Daniel P. Tokaji is the Charles W. Ebersold and Florence Whitcomb Ebersold Professor of Constitutional Law at The Ohio State University Moritz College of Law.
A constitutional standard for partisan gerrymandering is the holy grail of election law. For decades, scholars and jurists have struggled to find a manageable standard for claims of excessive partisanship in drawing district lines. Most of these efforts have focused on the equal protection clause. But as Justice Anthony Kennedy suggested in Vieth v. Jubelirer, the First Amendment provides a firmer doctrinal basis for challenging partisan gerrymandering. An established line of precedent understands voting as a form of expressive association protected by the First Amendment. These cases offer a nuanced standard that would avoid the undesirable result of rendering any consideration of partisan consequences unconstitutional.
The right of expressive association
There is an obvious difficulty in relying on the First Amendment in partisan-gerrymandering cases: The Supreme Court has never considered voting a form of protected speech. It has, however, long recognized that voting is a form of protected association, at least in certain contexts. Before getting to those cases, it’s helpful to examine the roots of the right of expressive association.
The original associational-rights cases involved groups like the NAACP and the Communist Party that were extremely unpopular – one might even say persecuted – in many parts of the country. In NAACP v. Alabama ex rel. Patterson, for example, the Supreme Court invalidated a requirement that the NAACP disclose its membership list. Justice John Marshall Harlan II’s opinion for the court remarked that “the freedom to engage in association for the advancement of beliefs and ideas is an inseparable aspect of the ‘liberty’ assured by the Due Process Clause of the Fourteenth Amendment, which embraces freedom of speech.” In other words, expressive association is a necessary corollary of free speech.
The right of expressive association is closely linked to the First Amendment’s prohibition on content and viewpoint discrimination. As Justice Antonin Scalia put it in one of his last dissenting opinions, “the First Amendment is a kind of Equal Protection Clause for ideas.” It prohibits the government from abusing its authority to suppress disfavored points of view. Most importantly, it restricts the dominant political group’s authority to diminish the voices of those who might challenge their grip on power.
Thus, in the first generation of association cases, disfavored groups like the NAACP and the Communist Party relied on the First Amendment to prohibit the government from taking adverse action against them and their members. These cases rest partly on the individual liberty interest identified in NAACP v. Alabama. But they’re also grounded in a larger vision of how democracy should function.
A leading example is the line of patronage cases that began with Elrod v. Burns, in which the Supreme Court struck down the practice of firing public employees who weren’t members of the Democratic Party, which controlled Cook County, Illinois. After describing the harm to individual employees, the plurality turned to systemic concerns arising from this practice: “It is not only belief and association which are restricted where political patronage is the practice,” wrote Justice William Brennan. “The free functioning of the electoral process also suffers.” Discrimination against non-party members tended to “starve political opposition,” thus “tip[ping] the electoral process in favor of the incumbent party.” In other words, party-based discrimination distorts the political process, entrenching the dominant party in power while subordinating its chief rival.
Voting as association
Partisan gerrymandering effects a comparable systemic harm, albeit through a different mechanism. By manipulating district lines, the dominant party can entrench itself in power even when the political winds shift. The increasingly sophisticated technology that line-drawers have at their disposal exacerbates the problem. It allows the dominant party to capture a large percentage of seats while ensuring that its majority will hold in both bad times and good.
That’s true not only in Wisconsin, from which Gill v. Whitford emerges, but in other states that would be competitive but for gerrymandering. Take my own state of Ohio. Although Ohio is a consummate purple state in presidential elections, Republican mapmakers drew lines there in 2011 that give their party a supermajority of districts – three-quarters of the state’s congressional delegation and roughly two-thirds of its state legislative districts. These districts were drawn with the goal of creating a firewall that would ensure Republican control even in a strong Democratic year. And they’ve been a spectacular success, ensuring Republican control of the Ohio state legislature throughout the current decade. Because they diminish the power of the non-dominant party in a manner that’s both substantial and enduring, excessive partisan gerrymanders violate the right of expressive association.
Still, one might argue that compelled disclosure and patronage are very different from redistricting. In the original association cases, particular individuals were harmed – discouraged from or punished for affiliating with disfavored groups. Moreover, those cases don’t directly involve voting. It’s a leap, one might argue, to hold that the right of association is implicated when voters, candidates and parties associate through the electoral process.
