RECOMMENDED FOR PUBLICATION
                                    Pursuant to Sixth Circuit I.O.P. 32.1(b)
                                           File Name: 20a0314p.06

                      UNITED STATES COURT OF APPEALS
                                       FOR THE SIXTH CIRCUIT



 CHAD THOMPSON; WILLIAM T. SCHMITT; DON                          ┐
 KEENEY,                                                         │
                        Plaintiffs-Appellees,                    │
                                                                 │
                                                                 │
         v.                                                       >        No. 20-3526
                                                                 │
                                                                 │
 RICHARD MICHAEL DEWINE, in his capacity as the                  │
 Governor of Ohio; LANCE HIMES, in his official                  │
 capacity as the Interim Director of the Ohio                    │
 Department of Health; FRANK LAROSE, in his official             │
 capacity as Ohio Secretary of State,                            │
                                Defendants-Appellants.           │
                                                                 ┘

                            Appeal from the United States District Court
                           for the Southern District of Ohio at Columbus.
                     No. 2:20-cv-02129—Edmund A. Sargus, Jr., District Judge.

                                Decided and Filed: September 16, 2020*

              Before: SUTTON, McKEAGUE, and NALBANDIAN, Circuit Judges.

                                           _________________

                                                COUNSEL

ON BRIEF: Benjamin M. Flowers, Michael J. Hendershot, Stephen P. Carney, Shams H. Hirji,
OFFICE OF THE OHIO ATTORNEY GENERAL, Columbus, Ohio, for Appellants. Mark R.
Brown, CAPITAL UNIVERSITY LAW SCHOOL, Columbus, Ohio, Oliver B. Hall, CENTER
FOR COMPETITVE DEMOCRACY, Washington, D.C., Jeffrey T. Green, SIDLEY AUSTIN
LLP, Washington, D.C., Naomi A. Igra, Stephen Chang, Jennifer H. Lee, Tyler Wolfe, SIDLEY
AUSTIN LLP, San Francisco, California, for Plaintiffs-Appellees. Anne Marie Sferra,
Christopher N. Slagle, Bryan M. Smeenk, BRICKER & ECKLER LLP, Columbus, Ohio, Paul
A. Zevnik, MORGAN, LEWIS & BOCKIUS LLP, Washington, D.C., for Amici Curiae.

         *This decision was originally filed as an unpublished opinion on September 16, 2020. The court has now
designated the opinion for publication.
 No. 20-3526                           Thompson v. DeWine                                     Page 2


                                        _________________

                                             OPINION
                                        _________________

        PER CURIAM. The COVID-19 pandemic has upended life in many ways. In response
to the unfolding public health crisis, states across the country imposed various orders in hopes of
containing the virus. Ohio, for its part, asked its citizens to stay at home and restricted the size of
gatherings.

        This case, which we’ve seen before, involves the intersection of COVID-19, the state’s
responses to that pandemic, and some of Ohio’s conditions that must be met before a ballot
initiative can get on the ballot for Election Day. See Thompson v. DeWine, 

, 806
(6th Cir.) (per curiam), mot. to vacate stay denied, --- S. Ct. ----, No. 19A1054, 

(2020).

        Plaintiffs say that Ohio’s ballot initiative conditions are unconstitutional as applied
during this pandemic and request that the federal courts relax them, at least for the time being.
Plaintiffs’ challenge is a curious one. There is no question that Ohio’s ballot initiative conditions
are, standing alone, constitutional, there is no question that Ohio is not responsible for
COVID-19, and Plaintiffs are not challenging Ohio’s restrictions on public gatherings and the
like, which Ohio imposed to address the pandemic—so we assume those are constitutional as
well.   And yet, Plaintiffs contend that when you put all of this together, in effect, two
constitutional rights plus one outside catalyst make one constitutional wrong. The district court
agreed and granted a preliminary injunction. We stayed that order because we disagreed. And
now, because we still disagree, we reverse the district court’s grant of a preliminary injunction.

                                                  I.

        To get an initiative on a municipal ballot, Ohio requires the ballot’s proponents to gather
signatures totaling at least ten percent of the number of electors who voted for governor in the
municipality’s previous election. Ohio Rev. Code Ann. § 731.28. The signatures must be
original and affixed in ink, and the petition’s circulator must witness them.

