Section 2 of the Voting Rights Act of 19651 (VRA) provides that no state practice “shall be imposed . . . in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race.”2 In the years post–Shelby County v. Holder,3 section 2 is perhaps the most powerful instrument for challenging racially discriminatory voting laws.4 The heightened interest in bringing vote denial claims under section 25 has forced courts to come up with a manageable standard for evaluating them. But while some courts have mostly settled on a test, it lacks the Supreme Court’s imprimatur and many argue it’s far too broad to stand.6 Recently, in Democratic National Committee v. Hobbs,7 the Ninth Circuit, sitting en banc, enjoined two Arizona laws on the grounds that they violated section 2 of the VRA.8 In its opinion, the court implicitly rejected the Department of Justice’s (DOJ) proposed framework for vote denial claims. This framework would require plaintiffs to show the challenged law produced both a disparate impact and a substantial or material burden on voting. But a material burden standard improperly blends the constitutional standard for racially discriminatory election laws with the statutory requirements of the VRA. Though alternatives to the current standard are worth considering given the uncertainty around the doctrine, the DOJ’s proposed standard would hobble efforts to challenge a host of laws that the VRA was enacted to combat.
Shortly before the 2016 general election, the Democratic National Committee (DNC), along with the Democratic Senatorial Campaign Committee (DSCC) and the Arizona Democratic Party (ADP), filed suit in the United States District Court for the District of Arizona asking for an injunction preventing the State from enforcing two separate Arizona voting laws: the State’s out-of-precinct (OOP) policy and Arizona House Bill 2023 (H.B. 2023).9 The OOP policy required counties to cancel the votes (including those for statewide and federal office) on a provisional ballot cast outside of the voter’s assigned precinct,10 while H.B. 2023 made it a felony for a third party to collect a voter’s mail ballot.11 The plaintiffs claimed both laws had a disproportionate effect on minority voters in contravention of section 2 of the VRA.12 The plaintiffs also argued that H.B. 2023 violated the Fifteenth Amendment.13
For vote denial claims, plaintiffs can show a violation of section 2 in one of two ways. The first is to show the challenged law was enacted with the intent to discriminate.14 This test — called the “intent test” — is coextensive with the guarantees of the Fifteenth Amendment.15 The second is to show that the law produced a discriminatory result.16 Most courts, including the Ninth Circuit, have applied the aptly named “results test,”17 the first step of which asks whether there is a “disparate burden on members of the protected class.”18 If so, the second step asks “whether, under the ‘totality of the circumstances,’ there is a relationship between the challenged ‘standard, practice, or procedure,’ on the one hand, and ‘social and historical conditions’ on the other.”19 The second step is evaluated under the so-called “Senate factors,” a list of nine considerations in the Senate Report accompanying the 1982 version of the VRA.20
Following a ten-day bench trial, the district court denied both of the VRA claims under the results test.21 With regard to the OOP policy, the district court found that the plaintiffs did not carry their burden under step one because they failed to show either that the policy caused a disproportionate number of minority voters to go to the wrong precinct or that the racial disparities led to “meaningful inequality” in the political process.22 The district court also found that the impact of H.B. 2023 was too trivial a burden to constitute a violation of section 2 under step one.23
The Ninth Circuit affirmed.24 Writing for the panel, Judge Ikuta held that the district court appropriately found that the plaintiffs did not meet their burden under the VRA.25 The panel, like the district court, found that both claims failed at step one and that there was not enough evidence to show that the Arizona legislators voting for the passage of H.B. 2023 were motivated by racial animus.26 The plaintiffs filed a petition for rehearing en banc, which the Ninth Circuit granted.27
