(1) The durability of partisan gerrymandering effects
The approach to partisan gerrymandering based on symmetry allows us to deal not only with the issue of severity in the form of the magnitude of partisan bias, but also with the issue of the expected durability of partisan gerrymandering effects that was raised by the Bandemer plurality.47 The Bandemer Court required that gerrymandering be durable in its effects before it could rise to the level of a constitutional violation (the redistricting must “consistently degrade” a voter's or group of voters' influence on the political process as a whole). In effect, under that standard, plaintiffs must show that they have little or no hope of overcoming the plan’s discriminatory bias.
It is sometimes claimed that we do not need to worry about a redistricting’s partisan impacts because these will “wear off” over the course of a decade. In general, that need not be true. It is certainly true, of course, that those attempting gerrymandering do not always achieve the effects they intend. And, if they make a mistake gerrymandering, the partisan effects of their efforts can wear off over time, or even be reversed.48 But that gerrymandering does not always work, or even that it sometimes fails over time, certainly does not mean that there can be no such thing as a successful partisan gerrymander. The methodology we propose allows us to distinguish those cases in which a gerrymandering might have been attempted but was not very well done from those cases in which the partisan bias imposed by gerrymandering is expected to be both substantial and long-lasting. While measures of partisan bias involve “if, then” scenarios rather than crystal ball gazing, there are special circumstances in which it may be possible to anticipate durable partisan biases that lock-in particular outcomes or ranges of outcomes.
Journalistic accounts of partisan gerrymandering often describe it as a process of packing one’s opponents into as few districts as possible and seeking to win the remaining districts by the barest of margins. While there is a good deal of truth in this portrait, more sophisticated analyses have shown it to be flawed. “Efficient” partisan gerrymandering is forward looking, and seeks to take into account the magnitude of longer run “electoral tides.”49 Thus, skilled gerrymanderers draw districts for the party intended to be the beneficiary of the gerrymander with expected margins that are large enough to insulate those districts from expected changes over times in voter preferences. That is, they attempt to draw districts so as to result in high levels of partisan bias in their favor and low levels of electoral responsiveness.
When we have skillful partisan gerrymandering, we see low levels of electoral responsiveness being implemented by the party doing the gerrymandering protecting incumbents of its own party with safe seats, and packing partisans of the opposing party into seats that can be confidently expected to be very safe for the minority party candidate in the district over the course of a redistricting decade. Thus, when we have sufficiently skillful partisan gerrymandering, rather than the constituents choosing their legislator, those drawing the lines choose each legislator’s constituents in such a fashion that the non-term-limited legislator can expect to be safe as long as the legislator chooses to run for reelection, unless there is a truly enormous voter swing.50 Indeed, if they are given a district free of competition from a fellow legislator of their own party, sometimes self-interested incumbent legislators of the minority party are very happy with the results of a partisan gerrymander for the other party because their own seats have been made so safe.51 But party leaders of the minority party, especially those from outside the legislature, are likely to be left very unhappy because the way in which lines have been drawn essentially freezes in the minority party’s status as a minority party. Fortunately, the same concepts of bias and responsiveness, and the methodology used to measure them that we have described, can be used to assess both the bias of a plan and the expected durability of its effects.
(2) Hypothetical, versus directly observed, unequal treatment
Justice Kennedy makes it clear in his opinion that he is bothered by the fact that estimates of partisan bias commonly rely on projecting previous information from previous elections into the future: “Even assuming a court could choose reliably among different models of shifting voter preferences, we are wary of adopting a constitutional standard that invalidates a map based on unfair results that would occur in a hypothetical state of affairs.”52 Consequently, as noted earlier, his opinion suggests that litigation might be better delayed until levels of partisan bias could be observed directly from one or more elections under a given plan. We do not wish to discourage partisan gerrymandering law suits being brought after an election had taken place under a new plan, since then any harm from the redistricting plan can in fact be judged with somewhat less uncertainty after the election. Moreover, if the Court required partisan symmetry to be part of the judicial standard of partisan gerrymandering only after the first election, redistricters would surely anticipate this in drawing the districts in the first place, especially since it is so easy to assess the plan before the election. Nevertheless, there exist a number of reasons to think that the plurality language in Bandemer53 allowing for reliance on either an “actual” or a “projected” history of disproportionate results is to be preferred to an insistence that only evidence from elections conducted under a plan be regarded as probative. We outline those reasons here.
