CENTRAL MAINE POWER COMPANY VERSANT POWER ENMAX CORPORATION MAINE PRESS ASSOCIATION MAINE ASSOCIATION OF BROADCASTERS JANE PRINGLE KENNETH FLETCHER BONNIE GOULD BRENDA GARRAND LAWRENCE WOLD v. MAINE COMMISSION ON GOVERNMENTAL ETHICS AND ELECTION PRACTICES WILLIAM SCHNEIDER DAVID HASTINGS III SARAH LECLAIRE DENNIS MARBLE BETH AHEARN AARON FREY (2025)

In 2023, Maine voters passed by ballot initiative “An Act to Prohibit Campaign Spending by Foreign Governments” (“the Act”) with the expressed purpose of prohibiting foreign governments and “foreign government-influenced entit[ies]” from contributing to or otherwise influencing candidate elections and ballot initiatives.¹ Me. Rev. Stat. Ann. tit. 21-A, § 1064 (2024). To accomplish the Act's aim of preventing what its supporters refer to as “foreign interference” in elections, the Act also requires media platforms to conduct due diligence to ensure that they do not distribute a public communication that violates this prohibition. Those who violate the Act may be subject to civil penalties, criminal penalties, or both.

Several companies and individuals, including Central Maine Power (“CMP”), Versant Power and Enmax Corporation (“Versant”), Maine Press Association and Maine Association of Broadcasters (“Press and Broadcasters”), and several individuals (“Electors”) filed suit against state officials and entities responsible for enforcing the Act (collectively, “Maine”), including the Maine Commission on Governmental Ethics and Election Practices (the “Ethics & Election Commission”). The challengers contended that the Act was facially invalid under the First Amendment and thus moved for a preliminary injunction enjoining the Act in its entirety. The district court granted the preliminary injunction, and Maine appealed. We affirm.

I. Factual Background

The Act aims to limit the influence of foreign governments in Maine's elections, including both candidate elections and referenda. Tit. 21-A, § 1064. The Act was overwhelmingly popular with voters, 86% of whom approved it as a ballot question after other attempts to enact similar legislation failed. To explain why the Act has proven controversial despite its support among Maine voters, we first sketch some of the specific facts leading up to the ballot question before turning to the language of the Act itself.

CMP and Versant, two of the plaintiffs here, are the two primary utility companies operating in Maine. The present case stems in large part from a contentious fight over the construction of an energy transmission line that, if completed, would run through the state of Maine, thereby connecting Canadian electricity to Massachusetts. The project, known as the “CMP Corridor,” is a joint project between two companies: CMP and H.Q. Energy Services (U.S.) Inc. (“HQUS”), which is a subsidiary of a Canadian public utility called Hydro-Québec. Unfortunately for supporters of the CMP Corridor, the project was unpopular with many Maine voters who, in multiple elections, aimed to stop its development through ballot initiatives. Maine voters also considered (but ultimately rejected) a ballot initiative that proposed to seize CMP's and Versant's assets through eminent domain and replace the companies with a quasi-governmental entity.

The companies that would have been negatively impacted by these ballot initiatives -- including CMP, HQUS, and Versant -- opposed their passage. They did so, in part, by contributing substantial amounts of money to political action committees and ballot question committees. Specifically, between 2013 and 2023, CMP and its affiliates contributed nearly $73 million combined, and HQUS contributed around $22 million. Versant contributed over $16 million in just the time between August 2020 and the end of 2023.

These campaign contributions were substantially higher than other corporate contributions in the state, and some Maine voters and legislators took issue with these companies’ contributions based on, to varying degrees, the companies’ foreign ownership. For example, at first glance, CMP might seem an unlikely target since it is incorporated in Maine and has operated there for over 125 years. Despite its ties to the state, however, CMP's parent company is wholly owned by a publicly traded company, Avangrid Inc. (“Avangrid”). At the time of this suit's filing, 81.6% of Avangrid's shares were owned by a Spain-based corporation, Iberdrola, S.A.² Further, Qatar's sovereign wealth fund owned somewhere between 7 to 11% of CMP in light of its 8.7% ownership interest in Iberdrola, S.A. and 3.7% ownership interest of Avangrid. HQUS, for its part, is wholly owned by the province of Québec in Canada.

