Direct Democracy Denied: The Right to Initiative During a Pandemic

I. Introduction

Putting aside the Supreme Court’s controversial decision in Republican National Committee v. Democratic National Committee,1 the case overextending the date for receipt of absentee ballots in the April 2020 Wisconsin primary, many (although not all) courts have done a fairly good job protecting voting rights during the COVID-19 pandemic. From easing candidate and party signature requirements for ballot access,2 to temporarily eliminating witness or notarization3 requirements for casting an absentee ballot, to interpreting the excuse provisions in for-cause absentee ballot laws to cover voters without coronavirus immunity who fear voting in person, courts have recognized that election laws that ordinarily do not burden voters can become burdensome in a pandemic. Courts have interpreted such laws to avoid disenfranchisement, and have even, in certain cases, temporarily suspended or altered them.

That welcome thumb on the scale favoring voters, however, has not extended uniformly to claims for the easing of signature gathering rules by ballot measure proponents. In four cases I examine,4 courts have rejected the demands of initiative proponents to ease requirements to qualify a measure for the ballot—such as allowing electronic instead of “wet” (in person) signatures—and easing witness requirements, total number of signatures required, or geographic requirements for signature collection. In just one case, Thompson v. DeWine, a federal district court ordered Ohio to alter its procedures for qualifying proposed measures for the ballot, including allowing the acceptance of electronic signatures. The decision, however, was put on hold by the Sixth Circuit in a stay order that was very dismissive of the rights of direct democracy and that portends bad things to come.5

In this short analysis, I argue that some of the reasons courts and states have offered against easing ballot measure qualification requirements during a pandemic are weak, and that the district court in Thompson was right to see that normal ballot qualification rules can impose a severe First Amendment burden on direct democracy participants under pandemic conditions. The problem, as illustrated by the Thompson case, is fashioning appropriate relief consistent with principles of federalism and separation of powers. It is difficult to craft a remedy that both puts the plaintiffs in the position they would have been in had there been no pandemic but that also does not usurp the state’s general role in enforcing its election rules or undermine sound principles of election administration and fairness.

II. Direct Democracy as Inferior Democracy

States and local governments do not have to offer voters the devices of direct democracy such as the initiative, referendum, or recall as a means of supplementing the regular legislative process, but when they do, activity related to these direct democracy measures is protected by the First Amendment. In Meyer v. Grant, for example, the Supreme Court struck down a Colorado rule that barred paying ballot petition circulators. “The refusal to permit appellees to pay petition circulators restricts political expression in two ways: First, it limits the number of voices who will convey appellees’ message and the hours they can speak and, therefore, limits the size of the audience they can reach. Second, it makes it less likely that appellees will garner the number of signatures necessary to place the matter on the ballot, thus limiting their ability to make the matter the focus of statewide discussion.” And in Buckley v. American Constitutional Law Foundation, the Court struck down, among other things, a Colorado law requiring petition circulators to be registered voters and to wear badges indicating whether or not they were paid circulators.

As lower courts have considered First Amendment challenges to various aspects of laws regulating ballot petitions and direct democracy, they typically have applied a sliding scale balancing test, commonly known as the Anderson-Burdick test, which requires the state to justify election laws imposing severe burdens on plaintiffs under strict scrutiny and less severe burdens under looser standards of review.6

Ballot circulation is generally an in-person activity, in which a ballot circulator typically sets up a table in a popular area, such as outside of a supermarket, and asks voters to sign petitions to put the measure on the ballot. To qualify a measure to appear on the ballot for voter approval, ballot circulators must comply with standards such as collecting a minimum number of signatures in a finite period of time.

Signature collection that is usually not burdensome in ordinary times has become extremely burdensome during the pandemic when states put in place orders confining people to their homes except for essential activities. Even in areas without formal orders, signature collection can be very difficult as health experts have cautioned against unnecessary close contact with other people and with shared surfaces such as the pens and clipboards that are typically used to collect signatures.