First, as noted above, use of a partisan symmetry test does not involve predicting actual election outcomes in the form of vote shares. The redistricting plan itself, which is the document that to be declared constitutional or not, exists prior to the election and observing it involves no hypotheticals or counterfactuals. Moreover, the effect of the redistricting plan can be directly estimated before the election based on conditional, statistically grounded assertions about what would happen to outcomes if any particular aggregate vote share percentage were achieved by a given political party’s candidates. Measurements by experts about partisan bias in a redistricting plan are thus not efforts to divine the future. For example, in a single district election, the claim that “electoral rules declare whichever candidate who garners a plurality of the vote the winner of the seat” is an example of an if-then statement, and requires no prediction about or knowledge of which party will win the vote plurality. Similarly, the statements required in ascertaining partisan bias are statements about what level of asymmetry can be expected in translating the votes for the Democratic and Republican candidates in a legislature into seats in that legislature if, in the future, there were to exist a particular level of voter support for each party. The level of support that voters give to the parties is the voter’s own business and is not stipulated ex ante as part of this definition of fairness. Rather, by examining all the relevant data and the potential seat divisions that would occur for particular vote divisions, social scientists can compare the potential scenarios and determine the consequences for partisan bias of a map, separating out other potentially confounding factors. The question is not whether a particular party will win; it is whether the redistricting plan on the table has stacked the deck to such a degree that the plan burdens the other party’s “rights of fair and effective representation.”54
Second, as also noted above, the assumptions used are relatively weak ones, involving only relative propensities of the voters within a particular geographic area in a state to vote for a given political party as compared to voters within some other geographic area. It is an empirical question as to whether units of geography can be distinguished in their partisan propensities, and thus whether partisan gerrymandering is even possible. But if partisan gerrymandering is possible, then systematic measurement of it is possible. Just as racially polarized voting is a “linchpin” of racial vote dilution claims, even though vote dilution and racial polarization are completely distinct concepts, so the empirical claim that partisan propensities exist in a jurisdiction that are strong enough to allow us to roughly rank order units of geography in their expected partisanship is a necessary predicate of any partisan gerrymandering claim. 55
Third, when we look at actual evidence about the predictive power of this methodology we are reassured about its usefulness. For example, in expert witness testimony in a California congressional districting case of the 1990s by Bernard Grofman, specific claims were made in 1994 about a “Democratic lock” that would be expected to last the rest of the decade barring substantial changes in California statewide voting patterns. These observations turned out to be highly accurate.56 Similarly, predictions given in expert witness testimony in Federal District Court by Gary King for every district election in the Ohio State House and Senate also proved highly accurate .57
Fourth, the use of this methodology by the experts for opposing sides who testified about partisan gerrymandering in the cases consolidated into LULAC (Professor Alford and Professor Gaddie) were in remarkable agreement about the partisan implications of the plans whose partisan bias they investigated. As Justice Stevens points out “According to Professor Gaddie, the State’s expert, Plan 1374C gives Republicans an advantage in 22 of 32 congressional seats. The plaintiffs’ expert … agreed. [That expert] added that in his view, the only surprise from the 2004 elections was ‘how far things moved’ toward achieving a 22-10 pro-Republican split ‘in a single election year.’”58
While such congruence among experts for opposing sides cannot prove that the methodology for determining partisan symmetry is valid for predictive purposes, it enhances our confidence that courts can find credible expert witness testimony to give reliable estimates that can be helpful in resolving partisan gerrymandering cases. The assertions that using assessments of hypotheticals based on previous election results (and some statistical calculations) to calculate levels of partisan bias is beyond the competence of expert witness testimony, or results in claims whose credibility cannot be evaluated by courts, is simply wrong. In cases involving racial vote dilution claims under the 14th Amendment or under Section 2 of the Voting Rights Act of 1965 as amended in 1982, for almost forty years, expert witnesses have been routinely making use of hypotheticals in a very similar way to what would be required in calculations of partisan bias. What experts in these racial cases have done is use previous election outcomes and knowledge of racial demography to calculate the hypothetical minority population proportion needed to provide minorities a realistic opportunity to elect candidates of choice with votes entirely from within the minority group, and/or the hypothetical minority population proportion needed to provide minorities a realistic opportunity to elect candidates of choice if there is a given level of reliable white crossover voting.
Fifth, in voting rights cases, courts commonly make what we might think of as “presumptive” findings of statutory or constitutional violations. For example, in looking at plans under the Voting Rights Act, it is almost inevitable that a plan will be evaluated for Section 5 compliance before any elections have been held under the plan. There is no obvious reason why an analogous rule should not apply to partisan gerrymandering.
Finally, and relatedly, there are also good reasons to think that a gerrymandering standard that is addressed to prospective harms rather than to harms already committed can be lawful. Consider, for example, a claim that a contractor has violated fiduciary obligations or been guilty of wrongful neglect by using a cement mix that is likely to crumble over time. We would not want to have to wait until the building collapsed to compel a remedy. Nor would we need to know exactly when the disaster was going to occur as long as its occurrence could be reliably predicted. Moreover, in the context of redistricting, once lines have been redrawn to favor a given party, the electoral advantage of incumbency will be conferred on the winners, who will then carry this advantage into new elections, which may make it harder to undo the damage from an illegal partisan gerrymander than if the problem were avoided in the first place. In addition, the minority party “will surely have a more difficult time recruiting strong candidates, and mobilizing voters and resources, in these …safe districts.”59