Political Gerrymandering

By far the most significant form of gerrymandering is partisan gerrymandering. It is painfully ubiquitous and yet can be covert, hidden in plain view. More importantly the advent of the computer and the growing polarization of the electorate have enhanced its importance. Already the House of Representatives has become virtually safe for most incumbents and in recent elections experienced less turnover than the Senate, which has far fewer members and only one-third of its members up for reelection.

But its ubiquity and its subtleties make politically gerrymandering difficult to judicially control. Courts have been leery of reviewing and revising each and every redistricting. Can there be both threshold indicators of severe distortions and, more importantly, can courts be given the tools to distinguish extravagantly unfair behavior from ordinary, political indulgence by the majority?

Evolving Supreme Court Views on Partisan Gerrymandering

The Supreme Court first ruled on the subject of partisan gerrymandering in 1986 in the case of Davis v. Bandemer (1986). Three justices (Warren Burger, Sandra Day O'Connor, and William Rehnquist) argued partisan gerrymandering was non-justiciable. Two (Lewis Powell and John Paul Stevens) argued that it was justiciable and set out specific standards for measuring it using traditional Fourteenth Amendment standards criteria for establishing the presence of partisan gerrymandering, and would have invalidated the Indiana district being reviewed. The controlling opinion (written by Byron White, and joined by Harry Blackmun, William J. Brennan, and Thurgood Marshall) held that partisan gerrymandering was justiciable, that is to say, reviewable by a court. Their criteria for determining whether a plan was unconstitutional, however, focused on the extent of harm to those claiming they were unfairly disadvantaged. Specifically, they held that a districting plan should be invalidated for partisan gerrymandering only if claimants could show that they were “completely shut out of the political process.” To prevail, complainants had to show “continued frustration of the will of a majority of the voters, or a denial to a minority of voters a fair chance to influence the political process.” Given this stringent language, lower court judges have gone through the motions and heard copious testimony, but arrived at a foregone conclusion.

The Supreme Court revisited the subject of partisan gerrymandering in 2004 in the case of Vieth v. Jubilirer. Justice Antonin Scalia, in a plurality opinion which Rehnquist, O'Connor, and Clarence Thomas joined, would have overruled Bandemer and noted that two decades had burdened the courts but produced no results. Scalia was particularly skeptical of attempts to extend equal protection claims in this area beyond the “strict scrutiny” of race. Consequently, he developed in detail greater judicial appropriateness for claims of racial gerrymandering, especially as a practical matter. “Even within the narrower compass of challenges to a single district, applying a ‘predominant intent’ test to racial gerrymandering is easier and less disruptive. The Constitution clearly contemplates districting by political entities, see Article I § 4, and unsurprisingly that turns out to be root-and-branch a matter of politics.”

He was also dismissive of the proposals of his dissenting colleagues, which he deemed to be inadequate, as well as quite different, among the dissents. But he extended his analysis beyond the particular controversy addressed by this case and concluded that the standards articulated in Davis v. Bandemer also were inadequate. The need, under that standard, to demonstrate both intentional discrimination and an actual discriminatory effect upon a group was, according to the historical record, not easily capable of being practically applied. Justices Stephen Breyer, Ruth Bader Ginsberg, David Souter, and Stevens argued that partisan gerrymandering was justiciable and followed the earlier approach of Powell and Stevens in Bandemer in determining its presence. They differed amongst themselves in several specifics and measures, however, resulting in three separate dissents, and Scalia taunted their own failure to agree on a litmus test or indeed, on any approach. (Breyer suggested in rebuttal that the minority sought to broaden the discussion, there being no advantage to airing at a common dissent.)

The key vote was that of Justice Anthony Kennedy, who found no basis for invalidating the Pennsylvania districting at issue (thus joining the Scalia plurality on the merits) but did not wish to overrule Bandemer. The Court later sent the subsequent Texas redistricting, (orchestrated by then House Majority Leader Tom Delay, R-TX) back for a rehearing presumably because Kennedy still thought Bandemer has some potential. But Kennedy was quite pointed in Vieth about the missing conditions that need to be fulfilled in order for this particular type of controversy to be justiciable. First, a set of principles for drawing electoral boundaries that are both neutral and comprehensive must be established. Second, a system of rules to limit the scope of judicial intervention in such cases (giving proper deference to the legislative process) must also be in place.