As in cases involving other election laws that have become burdensome in light of the COVID-19 pandemic,7 proponents of state and local ballot measures have asked federal courts for relief from the usual election rules in light of the difficulty of meeting those requirements. For the most part, courts have been unsympathetic to the claims of ballot measure proponents even while some courts have granted relief to minor political parties and candidates who also need to collect signatures to remain on the ballot.8

For example, in Morgan v. White, plaintiffs were Illinois registered voters who wished to circulate a constitutional amendment referendum on the Illinois Democracy Amendment and an initiative for a local government referendum in Evanston, Illinois. As is typical, Illinois requires ballot circulators to collect a certain number of signatures on paper. The state also requires ballot circulators to add a sworn statement “certifying that the signatures on that sheet . . . were signed in his or her presence.” Circulators must file the original signed witness sheets with election authorities.

Plaintiffs sought a court order modifying some of these rules and extending the deadline for submitting signatures. The court held that the plaintiffs lacked standing because they had not circulated petitions for the first 16 months of the 18-month period to collect signatures for the upcoming election. Without engaging in this circulation, they could not show they were injured by the existing ballot access rules. Plaintiffs argued they were waiting for a reform commission to first make recommendations about election reform before they decided to circulate the petition on the same topic, but the court concluded that “[n]othing in the record supports an inference that, absent [the stay-at-home order of the Illinois governor], Plaintiffs would have been able to collect necessary signatures in the weeks between the issuance of the order and [the deadline].”

The court further held that even if the plaintiffs had standing, they should lose on the merits. After stating that “there is no constitutional right to place referenda on the ballot,” the court distinguished cases in which other courts had loosened requirements for minor parties and candidates to appear on the ballot as somehow more worthy of protection. Despite the flexible Anderson-Burdick balancing test, the court applied rational basis review, which is easy for the state to satisfy and get its restrictions upheld.

Accordingly, the court concluded that although the challenge to requiring a wet signature and witness statements was more weighty given COVID-19, there should be no relief for plaintiffs because “[t]hese circumstances are caused by the virus itself . . . not by state law. It is only when state law prevents certain individuals from circulating petitions that First Amendment harms are implicated.” Finally, the court concluded that it was plaintiffs’ own dilatory conduct, rather than state law, that was responsible for their predicament.9

Similar reasoning appears in the other two cases rejecting arguments to loosen ballot access requirements in light of the virus. In Bambenek v. White,10 a different federal district court reviewing another challenge to Illinois rules for ballot measures followed the reasoning in Morgan. The Bambenek court distinguished cases giving relief to political parties collecting signatures, cases which “concern[ ] placing candidates on the ballot, which implicate[ ] unique constitutional concerns, as opposed to this case which involves placing a proposed constitutional amendment and various referenda on the ballot and therefore does not implicate precisely the same constitutional concerns.” The court, citing Morgan, described that court as distinguishing candidate ballot access from a “state-created right to non-binding ballot initiatives”—even though plaintiffs were proposing a binding ballot measure, not an advisory one.11 The court also found that easing the ballot access rules for initiative proponents would place a high burden on state defendants,12 and that the plaintiffs waited too long to file suit.  

In Arizonans for Fair Elections v. Hobbs, plaintiffs who wished to circulate ballot measures for the 2020 general election sought to change ballot measure qualification rules. The plaintiffs asked the court to permit them to submit signatures electronically through a system used by Arizona candidates to submit qualifying signatures for their candidacies. The court rejected the relief on a number of grounds. First, the plaintiffs had challenged only Arizona’s statutory rules concerning the ballot measure qualification process, and not the state constitutional provisions requiring, among other things, in-person signature qualification. Given this limited challenge, the court concluded that the relief plaintiffs sought would not help them. Second, the court characterized plaintiffs’ conduct as dilatory, in that other ballot measure committees had gathered enough signatures but “the two committees in this case didn’t start organizing and gathering signatures until the second half of 2019.” Further, given that the pandemic could lift before the July 2020 deadline for the submission of signatures, the court held that plaintiffs did not demonstrate that “Arizona law creates a severe burden that would prevent a reasonably diligent initiative committee from placing its proposed initiative on the ballot.”

