Democracy demands participation, usually in the form of voting. But, with voters as with mice, even the best-laid plans can go askew.1×1. Robert Burns, To a Mouse (1786), reprinted in The Best Laid Schemes: Selected Poetry & Prose of Robert Burns 47, 48 (Robert Crawford & Christopher MacLachlan eds., Princeton Univ. Press 2009); cf. John Steinbeck, Of Mice and Men (1937). Rather unlike in the case in Burns’s poem, the burden of unexpected barriers to voting often falls unequally across voters, disproportionately affecting poor voters and voters of color. See Daniel Weeks, Why Are the Poor and Minorities Less Likely to Vote?, The Atlantic (Jan. 10, 2014), https://www.theatlantic.com/politics/archive/2014/01/why-are-the-poor-and-minorities-less-likely-to-vote/282896 [https://perma.cc/S7V4-U6DD]. Recently, in Mays v. LaRose,2×2. 951 F.3d 775 (6th Cir. 2020). a Sixth Circuit panel upheld an Ohio law that provides an exception to the state’s deadline for absentee ballot applications to electors unexpectedly hospitalized, but fails to do the same for electors unexpectedly jailed.3×3. See id. at 780. Under the banner of judicial deference, the court rested its decision on an exercise of judicial imagination not warranted under the Anderson–Burdick framework used to assess many voting rights claims.4×4. See id. at 783–84; see also Burdick v. Takushi, 504 U.S. 428, 434 (1992); Anderson v. Celebrezze, 460 U.S. 780, 789 (1983). In accepting Ohio’s justification for its disparate treatment of hospitalized and late-jailed electors, the court relied on its own determination of the required scope of the opportunity to vote owed electors, in defiance of Sixth Circuit precedent that sought to prevent states from discriminating among voters.
The Ohio Code requires an absentee ballot application to be delivered to the Secretary of State no later than noon three days before an election if mailed, or by 6:00 p.m. on the Friday preceding the election if delivered in person.5×5. Mays v. LaRose, No. 18-cv-1376, slip op. at 5 (S.D. Ohio Nov. 6, 2019) (citing Ohio Rev. Code Ann. § 3509.03(D) (LexisNexis 2013)). In practice, however, the Secretary of State accepts all applications handed in by noon three days before an election, whether hand delivered or mailed.6×6. Id. A distinct provision of the code provides accommodations for disabled or confined electors, a category that includes by statutory acknowledgement both electors who are hospitalized and those who are incarcerated.7×7. Id. at 5–6 (citing Ohio Rev. Code Ann. § 3509.08(A) (LexisNexis 2020)). This provision permits the Board of Elections either to mail or to deliver an absentee ballot to electors who submit an absentee ballot application by noon on the third day before an election.8×8. Id. (citing Ohio Rev. Code Ann. § 3509.08(A)). In addition to these accommodations, Ohio provides a limited exception to the noon deadline for electors who are unable to travel to the polls due to the unexpected hospitalization of themselves or their minor children, allowing such electors to apply for an absentee ballot by 3:00 p.m. on Election Day.9×9. Id. at 7 (citing Ohio Rev. Code Ann. § 3509.08(B)).
Tommy Ray Mays II and Quinton Nelson Sr. were arrested the weekend before Tuesday, November 7 — Election Day — 2018.10×10. Mays, 951 F.3d at 780. Unable to post bail, they realized they would be imprisoned through Election Day and thereby prevented from voting in person, as they had planned.11×11. See id. at 780–81. The generally applicable deadline for voters to request an absentee ballot had passed, and neither Mays nor Nelson qualified under the hospitalization exception.12×12. See id. On Election Day, the pair brought a class action suit, challenging the law on equal protection and First Amendment grounds.13×13. Id. at 781. The district court granted the plaintiffs a temporary restraining order, allowing Mays and Nelson to cast their ballots in the 2018 elections.14×14. Id.
Nearly a year after the plaintiffs’ initial filing, the district court considered their suit’s merits.15×15. See Mays v. LaRose, No. 18-cv-1376, slip op. at 2 (S.D. Ohio Nov. 6, 2019). The district court certified the class and granted summary judgment in favor of the plaintiffs.16×16. Id. Determining the burden placed on the plaintiffs’ right to vote to be between “slight and severe,”17×17. Id. at 33. Judge Watson rejected the State’s justifications for the law.18×18. See id. at 34–38. Though the justifications provided a coherent explanation for the regulation’s existence, in the court’s view, they failed to justify the State’s disparate treatment of two groups of unexpectedly confined electors — those hospitalized and those jailed.19×19. See id. at 35–36, 42. In the absence of a sufficiently strong reason, the court held that the “Equal Protection [C]lause prevents the State from granting special treatment to one class of voters . . . without extending that same treatment to similarly situated voters.”20×20. Id. at 43.