Ultimately, Justice Kennedy on the merits agreed with the majority, though with a certain amount of reluctance.

Perhaps the most important long-term effect of Vieth is that the number of justices willing to start doing something serious about partisan gerrymandering has grown from two (Powell and Stevens) in Bandemer to four (Breyer, Ginsberg, Souter, and Stevens) in Vieth. Furthermore, Rehnquist and O'Connor, who dissented in Bandemer and who joined Scalia, have been replaced by new justices—John Roberts and Samuel S. Alito—with unknown views on this issue.

In any event it is clear from all opinions that Bandemer is on probation. As Judge Patrick Higginbotham's decision in the remanded Texas case also indicates, there is frustration with a standard that goes through copious detail but never produces results. Bandemer must be honed and perfected or it will be overruled or ignored. Yet the issue it raises grows in significance and its danger to democratic processes waxes almost daily.

Given the plain text of Kennedy's opinion in Vieth, it was a bit of an eyebrow-raiser for the Supreme Court to remand the Texas redistricting case to the original three-judge court to adjudicate in the light of Vieth. A remand, unlike certiorari, takes five votes and almost certainly required Kennedy's agreement. Given his plea for time and the desirability of more play within the court system, it seems unlikely he would hasten the flow back up to the Supreme Court in order to kill the Bandemer precedent.

The lower-court judges were a bit bemused by the order to revisit their previous decision. Higginbotham's majority opinion had on the first round explicitly challenged the Supreme Court to either forge effective standards or give up ineffectual time-wasting reviews. Scalia's opinion built on that to urge ending the Bandemer precedent. But the Court now was presumably asking them to fact-find on the basis of a further subtle erosion in Kennedy's opinion, which in any event was not more precise than Bandemer.

Higginbotham wisely chose to follow Kennedy's swing-vote views. He specifically found the Texas variations less partisan than those upheld in Vieth. The “special concurrence” by Judge T. John Ward also found that within the terms of the remand, the three-judge court had no choice but to accept the new redistricting. Though he disagreed with Kennedy's standards, that was not at issue under the remand instructions.

Remands are made for many reasons. The most obvious is when a new Supreme Court decision might alter a lower court ruling or the Supreme Court wishes a more refined analysis reflecting its new precedent. But in truth, there is precious little new in Vieth, so the Justices were requiring a largely repetitive exercise. Remands may also involve clarification of facts or of the law involved at the state level. This does not appear on the face of the instructions in this instance but may have been sought. Additionally the Justices may be hoping for better briefs or new theories.

When Higginbotham's opinion seemed to provide no new targets, many assumed certiorari would be denied. But much happened out of court to transform an interesting case into front page news. It was revealed that the Department of Justice experts in charge of reviewing districting under the Voting Rights Act of 1965 had unanimously concluded that the Texas plan was in violation of that law (Eggen, 2006). They had been overruled by politically-appointed officials, an act fully within their authority but rare. (The Voting Rights Act requires Department of Justice approval of redistricting in states with specified histories of vote-denial in the past. Justice Hugo Black thought this unconstitutional and its perennial reenactment by Congress remains controversial.) Attorney General Alberto Gonzales, defending the action of his predecessor, noted the three-judge court had upheld the redistricting, but this is somewhat disingenuous. That precise issue was not before the court, and the internal administrative overruling was not known to the judges, or the plaintiffs. In a sense, Texas drew a bye on the issue as it was assumed the Department of Justice had done the analysis with full rigor. (This late disclosure came after several key voting rights professionals in the Department retired and presumably leaked the information.) It also was revealed that the Department excluded the division from review in several key cases, assigning the entire process of approval to the higher (political) levels of the Department. In approving Georgia's picture-identification requirement for voting, doubts raised by the special voting division were overruled the same day, so that the entire process was telescoped into hours.