Versant, like CMP, is incorporated in Maine and has operated exclusively there for a century. But Versant's parent company is wholly owned by a foreign entity -- the City of Calgary in Alberta, Canada. Versant's operations, however, are subject to domestic control pursuant to a stipulation approved by the Maine Public Utilities Commission that prohibits Calgary from participating in operations or management decisions.

The companies’ ownership structures led their opponents to characterize the companies’ campaign spending as foreign interference in domestic elections. Believing such foreign involvement improper, opponents set out to ban it, including by supporting legislation that would prohibit political spending by companies “influenced” by foreign governments or companies. Several attempts to enact legislation failed, in part due to concerns by Maine's Governor that such restrictions might be unconstitutional. Finally, the Act was submitted as a ballot question in 2023 and approved by Maine voters.

II. The Act's Language

The Act states that “[a] foreign government-influenced entity may not make, directly or indirectly, a contribution, expenditure, independent expenditure, electioneering communication or any other donation or disbursement of funds to influence the nomination or election of a candidate or the initiation or approval of a referendum.” Tit. 21-A, § 1064(2). A “foreign government-influenced entity” is defined as:

(1) A foreign government; or

(2) A firm, partnership, corporation, association, organization or other entity with respect to which a foreign government or foreign government-owned entity:

(a) Holds, owns, controls or otherwise has direct or indirect beneficial ownership of 5% or more of the total equity, outstanding voting shares, membership units or other applicable ownership interests; or

(b) Directs, dictates, controls or directly or indirectly participates in the decision-making process with regard to the activities of the firm, partnership, corporation, association, organization or other entity to influence the nomination or election of a candidate or the initiation or approval of a referendum, such as decisions concerning the making of contributions, expenditures, independent expenditures, electioneering communications or disbursements.

Id. § 1064(1)(E). “Foreign government-owned entity,” in turn, is defined as “any entity in which a foreign government owns or controls more than 50% of its equity or voting shares.” Id. § 1064(1)(F).

Subsection 2's prohibition on campaign spending is supplemented by the Act's following three subsections. Subsection 3 prohibits “knowingly solicit[ing], accept[ing] or receiv[ing] a contribution or donation prohibited by subsection 2.” Id. § 1064(3). Subsection 4 prohibits “knowingly or recklessly provid[ing] substantial assistance” in the making of “a contribution or donation” or “an expenditure, independent expenditure, electioneering communication or disbursement prohibited by subsection 2,” as well as in the “solicitation, acceptance or receipt of a contribution or donation prohibited by subsection 2.” Id. § 1064(4). Subsection 5 prohibits “structur[ing] or attempt[ing] to structure” a transaction “to evade the[se] prohibitions and requirements.” Id. § 1064(5).

A knowing violation of subsections 2 through 5 is a class C crime under Maine law. Id. § 1064(9). Class C crimes are punishable by up to five years’ imprisonment. Me. Rev. Stat. Ann. tit. 17-A, § 1604(1)(C) (2024).

The Act also requires that specified media platforms conduct due diligence to ensure that they do not distribute any prohibited communications (e.g., election-related communications paid for by a foreign government-influenced entity):

Each television or radio broadcasting station, provider of cable or satellite television, print news outlet and Internet platform shall establish due diligence policies, procedures and controls that are reasonably designed to ensure that it does not broadcast, distribute or otherwise make available to the public a public communication for which a foreign government-influenced entity has made an expenditure, independent expenditure, electioneering communication or disbursement in violation of this section. If an Internet platform discovers that it has distributed a public communication for which a foreign government-influenced entity has made an expenditure, independent expenditure, electioneering communication or disbursement in violation of this section, the Internet platform shall immediately remove the communication and notify the [Ethics & Election Commission].

Tit. 21-A, § 1064(7); see also id. § 1001(1).

The Act also requires that whenever a foreign government-influenced entity “finance[s] a [permissible] public communication ․ to influence the public or any ․ local official or agency” concerning government policy or government relations with a foreign country or political party, that communication must contain the words: “Sponsored by [the name of the entity],” followed by the label of a “foreign government-influenced entity” or a “foreign government.” Id. § 1064(6).

A violation of the Act may be penalized by a fine of not more than $5,000 or “double the amount of the contribution, expenditure, independent expenditure, electioneering communication, donation or disbursement involved in the violation, whichever is greater.” Id. § 1064(8). The Ethics & Election Commission has the discretion to assess the penalty and “shall consider, among other things, whether the violation was intentional and whether the person that committed the violation attempted to conceal or misrepresent the identity of the relevant foreign government-influenced entity.” Id.