As it turns out, the Supreme Court made this leap long ago. For almost a half-century, the court has recognized that voting is a form of association protected by the First Amendment. The first voting-as-association case was Williams v. Rhodes, which challenged Ohio’s ballot-access requirements for new political parties like George Wallace’s American Independent Party. Justice Hugo Black’s opinion for the court relied on both the First Amendment right of association and equal protection to strike down this requirement. Ohio’s onerous rules for adding new parties to the ballot gave “the two old, established parties a decided advantage … plac[ing] substantially unequal burdens on both the right to vote and the right to associate.” In Williams, the Supreme Court thus stressed the risk of dominant parties using voting rules to entrench themselves in power, thereby harming non-dominant parties and their supporters.
Several years later, in Anderson v. Celebrezze, the Supreme Court again relied on the right of association to invalidate another ballot-access rule in Ohio, this time one that would have kept John Anderson from running as an independent presidential candidate in 1980. Justice John Paul Stevens’ opinion for the majority recognized that there’s no “litmus-paper test” to separate valid and invalid restrictions on voting and association. Rather, the court should weigh the “character and magnitude” of the burden on voting and association against the state’s asserted interests. Although “reasonable, nondiscriminatory restrictions” can usually be justified by important regulatory interests, a stronger justification is required for more serious burdens, including ones that discriminate against outsider candidates and their supporters.
A subsequent case involving write-in voting in Hawaii, Burdick v. Takushi, reaffirmed Anderson’s “flexible” standard while clarifying that strict scrutiny applies only if the burden on voting and association is “severe.” Other cases like Tashjian v. Republican Party of Connecticut and Washington State Grange v. Washington Republican Party apply this standard to electoral rules that burden the associational rights of major parties and their adherents. The Anderson-Burdick balancing test is now used in constitutional challenges to a wide variety of election laws, including ballot access, blanket primaries and even voter ID. What’s not commonly recognized is that this legal standard originated in voting-as-association cases.
Applying the voting-as-association standard
It’s true that the Supreme Court hasn’t yet applied the Anderson-Burdick standard to partisan gerrymandering. In fact, the court has been maddeningly unclear about what legal standard should apply in these cases. But for several reasons, the standard emerging from the voting-as-association cases provides the best fit for evaluating partisan gerrymandering claims.
The first is that the First Amendment right of association best captures the type of injury alleged, specifically the lasting harm to non-dominant political parties and their adherents arising from the dominant party’s self-entrenchment. Expressive-association cases have long focused on such harms, not only to the non-dominant party but to our political system. By contrast, equal protection law doesn’t accord any special status to political party affiliation. Unlike race or sex, party affiliation isn’t a protected class under the equal protection clause.
The second advantage of relying on the established voting-as-association standard is that it focuses on effects rather than intent. Recall that the Anderson-Burdick standard requires courts to weigh the “character and magnitude” of the burden on voting and association against the state’s asserted interests. An intent to harm the non-dominant party may be relevant, but it isn’t required. That’s a good thing, because intent is notoriously hard to prove or disprove, especially in redistricting cases. The Shaw v. Reno line of racial-gerrymandering cases exemplifies this difficulty. Under those cases, the pivotal question is whether race was the “predominant factor” in drawing a particular district. Twenty-four years after Shaw, the Supreme Court is still struggling to explain what this means. An effect-based test is preferable. Though Anderson-Burdick is hardly a bright-line rule, its balancing standard has proven manageable in other voting contexts and can be adapted to partisan-gerrymandering claims.
This brings me to a third advantage of relying on the voting-as-association cases to assess partisan gerrymandering claims: It provides a nuanced legal standard. The Anderson-Burdick balancing approach would allow lower courts to sort through the evidence, striking down the most egregious and unjustified partisan gerrymanders without categorically prohibiting any consideration of party affiliation when drawing districts. Of course, partisan-gerrymandering claims demand hard judgments. There are no bright lines here. But the legal standard that the Supreme Court has long used in voting-as-association cases provides the best fit for partisan-gerrymandering cases like Gill v. Whitford.