How much or how little all this influenced the Justices may (or may not) become clear as the case unfolds. The Supreme Court granted certiorari and set an expedited March 1 hearing date, recognizing the problem of looming elections in Texas and perhaps affecting redistricting in Georgia as well. The announcement was the banner story in leading newspapers that day.

The New Political Realities of Partisan Gerrymandering

Partisan gerrymandering is much more serious now because it is demonstrably more effective than in the past. Students of the subject often discounted its impact because it was, they thought, both self-correcting and rather more difficult to carry off in the face of a mobile society, a fickle electorate, and the intricacies of census data with bewildering and not easily available aggregated and disaggregated forms. Gerrymandering, they pointed out, was a long-practiced art. (Eldridge Gerry, after whom the practice is named, was a Founding, if dissenting, Father who left the Constitutional Convention without signing, opposed ratification, and subsequently had a decades-long political career.) Yet many efforts to rig districts have boomeranged, and the technique was visibly less effective and had a less enduring effect than benign neglect, failure to redistrict, or creating some limited electorate or a rigged weighting system.

The “self-correcting” argument is put forward in O'Connor's Bandemer dissent as an argument against judicial oversight. Because it has created more seats in which it hopes to win relatively narrow victories, the same swing in overall voting strength will tend to cost the legislative majority more and more seats as the gerrymander becomes more ambitious. This Pollyannaish expectation that gerrymanderers will systematically overreach and the gerrymander collapse or even boomerang is not based on recent experience. Gerrymanders sometimes get over-greedy, but they learn quickly. Embezzlers sometimes steal so much the bank fails, but most are exposed by rigorous auditing. A wise gerrymanderer—and there are now national expert consultants—will allow margins for error. And in any event even transitory gerrymanders create new incumbents with a distinct advantage.

Complexity has historically been the real deterrent. But today's computers effortlessly run through hundreds of alternative plans which would have required tedious human manipulation by expensive experts. The census and other data are usually obtainable by a few clicks of the mouse. Ingenious programs can match past voting patterns to current residential locations. Mobility remains a problem but past patterns of conversion of new arrivals can be estimated, though with problematic results. The outflow of one's partisans is probably more calculable than future votes of not-yet-arrived residents, but past history helps on that issue, too.

Reinforcing predictability has been the increased polarization of American politics. There is a cyclical process involved. Polarized parties promote “regularized” voting; a predictable electorate allows easier gerrymandering; non-competitive districts promote lower turnout, which is usually of ideological voters. The voters can break this cycle but it seems unlikely under current conditions.

The Anglo-American tradition of a single-member district and “winner take all” elections tempt legislatures engaged in districting to gerrymander. Indeed, in a sense the system is a gerrymander since all the votes of a minority and those of a majority in excess of 50% do not contribute to further representation and are said to be “wasted.” Under most neutral conditions a party with a system-wide majority can expect to garner an even larger number of seats than its share of the popular vote. Political scientists and electoral sociologists generally applaud such a bonus for the majority party as facilitating government. Indeed, even “proportional” representation systems universally are constructed to eliminate tiny parties (by requiring threshold minimums) and to give larger parties a premium to help them construct a viable government, though not as dramatic a premium as under the single-member district electoral system. (Another feature of single-member winner-take-all systems is that it seems to buttress the two-party system and is therefore able to skip the need for coalition building after elections.)

The majority-enhancing effect of the single-member district results from a statistical phenomenon: as a party's share of the popular vote increases it is also increasingly and disproportionately likely to gain a majority in each electoral district. The British long thought they had uncovered the basic relationship between popular votes and seats won in the “cube rule,” namely

A 60% popular vote would produce (3/5)³ seats, roughly 77% of the seats, for a landslide outcome. Like many rules-of-thumb, experience shows the neat relationship isn't there, but the general notion remains correct. Parties that “go for broke” and lose badly in the popular vote are punished even more severely in their allocation of legislators. Arguably this leads to more moderate parties.

