Mich. State A. Philip Rand Inst. v. Ruth Johnson

NOT RECOMMENDED FOR PUBLICATION File Name: 18a0461n.06 No. 18-1910 UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT MICHIGAN STATE A. PHILIP RANDOLPH ) FILED INSTITUTE; MARY LANSDOWN; ERIN ) Sep 05, 2018 COMARTIN; DION WILLIAMS; COMMON ) DEBORAH S. HUNT, Clerk CAUSE ) ) Plaintiffs-Appellees, ) ON APPEAL FROM THE ) UNITED STATES DISTRICT v. ) COURT FOR THE EASTERN ) DISTRICT OF MICHIGAN RUTH JOHNSON, in her official capacity as ) Michigan Secretary of State ) ) Defendant-Appellant. ) BEFORE: BOGGS, KETHLEDGE, and DONALD, Circuit Judges. BOGGS, Circuit Judge. This litigation commenced in 2016 with the filing of plaintiffs’ complaint attacking recently passed legislation, PA 268, which eliminated Michigan’s straight- ticket voting option. PA 268 aligned Michigan with the large majority of states that require voters to vote individually for numerous partisan offices, rather than allowing them to make one mark to select a party’s slate for many offices. Discovery was completed by September 29, 2017. The district judge then conducted a bench trial, though no witnesses were heard. The litigation was essentially conducted on voluminous documents and depositions. The district court issued a lengthy opinion on August 1, 2018, enjoining application of the Michigan law on two constitutional grounds and as a violation of Section 2 of the Voting Rights Act (“VRA”). Defendants filed a notice of appeal on August 13 and, pursuant to Fed. R. App. P. No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson 8(a)(1)(C), moved for a stay in the district court on August 14. On August 23, the district court denied that motion. On August 30, defendant-appellant Ruth Johnson, the Michigan Secretary of State, moved this court to stay or immediately reverse the district court’s order so that Michigan can proceed with its November election under PA 268 as the Legislature intended. The stay factors are four-fold: (1) the likelihood that the party seeking the stay will prevail on the merits—which, in the case of staying a permanent injunction, constitutes the likelihood of reversal; (2) the likelihood that the moving party will be irreparably harmed; (3) the prospect that others will be harmed by the stay; and (4) the public interest in the stay. Michigan Coal. of Radioactive Material Users, Inc. v. Griepentrog, 945 F.2d 150, 153 (6th Cir. 1991); see also Crookston v. Johnson, 841 F.3d 396, 398 (6th Cir. 2016); Coalition to Defend Affirmative Action v. Granholm, 473 F.3d 237, 252 (6th Cir. 2006). The likelihood of success is perhaps the most important factor. Of course, the plaintiffs bear the burden of proving their claims in this case. That means, for purposes of the motion here, that the Secretary must show a likelihood that the plaintiffs have not borne that burden. As stated above, the movant must show a likelihood of reversal. Michigan Coal. of Radioactive Material Users, 945 F.2d at 153. On appeal, we review the district court’s legal conclusions de novo and its factual findings for clear error. Ibid. “[A]ny time a State is enjoined by a court from effectuating statutes enacted by representatives of its people, it suffers a form of irreparable injury.” New Motor Vehicle Bd. of Cal. v. Orrin W. Fox Co., 434 U.S. 1345, 1351 (1977) (Rehnquist, J., in chambers). The public interest is in the proper application of federal and constitutional statutory provisions, and in “the 2 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson will of the people of Michigan being effected in accordance with Michigan law.” Coalition to Defend Affirmative Action, 473 F.3d at 252. The district court’s opinion is extensive, but its underpinnings are quite weak. As the district court noted, a number of states have eliminated straight-ticket voting in recent years. The first state to do so, Massachusetts, adopted the so-called “office block” ballot in 1888. More than twenty-five states did not permit straight-ticket voting before 1994. Currently, forty states other than Michigan do not permit straight-ticket voting,1 including Democratic-leaning states such as Massachusetts, California, New York, and Washington. Rhode Island—which, like Michigan, lacks early voting and no-excuse absentee voting—likewise banned straight-ticket voting in 2014. Act of July 1, 2014, 2014 R.I. Pub. Laws chs. 279, 280. As a logical matter, whichever party is in the minority in an area often favors the office-block system, hoping that some of its candidates may prevail if voters are “nudged” to consider their individual merits against a generally adverse partisan tide. So the public interest in allowing the Michigan legislature to make a public policy decision affecting a debate that has continued in America for over a century is quite weighty. The district court’s decision rests almost entirely on conclusions about the practical effect of requiring individual consideration of each office. First, the district court found that voting individually for each of the approximately eighteen partisan offices on Michigan’s ballot, rather than casting one straight-ticket vote, would take a voter three additional minutes or, alternatively, add 25% to the voter’s total voting time. Second, this increase in voting time for many individual 1 Straight Ticket Voting States, Nat’l Conf. St. Legislatures (Jan. 8, 2016), http://www.ncsl.org/research/elections-and-campaigns/straight-ticket-voting.aspx. 3 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson voters would in turn increase the time that it takes all voters in a precinct to vote once they arrive at the polling place. In other words, eliminating straight-ticket voting would make casting an individual ballot take longer, causing delays for all voters at a polling station. As to the first conclusion, the three-minute increase is essentially pulled out of the air. Christopher Thomas, the then-Michigan Director of Elections, simply stated that he and Bureau of Elections staff, in consultation with county and municipal clerks, estimated that figure. We have no information about how they reached that estimate or what data they considered, making it difficult to assess its reliability. As for the 25% figure—which the district court mentions frequently—its flaws are obvious. Associate Professor Theodore Allen, the plaintiffs’ expert and the source of the number, explained how he determined it: I personally timed myself on the whole ballot and the other one and I came up with that kind of ballpark and then I learned about other ballots that were much longer and there were 32 partisan races and so I wanted to come up with something that was simple, that was understandable, and that’s where I came up with that. Theodore Allen Dep. 240:22–241:3. He also had plaintiffs’ counsel time herself. Finally, he said: [O]ther people had mentioned reports from the experts and so it said that this would be more than one minute and it could be up to five minutes. So I mean . . . I heard about it, and so from that whole discussion I decided to come up with a simple formula that would in my mind be conservative. Theodore Allen Dep. 241:8–14. In any event, any number of policy decisions might influence the length of time it takes an individual voter to vote, in addition to, obviously, each voter’s own decisions. Having judicial elections increases that time. Having non-partisan elections, which cannot be affected by the 4 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson choice to have straight-ticket voting or not, adds to the time to vote. Allowing citizens to vote on legislation or propositions, by whatever means, adds to voting time. Deciding how many local offices to elect can add to that time: some states elect coroners, jailers, drain commissioners, and surveyors on partisan ballots. All of these are policy choices a state may legitimately make, and yet all would be subject to attack if individual voting time were a consideration that courts could use to strike down legislation. All these choices reflect a deliberate determination that it is better if voters are encouraged or required to make individual assessments of candidates, rather than mass choices. That may be a bad determination. Some political scientists think so. But most American states have made the opposite choice, not only as to straight-ticket voting for partisan races, but by allowing and requiring choices on a large number of other ballot items. That choice is not an arbitrary one, nor is it a “tenuous” interest. See Michigan State A. Philip Randolph Inst. v. Johnson, No. 16-cv- 11844, 2018 WL 3769326, at *37 (E.D. Mich. Aug. 9, 2018). The second basic pillar of the district court’s decision is an increase in total wait time. Once again, the district court’s findings lack rigor. The closest that the court came to providing even an estimate of the increase in wait time is where it mistakenly stated that Allen “found that the eradication of straight-ticket voting would increase wait times by 25% or more for every voter who previously voted a straight-ticket.” Michigan State A. Philip Randolph Inst., 2018 WL 3769326, at *9. The court then repeated this error in its denial of the Secretary’s Emergency Motion for a Stay Pending Appeal. See Michigan State A. Philip Randolph Inst. v. Johnson, No. 16-cv-11844, 2018 WL 4024895, at *4 (E.D. Mich. Aug. 23, 2018). Allen found no such thing. 