A Sixth Circuit panel reversed and granted summary judgment in favor of the appellants.21×21. Mays, 951 F.3d at 780. Writing for the panel, Judge Nalbandian22×22. Judge Nalbandian was joined by Judges Thapar and Merritt. first held the Ohio law did not violate the Equal Protection Clause, determining that the State’s justifications for denying accommodations to late-jailed electors “outweigh[ed] the moderate burden” the law placed on the electors’ ability to vote.23×23. Mays, 951 F.3d at 783. Judge Nalbandian began the opinion’s equal protection analysis by noting that binding Sixth Circuit precedent required the court to analyze the case under the Supreme Court’s Anderson–Burdick framework.24×24. See id. (citing Obama for Am. v. Husted, 697 F.3d 423, 429 (6th Cir. 2012)). The court used the same framework to decide the plaintiffs’ First Amendment claims. Id. at 791–92. The court remarked in a footnote that the challenge sat uneasily within this framework, which tasks the court with weighing the burden a law places on a plaintiff’s right to vote against the state’s interests in enacting the law.25×25. Id. at 783 n.4. Courts apply a sliding scale of scrutiny, determined by the weight of the burden. See Burdick v. Takushi, 504 U.S. 428, 434 (1992); Anderson v. Celebrezze, 460 U.S. 780, 789 (1983). Courts use the Anderson–Burdick framework to evaluate both First Amendment and Fourteenth Amendment voting rights claims, just as the Mays court did. See Daniel P. Tokaji, Voting Is Association, 43 Fla. St. U. L. Rev. 763, 763 (2016) (citing Obama for Am., 697 F.3d 423). On the issue of burdens, the court voiced skepticism regarding the burden’s form, noting that the fact that Ohio makes it easier for some voters to vote does not “make it harder to vote for electors that don’t get the same benefit.”26×26. Mays, 951 F.3d at 784 n.4. Setting aside these concerns, the court proceeded, acknowledging the force of binding circuit precedent in the form of Obama for America v. Husted.27×27. 697 F.3d 423; see Mays, 951 F.3d at 783–84, 784 n.4 (“[W]e are bound by Obama for America . . . .” Id. at 784 n.4.).
Within the Anderson–Burdick framework, the court first set out to determine the severity of the burden placed on the plaintiffs’ right to vote.28×28. Mays, 951 F.3d at 784. On this initial point, the State seemed to have lost out. In its briefing, the State had argued that the court should consider the burden from the perspective of all electors, emphasizing the law’s selective burdening of the slim group of “only those voters who are jailed after the Saturday-noon deadline for absentee-ballot applications and remain jailed through Election Day.”29×29. Opening Brief of the Secretary of State of Ohio at 27, Mays, 951 F.3d 775 (6th Cir. 2020) (No. 19-4112). The court, however, insisted on considering the law’s “burdensome effects . . . from the perspective of only affected electors — not the perspective of the electorate as a whole.”30×30. Mays, 951 F.3d at 785.
But the State need not have lost heart — lose the battle, win the war. On the determination of the magnitude of the burden placed by the State on the plaintiffs’ right to vote, the court considered the entirety of the opportunities for voting available to the plaintiffs.31×31. Id. Because the plaintiffs had been part of the general electorate for all but the weekend before Election Day, the court reasoned, they had foregone a substantial number of opportunities to exercise the franchise prior to their arrest.32×32. Id. at 786. As a result, the court determined the deadline placed a “moderate” burden on the plaintiffs’ right to vote.33×33. Id.
Turning to the State’s justifications for the law,34×34. Id. at 787. the court credited the State’s argument that logistical considerations justified disparate treatment of hospitalized electors and late-jailed electors.35×35. See id. at 787–91. The court emphasized that providing accommodations to late-jailed voters would detract from Election Boards’ “ability to accomplish . . . tasks necessary to preserving the integrity of the election process . . . to a much greater degree than allowing for late applications from hospital-confined electors.”36×36. Id. at 787. And, having recognized the draw of the State’s logistics-based argument, the court then rejected the plaintiffs’ argument that jail-confined electors and hospitalized electors were similarly situated for the purposes of the Equal Protection Clause.37×37. Id. at 788. The court identified as a relevant difference between the two groups the location of their confinement38×38. Id. and highlighted appellate courts’ longstanding recognition that “[p]risoners are not similarly situated to non-prisoners.”39×39. Id. (alteration in original) (quoting Roller v. Gunn, 107 F.3d 227, 234 (4th Cir. 1997)).