Clearly though, this effect (which Backstrom, Robins, and Eller (1978) call the “balloon effect”) makes detecting gerrymandering more difficult since single-member districting is inherently (and sensitively) responsive to the distributive aspect of voter response, in a way antithetical to the equal effect of each vote. To counter this problem, Backstrom, Robins, and Eller proposed a corrective measure for the “ballooning” effect of single-member districts to distinguish systemic effects from gerrymandering. The courts have not embraced the measure, perhaps because it is counterintuitive and not easily grasped by statistically challenged legalists.

Conscious partisan gerrymandering tries to minimize the wasted votes of one's party and maximize wastage for the opposing party. The trick is either to densely pack the opposition into their strongholds with overwhelming majorities and/or scatter their voters into districts where they are a permanent minority. A combination of these techniques (including irrelevant minority enclaves tacked on to safe districts) further fritter away rival votes. This leaves a maximum number of safe districts of supporters with reasonable margins for contingencies like voter unrest or sudden population changes.

There is nothing unique to either party about use of gerrymandering, particularly when control of state legislatures is at stake. Philip Burton of California, a liberal Democrat from the Bay area, is generally regarded as the father of the modern art of computer-based gerrymandering. A transitional figure, he relied on his unequaled grasp of California voting patterns more than on electronic aids, which were rather less accessible in his day. And Judge Higginbotham quotes the knowledgeable Michael Barone, on what some have called “the great Texas gerrymander of '91” authored by the Democrats: “The plan carefully constructs Democratic districts with incredibly convoluted lines and packs heavily Republican suburban areas into just a few districts” (Almanac of American Politics 2004, 1448).

The most controversial gerrymanders have in recent years been Republican mainly because Southern states have only gradually shifted their local politics—including governorships and legislatures—to Republican dominance to match their established national presidential allegiance. As Southern legislatures moved state-by-state into the Republican column, long-time Democrat gerrymanders were replaced by Republican handiwork, often with dramatic shifts. It was just such a shift that resulted in the Texas redistricting case, and entangled Tom Delay in ethical issues and finally criminal indictment.

Texas-Sized Redistricting

The Texas redistricting case which was finally heard at the Supreme Court level as League of United Latin American Citizens v. Perry (2006) has all the earmarks of a successful Michael Moore movie. Darkly comedic, ultimately deadly serious, it grew out of unfortunate partisan gerrymandering politics openly practiced. In 1991, the Democrats passed the obvious gerrymander mentioned above, a plan designed by Bob Mansker, aide to Congressman Martin Frost (D-TX), who was later ironically a plaintiff in Henderson v. Perry, as the case was heard on remand. It was challenged and somewhat modified in extensive litigation but still superbly served to maintain Democratic advantage even as they lost popular support. In 1992, the Democrats gained a 50–48% margin in the popular vote but won 21 of 30 congressional seats. Subsequent elections resulted in Republican popular vote majorities of 10% or more, but Democratic majorities held in the congressional delegation.

After the 2000 census, the divided Texas legislature was unable to agree on a redistricting plan and so the federal courts redistricted, accommodating especially the two-seat growth in congressional strength occasioned by the state's spurt in population. According to Judge Higginbotham, the court plan conservatively incorporated many aspects of the Democratic gerrymander because it could not find an objective basis to undo the acts of the Democratic legislature. The court basically created new districts in the area of growth and tried to respect governmental subdivisions. Higginbotham characterized the results as “unchanged” except for the two new seats. Nonetheless, the court's “small” changes were important. The popular vote and the allocation of congressional seats was for the first time in recent Texas history basically congruent—17 Democrats to 15 Republicans—though the Republicans had a slight majority in the popular vote.

Republican forces led by Tom Delay realized there was an opportunity to gain an additional half-dozen votes by doing unto the Democrats what had been done to them over the years. The key was control of both houses of the legislature. An all-out effort including Washington-based solicitation of funds was undertaken. In their zeal to accomplish this, the Delay forces are alleged to have violated Texas law against use of corporate funds. And in their bravado, aides supposedly boasted of their cleverness in outflanking the law. This has resulted in a number of indictments but at this writing no convictions.