20-3526                                   Thompson v. DeWine                                                Page 3


And the initiative’s proponents must submit these signatures to the Ohio Secretary of State at
least 110 days before the election.1

here are three Ohioans hoping to get initiatives on local ballots to decriminalize
marijuana.2        They argue that Ohio’s ballot initiative requirements, as applied during the
COVID-19 pandemic and given Ohio’s stay-at-home orders and other pandemic restrictions,
violate the First and Fourteenth Amendments. So they asked the district court to enjoin Ohio
from enforcing the ballot initiative requirements. The district court agreed, at least in part.
It granted plaintiffs’ request for a preliminary injunction, enjoining Ohio from enforcing some of
its ballot access requirements. And it ordered Ohio to accept electronically signed and witnessed
petitions, extended the deadline for petition submission, and told Ohio to come up with a system
that would “reduce the burden on ballot access.”3 Thompson v. DeWine, --- F. Supp. 3d ----,
No. 2:20-CV-2129, 

, at *21 (S.D. Ohio 2020) (quotation omitted).

         Ohio asked us to stay the district court’s injunction while its appeal was pending.
We did.      

.                 We reasoned that Ohio’s compelling interests in
preventing fraud and ensuring a fair and orderly signature verification process outweighed the
intermediate burden the requirements imposed on plaintiffs’ First and Fourteenth Amendment
rights.

Now, we review whether a preliminary injunction was warranted in the first
place. For reasons we’ll discuss below, we don’t think it was. We thus reverse the district
court’s grant of a preliminary injunction.

                                                            II.

         This case comes to us on appeal from an order granting an injunction. So we have
jurisdiction under 28 U.S.C. § 1292.                 We review a district court’s grant of a preliminary


         1This     date has already passed. But Ohio doesn’t argue that the case is moot. And we are satisfied that we
still have jurisdiction despite the date’s passing. Plaintiffs ask us to place their initiative directly on the ballots—and
that relief is still available, in theory, until Ohio prints its first round of ballots.
         2Our   original stay order covered these Plaintiffs and two Intervenor-Plaintiffs who sought to get proposed
constitutional amendments on Ohio’s November ballot. The Intervenor-Plaintiffs have since withdrawn from this
litigation. See Order Granting Mot. to Withdraw by Intervenors-Appellees.
         3The    court upheld Ohio’s signature quantity requirement.
 No. 20-3526                                  Thompson v. DeWine                                              Page 4


injunction for abuse of discretion, “subjecting factual findings to clear-error review and
examining legal conclusions de novo.” Daunt v. Benson, 

, 406 (6th Cir. 2020).

         “A plaintiff seeking a preliminary injunction must establish that he is likely to succeed on
the merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the
balance of equities tips in his favor, and that an injunction is in the public interest.” Winter v.
Nat. Res. Def. Council, Inc., 

, 20 (2008). When we evaluate these factors for an
alleged constitutional violation, “‘the likelihood of success on the merits often will be the
determinative factor.’” Obama for Am. v. Husted, 

, 436 (6th Cir. 2012) (quoting
Jones v. Caruso, 

, 265 (6th Cir. 2009)). So we start there.

                                                          A.

         If this all sounds familiar, that’s because it is. In staying the district court’s preliminary
injunction, we went through the factors above and concluded that Plaintiffs aren’t likely to
succeed on the merits. 

. We still think so.

         The First Amendment doesn’t guarantee the right to an initiative. Taxpayers United for
Assessment Cuts v. Austin, 

, 295 (6th Cir. 1993). But once the people of a state, in
their sovereign authority, decide to allow initiatives, “the state may not place restrictions on the
exercise of the initiative that unduly burden First Amendment rights.”