Sitting en banc, the Ninth Circuit reversed.28 Writing for the majority, Judge Fletcher enjoined both the OOP policy and H.B. 2023’s ban on ballot collection under section 2 of the VRA.29 The court first found that the OOP policy ran afoul of section 2’s results test. Under the first step of the test, the court accepted the expert report introduced at trial and concluded that the OOP policy produced a disparate burden on minority voters.30 Judge Fletcher noted that Arizona had a notably high rate of rejection for provisional ballots, and that many were rejected for having been cast out of precinct.31 The court gave three reasons for this high rate of rejection: “[F]requent changes in polling locations; confusing placement of polling locations; and high rates of residential mobility.”32 Each of these factors had a much greater effect on minority voters, and therefore, the OOP policy itself disproportionately impacted minority voters.33 The court also rejected the district court’s conclusion that the OOP policy did not cause the discriminatory result.34
The majority also found that the policy violated the second step of the results test. The court examined the Senate factors the district court considered at trial.35 Judge Fletcher noted that the fifth (the effects of discrimination on minority access to the ballot) and ninth (the justification given for the policy) factors were “particularly important” to the analysis.36 The court emphasized Arizona’s history of racial disenfranchisement, low turnout of minority residents compared to their white counterparts, racially polarized voting patterns, other occurrences of racial discrimination, racialized campaign rhetoric, and general unresponsiveness to concerns of minority residents, as well as the lack of diversity of Arizona’s elected officials and weakness of Arizona’s justification.37
Turning to H.B. 2023, the majority found that the ban on ballot collection also had a disproportionate effect on minority voters.38 Judge Fletcher wrote that “[i]n urban areas of heavily Hispanic counties, many apartment buildings lack outgoing mail services” and that “[o]nly 18 percent of American Indian registered voters have home mail service.”39 Moreover, issues such as mail security, lack of accessible transportation, and community isolation make third-party ballot collection all the more important to minority communities.40 And because the majority found that the number of ballots collected from minority voters was sufficient to constitute “meaningful inequality,” it held that the ban failed the first step of the results test.41 The court concluded the ban also failed step two for largely the same reasons as the OOP policy did. In its analysis of Senate factor nine, it noted that the State’s justifications for the law, preventing voter fraud and bolstering “public confidence in election[s],” were tenuous.42 There was no evidence of fraud in Arizona’s history, and Arizona already criminalized fraudulent ballot collection.43 Therefore, the ban violated section 2’s results test.44
The court also found that H.B. 2023 violated the intent test under section 2, as its passage was partially motivated by a discriminatory purpose.45 The court wrote that there were two reasons that Republican legislators passed H.B. 2023: “[U]nfounded . . . allegations” of voter fraud promoted by State Senator Don Shooter and a “racially-tinged” video.46 The “LaFaro Video” pictured a man who appeared to be of Hispanic heritage depositing ballots into a drop box.47 Nothing showed the man in the video was engaged in fraud, but the video was widely shared and even featured in a political ad.48 The court found that even though the legislators may have sincerely believed the ban was necessary to address ballot fraud, that belief was largely influenced by (and thus attributable to) the LaFaro Video and false and racially motivated allegations of fraud.49 And since the court had already concluded the law produced a discriminatory effect, it held that the plaintiffs had carried their burden under the intent test.50 Under Arlington Heights,51 Arizona had to show that the law still would have been adopted despite the discriminatory motivating factor.52 But Judge Fletcher found that the district court had already concluded that the law “would not have been enacted without Senator Shooter’s and LaFaro’s false and race-based allegations of voter fraud.”53 Therefore, H.B. 2023 ran afoul of section 2 of the VRA as well as the Fifteenth Amendment.54
Judge O’Scannlain dissented.55 He would not have found that the district court clearly erred in making its determination.56 Starting with the OOP policy, the dissenters disagreed that precedent required only that a plaintiff show a discriminatory effect.57 In their view, showing that the affected population had “substantial difficulty” in electing preferred representatives is critical to establishing a violation of section 2.58 And while the majority wrote that Congress — as well as Supreme Court precedent — had made clear that requirement applies only to vote dilution claims (like challenges to gerrymandering), rather than vote denial claims,59 the dissent replied that if only disparate impact were necessary, there would be virtually no limit to courts’ ability to enjoin any law for that reason.60 Furthermore, to Judge O’Scannlain, the alleged burden stemmed not from the consequence of