By a narrow margin the effort to take both houses of the Texas legislature was successful. But traditions made it necessary to have Democrats present for a quorum to implement the victory. Pointing to the fact that re-redistricting has fallen out of political favor, the Democrats en masse tried to block action by going out-of-state so they could not be arrested and forced to return to provide a quorum. In a stunning lapse of judgment, Delay secured the assistance of federal aviation and security officials to try to locate Democratic legislators en route to their exile. When rebuked by the House Ethics Committee—not for the first time—Delay and his supporters forced out the Republican Committee Chairman and attempted a rule change that would give either party a veto over consideration of an ethics issue. Only when the Democrats refused to serve under those conditions was the rule change abandoned.

Ultimately a Democrat in the Texas legislature tired of hasty trips to Oklahoma and provided the quorum which passed the Republican gerrymander. In the 2004 election the new plan resulted in 21 Republican and 11 Democratic congressmen. The six-seat switch was seen as insurance against a shift in House control given the relatively small number of competitive seats in play. By essentially doubling their margin in the House, the Republicans' complex maneuver was politically vindicated.

Naturally, the districting plan was challenged. After it was upheld by a federal three-judge court it was appealed to the U.S. Supreme Court. Subsequent to the Vieth decision, it was remanded for proceedings under the new “clarifications.” Higginbotham read the remanding as instructing him to use Kennedy's opinion as controlling and it is difficult to see any other reason for the Supreme Court action. Though Higginbotham had earlier in effect challenged the Supreme Court to “make an honest decision” out of Bandemer by reversing it, or giving it teeth, Higginbotham treated his new task as one of finding the Texas situation more offensive or objectively discriminatory than the “unbroken string” of cases where the courts have refused to apply Bandemer. Most particularly, he compared Henderson (2005), the Texas situation, with the Pennsylvania districting which Kennedy had reluctantly upheld. The major plaintiff contention—excessive or exclusive partisanship—was Steven's test, but rejected by Kennedy. The essential problem Higginbotham suggests is that virtually all districting is partisan but also involves some “neutral” principles. Beyond that, most choices are subjective. Without some logical or metric limit the judge is helpless.

Obviously, the salient difference in the Henderson case was its “voluntary” redistricting of a post-census districting, albeit one originally generated by a court. That fact was self-evident in the Supreme Court filing and had been dismissed as legally irrelevant in the first lower court decision by Higginbotham.

Even if the first redistricting had been legislative, however, there is nothing in the Constitution and laws of the U.S., or for that matter Texas law, prohibiting re-redistricting. It would have been highly unusual for a three-judge court to have created a new per se rule and the panel refused to do it. It is frankly difficult to see what else the Supreme Court might have expected. The most interesting new development at the remanded proceeding was the amicus curiae of the University of Texas Law Professors which won the moral support of Judge Ward.

The argument of Professors Lucas Powe and Sanford Levinson is based on a thoughtful and subtle analysis of population figures in an article by Levinson (2001), a major figure in linguistic analysis of the law. Population figures, like the exact amount of the national debt at any moment in time, or one's height, or many other commonly used conventions, are best estimates or close approximations. The census misses some people, and over-counts others. By the time census results are announced, deaths, births and migration makes them wrong in yet another way.

The Constitution and laws say that the census enumeration will be accepted for all apportionment purposes and, as our best set of figures, that is appropriate, Levinson and Powe suggest, for the first bite of the apple. But if a state chooses to re-redistrict, the rule of “one-person, one-vote” dictates that they come up with more recent population figures. This would not prevent re-redistricting, but the added burden to the state would be a reasonable accommodation to the equal protection clause.

This ingenious argument is not without its problematics. More recent census enumerations do not exist and Congress has forbidden the use of estimates in apportionment. Perhaps a state could hire the Census Bureau for such a purpose and perhaps that would meet the constitutional and legal requirements. But none of this could occur in time for the next congressional races, and, as Higginbotham argues, such a ruling would in effect restore the court-ordered plan. If he were not on the Supreme Court's short-leash, Judge Ward would have hazarded these problems which in the event could only ensue from Supreme Court action.