III. Procedural History

On December 12, 2023, after the ballot initiative passed, CMP filed this lawsuit against Maine and moved for a preliminary injunction. As relevant here, CMP alleged that the Act violated the First Amendment and that any constitutional provisions could not be severed. Versant soon filed a complaint making these same arguments and arguing that federal law preempted the Act, in addition to raising other claims. The Press and Broadcasters then filed a complaint challenging subsection 7 -- the due diligence requirement -- as unconstitutionally vague and a violation of the First Amendment. Finally, the Electors filed a complaint bringing a variety of federal and state constitutional claims based on their rights as voters to receive and consider political speech. The cases were soon consolidated and each complainant, like CMP, moved for a preliminary injunction.

After reviewing the submissions of the parties and amici curiae, the district court held a hearing on February 23, 2024 and granted the motions for a preliminary injunction a few days later, shortly before the Act was scheduled to go into effect. Cent. Me. Power Co. v. Me. Comm'n on Governmental Ethics & Election Pracs., 721 F. Supp. 3d 31, 37 (D. Me. 2024). Given the compressed time frame, however, the district court issued the preliminary injunction based on only CMP's and Versant's motions. Id.

In its order, the district court noted that First Amendment facial challenges based on overbreadth “succeed if ‘a substantial number of the law's applications are unconstitutional, judged in relation to the statute's plainly legitimate sweep.’ ” Id. at 49 (cleaned up) (quoting Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 449 n.6 (2008)). The court laid out the familiar four-part framework for deciding whether to grant a preliminary injunction, noting that the first factor of likelihood of success on the merits is the most important part of the analysis in the First Amendment context. Id. at 42 (citing Sindicato Puertorriqueño de Trabajadores v. Fortuño, 699 F.3d 1, 10 (1st Cir. 2012) (per curiam)).

The district court then turned to its analysis and began with Versant's claim that the Act was preempted by federal law. Id. As to that argument, it held that, to the extent that the Act covered foreign spending in elections for federal office, it was likely preempted by the Federal Election Campaign Act's (“FECA”) preemption provision. Id. at 42-43; see 52 U.S.C. § 30143(a). In contrast, the district court held that the Act was likely not preempted as to state and local elections. Cent. Me. Power Co., 721 F. Supp. 3d at 43-49.

The district court then turned to the merits of the challengers’ First Amendment arguments with respect to (1) referenda and (2) state and local candidate elections. As a threshold matter, the parties disagreed about the applicable level of scrutiny. Maine argued that the more lenient standard of “exacting scrutiny” applied, but the district court agreed with the Act's challengers that the Act was properly subject to strict scrutiny.³ Id. at 50 (citing Fortuño, 699 F.3d 1). The district court thus held that, to prevail, Maine had to show that the Act furthered a compelling interest and was narrowly tailored to achieve that interest. Id. (citing Citizens United, 558 U.S. at 340).

With respect to the former, the district court found that Maine had a compelling interest in limiting foreign government influence in state candidate elections and assumed without deciding that Maine also had a compelling interest in limiting foreign government influence in state referenda. Id. at 50-51. In contrast, the district court found that Maine lacked a compelling interest in limiting the appearance of foreign government influence in state elections. Id. at 51-52.

Accordingly, the district court next considered whether the Act was narrowly tailored to achieve the state's compelling interest in limiting foreign government influence in state candidate elections and referenda. On that point, the district court held that subsection 2's prohibition on spending in state elections by foreign governments was likely narrowly tailored because federal law provides no protections against foreign government spending in state referenda (as opposed to state candidate elections). Id. at 52; see tit. 21-A, § 1064(1)(E)(1). However, the district court held that subsection 2's prohibition on spending by entities with at least 5% foreign ownership was likely not narrowly tailored. Cent. Me. Power Co., 721 F. Supp. 3d at 52-53; see tit. 21-A, § 1064(1)(E)(2)(a). The district court reasoned that the 5% definition seemed arbitrary and had the impact of prohibiting political speech from corporations with potentially 95% U.S. citizen ownership. Cent. Me. Power Co., 721 F. Supp. 3d at 52-53.