5 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson As the district court properly recognized elsewhere in its own opinion, Allen’s 25% figure concerns voting time, not wait time. And critically, there is no necessary connection between these two values. If there are a number of open voting booths, for instance, then an increase in voting time would not affect wait time. At best, the district court referenced static wait-time figures; but these figures are insufficient for a variety of reasons. Most obviously, the fact that there is a wait time at a polling station does not answer what the cause of that wait is. And without knowing that, one cannot conclude that increased voting times will translate to increased wait time. As discussed below, all the data used to predict hypothetical wait times and their increase without straight-ticket voting derives from simulations based on observations (made by volunteers apparently recruited and trained by plaintiffs or their associates) at thirty-one Michigan precincts on Election Day 2016. Allen admits in his report that he “adjusted” the statistics—using a “simulation” about which we know nothing—due to recording errors by the volunteers. Beyond these obvious statistical shortcomings, the district court took almost no notice of the Michigan voting process. Once voters are checked in as eligible voters and given a ballot, they may proceed to one of a number of booths to mark their ballots. There is not a single file of voters queuing through a single location, as was, for example, the case in many states that used mechanical voting machines, had only one machine per precinct, and thus permitted only one voter at a time to vote.2 2 Most of the selected precincts are noted as using “Paper” ballots as just described. A few precincts, especially the three Detroit precincts, are noted as “Electronic.” This does not seem to have affected the progression from registration table to a voting booth in any way, though it is not clear if each booth had a dedicated device, or if the voter was issued a device. In any event, there is no indication from any of the parties, observers, or reports, that 6 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson The thirty-one precincts included in the record had between three and twenty-eight booths per precinct, generally related to the number of registered or anticipated voters. Obviously, any individual voter is not delayed at all in voting if there is an empty booth available when given a ballot, regardless of how long it takes the individual to vote—and, of course, the individual is entitled to ponder his or her choices. The volunteers observed and recorded data in the following manner. Every half-hour on the half-hour, a volunteer would note a voter (thus twenty-seven voters at most), who would then record that voter’s passage through the voting process: • Time of arrival on the line; • Number of people in line at that time; • Time of arrival at the registration table; • Time left the registration table; • Time entered the voting booth; • Time left the voting booth; • Time left the polling place. The volunteer also recorded the total number of voting booths at the polling place. Very significantly, the volunteer was not asked to record whether all the booths were full at any given time. In his deposition, Allen gave rather evasive explanations for not having the volunteers record this data—which would seem to be the most relevant fact for the current litigation. If booths were always available, then there is no possible wait-time problem. If only one or two were unoccupied out of a large number, then there is probably not a problem. If all booths were filled, or only one or two booths were unoccupied out of a small number of booths, it would be more plausible that this affected the observed wait times, though it could have affected, for good or ill, the length of time a voter took to record a vote. 7 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson extra voting time for an individual voter could cause a significant clog. But no such data was recorded. Again, to be clear, there is no evidence, from any source, as to the actual status of booths, as to how many were occupied at any time. Allen was quite cavalier in denying the need for that data: [M]y process and my sheet generation relate to my experience of doing simulation applications in many instances. . . . I would say probably I’ve done it more in non-election system environments. . . . This issue of whether booths are free or not is rarely relevant to things my clients are asking me about, and in a way you can just derive it from other considerations using math or even simulation. Theodore Allen Dep. 31:1–16. It would seem that the volunteers could very easily have recorded this crucial information, if Allen wanted to have actual data, rather than relying on his simulations. Similarly, it seemingly would have been very easy to have the affected election officials give some indication, by affidavit or deposition, of whether booths were fully occupied ever, some of the time, or all the time, and whether they delayed checking in voters if all booths were occupied. Yet there is no such evidence, only “simulations and adjustments.” For example, one of the major adjustments was that Allen reduced the number of available booths—in exactly the precincts with the longest wait times—“to adjust the number of location booths to tailor to the precinct level model,” and “to calibrate the simulation with the measured average waiting times.” These adjustments were quite large. As stated in the report, “[t]he adjustments were: 8 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson Detroit 1-271[:] 13, Flint[:] 5, and Saginaw[:] 11.” It remains unclear whether these “adjustments” are the number of booths that were arbitrarily removed, or the number of booths remaining after the adjustment. In either case, it would make a very big difference. The number of booths actually counted by the on-the-scene observers, and the number of registered voters in that particular precinct, were: Precinct Number of Number of Registered Booths Voters Detroit 1-271 15 1,505 Flint 28 1,714 Saginaw 25 2,772 Therefore, the model either used 2, 23, and 14—if the number of booths used was the original number minus the “adjustment”—or it used the “adjustment” numbers of 13, 5, and 11 as stated above. Either would be problematic, as “modeling” that the Detroit precinct had only two booths for over 1,500 registered voters seems unlikely, and yet, reducing Flint from 28 booths to only 5 for over 1,711 registered voters also seems excessive. So any adjustment to the booths actually observed seems unwarranted. Given the lack of data on booth availability, as well as Allen’s adjustments, the plaintiffs’ prediction of widespread bottlenecks caused by voters being unable to reach a booth because other voters were puzzling over additional choices seems implausible. Their likelihood of success on the merits of that claim is small. Moreover, actual use of straight-ticket voting is strongly correlated to the partisanship of an area, but not to race directly. As one might expect, strong partisans are more likely to abhor 9 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson voting for any candidates of the opposite party and more likely to use straight-ticket voting. Some of the evidence showed that strongly Republican areas had high rates of straight-ticket voting as well, though not as high as areas that were strongly Democratic and largely African-American. But there simply were few or no Republican areas that were as strongly Republican as the strong Democratic areas were Democratic. The Michigan House Fiscal Agency prepared a document analyzing proposed legislation eliminating straight-ticket voting (“STV”), which is included in the record as an exhibit to plaintiffs’ complaint. It compared STV rates in several of the large counties in the state in the election of 2014. Its chart showed that Ottawa County (which is almost wholly white and frequently the most Republican county in the state) used STV at a 60% rate. Wayne County, which includes Detroit, and many areas outside it, (and is usually the most Democratic county in the state) used STV at a 58% rate. It is clear that the rate of STV use in the areas of Detroit with the densest population of African-Americans is quite a bit higher, but that is a matter of degree, not of kind, and is clearly linked to strong partisanship, rather than race. The alleged evils of eliminating Michigan’s straight-ticket system seem unlikely to outweigh the ability of a state to make a public policy choice common across all fifty states. This makes it unlikely that the plaintiffs will prevail on the merits. The irreparable harm to voters in taking what would be at most very small additional time to register their choices, an additional time largely within the control of the voter, is very small. And the public interest in allowing states to control their own elections is quite strong, as the Constitution itself makes clear. See U.S. Const. art. I, § 4. 10 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson Our circuit’s most recent case raising similar issues was Ohio Democratic Party v. Husted, 834 F.3d 620 (6th Cir. 2016). That case involved Fourteenth Amendment and Voting Rights Act challenges to Ohio’s restrictions on early voting. Our court dissected the type of showing that a plaintiff must make, the ways of analyzing a state’s interest in its voting practices and democratic policy choices, and the role of the district court in requiring rigor on both sides of the equation. We cautioned district courts not to casually “become entangled, as overseers and micromanagers, in the minutiae of state election processes, without careful consideration.” Id. at 622. We also emphasized, citing Burdick v. Takushi, 504 U.S. 428, 434 (1992), that the rigorousness of the inquiry into state election law depends on the degree to which the challenged practice actually burdens protected voting rights. Ohio Democratic Party, 834 F.3d at 626–27. With those admonitions in mind, and in light of the factual discussion above, we turn to the three specific legal challenges raised. And we do so mindful of the cogent point made in the plaintiffs’ response to the stay motion—the Secretary must show a likelihood of success on each of the three claims raised. See Resp. of Pl.-Appellees to Def.