What Is to be Done?

After three decades, the evidence on partisan gerrymandering seems pretty clear. As Issacharoff and Karlan (2004, 574) sententiously summarize the literature: “in the original Constitution the Senate was picked by the state legislators and the House was chosen by ‘the people’ but that after a process of amendment and political adaptation the houses have been inverted: now the people pick the Senate, and the state legislatures through gerrymandering, pick the House.”

Remedies are more tentative and murky. The growth of such gerrymandering is almost as visible as the decline in competitiveness. The linkage seems also clear, since partisan gerrymandering is motivated by an intent to create a maximum ensured seats, and the methods for creating them result in abundantly guaranteed seats for opponents, but with their majorities supersized, in a smaller number of districts.

Obviously, in a democratic society representational issues are best settled in a broad, popular way, though not necessarily in a referendum or a legislature. Courts are a last resort, but even if courts are “umpires and not players” as in Chief Justice Roberts' confirmation mantra, they should be enforcing basic rules of the game. The original reluctant intrusion of courts in Baker v. Carr occurred because rural politicians in areas losing population could maximize political power by ignoring census results, much as Christian minorities in Lebanon with a low rate of natural increase vetoed any new census for half-a-century. Justice Felix Frankfurter, wringing his hands and warning of future problems, suggested that the problem would eventually be resolved by “searing the conscience” of the legislature. But such manifestations of conscience are infrequent and slow in the face of the prospect of giving up power to those whose only weapon is a clear claim of unenforceable entitlement.

A decline in polarized and predictable voting by the electorate would make partisan gerrymandering much more difficult. These things are to some degree cyclical. But the present party linkages to donors and fundraising lobbyists who help perpetuate a brand name incumbency through media access suggest such shifts unlikely and that parties could more easily absorb (and minimize) voter dissatisfaction.

A number of states—a half dozen or so—have tried to resolve matters by establishing electoral commissions. The tacit bargain is in most cases to give the bulk of the membership to the two parties. This is a two-edged sword because it further institutionalizes the existing parties independent of electorate choice and also makes commissioners likely to be partisan at crunch time. It also makes the choice of the “tie-breaking” commissioners more difficult and complex. Still it avoids macho-sloganizing legislators openly, deriding even-handed districting, and usually but not invariably ends the work of a behind-the-scenes architect of a fully-rigged gerrymander. Experience suggests that commissions do not end but can moderate partisan activities.

Since gerrymandering is not limited to one party or one ideology, there is a public interest served by curbing abuses by any of the perpetrators. It certainly is not edifying, or equitable, to argue as Judge Higginbotham occasionally suggests, that it is proper for Texas Republicans to gerrymander because Texas Democrats have long done so. Still less reasonable is any suggestion that, e.g., California Democrats got away with it, so Pennsylvania Republicans should get their bite of the apple. (The bipartisan nature of the abuse is well-indicated, as Gordan Baker (2005) reminds us by the Republican National Committee supporting the Democrats in Bandemer and the California Democratic congressional members filing an opposite amicus curiae brief opposing judicial review of alleged partisan gerrymanders.)

But the Republican Party's interest in judicial intervention has waned as breakthroughs in legislatures in the South continue, and as the “trickle-down” effects from presidential Republican voting wend their way through the system to gain control of legislatures. As each party moves from victim to perpetrator its theory of ox-goring shows flexibility. Nonetheless, there remains considerable bipartisan interest in restraining gerrymandering and restoring competitive districting. Part of that effort has reflected in commissions which are found in a surprising array of states of various political hues. The other main target is toward establishing standards of districting which would serve both as political heuristics for boundary-drawers and could generally aid courts as well.

So far the results from “independent” commissions have been underwhelming. In six states using such commissions, the bipartisanship built into the system has cemented the hold of incumbents. Lack of accountability has insulated commissions from the consequences of not following required transparency; using the judiciary to nominate “tie breakers” dilutes the likelihood of effective subsequent court review and control.