Though the district court had refrained from deciding the First Amendment issue, the Sixth Circuit panel exercised its discretion and proceeded.40×40. Id. at 791. Relying on the reasoning offered in its equal protection Anderson–Burdick analysis, the court determined that the State’s interest in election administration outweighed the “much lower burden” placed on the plaintiffs’ First Amendment rights by the generally applicable absentee ballot deadline.41×41. Id. at 792.
Lastly, the Sixth Circuit reversed the district court’s certification of the plaintiffs’ class. The court found that the class satisfied neither the commonality nor the typicality requirement of Federal Rule of Civil Procedure 23,42×42. Id. at 793–94. as it included both individuals arrested before the noon absentee ballot application deadline and those arrested after.43×43. Id. at 794. The court concluded that the two groups had “substantially different claims” because only those arrested after the deadline were treated differently than unexpectedly hospitalized electors under Ohio law.44×44. Id. at 793; see id. at 793–94.
In Mays, the Sixth Circuit yielded too quickly to Ohio’s cry of logistics. In accepting the State’s argument that resource constraints meant providing applications to late-jailed voters would prevent boards of elections from completing other tasks “necessary” to election administration,45×45. Id. at 788. the panel unfaithfully applied circuit precedent. Eight years prior, in Obama for America v. Husted, the Sixth Circuit had recognized that equal protection challenges to voting rights laws like the one at issue in Mays demand two modes of analysis,46×46. Obama for Am. v. Husted, 697 F.3d 423, 428 (6th Cir. 2012) (describing two applications of the Equal Protection Clause: when a state “classifies voters in disparate ways,” and when a state “places restrictions on the right to vote”). guided in part by the Supreme Court’s reasoning in Bush v. Gore.47×47. 531 U.S. 98 (2000) (per curiam); see Obama for Am., 697 F.3d at 428. Despite claiming to adhere to the principles set forth by Obama for America, however, the Mays court deviated from this understanding. Rather than observe the Obama for America framework, the Sixth Circuit in Mays relied solely on a conventional application of the Anderson–Burdick framework, which is unsuited to address the challenges posed by laws like Ohio’s.
In Obama for America, the court entertained an equal protection challenge to an Ohio voting statute that extended early in-person voting for military voters but not otherwise.48×48. Obama for Am., 697 F.3d at 426–27. The Sixth Circuit found that Ohio’s change to its early voting regime likely violated the Equal Protection Clause for purposes of a preliminary injunction.49×49. Id. at 436. Relying heavily on the Supreme Court’s analysis in Bush v. Gore, the court applied a more demanding version of the Anderson–Burdick test, combining the conventional Anderson–Burdick approach with Bush v. Gore’s skepticism of classifications that arbitrarily differentiate between voters.50×50. See id. at 428 (citing Bush, 531 U.S. at 104–05); see also Richard L. Hasen, Essay, The 2012 Voting Wars, Judicial Backstops, and the Resurrection of Bush v. Gore, 81 Geo. Wash. L. Rev. 1865, 1883 (2013) (“The appellate court seemed to meld a stricter Bush v. Gore voting as a ‘fundamental right’ standard with the flexible Anderson–Burdick balancing test . . . .” (citing Obama for Am., 697 F.3d at 428–32)); Michael T. Morley, Prophylactic Redistricting? Congress’s Section 5 Power and the New Equal Protection Right to Vote, 59 Wm. & Mary L. Rev. 2053, 2111 (2018) (remarking on the Obama for America court’s reliance on Bush v. Gore). By contrast, the Mays court failed to cite Bush v. Gore once. This revised version of the Anderson–Burdick test led the Obama for America court to undertake a searching examination of the state’s justifications for the law at issue51×51. See Obama for Am., 697 F.3d at 432–36. — an inquiry notably absent in Mays.