-Appellant’s Emergency Mot. For Stay of Permanent Inj. Pending Appeal 6. The district court’s injunction is based on each of the three counts, and if the Secretary has little likelihood of success on any one of them, the stay must be denied. Count I — The Anderson-Burdick Framework As this court noted in Ohio Democratic Party, “[e]lection cases rest at the intersection of two competing interests[,]” namely, an individual’s right to vote versus a state’s prerogative to 11 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson regulate the right to vote. 834 F.3d at 626. To resolve these competing interests, we apply the Anderson-Burdick framework, which directs us to engage in a three-part analysis: [First, the court must] consider the character and magnitude of the asserted injury to the rights protected by the [Constitution] that the plaintiff seeks to vindicate. Second, it must identify and evaluate the precise interests put forward by the State as justifications for the burden imposed by its rule. [Third], it must determine the legitimacy and strength of each of those interests and consider the extent to which those interests make it necessary to burden the plaintiff’s rights. Id. at 626–27 (quoting Green Party of Tenn. v. Hargett, 791 F.3d 684, 693 (6th Cir. 2015)) (second alteration in original). Importantly, the rigorousness of this inquiry turns on the severity of the burdens imposed by the state on the plaintiffs’ First and Fourteenth Amendment rights. Burdick, 504 U.S. at 434. Where the restriction on a plaintiff’s right to vote is severe, the state’s “regulations survive only if ‘narrowly drawn to advance a state interest of compelling importance.’” Ohio Democratic Party, 834 F.3d at 627 (quoting Burdick, 504 U.S. at 434). In contrast, where the regulations are “minimally burdensome and nondiscriminatory, rational-basis review applies, and the regulations will usually pass constitutional muster if the state can identify ‘important regulatory interests’ that they further.” Green Party of Tenn. v. Hargett, 767 F.3d 533, 546 (6th Cir. 2014) (quoting Burdick, 504 U.S. at 434). Regulations that fall between these extremes—i.e., those that “impose a more-than-minimal but less-than-severe burden”—require us to “weigh[] the burden on the plaintiffs against the state’s asserted interest and chosen means of pursuing it.” Ohio Democratic Party, 834 F.3d at 627 (citation omitted). Although we review the district court’s factual findings for clear error, “the district court’s characterization of the resultant burden . . . is not a factual finding, but a legal determination 12 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson subject to de novo review.” Id. at 628. The district court made such a legal error here. As discussed above, the plaintiffs likely have not shown that PA 268 imposes more than a minimal burden on the ability to vote. And Michigan (like forty other states, presumably), has shown far more than a rational basis for its decision. When balancing the Anderson-Burdick factors, the plaintiffs are unlikely to succeed in demonstrating, as a matter of law, that the speculative burdens noted above outweigh a deliberate choice as to our system of governance. Count II — Intentional Discrimination As the Supreme Court made clear in Village of Arlington Heights v. Metropolitan Hous. Dev. Corp., 429 U.S. 252, 264–65 (1977), “official action will not be held unconstitutional solely because it results in a racially disproportionate impact.” Rather, to establish a violation of the Equal Protection Clause, “[p]roof of racially discriminatory intent or purpose is required . . . .” Id. at 265. That said, a plaintiff need not prove that the challenged action stemmed solely from racially discriminatory intentions. Ibid. To determine whether discrimination was a motivating purpose behind the challenged action, we engage in a “sensitive inquiry into . . . circumstantial and direct evidence of intent[.]” Ibid. Specifically, we consider the following “evidentiary sources”: [1] the historical background of the decision, particularly if it reveals a series of official actions taken for invidious purposes; [2] the specific sequence of events leading up the challenged decision; [3] departures from the normal procedural sequence; [4] substantive departures, particularly if the factors usually considered important by the decisionmaker strongly favor a decision contrary to the one reached; [5] and the legislative or administrative history, especially where there are contemporary statements by members of the decisionmaking body, minutes of its meetings, or reports. 13 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson Spurlock v. Fox, 716 F.3d 383, 397 (6th Cir. 2013) (alterations and internal quotation marks omitted). These factors overlap and are non-exhaustive. Ibid. In concluding that PA 268 contravened the Equal Protection Clause’s prohibition on intentional discrimination, the district court’s analysis again lacked rigor. For instance, in assessing the “historical background” factor, the court devoted half of its discussion solely to historical voting patterns, not to the historical background of the legislature’s decision. See Michigan State A. Philip Randolph Inst., 2018 WL 3769326, at *28. And where the court did discuss the legislative decision to enact PA 268, the court confined itself to discussing the views of a supporter who was not a legislator and who gave race-neutral reasons for supporting the bill, namely, that PA 268 is good policy and that it would help Republicans. Ibid. Neither reason, of course, indicates discriminatory intent. The remainder of the district court’s analysis is equally facile. In discussing the “specific sequence of events” factor, the court did not discuss any event, let alone a sequence of events, that led up to the decision. Instead, it simply noted that straight-ticket voting previously existed in Michigan, then cited the characterizations of two Democratic state senators, who presumably opposed the bill’s passage. Id. at *29. It should go without saying that we do not judge the intention of a bill’s supporters by the characterization of its opponents. And as for the “substantive departure” factor, the district court does not outline what is an “ordinary legislative action,” let alone how the passage of PA 268 departed from it. Instead, the court makes the non sequitur that because a lobbyist for clerk associations “encouraged clerks to ‘[t]ell [legislators] how many of [the clerks’] voters, both Republicans and Democrats, avail themselves of the straight-party 14 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson option[,]’ . . . the record suggests that the Michigan Legislature substantively departed from its usual conduct in enacting PA 268.” Id. at *30. For a lobbyist’s exhortations to be evidence of the conduct of other parties is quite a stretch. And the statements of those opposing any bill cannot be taken as a sign that the proponents intended the predicted parade of horribles. Such legislative arguments are just that—arguments, not facts. The parties clash mightily on whether Supreme Court political reapportionment cases can provide any guidance for cases such as ours, in which the challenged actions may seem to be affected by the intersection of racial and partisan considerations. In such cases as Thornburg v. Gingles, 478 U.S. 30 (1986), and Shaw v. Reno, 509 U.S. 630 (1993), the Court was quite clear that political action taken to advantage or disadvantage a political party or faction does not in itself show racial discrimination. That issue is, of course, relevant here, where there is no facial distinction in how PA 268 treats voters, as opposed to the gerrymandering cases, where the drawing of district lines does affect voters of different races differently in every instance. There is strong reason for importing the Supreme Court’s analysis to cases such as ours. In her emergency motion requesting a stay, the Secretary argues that, because partisan discrimination is a defense to a racial-gerrymandering claim, the district court erred by “conflating racial and political discrimination simply because there is an alleged correlation between Democratic Party affiliation and the African-American population.” When the Secretary raised this same argument below, the district court rejected it, holding that the court applies a standard in gerrymandering cases that is different from the standard applied in “voting rights cases.” Michigan State A. Philip Randolph Inst., 2018 WL 4024895, at *5. Specifically, a plaintiff alleging 15 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson intentional discrimination in a gerrymandering case must prove that race was the “predominant consideration” when drawing district lines. Ibid. In contrast, a plaintiff alleging intentional discrimination by any other government conduct need only prove that race was a “motivating factor.” Id. at *6. Although a plaintiff challenging a gerrymander must meet a higher standard than a plaintiff challenging other government conduct (including a voting law), both of plaintiffs’ claims arise under the Equal Protection Clause of the Fourteenth Amendment. Compare Miller v. Johnson, 515 U.S. 900, 904 (1995) (gerrymander), with Arlington Heights, 429 U.S. at 265–66 (zoning laws). Thus, the Supreme Court’s interpretation of the Equal Protection Clause in cases challenging a gerrymander are relevant in cases challenging a voting law. And in the gerrymandering cases, the Supreme Court has repeatedly warned against using party affiliation and race as proxies. To do otherwise, the Court has said, would be to “reinforc[e] the perception that members of the same racial group—regardless of their age, education, economic status, or the community in which they live—think alike, share the same political interests, and will prefer the same candidates at the polls.” Shaw, 509 U.S. at 647. The Court has “rejected such perceptions elsewhere as impermissible racial stereotypes.” Ibid. (listing cases). But we need not resolve that conflict at this stage. We need only say that, in combination with the other factors mentioned, in the ultimate disposition of this appeal, the Secretary is very likely to show that the plaintiffs have not met their burden of demonstrating that PA 268 was racially motivated. 16 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson Count III — Voting Rights Act In 1982, following the Supreme Court’s decision in City of Mobile v. Bolden, 446 U.S. 55 (1980), Congress amended the Voting Rights Act to make clear that intentional discrimination was not a prerequisite for a claim brought under the statute. Ohio Democratic Party, 834 F.3d at 636. As amended, Section 2(a) of the VRA prohibits a state from imposing a voting qualification or prerequisite to voting or standard, practice, or procedure … 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 or color[.] 