But it does allow some mitigation of one-party partisan gerrymandering, even as it moves the system more toward O'Connor's fears of bipartisan incumbents protecting gerrymandering. But as we suggest here and in our conclusions, the greatest hope is that such commissions will develop—should be required to develop—more precise, and therefore reviewable, standards for apportionment to replace computer-generated ad hocery. Commissions must also be forced to hold open meetings and teeth should be put into those laws, to prevent reported evasion by “face-to-face” negotiations.

The sample from history—the pre Baker attempt to “sear the consciences of the legislatures”—is discouraging to expectations of relying on normal politics to cure malaportionment. A minimalist argument for Justice Kennedy's effort to preserve Bandemer is that it strengthens the hand of reformers at the political level who can argue that if efforts to clean up the act fail, the courts would be pushed into action. The American system abounds with—arguably rests upon—many residual checks and balances seldom utilized but which serve as reminders to power-holders that they need to be circumspect.

But the case for court authority is broader. The role for the judiciary to intervene where protection of the rules of the game, where the heart of the democratic process is at stake, and not leave it to the vagaries of transitory majorities is the essence of the footnote four of Carolene Products Case (1938), the blueprint for some abnegationist and all activist courts. Leaving legislatures to structure and restructure districts with no restraint is not realistic and an open invitation to abuse. It is not hard to conjure up extreme examples of creating special electorates of the type mentioned in Baker and that are best protected by Bandemer.

Regardless of the immediate problem of overruling Bandemer (which would require Kennedy or one of the four other supporters recanting), any effective control of partisan gerrymandering still is stymied over the issue of standards for judicial intervention.

As it is, Kennedy has provided life support for dealing with the problem but his Vieth opinion is skeptical, rather than optimistic about sustaining the Bandemer precedent. After all, in two decades there has never emerged any useful standard in the view of judges trying to implement it. The Bandemer test has, as Scalia and Higginbotham point out, forced court after court to go through statistics and arguments and come up empty. Forced to hear these cases, judges, however sympathetic, have reluctantly or avidly found the alleged gerrymandering unproven by Bandemer standards.

Over these two decades, scholars and lawyers have not helped the courts develop useful tests to distinguish “normal” clever redistricting from outrageous partisanship. They have piled on data by the pound in briefs, amicus briefs, and journal articles about the effects of gerrymandering and the precipitous drop in competitiveness in the House of Representatives, the chamber of the people now virtually immune from popular control. But they have not helped find workable judicial standards, to help courts make decisions or even to distinguish between cases worthy of judicial time and those that are intrinsically incorrigible or even present valid redistricting. Until that happens—until plaintiff lawyers or election analysts find some standards paralleling the efforts of Professors Levinson and Powe in the Texas controversy—Kennedy's lifeline will sooner or later be withdrawn.

It must, however, be noted that finding such standards, given the Court's own strictures is difficult indeed. The Courts' caution is clear in the Bandemer plurality opinion, which severely limits any possible application with its almost impossible standard for intervention. And Justices William Brennan and Marshall, hardly timid souls, were half that plurality. It was a daunting task then and a daunting task today.

In particular, the court has taken to heart O'Conner's argument that tackling partisan gerrymandering will lead to proportionate representation. While a full spelling out of “one-person, one-vote” and a strict “no-wasted votes” policy could lead down that slope, there is no reason to begin to contemplate the courts making, or being allowed to make, that decision. No proportional scheme in use is designed to totally eliminate wasted votes and our courts would not have the moral authority to change a system that is deeply intertwined with our whole political structure. There is no basis for courts restructuring the political order or the Electoral College or the Senate or the single-member district. Just as the Court has had to make a reasonable accommodation with not-completely equal districts (including allowance for existing boundaries), it would stop reviewing gerrymanders long before any dismantling of single-member districts was even broached. But this reality does not answer the Justices' real or imagined fear. Any measures they will accept will have to alleviate fears that the camel's nose will ultimately bring in the camel of proportional representation.