The court also expressed separation of powers and federalism concerns, raising doubt about the power of the federal court “to rewrite state election laws that have been in place since the 1910s.” Finally, the court was concerned about the “array of granular policy choices this Court would need to make in order to effectively implement[ ] relief for the plaintiffs.” As in Morgan, the Arizonans for Fair Elections court distinguished cases granting relief for plaintiffs wishing to vote during the pandemic on grounds that those were “vote restriction” cases to which a “reasonable diligence” standard would not apply.

Some of the points made by these three courts are meritorious.13 For example, a ballot measure proponent who challenges statutory requirements but fails to challenge applicable state constitutional requirements for in-person signature gathering cannot expect a federal court to be able to grant effective relief. And an initiative proponent who would have no chance of qualifying a ballot measure under normal conditions should not expect relief from a federal court to account for increased difficulty created by the pandemic. A court should put plaintiffs in the rightful position, which is the position they would have been in had there been no wrong.14

But these cases are deeply problematic. They denigrate the rights that ballot measure proponents are seeking to vindicate in these cases by unfavorably comparing them to the rights of candidates to get on the ballot. Once a jurisdiction offers direct democracy options to the public, severe burdens on that right should be subject to closer scrutiny. The courts’ arguments seeking to distinguish candidate ballot access from ballot measure ballot access are unsupported by any reasoning. Both involve voting rights and both involve severe burdens caused by the pandemic. If the external shock of COVID-19 is enough to justify judicial changes in candidate signature requirements, it should for ballot measures as well.15 Cases in both areas should be judged under the flexible Anderson-Burdick standard.

The ballot measure cases also appear to impose too high a standard for ballot measure proponents’ standing requirements and for obtaining relief. As the Bambenek court recognized, “signature collection efforts for referenda drives . . . tend to ramp up in the final weeks.”16 That ballot measure proponents did not seek to gather signatures during the entire window for signature gathering should not be evidence of dilatory conduct. A court should instead ask if plaintiffs had a realistic chance of qualifying a measure for the ballot had there been no pandemic.

Once a court is convinced that initiative proponents would have had a realistic chance of qualifying a measure for the ballot but for the pandemic, the court needs to turn to more difficult questions about court competency to rewrite rules and separation of powers concerns, issues I address in the next section.

III. Fashioning Appropriate Direct Democracy Relief in a Pandemic

I am much more sympathetic to the constitutional analysis of the federal district court in Thompson v. DeWine, the case brought by proponents of a ballot measure in Ohio. But the case demonstrates the difficulty of federal courts appropriately granting relief, especially given election administration issues and given separation of powers and federalism concerns as raised in the Arizonans for Fair Elections case. And given the Sixth Circuit’s analysis granting a stay in Thompson, the case likely will soon be reversed.

As to constitutional rights, the district court in Thompson agreed that once a state offers direct democracy devices, limits on the practice may violate the Constitution and are subject to Anderson-Burdick balancing. The court also recognized that it was appropriate to take the COVID-19 pandemic into account when balancing, and that normally non-severe ballot access rules become severe under pandemic conditions. The state’s stay-at-home orders and other restrictions have “pulled the rug out from under [plaintiffs’] ability to collect signatures,” creating a severe burden on plaintiffs’ rights. This synergistic approach is fairer than the Morgan court’s nonsensical view that these problems were “caused by the virus itself . . . not by state law.”

Once the court concluded that Ohio’s ballot access laws imposed a severe burden on plaintiffs’ rights and found other factors meriting injunctive relief, the question became one of remedy. The court recognized that state officials were without power to waive those requirements for ballot access contained in the state constitution, and so while it deferred to state officials to some extent, it ordered the state “to accept electronically-signed and witnessed petitions collected through the on-line signature collection plans” that the plaintiffs had proposed for their ballot measure drives. The court also extended the deadlines for collection of signatures and ordered the parties to meet as to appropriate standards for electronic signature collection.17

The Thompson court did a good job of trying to put the plaintiffs in the position they would have been in if there had been no pandemic. And following the Sixth Circuit precedent in the Esshaki case, which extended the petitioning deadline for political parties,18 the district court gave state officials maximum flexibility to cure the constitutional defects created by the confluence of the coronavirus and state law.

