers are relying on the court's traditional stance that partisan gerrymandering--unlike gerrymandering to preserve racial inequality or rural power--is not a justiciable issue.
Part of me wishes the plaintiffs luck. Over the past couple of decades, a succession of gerrymanders has created so many safe one- party seats in Congress and in legislatures that entire election years go by with very few legitimately competitive contests anywhere in the country.
On the other hand, when I think about the clumsiness with which courts have treated redistricting--imposing illogical rules of rigid population equality, handing down vague and conflicting decisions on race--the idea of judges trying to define "excessive partisanship" strikes me as a recipe for chaos.
Somewhere between these extremes, however, there are things the Supreme Court could constructively do. It could start by adopting the principle that redistricting is an inherently problematic activity in which self-interested political actors determine their own fortunes. The best way to do it is to avoid doing it very often. When new census numbers come out, states have a constitutional obligation to redraw congressional and legislative district lines. After that, they should forget about them until the new census. Redrawing lines in the middle of the decade, as happened recently in Texas and Colorado, is an extra-legal form of mischief that might actually be curable by judicial decree.
Second, the court could relax the requirement that districts be 99 percent equal in population. Not only is this rule mathematically absurd, given the imprecision of the census, but it offers legislatures an excuse to tinker endlessly with district boundaries to meet a meaningless numerical standard. Allowing districts within a state to vary in population by, say, 10 percent, would disenfranchise no one and let mapmakers leave well enough alone in most places, keeping the districts the same unless they had become grossly disproportional.
Finally, the justices could go back and read Frankfurter on this subject. "To charge courts with the task of accommodating the incommensurable factors of policy that underlie these mathematical puzzles," he wrote in Baker v. Carr, "is to attribute, however flatteringly, omnicompetence to judges."
That, as we all know, is a quality judges don't possess on any subject. On this one, competence would be a step forward.