Amendment challenges to nondiscriminatory, content-neutral ballot
initiative requirements under the Anderson-Burdick framework.”4 

;
see Burdick v. Takushi, 

, 434 (1992); Anderson v. Celebrezze, 

, 788

         4Although   Ohio recognizes this, it also argues that “[l]aws regulating ballot access for state initiatives do
not implicate the First Amendment at all.” (Appellants’ Br. at 26.) But as Ohio admits, that’s not the law in this
circuit. (Id. at 29–30.) And “until this court sitting en banc takes up the question of Anderson-Burdick’s reach, we
will apply that framework in cases like this.” 

n.2. Still, we note that at least two other
courts of appeals take Ohio’s position. See Initiative & Referendum Inst. v. Walker, 

, 1099–100 (10th
Cir. 2006) (en banc); Marijuana Pol’y Project v. United States, 

, 85 (D.C. Cir. 2002). “And this court
has often questioned whether Anderson-Burdick applies to anything besides generally applicable restrictions on the
right to vote.” 

n.2 (collecting cases). So there’s a circuit split on the applicability of
Anderson-Burdick to laws regulating ballot access for initiatives. This has caused “predictably contrary conclusions
as to whether and to what extent States must adapt the initiative process to account for new obstacles to collecting
signatures.” Little v. Reclaim Idaho, --- S. Ct. ----, No. 20A18, 

, at *1 (2020) (Roberts, C.J.,
concurring in the grant of a stay). That said, “the [Supreme] Court is reasonably likely to grant certiorari to resolve
the split presented by this case on an important issue of election administration.”

v. DeWine                                            Page 5


(1983). Under that framework, the level of scrutiny we apply to “state election law depends
upon the extent to which a challenged regulation burdens First and Fourteenth Amendment
rights.” 

. When the burden is severe, the state must narrowly draw the
regulation to serve an “interest of compelling importance.”

when
the law imposes “reasonable, nondiscriminatory restrictions,” we subject it to rational-basis
review.

one more layer to Anderson-Burdick. A challenged law imposes an intermediate
burden when the burden is somewhere between severe on the one hand and reasonable and
nondiscriminatory on the other. Kishore v. Whitmer, --- F.3d ----, No. 20-1661, 

, at *2 (6th Cir. 2020). When the burden is intermediate, we weigh it against “the
precise interests put forward by the State as justifications for the burden imposed by its rule.”

; see also 

. In doing so, we consider
“the extent to which those interests make it necessary to burden the plaintiff’s
rights.” 

(quoting 

). It’s this level of
scrutiny that we apply to Ohio’s laws here.5

        1. The Burden

        We see no reason to depart from our previous holding that Ohio’s ballot-access
restrictions impose, at most, only an intermediate burden on plaintiffs’ First Amendment rights,
even during COVID-19.6

If anything, the interim between our stay order and
now has reinforced our holding. The federal circuit tide has turned against Plaintiffs. The Eighth

        5In    a surreply, Plaintiffs expand on their previous argument that Ohio—by failing to answer Plaintiffs’
complaint or file a Rule 12 motion—“admitted” Plaintiffs’ claim from the complaint that it was “impossible” for
them to collect signatures. See Fed. R. Civ. P. 8(b)(6). If this were true, perhaps stricter scrutiny would be
appropriate. But we don’t think “impossibility” here is a factual allegation that can be admitted in pleadings. See
Ohio Democratic Party v. Husted, 

, 628 (6th Cir. 2016) (collecting cases); Bright v. Gallia County,

, 652 (6th Cir. 2014) (explaining, in the context of a motion to dismiss, that “legal conclusions
masquerading as factual allegations” don’t turn legal questions into factual ones (quotations omitted)). And “a
defendant’s failure to deny conclusions of law does not constitute an admission of those conclusions.” 5 C. Wright
& A. Miller, Federal Practice & Procedure § 1279 (3d ed.). In any event, Ohio has consistently argued, both before
the district court and before us, that it wasn’t impossible for Plaintiffs to collect signatures.
        6Plaintiffsargue that our stay order “carries limited weight.” (Appellees’ Br. at 24 n.29.) We don’t need to
decide the precedential weight to give to that order. But it’s worth noting that we’ve since relied on it as “binding
precedent.” Hawkins v. DeWine, 

, 604 (6th Cir. 2020).
 No. 20-3526                                Thompson v. DeWine                                          Page 6


Circuit, for instance, held that Arkansas’s “in-person signature requirement, while implicating
the First Amendment, imposes less-than-severe burdens on the plaintiffs’ rights and survives the
applicable lesser scrutiny.” Miller v. Thurston, 