But these cases do raise a host of difficult problems of election administration, federalism, and separation of powers. To begin with, state deadlines for ballot qualification are necessary so that election officials have time to adequately determine whether measures qualify for the ballot, and if they do, to properly prepare them for inclusion in ballot materials. As a court extends deadlines, it puts more pressure on election officials to do more in less time under their own pandemic-related stresses. Having state officials process electronic ballot measure signatures under pandemic conditions when they have not done so before may be difficult. Before a court grants relief to ballot measure proponents, it should ensure that the new procedures will not unduly interfere with the other responsibilities of election officials.

Second, as the courts rejecting relief for ballot measure proponents have noted, having a federal court intervene in state and local ballot measure machinery is both intrusive and difficult work.19 There is no set way for courts to ease rules. States might consider electronic submission of ballots (in states that already have some such system in place), or a reduction in the number of required signatures, or an end to witness requirements, or an extension of deadlines or some other relief. Unlike a legislature that can consider input from a variety of sources on how to best balance these kinds of questions, courts can only take advice offered by the party, by witnesses, or by others submitting formal amicus briefs. The chances are high that in setting forth detailed changes in procedures a court will get things wrong.

So how best to engage in this balancing? The Thompson court seems to have the balance right. While it put a thumb on the scale favoring plaintiffs’ rights when it came to the question of whether or not there is a constitutional violation, it offered deference to the state over how to best remedy the loss of plaintiffs’ constitutional rights. It ordered the parties to meet and confer as to how to best put the electronic signature collection mechanism into place, and the court should continue to be open to state concerns about burden, cost, and availability of a remedy.

A Sixth Circuit panel, granting a stay of the district court’s order for the state of Ohio, viewed the matter very differently. The court applied the Anderson-Burdick framework as binding circuit precedent, but left open the possibility of an en banc change of standard going forward.20 Dismissing the realities of how the pandemic had essentially ended successful petitioning activity,21 the court held the law imposed only a minor burden on plaintiffs. It declared that “just because procuring signatures is now harder (largely because of a disease beyond the control of the State) doesn’t mean that Plaintiffs are excluded from the ballot.” The court did not explain how anything short of full exclusion of the plaintiffs from the ballot could count as merely a minor burden. The court rejected the district court’s order even assuming an intermediate burden, holding the state’s rules appeared justified on antifraud and election administration grounds even in the midst of a pandemic.22

The Sixth Circuit panel distinguished the Circuit’s earlier Esshaki case, which had eased ballot access rules for party and candidate qualification in Michigan, on grounds that Ohio’s stay-at-home order did not formally ban First Amendment activity like petition circulation. It noted that Ohio was beginning to lift its stay-at-home order, suggesting without evidence that petition circulators would have an easier time collecting signatures in Ohio than in Michigan as the pandemic spread in both states.

Finally, the Sixth Circuit, turning to the federalism and separation of powers issues that the district court flagged before granting plaintiffs relief, held that the district court exceeded its powers in granting plaintiffs a remedy:

The broader point is that the federal Constitution provides States—not federal judges—the ability to choose among many permissible options when designing elections. And because that’s where the decision-making authority is, federal courts don’t lightly tamper with election regulations. These concerns are magnified here where the new election procedures proffered by Plaintiffs threaten to take the state into unchartered waters. It may well be that the new methods for gathering signatures and verifying them proposed by Plaintiffs (using electronic signatures gathered online by third parties and identified by social security number) will prove workable. But they may also pose serious security concerns and other, as yet unrealized, problems. So the decision to drastically alter Ohio’s election procedures must rest with the Ohio Secretary of State and other elected officials, not the courts.23

The decision of the Sixth Circuit is unfortunate. The district court’s resolution of the difficult issues posed by ballot circulation during a pandemic may not have been perfect, but they demonstrate the efforts of a court that both takes the First Amendment rights of ballot measure proponents seriously and considers the real costs of a federal court changing election rules in the midst of a pandemic. As long as the court provides some flexibility for the legislature to negotiate a constitutional alternative with the plaintiff, as did the district court, the intrusion on legislative prerogatives is tolerable to protect voting rights during a pandemic. A court sensitive to federalism and separation of powers concerns can craft targeted relief to protect the right to ballot circulation.