The Obama for America court rested its reasoning on broad principles of equality pertinent to the equal protection analysis of voting laws.52×52. See id. at 429–36. Professor Samuel Issacharoff has recognized the importance of proffered reasons in the equal protection analysis inaugurated by Obama for America. Samuel Issacharoff, The Supreme Court, 2012 Term — Comment: Beyond the Discrimination Model on Voting, 127 Harv. L. Rev. 95, 104–07 (2013). Issacharoff notes: [T]he new equal protection of the right to vote expanded judicial scrutiny beyond the constricted categories of outright denial of the franchise and protection of vulnerable minorities against mistreatment on account of race or some other specified characteristic. . . . In practical terms, this doctrine meant that Ohio was free to alter the conduct of elections, but that the combination of a suspected constriction of voting opportunities and a lack of substantial reasons for it would be constitutionally fatal. Id. at 105. But see Edward B. Foley, Essay, Voting Rules and Constitutional Law, 81 Geo. Wash. L. Rev. 1836, 1846 (2013) (describing Obama for America’s equal protection jurisprudence as based on “selective non-retrogression”); Richard L. Hasen, Essay, Race or Party: How Courts Should Think About Republican Efforts to Make It Harder to Vote in North Carolina and Elsewhere, 127 Harv. L. Rev. F. 58, 73 (2014) (describing Judge White’s Obama for America concurrence as adopting “nonretrogression”). In order to meet the demands of the Equal Protection Clause, a state’s discriminatory burdening of an individual’s right to vote must be due to reasons that explain both the existence of the burden and its discriminatory nature.53×53. See Obama for Am., 697 F.3d at 432. The Obama for America court was clear in its worry that to demand otherwise would be to permit states “to pick and choose among groups of similarly situated voters to dole out special voting privileges.”54×54. Id. at 435. The Sixth Circuit also remarked: If the State had enacted a generally applicable, nondiscriminatory voting regulation that limited in-person early voting for all Ohio voters, its “important regulatory interests” would likely be sufficient to justify the restriction. However, Ohio’s statutory scheme is not generally applicable to all voters, nor is the State’s justification sufficiently “important” to excuse the discriminatory burden it has placed on some but not all Ohio voters. Id. at 433–34 (citation omitted) (quoting Burdick v. Takushi, 504 U.S. 428, 434 (1992)); see also Richard Pildes, Early Voting and Constitutional Law, Election L. Blog (Nov. 27, 2012, 4:25 PM), http://electionlawblog.org/?p=44801 [https://perma.cc/LX3J-38PB] (“[W]hat Ohio wanted to do the weekend before the election was to have polling locations that were open, but to turn most voters away while letting a select group of voters through the doors.”).
This caution rests on good reason. By itself, the Anderson–Burdick framework is ill-suited to prevent a state’s institution of underinclusive accommodations. The test’s focus on burdens obscures voters’ right to equal treatment by the state, as distinct from the right to the opportunity to vote.55×55. In this sense, the Anderson–Burdick framework’s exclusive focus on burdens functions as a threshold test — as long as a state refrains from restricting a voter’s opportunity to vote beyond some threshold, a state’s differential provision of the opportunity to vote fails to offend the Equal Protection Clause under the Anderson–Burdick approach. However, separate from an interest in the opportunity to vote, voters retain an interest in equal treatment in the provision of such opportunities. See Bush, 531 U.S. at 104–05 (“The right to vote is protected in more than the initial allocation of the franchise. Equal protection applies as well to the manner of its exercise. Having once granted the right to vote on equal terms, the State may not, by later arbitrary and disparate treatment, value one person’s vote over that of another.”). But see Joseph Fishkin, Equal Citizenship and the Individual Right to Vote, 86 Ind. L.J. 1289, 1324 (2011) (defending the Anderson–Burdick framework). In viewing the burdens on a group of voters in isolation, the Anderson–Burdick test fails to appreciate that, in some cases, the broader context of a given restriction on voting is relevant. As the Obama for America court articulated, in some cases, the reigning equal protection worry is not that the state is unfairly undercutting a group’s right to vote, but that the state is impermissibly favoring that of another group; hence the court’s application of a standard distinct from the orthodox variety of Anderson–Burdick.56×56. Professor Richard Hasen has described the Obama for America court’s standard as “read[ing] Bush v. Gore as a kind of free-floating license to do equity in election cases.” Hasen, supra note 50, at 1896; see also Samuel Issacharoff & Richard H. Pildes, Politics as Markets: Partisan Lockups of the Democratic Process, 50 Stan. L. Rev. 643, 700–01 (1998) (arguing in favor of an approach to voting rights that looks to background conditions rather than solely focuses on individual rights). Because voting is, in part, a matter of relative power, both the excessive burdening and the privileging of groups’ right to vote matter.