, 741 (8th Cir. 2020); see also
Libertarian Party of Pa. v. Governor of Pa., 813 F. App’x 834, 835 (3d Cir. 2020) (mem.)
(holding that Pennsylvania’s ballot-access law, which includes a signature requirement,
“survives intermediate scrutiny because it serves the Commonwealth’s legitimate and
sufficiently important interests in ‘avoiding ballot clustering, ensuring viable candidates, and the
orderly and efficient administration of elections.’”). And in Morgan v. White, the Seventh
Circuit said that if Illinois wanted to just skip referenda for the year, “there is no federal
problem”: “Illinois may decide for itself whether a pandemic is a good time to be soliciting
signatures on the streets in order to add referenda to a ballot.” 

, 652 (7th Cir. 2020).

          And in addition, the Supreme Court stayed two injunctions against state enforcement of
ballot access restrictions. Little v. Reclaim Idaho, --- S. Ct. ---, No. 20A18, 

(2020); Clarno v. People Not Politicians, --- S. Ct. ----, No. 20A21, 

(2020).
And the Court left our previous ruling in place. Thompson, --- S. Ct. ----, 

(2020).

          Even without those developments, Plaintiffs still faced an uphill battle. We noted in our
stay order that “[a]t bottom, a severe burden excludes or virtually excludes electors or initiatives
from the 

. But Ohio’s ballot access laws don’t do that.

the pandemic, “Ohio specifically exempted conduct protected by the First
Amendment from its stay-at-home orders.”

signatures for petitions.7
Even if that was unclear at first, Ohio made it clear by April 30—which gave Plaintiffs months to
gather signatures. Ohio Dep’t of Health, Director’s Order that Reopens Businesses, with
Exceptions, and Continues a Stay Healthy and Safe at Home Order ¶ 4 (April 30, 2020).




          7Plaintiffsargue that Ohio’s First Amendment exception to its stay-at-home orders was “too vague to
alleviate the burden on Thompson.” (Appellees’ Br. at 31.) We confronted that argument head on in Hawkins and
rejected it. 

(“[T]he orders explicitly exempt First Amendment protected speech, and it is
well-established that the act of collecting signatures for ballot access falls under that ambit.”).
 No. 20-3526                           Thompson v. DeWine                                    Page 7


        And even if prospective signatories were deciding to stay home or avoid strangers—thus
reducing Plaintiffs’ opportunities to interact with them—we don’t attribute those decisions to
Ohio. “[W]e must remember, First Amendment violations require state action.” 

. So “Plaintiffs’ burden is less than severe” because Ohio hasn’t excluded or
virtually excluded them from the ballot. Id.; see Hawkins v. DeWine, 

, 607 (6th Cir.
2020)

        Plaintiffs argue that “total exclusion” from the ballot isn’t essential for finding a severe
burden. (Appellees’ Br. at 25.) But the cases Plaintiffs cite don’t support their theory. For
instance, they rely on our recent decision in Esshaki v. Whitmer to claim that the “combined
effect” of strictly enforced ballot access laws and stay-at-home orders can create a severe burden.
See 813 F. App’x 170, 171 (6th Cir. 2020). This language, they say, means that “total exclusion”
isn’t necessary to make out a severe burden. And for extra support they cite SawariMedia, LLC
v. Whitmer, where “neither this court, nor the district court applied a ‘total exclusion’ test to find
severe burden.” (Appellees’ Br. at 28); see 

(6th Cir. 2020).

        True, we held in Esshaki that “the combination of [Michigan’s] strict enforcement of the
ballot-access provisions and the Stay-at-Home Orders imposed a severe burden on the plaintiffs’
ballot access.” 813 F. App’x at 171. But Plaintiffs omit why we held that way. We later
clarified: “We held that there was a severe burden because Michigan’s Stay-at-Home Order
remained in effect through the deadline to submit ballot-access petitions, effectively excluding all
candidates who had not already satisfied the signature requirements (and predicted a shutdown).”
Kishore, --- F.3d ----, 

, at *3 (emphasis added). And Kishore’s explanation of
why we found a severe burden in Esshaki applies with equal force to SawariMedia. The
restrictions at issue there were “identical” to those in Esshaki. SawariaMedia, 

. So in finding a severe burden in both Esshaki and SawariMedia, we relied on the fact that
Michigan’s restrictions “effectively excluded” the plaintiffs from ballot access.