voting out of precinct, which is what the challenged law prescribed, but the precinct system itself.61 Turning to the law on ballot collection, the dissent found that there was no clear error by the district court in concluding the plaintiffs did not carry their burden to prove a violation of section 2 because even though minority voters relied more on ballot collection, the plaintiffs didn’t show that minority voters now had “less opportunity” to cast their ballots than did nonminority voters.62 Finally, Judge O’Scannlain criticized the majority’s analysis of the legislature’s intent and reasoned that since at least some legislators did not harbor racial animus when voting for H.B. 2023, the district court did not clearly err in its finding.63
Judge Bybee penned a second dissent.64 He wrote to stress the importance of regulations as part of a comprehensive election code necessary for a well-functioning democracy.65 Specifically, Judge Bybee mentioned the long history of precinct-based voting in states around the country,66 noting that “[n]owhere in its discussion of the ‘totality of the circumstances’ has the majority considered that Arizona’s OOP provision is a widely held time, place, or manner rule.”67 Moreover, the majority’s implicit requirement that Arizona count the portions of the ballot for which the voter would have otherwise been eligible undervalues local elections and overvalues statewide and national elections.68 With respect to H.B. 2023, Judge Bybee emphasized “[t]here is no constitutional or federal statutory right to vote by absentee ballot.”69 And laws of general applicability designed to target potential voter fraud ought to be evaluated against the backdrop of genuine concerns about election integrity.70 The U.S. Supreme Court granted certiorari.71
As judges and scholars argue in the pages of court opinions and law reviews about the applicability of the results test,72 Hobbs brings several of these disagreements to the fore. But while the court homed in on the analysis under step one, it passed on the opportunity to engage substantively with another of the chief disputes in the case. In briefings and oral arguments, the DOJ advocated a new, elevated evidentiary standard for plaintiffs under section 2.73 Concerned about the potential for any law, however ordinary, to be struck down, it argued that section 2 requires plaintiffs to show not only that the challenged law produces a disparate impact, but also that it “materially burdens” the affected group’s ability to cast a ballot.74 But the efficacy of the test hinges on its application and, more specifically, which obstacles are and aren’t considered “material.” In applying the “material burden” standard to the facts of Hobbs, the DOJ and Arizona demonstrated that this proposed standard conflicts with other antidiscrimination laws and that adopting it would undermine the purpose of section 2.
The DOJ introduced its proposal in a time of relative unease about the state of the VRA. Prior to Shelby County, the preclearance regime under section 5 of the VRA prevented many voting laws with racially disparate impacts from going into effect.75 As Judge Fletcher noted in the court’s opinion, a previous iteration of Arizona’s ban on ballotcollection was derailed in 2011.76 But after the Supreme Court effectively ended preclearance,77 litigants were left with two main options for challenging laws that made it more difficult for minorities to vote. The first was to argue that the law was unconstitutional under Anderson-Burdick, a relatively high bar for plaintiffs.78 The second was to bring a vote denial claim under section 2. Since section 2, like preclearance, provides a mechanism for challenging racially discriminatory laws, many experts thought it could be used to reverse at least some of the effects of Shelby County on minority voters.79 But while standard vote dilution claims are well established under section 2, vote denial claims are not. Five circuits have settled on the two-step results test to adjudicate these claims, but the Supreme Court has yet to give the test its stamp of approval.80 As such, lawyers continue to debate the benefits and drawbacks of the burgeoning test.81