52 U.S.C. § 10301(a) (emphasis added). The statute establishes that a voting practice or procedure can “result in” such a denial or abridgement even if there is no proof of discriminatory intent: A violation of subsection (a) is established if . . . the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. Id. § 10301(b). In determining whether a violation of Section 2(a) occurred, we consider the totality of the circumstances. Ibid. Complicating matters somewhat is the fact that the VRA encompasses two conceptually distinct claims—an important consideration, because our jurisprudence is not equally well developed for each. First, the VRA prohibits “vote dilution” by states, i.e., the practice of “den[ying] minorities an equal opportunity ‘to elect representatives of their choice[.]’” Ohio Democratic Party, 834 F.3d at 636 (quoting 52 U.S.C. § 10301(b)). As we have previously noted, “[t]he majority of cases interpreting Section 2 arose in the vote-dilution context, epitomized by the 17 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson Supreme Court’s decision in Thornburg v. Gingles . . . .” Ibid. Second, the statute prohibits the practice of “vote denial,” which, as the name suggests, involves “the denial of [the] opportunity to ‘participate in the political process.’” Ibid. (quoting 52 U.S.C. § 10301(b)). Critically, “[w]hile vote-dilution jurisprudence is well-developed, numerous courts and commentators have noted that applying Section 2’s ‘results test’ to vote-denial claims is challenging, and a clear standard for its application has not been conclusively established.” Ibid. The language of the statute alone can hardly be read to cover a change in ballot format that applies to all voters, that keeps no one away from the polls, and that prevents no one from registering their vote. So, faced with uncertainty in how to extend the language in the vote-denial context, we have applied a two-step framework when assessing such claims. Id. at 637. First, we examine whether a voting practice “resulted in an adverse disparate impact on protected class members’ opportunity to participate in the political process.” Ibid. To satisfy this element of the test, it must be shown not only that there is some statistical discrepancy between minority groups and whites, but that the challenged practice also “causally contributes to the alleged discriminatory impact by affording protected group members less opportunity to participate in the political process.” Id. at 638. Second, provided that the initial step is satisfied, we consider the totality of the circumstances, “potentially informed by the ‘Senate Factors’ discussed in Gingles.” Ibid. (emphasis added). More specifically, at this stage of the analysis, we consider whether the challenged voting practice “causes the discriminatory impact as it interacts with social and historical conditions.” Ibid. 18 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson The Gingles factors include: 1. the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process; 2. the extent to which voting in the elections of the state or political subdivision is racially polarized; 3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti- single shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group; 4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process; 5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process; 6. whether political campaigns have been characterized by overt or subtle racial appeals; 7. the extent to which members of the minority group have been elected to public office in the jurisdiction. Gingles, 478 U.S. at 36–37. Additional factors that “in some cases have . . . probative value” are: “whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group” and “whether the policy underlying the state or political subdivision's use of such voting qualification, prerequisite to voting, or standard, practice or procedure is tenuous.” Id. at 37. 19 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson Here the challenged practice is simply the adoption of the “office-block” system, requiring a vote for each individual office. In granting a permanent injunction, the district court held that factors one, three, and four were not relevant to this case. Michigan State A. Philip Randolph Inst., 2018 WL 3769326, at *33 n.16. The reasoning for this is not readily apparent, and the absence of any of those factors would seem to point against invalidation of PA 268. With respect to factor six, in support of its claim that “political campaigns in Michigan have been characterized by overt or subtle racial appeals,” id. at *35, the district court cited a comment made at a Tea Party meeting and an ambiguous comment by a former Republican state senator. The district court also drew upon the nationwide discussion of race-related issues to conclude that factor six favors the plaintiffs—a standard that, if accepted, would result in factor six always favoring plaintiffs bringing a VRA claim. As for factor seven, the court glossed over the facts that President Obama has won the state of Michigan twice and that many elected judges—including the Chief Justice of the Michigan Supreme Court—are African-American (to say nothing of two current United States Representatives and twenty-three members of the Michigan Legislature) in concluding that this factor is neutral. Id. at *36. In short, even if we eventually reach the second step and consider the Gingles factors, there is a significant likelihood that the result on appeal would point in the opposite direction to the district court’s analysis. 20 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson CONCLUSION As shown above, there are very serious problems with both the factual underpinnings and the legal analysis of the district court’s opinion. At this stage, we are not required to render a final decision on any of these points. In a stay motion, we examine the likelihood of success of the movant’s appeal (and thus the likelihood of reversal) in conjunction with the other stay factors. The Secretary has demonstrated a likelihood of reversal. For Count I, the record shows that the district court likely erred in reaching its legal conclusion that plaintiffs had proven that the burden imposed by PA 268 was “not severe, but also not minimal.” Johnson, 2018 WL 3769326, at *27. For Count II, the district court likely committed clear error when it concluded that the plaintiffs had demonstrated by a preponderance of the evidence that the Michigan Legislature intentionally discriminated against African Americans. And for Count III, even were we to reach the second step and analyze the Gingles factors, the district court likely committed clear error by finding that the plaintiffs had proven a violation of Section 2 of the VRA. We note that voter turnout is invariably less in midterm elections such as are impending than in the presidential election that generated most of the data, which should further reduce the likelihood of any significant harm. But when we balance all of the factors as laid out above, we hold that the balance favors granting the stay. The Secretary’s motion is hereby GRANTED. 21 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson KETHLEDGE, Circuit Judge, concurring. Election-law cases sometimes prompt strong reactions, driven (often expressly) by a sense that these laws are aimed specifically at reducing African-American turnout and thus in spirit no different from the laws of Jim Crow. I share the district court’s desire to strike down every law with that aim, but I respectfully disagree that the Michigan law here is one of them. That law, known as PA 268, simply requires voters to choose candidates for each office individually, rather than vote for a party slate en masse. Forty other states have already enacted the same kind of law, including most recently Rhode Island. None of them have ever been declared unconstitutional. And one can hardly dispute, as Judge Boggs makes clear, that the Michigan law, like all the rest, is supported by good policy reasons. Of course, some mix of policy and partisan reasons likely lay behind the enactment of PA 268. The same is likely true of most laws. But where the district court was most clearly mistaken, I respectfully submit, was in equating partisan motives with racial ones. See Mich. State A. Philip Randolph Inst. v. Johnson, No. 16-cv-11844, 2018 WL 3769326, at *30 (E.D. Mich. Aug 9, 2018). So far as I can tell, the Supreme Court has never equated those two things; to the contrary, in any number of cases, the Court has demanded more than partisan motives to support a finding of racial intent. See, e.g., Miller v. Johnson, 515 U.S. 900, 914-16 (1995); Shaw v. Reno, 509 U.S. 630, 647-48 (1993). Here, the record shows that, in Michigan, typically about 60% of straight-ticket voters are Democrats, with about 40% Republicans. That disparity might have made Republican legislators more open to the virtues of candidate-by-candidate voting than their Democratic colleagues were. But the record gives us no reason to think those legislative views would have been any different if, say, white Democrats voted straight-ticket in much higher percentages than 22 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson black Democrats do. What matters, so far as we can tell, is that more Democrats than Republicans use the straight ticket, not the racial composition of straight-ticket voters within the Democratic ranks. On this record, then, the remedy for anyone unhappy with PA 268 is not another constitutional ruling from the federal courts. The remedy instead is to wait the extra 20 minutes or so (according to the district court’s estimate) in line at the polls, and then vote to turn out the state legislators who supported the law. With these observations, I fully join Judge Boggs’s opinion. 