        Plaintiffs also cite Libertarian Party of Ky. v. Grimes.         That case noted that “the
‘combined effect’ of ballot-access restrictions can pose a severe burden.” 

, 575 (6th
Cir. 2016). Fair enough. But again, Plaintiffs read the case too narrowly. In fact, Libertarian
Party of Ky. explicitly stated—multiple times, at that—that the ballot access restrictions at issue
 No. 20-3526                           Thompson v. DeWine                                  Page 8


couldn’t be a severe burden because they didn’t “constitute exclusion or virtual exclusion.”

see

(“The hallmark of a severe burden is exclusion or virtual exclusion from the
ballot.”).

        Since our stay order, we’ve already had the chance to take another look at the
burden Ohio’s ballot access regulations impose. See 

; see also Kishore,
--- F.3d ----, 

, at *3. Hawkins involved a challenge to Ohio’s requirements for
running for President of the United States as an independent, which are virtually identical to
those 

(noting that Ohio requires independent presidential candidates to file
“a nominating petition with no fewer than 5,000 signatures,” which must be fixed in ink and
witnessed by the circulator). Relying on our Thompson stay order, we held that “the burden
imposed on Plaintiffs by Ohio’s ballot-access statutes—in light of the state’s response to the
pandemic—is an intermediate one.”

And in Kishore, we applied intermediate
scrutiny to Michigan ballot access regulations that were “comparable to the burdens imposed
upon the plaintiffs in Thompson and Hawkins.” --- F.3d ----, 

, at *3.

        To be sure, it may be harder for Plaintiffs to obtain signatures given the conditions.
But “just because procuring signatures is now harder . . . doesn’t mean that Plaintiffs
are excluded from the ballot.” 

. The burden Plaintiffs face here is
thus an intermediate one. That means we next weigh it against the interests Ohio puts forward to
justify its regulations.

        2. Ohio’s Justifications

        Ohio’s ballot access laws place an intermediate burden on Plaintiffs’ First and Fourteenth
Amendment rights. So the next step in the Anderson-Burdick framework is “a flexible analysis
in which we weigh the ‘burden of the restriction’ against the ‘state’s interests and chosen means
of pursuing them.’”        Schmitt v. LaRose, 

, 641 (6th Cir. 2019), cert. denied,

(2020).

        Ohio articulates two interests relevant to this appeal. The first relates to the ink and
attestation requirements: preventing fraud by ensuring the authenticity of signatures. There’s no
question this is a legitimate—indeed compelling—interest. “The State’s interest in preserving
 No. 20-3526                           Thompson v. DeWine                                     Page 9


the integrity of the electoral process is undoubtedly important.” John Doe No. 1. v. Reed,

, 197 (2010). And “states have a strong interest in ‘ensuring that [their] elections
are run fairly and honestly,’ as well as in ‘maintaining the integrity of [their] initiative process.’”

(quoting Taxpayers United for Assessment 

).

       So Ohio’s first interest is important—what about its second? Ohio says that its deadlines
allow it to verify signatures in a fair and orderly way, ensuring that interested parties have
enough time to appeal an adverse decision in court. This is also an important interest. Indeed,
“[s]tates may, and inevitably must, enact reasonable regulations of parties, elections, and ballots
to reduce election- and campaign-related disorder.” Timmons v. Twin Cities Area New Party,

, 358 (1997).

       3. The Balancing Test

       Finally, “[a]t the third step of Anderson-Burdick we assess whether the State’s
restrictions are constitutionally valid given the strength of its proffered interests.” 

; see Kishore, 

, at *4. Remember, this stage of the analysis is
flexible, and we give states considerable leeway to pursue their legitimate interests. Buckley v.
Am. Const. Law Found., 

, 191 (1999). And all that’s required for the State to win at
this step is for its legitimate interests to outweigh the burden on Plaintiffs’ First Amendment
rights. 