The Sixth Circuit, in contrast, has put a thumb on the scale favoring the state, denigrating the right to petition along the way, and minimizing the real costs that the pandemic has placed on democratic petitioning activity. The court weighed federalism and separation of powers issues too heavily as it dismissed the burdens on the right to petition as minor. Most importantly, the Sixth Circuit decision sends a disturbing signal about how some courts may approach burdens on fundamental voting rights during the pandemic.24

• 1I critique this opinion in Part II.A of Richard L. Hasen, Three Pathologies of American Voting Rights Iluminated by the COVID-10 Pandemic, and How to Treat and Cure Them (draft dated June 12, 2020).Part II.B of the Article describes the split among courts over protecting voting rights during a pandemic.
• 2Esshaki v. Whitmer, No. 20-1336, 2020 WL 2185553, *1 (6th Cir. May 5, 2020).
• 3After the ruling, the Oklahoma Legislature reinstated the notarization requirement.
• 4Arizonans for Fair Elections v. Hobbs, No. CV-20-00658-PHX-DWL, 2020 WL 1905747 (D. Ariz. Apr. 17, 2020); Bambenek v. White, No. 3:20-cv-3107, 2020 WL 2123951, *2 (C.D. Ill. May 1, 2020); Morgan v. White, No. 1-20-cv-02189 (N.D. Ill. May 18, 2020); Thompson v. DeWine, ___ F.3d. ___, 2020 WL 2702483 (6th Cir. May 26, 2020). Although beyond the scope of this article, a federal district court also rejected an attempt to relax signature requirements for a gubernatorial recall in Nevada. Fight for Nevada v. Cegavske, No. 2:20-cv-00837-RFB-EJY (D. Nev., May 15, 2020). The court distinguished cases relaxing signature requirements for candidate and party ballot qualification, calling them “very different” from qualifying a recall for the ballot. Id. (slip op. at 7–8).
• 5See Part III.
• 6The test is named after two Supreme Court cases, Anderson v. Celebrezze, 460 U.S. 780 (1983) and Burdick v. Takushi, 504 U.S. 428 (1992). For more on the test, see Hasen, Three Pathologies,at Part II.B; Nicholas Stephanopoulos, Election Litigation in the Time of the Pandemic, U. Chi. L. Rev. Online (2020). On lower court application of Anderson-Burdick to First Amendment ballot measure cases, see Thompson, 2020 WL 2557064, at *10 (“Importantly, this Court is bound by the Sixth Circuit, which has twice in the last two years applied the Anderson-Burdick framework to First Amendment challenges to Ohio’s statutory requirements for initiative petitions.”).
• 7See Hasen, Three Pathologies, at Part II.B (describing COVID-19-related election litigation).
• 8See, for example, Esshaki v. Whitmer, No. 20-1336, 2020 WL 2185553, *1 (6th Cir. May 5, 2020).
• 9The court also noted that changes in the laws would burden election administrators who need to determine if measures qualify for the ballot and then prepare them for inclusion. “Preventing Defendants from being able to fulfill these statutory duties [on time] not only imposes harm on them but also appears contrary to the public interest.”
• 10Bambenek v. White, 2020 WL 2123951 (C.D. Ill. May 1, 2020).
• 11The case that the Bambenek court erroneously relied on for this proposition involved a binding referendum, not an advisory voter plebiscite.
• 12As to those plaintiffs in the case seeking to place a measure on local ballots, the court noted that the deadline for signature submission was August 3, 2020 and that it was speculative whether the virus would affect signature gathering during the summer. Bambenek, 2020 WL 2123951, at *3.
• 13I address the federalism and separation of powers issues in Part III below.
• 14Douglas Laycock & Richard L. Hasen, Modern American Remedies 14–15 (5th ed. 2019).
• 15One answer to the argument that these laws infringe First Amendment rights is that ballot measure proponents could simply decide to propose a measure in a later election. But we would not accept such an answer if applied to a candidate kept off the ballot for lack of enough signatures during a pandemic. Just like a candidate expresses her First Amendment right by choosing to run in a particular election, so too do ballot proponents pick the times for attempting to place measures on the ballot.