23 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson BERNICE BOUIE DONALD, Circuit Judge, dissenting. This case involves one of our most sacred and fundamental individual rights—without which all others are meaningless—to exercise one’s constitutional right to vote. The majority finds that Michigan legislation, PA 268 does not disproportionately or negatively impact African American voters, that it does not violate the equal protection clause of the Fourteenth Amendment nor Section 2 of the Voting Rights Act. I dissent. Were we writing on a clean slate, without the backdrop of a historical legacy of disenfranchisement of African Americans, one might reasonably come to that conclusion. But this court’s conclusion, no matter how finely parsed nor eloquently written, ignores the 150 years of shameful and painful history of disenfranchisement, suppression, and dilution of African American voters and the overt and covert mechanisms used to achieve that objective. It also ignores the fact that Michigan began efforts to dismantle this practice immediately after the passage of the 1964 Civil Rights Act, whose very purpose was to outlaw discrimination and to increase participation by African Americans in American society. Voter disenfranchisement, suppression, and oppression of African Americans is woven into the fabric of America. The dispositive inquiry before us—at this stage—is whether the Secretary has met her burden in satisfying the stringent four-factored test to stay the permanent injunction issued by the district court. The majority finds that bar low. In the face of the Declaration of Independence, the U.S. Constitution, Civil Rights Statutes, and a developed body of jurisprudence, African Americans have had to wage a more than 200- year battle to get and keep the voting franchise. See Dred Scott v. Sandford, 60 U.S. 393 (1857) 24 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson (Citizenship is a prerequisite to voting). In December 1865, the Thirteenth Amendment was ratified which abolished slavery, but it did not provide citizenship or real protection for newly freed people. In 1868, after massive racial violence, the Fourteenth Amendment was passed and later ratified. This Amendment provided for citizenship, due process, and equal protections, among other things for “male persons” born or naturalized. In 1870, the Fifteenth Amendment to the United States Constitution was passed to prevent states from denying the right to vote on grounds of “race, color, or previous conditions of servitude.” U.S. Const. amend. XVI. Almost immediately, states began enacting laws and supporting practices to disenfranchise African American voters. In the immediate aftermath of Reconstruction, states of the former Confederacy immediately passed Jim Crow laws and amendments to disenfranchise black voters. They imposed poll taxes, literacy tests, and other restrictions to disenfranchise African Americans. Notwithstanding the explicit language of the Fourteenth and Fifteenth Amendments, federal courts, including the United States Supreme Court, generally sanctioned these discriminatory practices. See Giles v. Harris, 189 U.S. 475 (1903); Giles v. Teasley, 193 U.S. 146 (1904); contrast Smith v. Allwright, 321 U.S. 649 (1944); Lane v. Wilson, 307 U.S. 268 (1939) (where the court found a twelve-day one-time voter registration window to be discriminatory against black citizens and repugnant to the Fifteenth Amendment). It was not until 1964, ninety-six years after ratifying the Fourteenth Amendment and ninety-four years after ratification of the Fifteenth Amendment, that the poll tax was eliminated as a precondition to voting in a Presidential election. U.S. Const. amend. XXIV. Similar conditions 25 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson were prohibited to the states in 1966. Harper v. Virginia State Bd. of Elections, 383 U.S. 663 (1966). In 1965, Congress passed the Voting Rights Act, 100 years after the passage of the Thirteenth Amendment. This chronology illustrates the long struggle and disenfranchisement of African American voters. PA 268, as the District Court found, disparately impacts African American voters and therefore, constitutes a denial of equal protection and constitutes a violation of Section 2 of the Voting Rights Act. Since 1981, one hundred and twenty-seven years, Michigan has permitted the practice of straight-party voting. I begin with the—perhaps coincidental—timing in which the Michigan legislature embarked on its first attempt to eliminate straight-party voting—1964. This timing is indeed significant and warrants our attention to the historical practices employed by states to continuously impose massive and unconstitutional barriers to exclude African-Americans from the franchise. During the Reconstruction era, many states enacted laws with carefully selected disenfranchising crimes to disqualify a disproportionate number of African-American voters, to Jim Crow—marked by comprehensive and well-disguised systems of racialized social and legal control—the right to vote for African-Americans has been tumultuous. See e.g. Hunter v. Underwood, 471 U.S. 222, 233 (1985) (finding that an Alabama statute was intentionally adopted to disenfranchise blacks on account of race and that it had its intended effect). In response, in 1965, Congress enacted the Voting Rights Act “to address entrenched racial discrimination in voting, ‘an insidious and pervasive evil which had been perpetuated in certain parts of our country through unremitting and ingenious defiance of the Constitution.’” Shelby Cty. 26 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson v. Holder, 570 U.S. 529, 535 (2013) (quoting South Carolina v. Katzenbach, 383 U.S. 301, 309 (1966). Yet, this ideal has been historically difficult to attain, as many of the most “insidious” forms of discriminatory barriers affecting racial minorities are facially guised in neutrality. See Vanessa Williams, Ill-fated plan to close polling places in Georgia county recalled lingering prejudice, Wash. Post (Aug. 24, 2018), (Rudolph County Georgia Board of Elections’ proposal to close two-thirds of the polls in a predominantly African-American community just weeks ahead of a crucial mid-term election). I, unlike the majority, will not turn a blind eye to this historical relevance. I. Standard of review To stay a district court’s order for injunctive relief, “we consider the same four factors that are traditionally considered in evaluating the granting of a preliminary injunction.” Mich. Coal. of Radioactive Material Users, Inc. v. Griepentrog, 945 F.2d 150, 153 (6th Cir. 1991). “These well-known factors are: (1) the likelihood that the party seeking the stay will prevail on the merits of the appeal; (2) the likelihood that the moving party will be irreparably harmed absent a stay; (3) the prospect that others will be harmed if the court grants the stay; and (4) the public interest in granting the stay.” Id. (citations omitted). “These factors are not prerequisites that must be met, but are interrelated considerations that must be balanced together.” Id. (citing In re DeLorean Motor Co., 755 F.2d 1223, 1229 (6th Cir. 1985). On the merits, we review the district court’s order issuing a permanent injunction for abuse of discretion. Howe v. City of Akron, 801 F.3d 718, 753 (6th Cir. 2015) (citing United States v. Miami Univ., 294 F.3d 797, 806 (6th Cir. 2002)). In doing so, We have long recognized that the 27 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson “abuse of discretion standard of review is highly deferential.” Hardyman v. Norfolk & W. Ry. Co., 243 F.3d 255, 267 (6th Cir. 2001). We have repeatedly acknowledged that the abuse of discretion standard “is more than the substitution of the judgment of one tribunal for that of another.” NLRB v. Guernsey-Muskingum Elec. Co-op., Inc., 285 F.2d 8, 11 (6th Cir. 1960). In other words, “[u]nder the abuse of discretion standard, an appellate court may overturn a lower court's ruling only if it finds that the ruling was arbitrary, unjustifiable or clearly unreasonable.” Hardyman, 243 F.3d at 267 (alteration in original) (citing Plain Dealer Pub. Co. v. City of Lakewood, 794 F.2d 1139, 1148 (6th Cir.1986)). Because the determination of whether the movant is likely to succeed on the merits is a question of law, it is accordingly reviewed de novo. Liberty Coins, LLC v. Goodman, 748 F.3d 682, 689 (6th Cir. 2014). Considering each of these factors, I do not find a stay appropriate. A. Likelihood of Reversal on Appeal As the majority correctly notes, we must first consider whether the Secretary has demonstrated that “there is a likelihood of reversal” on appeal. While “the factors to be considered are the same for both a preliminary injunction and a stay pending appeal, the balancing process is not identical due to the different procedural posture in which each judicial determination arises.” Mich. Coal. of Radioactive Material Users, Inc., 945 F.2d at 153. With respect to the latter: [A] motion for a stay pending appeal is generally made after the district court has considered fully the merits of the underlying action and issued judgment, usually following completion of discovery. As a result, a movant seeking a stay pending review on the merits of a district court’s judgment will have greater difficulty in demonstrating a likelihood of success on the merits. In essence, a party seeking a stay must ordinarily demonstrate to a reviewing court that there is a likelihood of reversal. Presumably, there is a reduced probability of error, at least with respect 28 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson to a court’s findings of fact, because the district court had the benefit of a complete record that can be reviewed by this court when considering the motion for a stay. Id. (emphasis added). Therefore, to succeed in her request for an emergency stay of the district court’s permanent injunction, the Secretary must demonstrate a likelihood that the district court’s ruling was “arbitrary, unjustifiable or clearly unreasonable” as to each constitutional and statutory violation found, otherwise, we are precluded from entering a stay of the district court’s permanent injunction order. Id. The Secretary makes no such showing. Yet, the majority circumvents this standard of review by failing to accord the proper degree of “high deference” to the district court’s factual findings. The majority characterized the district court’s opinion as “extensive, but … weak and attacks the alleged differences in the proof. The district court considered the testimony of witnesses and made credible findings. Those findings should be accorded great deference. It must be noted that this comes on a motion for extraordinary relief where the case record has not yet been fully developed and discovery is incomplete. The secretary chose the timing of this action and strategically created the emerging conditions under which the district court and this court must act. The very pace of the litigation in the aftermath of the district court’s order has constituted manufactured a emergency. I find that the district court made detailed factual findings which amply support its conclusions of law. 1. Equal Protection Clause Challenge The majority embarks on an outcome-determinative de novo review of the district court’s factual findings. In doing so, it focuses a considerable amount of its attention on undermining the analysis conducted by one expert in the case, Professor Allen. Putting aside the impropriety of 29 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson dismissing the district court’s factual conclusions based on purported issues with a single expert’s opinions, the majority’s myopic attacks on the substance of Professor