Professor Peter Schuck is the author of an important 1987 article expressing doubt that the courts can successfully combat partisan gerrymandering. In a recent guest post on the Election Law Blog, Professor Schuck extends his skepticism to last month’s decision by a Wisconsin district court to strike down the state’s Assembly plan as an unconstitutional partisan gerrymander. As one of the attorneys in the Wisconsin litigation—and also as a longtime admirer of Professor Schuck’s work—I’d like to say a few words in response to his critiques.
First, Professor Schuck is correct that a plan’s partisan skew is sometimes the result of a state’s political geography—and not any intent to disadvantage the opposing party. As he puts it, “Increased clustering of partisans in discrete communities would assure many wasted votes even with no packing.” However, the skew of Wisconsin’s current Assembly plan is plainly not the product of geography. We know this because one of the plaintiffs’ experts, Professor Kenneth Mayer, designed a map that improved on the actual plan in terms of every traditional criterion, while also treating the parties far more symmetrically. Even more persuasively, Professor Jowei Chen used a computer algorithm to create hundreds of maps, all superior to the actual plan on every traditional measure. Every one of these simulated maps was more symmetric than the actual plan too, and many had no skew whatsoever.
Moreover, if political geography is responsible for a plan’s partisan distortion, then there is no liability under the plaintiffs’ proposed test. In that case, the distortion is justified and the test’s final prong is not satisfied. In the words of the district court, “the defendants [can] avoid liability if they can justify [the plan’s] effects on the basis of legitimate districting goals or Wisconsin’s natural political geography.”
Second, Professor Schuck states that “the test’s outcome can turn on slight changes in the distribution of votes among districts.” It’s true that small variations in votes received can have big effects on seats won; consider the mere two-vote difference between losing a seat 49-to-51 and winning it 51-to-49. But the possibility of such shifts is explicitly taken into account by the plaintiffs’ test, which requires a partisan asymmetry that is not just large but also durable. Political scientists routinely assess durability using what’s known as “sensitivity testing”—moving the parties’ vote shares in each district up and down by several percentage points, and seeing what happens to the parties’ seat shares as a result. If a plan’s asymmetry evaporates under sensitivity testing, then the asymmetry is not resilient and the plan is valid.
Third, sensitivity testing also resolves Professor Schuck’s concern that “independent voters, third-party voters, and party-switching might make past voting a less reliable guide to predicting future [voting].” That many voters do not affiliate with either major party, and that a few voters support third parties or switch from one major party to the other, are reasons why the major parties’ vote shares may change from one election to the next. These changes, though, are precisely what are captured by sensitivity testing. Again, if there is a plausible electoral environment under which a plan would become fair, then the plan would be permissible under plaintiffs’ test.
Fourth, Professor Schuck correctly notes that once the extent of partisan gerrymandering can be quantified, “it makes much turn on precisely how large a gap courts will tolerate.” But this is exactly the same issue courts confronted half a century ago in the context of malapportionment. Over a series of cases, the Supreme Court decided that population deviations above 10% are presumptively unconstitutional, while smaller deviations are usually lawful. The same sort of doctrinal move—not at once, but rather over several cases as courts grow more familiar with the area—should be possible in the partisan gerrymandering domain.
Fifth, Professor Schuck suggests that there’s more to representation than party; “[e]ven with a comfortable victory margin, a representative must still attend to the concerns of the other party’s voters.” Once upon a time, it was true that legislators’ voting records could not be predicted easily from their partisan affiliations. That era, though, is long past. In today’s hyperpolarized political environment, how a legislator votes is almost a perfect function of the legislator’s party. Legislators hailing from closely contested districts are virtually indistinguishable from representatives from the safest constituencies. For contemporary mapmakers, then, party is just about everything.
Sixth, Professor Schuck rightly observes that plaintiffs’ test would “protect bipartisan gerrymanders protecting both parties’ incumbents from competition.” That is a feature, though, not a bug, since it is only partisan, not bipartisan, gerrymandering that the Supreme Court has recognized as a constitutional harm. The Court has actually declared that “judicial interest should be at its lowest ebb when a State purports fairly to allocate political power to the parties in accordance with their voting strength.” Right or wrong, this sentiment is the law, and plaintiffs’ test is entirely consistent with it.
Finally, Professor Schuck argues that a viable cause of action for partisan gerrymandering would result in more litigation (“draw[ing] federal courts into new partisan battles”) and produce partisan winners and losers (“because one can usually predict which party will benefit from it”). However, redistricting litigation is already very common; in the current cycle, for example, 224 cases were filed in 42 states. Adding partisan gerrymandering to litigants’ quiver would just mean that these disputes could be settled based on the partisan factors that drive them—not shunted into other ill-fitting doctrinal categories. As for the possibility that judicial intervention might benefit one party and disadvantage its opponent, it’s important to remember that the status quo already—and unfairly—does so. Surely, eliminating an existing bias is different from, and more commendable than, creating that bias in the first place.