• 16Bambenek, 2020 WL 2123951, *3.
• 17Given the adjustments of deadlines and the acceptance of electronic signatures, the court found it unnecessary to reduce the state’s numerical and geographical requirements for signature collection.
• 18See Esshaki, 2020 WL 2185553.
• 19Having a state court, rather than a federal court, make these changes removes the federalism concern but the election administration and separation of powers concerns remain.
• 20In seeking initial en banc consideration of the district court’s order in Thompson, Ohio argued that Anderson-Burdick should not even apply to challenges to the mechanics of the initiative process. In the petition, the state conceded that existing Sixth Circuit authority requires use of the Anderson-Burdick balancing test to determine whether state ballot measure procedures are unduly burdensome. It asked for en banc review to reverse those precedents, arguing that D.C. Circuit and Tenth Circuit cases reject Anderson-Burdick in this context and pointing to a purported circuit split. In fact, the D.C. case does not mention the Anderson-Burdick test and instead held that federal legislation barring the District of Columbia from passing any laws on the subject of marijuana legalization did not violate the First Amendment. The Tenth Circuit case upheld against First Amendment challenge a provision of the Utah constitution requiring wildlife-related ballot measures to pass by a supermajority vote. This case, too, did not mention Anderson-Burdick, and neither case involved claims as in Thompson that the mechanics of the ballot measure process imposed a severe burden on ballot measure proponents in violation of the First Amendment.
• 21The court explained: “Plaintiffs’ claim effectively boils down to frustration over failing to procure as many signatures for their petitions (because of social distancing and reduced public crowds) as they would without the pandemic. But that’s not necessarily true. There’s no reason that Plaintiffs can’t advertise their initiatives within the bounds of our current situation, such as through social or traditional media inviting interested electors to contact them and bring the petitions to the electors’ homes to sign. Or Plaintiffs could bring their petitions to the public by speaking with electors and witnessing the signatures from a safe distance, and sterilizing writing instruments between signatures.”
• 22“Defendants claim the witness and ink requirements help prevent fraud by ensuring that the signatures are authentic. And the deadlines allow them time to verify signatures in an orderly and fair fashion, while also providing initiative proponents time to challenge any adverse decision in court. [¶] These interests are not only legitimate, they are compelling.”
• 23The court also cautioned against court orders issued too close to the election, citing the Purcell Principle and Purcell v. Gonzalez, 549 U.S. 1 (2006). For a critique of the principle as applied in the context of the pandemic, see Hasen, Three Pathologies, at 41–42. For a more general critique, see Richard L. Hasen, Reining in The Purcell Principle, 43 Fla. St. U. L. Rev. 427 (2015).
• 24As this Article went into production, the plaintiffs in Thompson v. DeWine filed an emergency application in the Supreme Court to vacate the Sixth Circuit’s stay. On June 25, 2020, the Supreme Court denied that application. In addition, a federal district court relaxed some Arkansas provisions related to the circulation of initiative petitions. Miller v. Thurston, 2020 WL 2617312 (W.D. Ark. May 25, 2020). The court applied the Anderson/Burdick framework and decided to allow petition circulators to collect individual signatures from voters, signed under penalty of perjury, rather than requiring circulators to witness signatures and attest to their authenticity. It rejected other requests of the plaintiffs, such as lowering the number of required signatures. Arkansas appealed and the Eighth Circuit has granted a stay pending appeal.