Allen’s opinions are misplaced. Professor Allen’s pedigree cannot be called into doubt. He holds a Ph.D. from the University of Michigan in Industrial and Operations Engineering, teaches classes entitled “Applied Waiting Line Analysis” and “Statistical Modeling, Queueing, and Lean Production” at Ohio State University, consults regularly with counties across the nation regarding resource decisions for elections, and has authored a textbook and several published articles specifically addressing wait times in election lines. More importantly, his opinions in this case draw upon and bespeak his experience. The majority expresses grave concerns about Professor Allen’s methodology: specifically, that he did not account for how many booths were in use at the time his team recorded data at the 31 precincts’ polling places. Despite all of the data points Professor Allen captured (from which one can surmise whether a bottleneck occurred at the registration table or while waiting for a voting booth), the majority contends that we can never know if the elimination of straight-ticket voting would increase voting times unless we know exactly how many booths were being used and how many booths were open at the time a voter was ready to vote. This concern is misplaced for two reasons: first, it does not explain what could cause lines at the voting booth other than full voting booths, and second, it disregards Professor Allen’s reliance on his own experience and other authoritative 30 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson materials in making the assumption that bottlenecks occur primarily at voting booths and not registration tables. First, given the data Professor Allen did collect, the majority does not explain satisfactorily why knowing how many booths are available helps in analyzing the impact of the statute on wait times. Professor Allen’s data distinguishes between wait times at the registration table and at the voting booth. That information is sufficient for our purposes. From it, one can determine exactly how long an individual had to wait to get into a voting booth after they finished registration, and the only plausible reason an individual would have to wait to vote after they have finished at the registration table is because there are no available booths. The majority offers no explanation to the contrary. The question of how many booths were available while individuals waited to vote is a red herring. Second, the majority ignores the fact that Professor Allen testified and opined in his expert report that, based on his extensive experience in this field, wait times are usually caused by bottlenecks at the booths and not the registration table. Professor Allen notes specifically in his report that if he were to consider bottlenecks at “other service processes like registration or [check- in] . . . , then the total waiting times could only increase.” By concluding that Allen failed to account for potential bottlenecks at other areas of the voting process, the majority has replaced Professor Allen’s decades of experience with their own conjecture. Next, the majority takes issue with Professor Allen’s correction of data that was improperly recorded by his team. As Professor Allen explained in his deposition, he encountered reporting from his team that did not match his instructions: “they wrote down not what I wished that they 31 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson would have written down, they wrote down what they believed was accurate, and in that sense it was accurate, but they didn’t understand my objectives.” He corrected the data to make the information uniform; it would have been problematic to leave the data as recorded. Of course, Professor Allen’s opinions are not perfect. Perhaps Professor Allen could have employed a more rigorous methodology from which to attain his estimate that eliminating straight-ticket voting would increase wait times by 25%. Nevertheless, perfection is not the standard. The majority overstepped its role by conducting a de novo review of Professor Allen’s opinions; and, further, the majority made a number of errors in its own analysis of the same. Moreover, at its core, “[t]he right to vote is a ‘precious’ and ‘fundamental’ right,” Obama for Am. v. Husted, 697 F.3d 423, 428 (6th Cir. 2012) (quoting Harper v. Va. State Bd. of Elections, 383 U.S. 663, 670 (1966), and it is clear that this right “is protected in more than the initial allocation of the franchise. Equal protection applies as well to the manner of its exercise.” Id. (quoting League of Women Voters of Ohio v. Brunner, 548 F.3d 463, 477 (2008)). Specifically, “[t]he Equal Protection Clause applies when a state either classifies voters in disparate ways, or places restrictions on the right to vote.” Id. (citations omitted). In analyzing a challenge to a voting restriction based on the Equal Protection Clause, we apply, as did the district court, the framework established by the Supreme Court in Burdick v. Takushi, 504 U.S. 428, 434 (1992), and Anderson v. Celebrezze, 460 U.S. 780, 788–89 (1983). See Green Party of Tenn. v. Hargett, 791 F.3d 684, 692 (6th Cir. 2015). Under the Anderson- Burdick test, we first consider the character and magnitude of the injury to the protected right. Id. at 693 (quoting Anderson, 460 U.S. at 789). Next, we identify and evaluate the interests proffered 32 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson by the state. “Finally, [we] must determine the legitimacy and strength of each of those interests and consider the extent to which those interests make it necessary to burden the plaintiff’s rights.” Ohio Democratic Party v. Husted, 834 F.3d 620, 626-27 (6th Cir. 2016). The level of scrutiny applied changes with the severity of the burden imposed. Green Party of Tenn., 791 F.3d at 692. Where the restriction of a right to vote is severe, “the statute will be subject to strict scrutiny and must be narrowly tailored and advance a compelling state interest.” Id. (quoting Burdick, 504 U.S. at 434). “On the other hand, ‘minimally burdensome and nondiscriminatory’ regulations are subject to a ‘less-searching examination closer to rational basis[,]’ and ‘the State’s important regulatory interests are generally sufficient to justify the restrictions.’” Ohio Democratic Party, 834 F.3d at 627 (quoting Ohio Council 8 Am. Fed’n of State v. Husted, 814 F.3d 329, 335 (6th Cir. 2016)). “If the burden lies somewhere in between, courts will ‘weigh[ ] the burden on the plaintiffs against the state's asserted interest and chosen means of pursuing it.’” Green Party of Tenn., 791 F.3d at 693. Based on its finding that PA 268 will disproportionately deter African Americans from voting, the district court applied the intermediate standard. In contrast, the majority, after undertaking a de novo-like review of the district court’s factual findings, substituting its own judgment for that of expert witnesses, and ignoring the objective statistical impact of PA 268 on African Americans, determined that rational basis was the correct standard. They then state that Michigan had shown “far more than a rational basis for its decision.” However, in coming to their conclusion, the majority pays only lip service to the applicable standard of review. Because it seems necessary, I remind the majority that, “[i]n 33 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson reviewing the district court’s findings for clear error, we may not substitute our judgment for that of the district court and ‘must uphold the [district] court’s account of the evidence if it ‘is plausible in light of the record viewed in its entirety.’” Ohio Democratic Party, 834 F.3d at 644 (quoting Lee v. Willey, 789 F.3d 673, 680 (6th Cir. 2015)). The majority, in its continued de novo review of the district court’s factual findings, reasons that because registration is also a component to voting, it is inconclusive whether increased wait times are attributable to the voting booth or the registration process. Therefore, the majority concludes that the district court erred in its finding that the elimination of straight-party voting will increase wait times, and thus disproportionately burden African-Americans. This reasoning misses the point. The record is clear that, with the lone exception of the Secretary’s expert, Paul Herrnson—who the district court found to lack credibility—it is indisputable that the elimination of straight-party voting will amount to increased wait times for voters in Michigan. Specifically, the district court relied on the expert testimony of Kurt Metzger, who, after studying all Michigan counties in 2012, 2014, and 2016, found that nearly half of all Michigan voters used the straight-party option in 2016. However, in communities where African- Americans constitute less than 40% of the voting-age population, the percentage of voters who use the straight-party option decreased to 46.5%. Yet, in communities where voting-age African- Americans are the majority demographic, 77.7% of voters use the straight-party voting option, and 34 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson 68% of voters used the straight-party option in communities where African-Americans made up 40 to 49.9% of the voting-age population.1 The majority attempts to refute Metzger’s finding of a strong association between straight- party voting and race by pointing to straight-party popularity in communities, such as Ottawa county, where African-Americans makeup no more than 1% of the population. The majority reasons that because 56% to 64% of voters in these communities also use the straight-party option, race is not the driving factor for straight-party voting. Yet, as highlighted by the district court, this reasoning is unconvincing as it fails to explain why 100 percent of majority African-American communities in Michigan experience exceptionally high straight-party voting rates. Accordingly, the district court reasonably relied on Metzger’s expert analysis in finding that “African-Americans are more likely to use the straight-party voting option and will be disproportionately affected by its elimination.” Despite this, the majority usurps district court’s role as fact-finder, and rests on its own conclusion that straight-party voting “is clearly linked to partisanship, rather than race.” (Maj. Op. at 10). I am not convinced. More importantly, the majority fails to so much as mention that the district court “clearly erred” in its factual findings. In her alternative argument, the Secretary asserts that even if eradicating straight-party voting imposes a substantial burden on Michigan voters, PA 268 is facially neutral, and thus, any burden imposed by it is a burden on all voters. Accordingly, the Secretary concludes that it does 1 The Secretary argues that the expert report of Paul Herrnson accurately concludes that PA 268 will not impact wait times. However, the district court rejected Herrnson’s testimony as completely unreliable as it contradicts all other expert reports indicating the PA 268 will impose increased wait times to some degree. 