. The method the State chooses to pursue its interests need
not be narrowly tailored.

much of the heavy lifting here. We’ve previously held, in multiple
cases, that the interests Ohio pursues through its ballot access laws “outweigh the intermediate
burden those regulations place on Plaintiffs.” Id.; 

; see also Kishore ---
F.3d ----, 

, at *3 (“On balance, the State’s well-established and legitimate
interests in administering its own elections through candidate-eligibility and ballot-access
requirements outweigh the intermediate burden imposed on Plaintiffs.”).             And “reasonable,
nondiscretionary restrictions are almost certainly justified by the important regulatory interests in
combating fraud and ensuring that ballots are not cluttered with initiatives that have not
 No. 20-3526                           Thompson v. DeWine                                    Page 10


demonstrated sufficient grassroots support.” Little, --- S. Ct. ----, 

, at *2
(Roberts, C.J., concurring in the grant of a stay).

                                          *       *       *

       In short, Ohio is likely to prevail on the merits—and that’s the most important part of this
analysis. Still, the remaining three preliminary injunction factors favor Ohio, too.

                                                  B.

       First, irreparable harm. “[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.”
Maryland v. King, 

, 

, 3 (2012) (quoting New Motor Vehicle Bd. of Cal.
v. Orrin W. Fox Co., 

, 1351 (1977) (Rehnquist, J., in chambers)). So “[u]nless the
statute is unconstitutional, enjoining a ‘State from conducting [its] elections pursuant to a statute
enacted by the Legislature . . . would seriously and irreparably harm [the State].’” 

(quoting Abbott v. Perez, --- U.S. ----, 

, 2324 (2018)). Because
we’ve already found that Ohio is likely to prevail on the merits here, it would cause the State
irreparable harm if we blocked it from enforcing its constitutional ballot access laws.

       Next, the balance of the equities.        “When analyzing the balance of equities, ‘[the
Supreme] Court has repeatedly emphasized that lower federal courts should ordinarily not alter
the election rules on the eve of an election.’” Kishore, --- F.3d ----, 

, at *4
(quoting Republican Nat’l Comm. v. Democratic Nat’l Comm., --- U.S. ----, 

,
1207 (2020) (per curiam)). Ohio will soon print ballots for overseas and military voting. Ohio
Rev. Code Ann. § 3509.01(B)(1). Because “federal courts are not supposed to change state
election rules as elections approach,” this factor also favors Ohio. 

.

       Finally, the public interest. It’s in the public interest that we give effect to the will of the
people “by enforcing the laws they and their representatives enact.”

So all four
preliminary injunction factors favor Ohio.
 No. 20-3526                           Thompson v. DeWine                                   Page 11


                                                 III.

       Finally, we note that the Federal Constitution gives states, not federal courts, “the ability
to choose among many permissible options when designing elections.”

tamper” with that authority.

to adapt or modify state law to changing
conditions—especially during a pandemic—rests with state officials and the citizens of the state.

       So while federal courts can sometimes enjoin unconstitutional state laws, we can’t
engage in “a plenary re-writing of the State’s ballot-access provisions.” Esshaki, 813 F. App’x at
172. Instead, “[t]he Constitution grants States broad power to prescribe the ‘Times, Places and
Manner of holding Elections for Senators and Representatives,’ which power is matched by state
control over the election process for state offices.” Clingman v. Beaver, 

, 586
(2005) (citations omitted).

       We don’t have the power to tell states how they should run their elections. If we find a
state ballot-access requirement unconstitutional, we can enjoin its enforcement.           See, e.g.,
Esshaki, 813 F. App’x at 172.         But otherwise, “state and local authorities have primary
responsibility for curing constitutional violations.” Hutto v. Finney, 

, 687 n.9
(1978); Esshaki, 813 F. App’x at 172 (holding that it “was not justified” for a district court to
extend the deadline to file signed petitions and order the state to accept electronic signatures).

       So when the district court here ordered Ohio to accept electronically signed and
witnessed petitions and extended the deadline for submitting petitions, it overstepped its bounds.
It effectively rewrote Ohio’s constitution and statutes and “intrude[d] into the proper sphere of
the States.”   Missouri v. Jenkins, 

, 131 (1995) (Thomas, J., concurring); see

(“[T]he district court exceeded its authority by rewriting Ohio law
with its injunction.”). Federal courts don’t have this authority.

                                                 IV.

       For these reasons, we reverse the district court’s grant of a preliminary injunction.