Indeed, in its opinion, the Mays court voiced concern about the fit of the Anderson–Burdick framework for equal protection claims;57×57. Mays, 951 F.3d at 783 n.4. such an awareness, however, failed to materialize into a faithful application of the Obama for America court’s reasoning. Rather than turn its attention to the worry of the state doling out privileges to favored groups, the Mays court relied on a narrow view of the burden on the plaintiffs’ right to vote. The court looked solely to the fit between the State’s justifications and the burden on the plaintiffs’ right to vote,58×58. See id. at 787–91. ignoring that such an inquiry neglects to consider situations in which the State has selectively privileged certain groups of voters.
If the Mays court had correctly applied the Obama for America standard, it would have joined the Obama for America court in dismissing election administration alone as an adequate justification for discriminatory burdens on voting rights.59×59. See Obama for Am., 697 F.3d at 432–34 (rejecting logistics as justification for burden). Consider, first, the optical illusion–like behavior of the State’s justifications in Mays. All tasks undertaken by boards of elections require the expenditure of resources — in that sense, any activity completed by Ohio’s regional election boards can be characterized as coming at the cost of other activities these boards could have opted to pursue with the “limited resources devoted to elections.”60×60. Mays, 951 F.3d at 787. Similarly, Ohio’s interest in preventing the overburdening of its election officials could not explain the resemblance between the State’s general absentee ballot application deadline and Swiss cheese — if overburdening is the concern, why not probe the empirical bases of the State’s selective deadline exceptions?61×61. Some courts have begun adopting this approach, frowning upon a state’s assertion of an interest in preventing voter fraud while failing to provide empirical evidence of voter fraud. See Thomas v. Andino, No. 20-CV-01552, 2020 WL 2617329, at *20 (D.S.C. May 25, 2020); DSCC v. Simon, No. 62-CV-20-585, 2020 WL 4519785, at *27 (Minn. Dist. Ct. July 28, 2020). Out of the possible ways the State could have opted to act in response to worries about resource allocation, the justification of logistics offered no explanation as to why it chose to act in the “precise”62×62. Crawford v. Marion Cnty. Election Bd., 553 U.S. 181, 203 (2008) (plurality opinion) (quoting Burdick v. Takushi, 504 U.S. 428, 434 (1992)). By contrast to the granular justification demanded in Obama for America, the Supreme Court in Crawford v. Marion County Election Board, 553 U.S. 181, acknowledged that the efficacy of Indiana’s voter identification law may be “debatable,” but upheld the restriction anyway, in part because of its general, nondiscriminatory veneer. Id. at 196; see id. at 203. fashion it did.
In determining whether hospitalized voters and late-jailed voters were similarly situated for the purposes of the Equal Protection Clause, the Mays court again fell short of the Obama for America standard. The court in Obama for America had considered whether military and nonmilitary voters were similarly situated.63×63. Obama for Am., 697 F.3d at 435. In its inquiry, the court looked to the underlying reasons provided by Ohio for the accommodation for military voters, in addition to arguments concerning the allocation of resources.64×64. See id. By contrast, the court in Mays asked not once for an explanation as to why Ohio provided the deadline extension for hospitalized voters.
In fairness, the Mays court did take care to distinguish Obama for America from the case at hand.65×65. See Mays, 951 F.3d at 790–91. The court rested its differentiation between the two cases on evidence presented by the parties involved in both cases, emphasizing that the State in Mays had demonstrated that some election boards potentially would have difficulty providing late-jailed voters with absentee ballot applications and that the Mays plaintiffs had not shown that “providing jail-confined electors the ability to request an absentee ballot at the last minute would ease the State’s burden on Election Day,” as the expansion of early voting in Obama for America had done.66×66. Id. These distinctions, however, failed to speak to the underlying “rationales”67×67. Id. at 790. of Obama for America, which required the court to ask whether a state’s interest in election administration justified its provision of a selective accommodation, rather than a nondiscriminatory provision.
As the facts of Mays suggest, time may well be the king of all men.68×68. Cf. William Shakespeare, Pericles act 2, sc. 3, l. 45. Mays and Nelson suffered a burden because of the timing of their arrest, a factor over which they had little control. Yet nowhere in democratic theory has time’s fickleness alone countenanced the installation of kings. The standard articulated by the Obama for America court emphasized that the state’s obligation to provide the franchise must proceed in a manner that treats equally the voters to whom the state is beholden. By contrast, the Mays court changed course for the Sixth Circuit’s treatment of equal protection challenges to election laws. The court’s approach, while perhaps more traditional, failed to capture the entirety of the equality interests at issue in Mays, highlighting the ways in which the basic Anderson–Burdick framework is ill-suited for analysis of underinclusive accommodations.