35 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson not run afoul of the Equal Protection Clause. This argument diverges from the principled purpose and application of the Equal Protection Clause in the voting rights context. The district court record is replete with evidence supporting its determination that—based on the conditions in Michigan as they exist today—the eradication of straight-party voting proffered by PA 268 would impose a substantial burden on the voting rights of African-Americans. Particularly, the district court relied on the testimony of several Michigan election officials and clerk asscociations that PA 268 would introduce substantial wait times, predicting a “doomsday- like situation.” Particularly, testimony from the then-Chair of the House of Elections Committee showed that longer wait times would be a “nightmare,” absent the enactment of no-reason absentee voting and would lead to “disaster.” Other elections clerks testified that they had “grave concerns over the lines that [PA 268] would create on voting day.” Finally, the district court relied on testimony from Christopher Thomas, who was the Director of the Michigan Bureau of Elections for thirty-six years. Thomas testified that, in his opinion, it would take at least three to five additional minutes to complete a ballot without a straight-party option. Applying the applicable law to the facts above, along with its finding that African- Americans vote the straight-party option at a significantly greater rate, the district court reasonably concluded that in those communities, African-Americans will disproportionately suffer a substantial increase in wait times, resulting in increased voter deterrence and incomplete ballots. The majority emphasizes that most states have eliminated straight-party voting in its argument that Michigan’s eradication of the same does not offend the Equal Protection Clause. However, we are obliged to ask, not whether other states have validly enacted similar laws, but 36 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson whether—under the jurisdiction-specific factors at play in the state of Michigan—the eradication of Michigan’s 127-year practice of permitting straight-party voting runs afoul of constitutional or statutory protections. Indeed, “comparing the isolated voting practice of one state with the isolated voting practice of another state is not always an apples-to-apples comparison.” Mich. State A. Philip Randolph Inst. v. Johnson, 833 F.3d 656, 665 (6th Cir. 2016) (“Johnson II”). In Johnson II, the Secretary petitioned this Court for an emergency stay of the district court’s July 21, 2016, preliminary injunction of PA 268. As we correctly reasoned in Johnson II, “[i]n assessing the burden imposed on voters by a state's electoral mechanisms, courts may undoubtedly consider whether the state's practices will cause long lines and delays at polling places and how these lines and delays may impact the right of a voter to cast his or her ballot.” Id. at 665 (citing Brunner, 548 F.3d 463, 477-78)). Specifically, the Johnson II court noted that while Michigan may indeed join most states in eliminating straight-party voting, it is of amount only a few states to do so while also prohibiting early voting and restricting absentee voting, making the average wait time at physical polling locations of tremendous significance to Michigan voters.” Id. Applying the appropriate standard of review, this Court held that the Secretary failed to establish a likelihood that the district court abused its discretion. Our reasoning in Johnson II applies with added force today: “[h]ow a state chooses to regulate the manner that a person must cast a vote undoubtedly impacts the individual right. Id. at 666. More importantly, the issue presented to us is not—as the Secretary and the majority conclude—simply the extra time that each straight-party voter will have to spend making additional selections on the ballot. Nor is it simply the longer lines at polling places resulting from 37 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson increased wait time. Rather, it is the fact, as supported by the record, that the longer lines will deter a disproportionate number of African-Americans from voting. Id. at 670 (Gilman, J., concurring). As a purely factual matter, the record reflects (1) that African-Americans in Michigan overwhelmingly vote in the democratic party; (2) that African-Americans in Michigan use straight- party voting at significantly higher rates; and (3) that eradicating straight-party voting would significantly increase wait times at the poll. It follows from logic that PA 268 will disproportionately deter African-Americans in Michigan from voting. The district court properly found that PA 268 imposes a substantial burden on African-Americans. Thus, the Secretary has not met her burden of establishing that the district court abused its discretion. Finally, as its third alternative argument against the district court’s finding of an Equal Protection violation, the Secretary asserts that analyzing the burden that PA 268 imposes on African-Americans “is not cognizable under the Fourteenth Amendment.” In support, the Secretary relies on Washington v. Davis, 426 U.S. 229, 242 (1976). However, Davis involved an alleged discriminatory employment practice, not a challenge in the voting context. Id. at 229. We analyze a challenge to a state’s law affecting the right to vote under the specific Anderson-Burdick framework. Thus, the Secretary’s argument must fail. In balancing the burden imposed on African-American voters with the state’s interest, the district court concluded that while the state’s purported interest in (1) encouraging voters to examine each candidate without regard to political party, (2) increasing the likelihood that voters will complete the non-partisan section of the ballot, and (3) reducing confusion about how to 38 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson complete a ballot was “reasonable” it was not “sufficiently weighty” to justify the effect that PA 268 imposes on the voting rights of African-Americans. (citing Obama for America v. Husted, 697 F. 3d 423, 433 (6th Cir. 2012)). Moreover, while a state may regulate the voting process, the Supreme Court has stated that “[t]he right to associate with the political party of one’s choice is an integral part of the basic constitutional freedom.” Kusper v. Pontikes, 414 U.S. 51, 57 (1973) (citing Williams v. Rhodes, 393 U.S. 23, 30 (1968)). The Secretary points to nothing apart from vague speculation that suggests that a voter will make a more informed choice in filling in each individual bubble rather than choosing to fill in one bubble for a straight-party vote. See Johnson II, 833 F.3d at 666. In light of the record, the district court’s conclusion was reasonable. Thus, the Secretary has not established that there is a likelihood that the district court abused its discretion in weighing the burden imposed on voters with the states interest. Moreover, the Secretary argues that the district court erred in concluding that the Michigan legislature intentionally discriminated against African-Americans, without first finding a showing of racial animus. Stated plainly, the Secretary asserts that the district court’s failure to make a specific finding of racial animus by the Michigan legislature precludes an Equal Protection Clause violation. Id. At best, the Secretary’s argument is misguided and is not supported under the law. The district court made factual findings that the overwhelming majority of African- Americans in Michigan staunchly support the Democratic Party. Importantly, the district court noted that Michigan legislatures recognized the straight-party voting pattern of African-Americans in passing PA 268. The district court’s conclusion is buttressed by the Fourth Circuit’s recent 39 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson decision in N.C. State Conf. of the NAACP v. McCrory, 831 F.3d 204, 222 (4th Cir. 2016), holding that “intentionally targeting a particular race's access to the franchise because its members vote for a particular party, in a predictable manner, constitutes discriminatory purpose.” The majority flatly ignores what the Fourth Circuit noted that “[t]his is so even absent any evidence of race-based hatred and despite the obvious political dynamics.” Id. at 222-23. Thus, the Fourth Circuit concluded—as did the district court here—that “[a] state legislature acting on such a motivation engages in intentional racial discrimination in violation of the Fourteenth Amendment and the Voting Rights Act.” Id. The Secretary further asserts that, in concluding that Michigan legislature enacted PA 268 with racially discriminatory motivation, the district court erred by considering statements made by Republican Party chair McDaniel that she supported PA 268 in part because eliminating straight- party voting “would help republicans win [elections],” and she “wanted to win.” The Secretary relies on Hunt v. Cromartie, 526 U.S. 541, 551-52 (1999) for her argument that political-based discrimination is a defense to, and defeats, a finding of racial discrimination. I am not persuaded that such a conclusion is supported by Hunt. The Secretary’s position conflates this Court’s analysis in the context of voting rights discrimination—as presented here—with that applied in the political gerrymandering context in Hunt. Moreover, it appears the Secretary’s political argument is a proxy for race discrimination. The district court considered more than just the statements of McDaniel. The district court also acknowledged statements by Senator Knollenberg, expressing his indifference to any wait 40 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson time that PA 268 could impose on voters and comparing voters who must wait in long