The Court has also been unwilling either in name or in reality to increase the list of suspect categories, and so that easy route has also been effectively closed. The multiplication of suspect categories and basic rights by the Warren Court has vexed the courts and created complexities difficult to bring into sensible patterns. This vexation limits the development of standards for partisan gerrymandering even among liberal judges who left alone might adopt such a course, but know colleagues would not be receptive.

An obvious solution is to expand the “one-person, one-vote” test, but that does implicate proportional representation and is usually coupled in plaintiff's brief with a plea that courts force a state to justify its methods of redistricting, which inferentially would be a “suspect category” expansion. It has been a non-starter at the district court level, and would be at the Supreme level.

The Backstrom-Robins-Eller's test, as we have noted, discounts the “balloon effect” of the single-member district. While Justice Stevens was beguiled by it in Karcher, he has lost interest. The Justices have indicated they regard it as a proportional representation surrogate, and suggest “winnability” in districting is inherent in the patterns of choice of the voters, as much as in patterns imposed by districters. They also seem to reject many similar measures known to electoral sociology which might help. This rejection has been facilitated by the plethora of measures, such as those of simple concepts like “compactness.” As is usually the case, differing measures have advantages and disadvantages in contrast of one against the other. Both sides in a dispute can hire equal-and-opposite statistical experts and it is often difficult to distinguish between appropriate and tendentious statistical evidence.²

Perhaps, this Gordian Knot can be untied by dividing review into two stages on analogy with the Texas professors' suggestion of a trigger event—re-redistricting—and subsequent requirements expected from the state. Moreover, Justice Breyer seems to embrace that approach as well. As a first step we suggest that plaintiffs might be required to show as a threshold matter an excessive departure from an agreed upon measure (perhaps Backstrom-Robins-Eller or some other winnability test) and to put forward a plan accomplishing better population results with greater respect for political boundaries. Only after these showings would there be any review.

Laying the groundwork for such an approach could include (1) states accepting broad standards such as those promulgated by the districting reform conference which would facilitate review by the courts; (2) the development of a winnability measure by scholars that ideally combines historical variations and “normal” variability; (3) new and creative measures that would aid the courts in the merits stage, such as compact measurement, and some measurement of off-setting inequalities, e.g., boundary infringement and population inequality.

Action by the states would be most helpful. States could directly or indirectly help the courts by choosing some of the intermediate standards that commit the state to principled districting, and this would allow the judges to review adherence to the state's own principles rather than force judges to invent them. Those principles could be embodied in constitutions or public referenda. But probably the most useful device would be action by commissions (permanent or temporary) which try to develop parameters for use in specific redistricting. Such commissions would have the time and patience to sort out suggested metrics and to reject the unworthy and sort out reasonable efforts, and even priorities among standards. This would maximize popular-based value choice and judicial fact-finding skills.

Rules limiting districting would not be after-the-fact made up decisions but orderly legal actions. As we have indicated there are a whole array of measures which would allow the courts to control excesses on gerrymandering without requiring them to become the real districters. Judge Higginbotham's arguments about the limited efficacy of judicial boundary-drawing because some value judgments are not justiciable is also an argument that limits must and can be drawn to prevent judicial intervention when inappropriate.

If the current court cannot be persuaded either to adopt a modest approach to the clear impasse, or even a more sweeping systematic approach of the type we present in the Appendix³

and which has drawn support from judges and lawyers, then the O'Connor-Scalia view will prevail, either de facto or by a formal overruling.

In the long run, we believe the courts cannot allow a permanent deterioration of the voting experience. There are of course dangers in the political thicket but there are greater dangers in not setting fair and equitable “rules of the game” and allowing the dominant players instead to make them up as they play on. Judge Richard Posner argued in Breaking the Deadlock (2001) that the Supreme Court had to abandon its non-political role, that otherwise Bush v. Gore would have devastated the country. The present art of gerrymandering threatens the creation of “rotting boroughs,” and waiting for Congress to remedy itself is much more of a civic danger than the lack of a solution to the disputed election of 2004.