These disagreements ostensibly prompted the DOJ to articulate a new test that would require plaintiffs to prove a challenged law produces both a disparate impact and a “material burden.” In briefings and oral arguments, the DOJ expressed concerns about the current test shared by judges and academics.82 Simply put, if the test is based primarily on disparate effect, it would give courts unbridled authority to invalidate a host of legitimate election regulations that disproportionately impact minorities because of socioeconomic conditions that have nothing to do with the challenged law. Because of these concerns, in its amicus brief in support of Arizona, the DOJ encouraged the Ninth Circuit to evaluate these laws under a new standard: “Rather, the burden a challenged rule imposes on the right to vote [under section 2] thus must be not only disproportionate, but also material to the voter’s ability to vote . . . .”83 According to the DOJ, if a law violates section 2 only when members of a protected class “have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice,”84 then a practice is unlawful only if it produces a disparate result and “creates an impediment to the ability to vote that is not offset by other opportunities to register or vote.”85 The DOJ also argued that circuits that have adjudicated vote denial claims have at least implicitly endorsed this principle by upholding laws that merely imposed inconveniences.86
If the DOJ is concerned that some courts may apply the results test too permissively, there’s no reason to believe its own test would fare any better. It’s not clear that antidiscrimination law is or should be concerned with whether one discriminatory policy is “offset” by some other opportunity to achieve racial equality. In Connecticut v. Teal,87 a case concerning racial discrimination under Title VII of the Civil Rights Act, the Supreme Court held that “a racial disparity at one stage of the promotion process, which bars certain minority employees from becoming supervisors, cannot be offset by racial balance after the process has concluded.”88 The same logic applies here. The burden should be evaluated on its own terms, not in the context of some overall election scheme.89 The only consideration that should matter is the racial disparity produced by the OOP policy and H.B. 2023, not the other opportunities the state provides for a potentially different set of minority voters to cast a ballot.
The DOJ’s proposed standard also conflates the section 2 and constitutional standards. The DOJ appeared to liken the material burden analysis to the severe burden analysis under Anderson-Burdick.90 Both the DOJ and the State of Arizona argued that the plaintiffs should lose under the section 2 analysis because the district court correctly concluded that the burden produced by the OOP policy and H.B. 2023 is neither more nor less onerous than other election regulations and therefore not substantial enough to constitute a violation of the VRA.91 Although a “material” and a “severe” burden are not linguistically the same, the DOJ’s application of its standard to the facts of Hobbs indicates that it would operate in a substantially similar way to Anderson-Burdick — a test known for its inconsistent application.92 The major difference is that the materiality test asks whether the state has offset the burden and the Anderson-Burdick test asks whether the state has an interest in regulating the activity at issue.93
This pseudoconstitutional standard proposed by the DOJ doesn’t comport with the history or text of the VRA. Section 2 was revised to include the results test in 1982 after City of Mobile v. Bolden.94 There, the Supreme Court held that section 2 was coextensive with the guarantees of the Fifteenth Amendment.95 Following that decision, Congress repassed the VRA and added language to section 2 instructing courts to use a results-based analysis.96 In doing so, Congress expressed its clear preference for a results-based VRA framework different from the intent-based constitutional framework courts generally applied. But the DOJ’s proposed framework would elide the congressional purpose of creating a VRA-specific standard by again integrating a constitutional framework (this time, Anderson-Burdick) into the section 2 statutory analysis.
There are reasonable questions as to whether the results test adopted by several circuits is a good one. It’s not impossible to imagine how the most basic election laws could come under scrutiny under this test. As some scholars, such as Professor Nicholas Stephanopoulos, have suggested, a burden shifting mechanism such as those required in other areas of disparate impact law has built-in safeguards that prevent invalidating every piece of legislation simply because it produces a disparate impact.97 But even so, in the results test, the section 2 inquiry is properly focused on the law’s impact on the ability to participate in the electoral process, rather than the materiality of the burden or the voter’s ability to find other ways to cast a ballot. In other words, in a section 2 case, the disparate impact analysis should be doing most of the work, not the evaluation about the burden the law imposes. The Supreme Court held in Thornburg v. Gingles98 that in drafting the VRA, Congress “ma[de] clear that a violation could be proved by showing discriminatory effect alone and . . . establish[ed] as the relevant legal standard the ‘results test.’”99 In that vein, the test the DOJ offers does not effectuate the intent behind the VRA.