lines to people “in countries who don’t have the right to vote:” [t]o those in countries who don’t have the right to vote, I assume how long it takes isn’t on their list of concerns . . . It is time that Michigan’s elections process becomes more about people, less about political parties, and even less about how long it takes to exercise one of our most fundamental rights. The Secretary correctly notes that “official action will not be held unconstitutional solely because it results in a racially disproportionate impact.” Spurlock v. Fox, 716 F.3d 383, 397 (6th Cir. 2013) (internal quotations omitted). However, “evidence of a policy’s disparate impact may be probative in determining whether the policymaker harbored a discriminatory intent.” Id. (citing Vill. of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 97 (1977)). Furthermore, a party need not show that discriminatory intent was “the dominant or primary” factor, only “that a discriminatory purpose has been a motivating factor in the decision.” Id. at 265-66. Upon consideration of the legislative history and substantial departures, the district court reasonably found that eliminating the democratic party’s success with straight-party voting— success especially driven by African-Americans residing in communities with high voting-age African-Americans—was a motivating consideration in the Michigan Legislature’s enactment of PA 268. Accordingly, the Secretary fails to demonstrate a likelihood of reversal on appeal as to counts I or II of the district court’s finding of Equal Protection Clause violations. 2. Voting Rights Act Challenge The majority correctly points out that in determining whether a violation of Section 2(a) of the VRA occurred, we must consider the totality of the circumstances. However, when reviewing the district court’s decision, the majority failed to apply this standard. Rather than showing how 41 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson the district court’s conclusion was clearly erroneous, the majority points to a select few “Senate Factors” and cites to new evidence contrary to the district court’s finding, concluding that on appeal the outcome would go in the opposite direction. While the majority blames the lack of jurisprudence for “vote-denial” claims under the VRA for its troubles in applying Section 2’s “results test”, it is the incorrect application of the standard of review and erroneous use of de novo review has led to the majority’s confusion. Although, as the majority points out, the jurisprudence regarding a vote-denial claim is has not been conclusively established, the Sixth Circuit has established guidance for evaluating these claims. The majority applies the elements set forth in Ohio Democratic Party v. Husted for vote- denial claims in evaluating the district court’s review of the vote-denial claim: (1) whether a voting practice “resulted in an adverse disparate impact on protected class members’ opportunity to participate in the political process” and (2) whether the challenged practice “causally contributes to the alleged discriminatory impact by affording protected group members less opportunity to participate in the political process”. 834 F.3d 620, 637 (6th Cir. 2016). While its approach is correct, the majority’s analysis loses sight of the totality of the circumstances in reviewing the district court’s findings. First, the majority concludes that the first element is not met because the change in ballot format applies to all voters, keeps no one away from the polls, and prevents no one from registering to vote. However, the requirement of the first element does not hang merely on the language of the statute, but rather the first element “requires proof that the challenged standard or practice causally contributes to the alleged discriminatory impact by affording protected group members 42 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson less opportunity to participate in the political process.” Id. At 638. As the district court found, the change will have a disproportionate impact on African American voters in Michigan, who use straight-party ballots at a significantly higher rate than other demographics, by creating longer wait times at the poles and thereby deterring voters from participating. Michigan State A. Philip Randolph Inst., 2018 WL 3769326, at *86. Next, the majority states that the claim fails on the second element because of evidence pointing in the opposite direction. However, the majority’s assessment does not adhere to the standard of review—clear error. To begin, the majority noted that the second element may potentially be informed by the “Senate Factors” discussed in Gingles v. Thornburg. 478 U.S. 30, 43-45 (1986). The majority then reviewed the district court’s assessment of the Gingles factors, noting that the district court held that factors one, three and four were not relevant to this case and that the absence of those factors would seem to point against invalidation of PA 268 and further discussing factors six and seven, pointing to evidence it finds contrary to the district court’s conclusion. Based on this assessment, and without discussion as to why the district court’s findings were clearly erroneous, the majority incorrectly concludes that there is a significant likelihood that the result of this analysis on appeal would point in the opposite direction to the district court’s analysis. Importantly, “Section 2, unlike other federal legislation that prohibits racial discrimination, does not require proof of discriminatory intent.” Moore v. Detroit Sch. Reform Bd., 293 F.3d 352, 363 (6th Cir. 2002). Rather, “a plaintiff need show only that the challenged action or requirement has a discriminatory effect on members of a protected group.” Id. 43 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson Moreover, “[t]he essence of a § 2 claim is that a certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities enjoyed by black and white voters to elect their preferred representatives.” Thornburg v. Gingles, 478 U.S. 30, 47 (1986). Guided by the Judicial Committee Report, the Supreme Court in Gingles listed several factors “that might be probative of a § 2 violation,” Id. at 36. Citing Gingles factors 2, 5, 6, 7, 8, and 9 as relevant, the district court concluded that the plaintiffs demonstrated that PA 268 would disproportionately impact African-American voters in Michigan because of their significantly higher usage of the straight-party option. The district court further found that this burden was linked to social and historical conditions in Michigan such as: historical discrimination in education resulting in lower literacy rates; discrimination in housing; and subtle racial appeals that have or currently produce discrimination against members of the protected class. The majority fails to articulate why Section 2 would not apply in this context, therefore they fail to demonstrate that the Secretary is likely of suceed on appeal as to the district court’s conclusion the PA 268 violates the Voting Rights Act. B. Irreparable Injury, Prospect that Others Will be Harmed and the Public Interest I now turn to the irreparable injury, harm, and public interest factors. This Court has acknowledged that “[w]hen constitutional rights are threatened or impaired, irreparable injury is presumed.” Obama for America Husted, at 436 (citing ACLU of Ky. v. McCreary Cty., Ky., 354 F.3d 438, 445 (6th Cir. 2003)). “A restriction on the fundamental right to vote therefore constitutes irreparable injury.” Id. Indeed, the Secretary correctly notes that “any time a State is enjoined by a court from effectuating statutes enacted by representatives of its people, it suffers a 44 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson form of irreparable injury.” Maryland v. King, 567 U.S. 1301, 133 (2012). The Secretary argues that “operation of PA 268 is the status quo” and that “a stay will not burden African-Americans.” Importantly, Michigan residents have enjoyed the option of straight-party voting for the past 127 years, since 1891. The abrogation of straight-party voting upends the status quo. Therefore, logic follows that maintaining the status quo in Michigan compels the continuance of Michigan’s 127-year practice of permitting straight-party voting. As this Court has already noted: This case does not involve the potential disruption of complicated election- administration procedures on the eve of Election Day; rather, denying the Secretary’s request for a stay here will merely require Michigan to use the same straight-ticket procedure that it has used since 1891. Johnson II, 833 F.3d at 669. On the contrary, s-taying the district court’s permanent injunction on PA 268—on the heels of an election—would likewise result in voter confusion that could, alone, impose a substantial burden on Michigan voters, particularly African-American voters. As we have been warned by the Supreme Court, “[c]ourt orders affecting elections, especially conflicting orders, can themselves result in voter confusion and consequent incentive to remain away from the polls. As an election draws closer, that risk will increase.” Purcell v. Gonzalez, 549 U.S. 1, 4-5 (2006) (per curiam). To the extent the state suffers any harm from continuing its 127-year practice of straight- party voting pending appeal, any prospective harm suffered by the state is minimal in comparison to the potential harm suffered by African-American voters in Michigan.2 Finally, the public 2 The Secretary’s two-week long wait to appeal the district court’s August 1, 2018 order further cuts against her argument of irreparable harm. While district court amended its order on August 9, 2018 to merely add a table of contents, this did not preclude the Secretary from appeals the district court’s decision sooner. 45 No. 18-1910, Mich. State A. Philip Randolph Inst. v Johnson interest warrants against disturbing Michigan’s 127-year practice of straight-party voting on the eve of Election Day. It is also worth noting that the Secretary’s significant lag in appealing the district court’s order, and in seeking an emergency stay in this Court strongly is contradictory to her newly-found sense of urgency in this matter. Because I disagree with the majority that the Secretary has met her burden of demonstrating that the district court was “arbitrary, unjustifiable or clearly unreasonable” as to each statutory and constitutional violation found, I respectfully dissent. 46