Similarly, the district court deemed insufficiently tailored subsection 2's application to entities in which a foreign government or foreign government-owned entity “[d]irects, dictates, controls or directly or indirectly participates in the decision-making process” of the entity “to influence the nomination or election of a candidate or the initiation or approval of a referendum.” Id. at 54-55 (alteration in original) (quoting tit. 21-A, § 1064(1)(E)(2)(b)). In the district court's view, this definition's focus on conduct initially seemed to fit the requisite interest more closely. Id. at 54. However, Maine defended the statutory text by referring to definitions in the Ethics & Election Commission's since-revised rules, which appeared to impermissibly broaden the Act's application beyond the “participation” requirement. Id. The court noted that the rules seemed to allow for the possibility that “influence” might encompass domestic corporations receiving unsolicited communications from foreign governments, which would then prohibit those domestic corporations from campaign spending. Id. at 55. The district court thus held that this definition was likely not narrowly tailored because it would “stifle” domestic speech regardless of actual foreign influence. Id.

In the end, because the district court determined that a substantial number of the Act's applications likely violated the First Amendment, and the remaining factors favored a preliminary injunction, it enjoined the Act in its entirety. Id. at 55-56. In doing so, the district court expressly noted Maine severability law but declined to sever given the expedited and preliminary nature of the proceeding; instead, the court reserved the issue for later consideration. Id. at 55.

Maine timely appealed, arguing that the district court abused its discretion as to its holdings regarding preemption, the applicable level of scrutiny, the state's compelling interest, and whether the Act was narrowly tailored. Maine also argued that the Act was not facially invalid, the injunction was overly broad, and the district court abused its discretion in reserving its decision on severability. Since March 21, 2024, the proceedings have been stayed pending appeal.

IV. Discussion

A. Preliminary Injunction Analysis

This court reviews the grant of a preliminary injunction for abuse of discretion. Fortuño, 699 F.3d at 10. “Under that rubric, findings of fact are reviewed for clear error and issues of law are reviewed de novo.” Id. (quoting Wine & Spirits Retailers, Inc. v. Rhode Island, 418 F.3d 36, 46 (1st Cir. 2005)). “[A] facial challenge to a statute presents a question of law ․” Id. at 11 (citing New Eng. Reg'l Council of Carpenters v. Kinton, 284 F.3d 9, 19 (1st Cir. 2002)).

Courts weigh four factors in considering whether to issue a preliminary injunction: “(1) the plaintiff's likelihood of success on the merits; (2) the potential for irreparable harm in the absence of an injunction; (3) whether issuing the injunction will burden the defendants less than denying an injunction would burden the plaintiffs and (4) the effect, if any, on the public interest.” Id. at 10 (quoting Jean v. Mass. State Police, 492 F.3d 24, 26-27 (1st Cir. 2007)). But the first factor -- likelihood of success on the merits -- is the “linchpin” of the analysis in the First Amendment context. Id. If the movants are likely to succeed, then “irreparable injury is presumed.” Id. at 11.

Facial challenges are “hard to win.” Moody v. NetChoice, LLC, 603 U.S. 707, 723 (2024). Although this “very high bar” is lowered somewhat in the First Amendment context, the standard is “still rigorous” and facial challenges are still “disfavored.” Id. at 723, 744. In the First Amendment context, “[t]he question is whether ‘a substantial number of the law's applications are unconstitutional, judged in relation to the statute's plainly legitimate sweep.’ ” Id. at 723 (cleaned up) (quoting Ams. for Prosperity Found. v. Bonta, 594 U.S. 595, 615 (2021)). In other words, a law “may be struck down in its entirety ․ only if the law's unconstitutional applications substantially outweigh its constitutional ones.” Id. at 723-24.

B. The Applicable Level of Scrutiny

In general, “[l]aws that burden political speech ordinarily are subject to strict scrutiny, requiring the government to prove that any restriction ‘furthers a compelling interest and is narrowly tailored to achieve that interest.’ ” Fortuño, 699 F.3d at 11 (quoting Citizens United, 558 U.S. at 340). Citing Fortuño, the district court reviewed the entirety of the Act under strict scrutiny. Cent. Me. Power Co., 721 F. Supp. 3d at 50. On appeal, Maine argues that the district court should have applied the somewhat more lenient standard of “exacting scrutiny,” which requires that restrictions on speech be “closely drawn to match a sufficiently important interest.” Daggett v. Comm'n on Governmental Ethics & Election Pracs., 205 F.3d 445, 454 (1st Cir. 2000) (internal quotations omitted) (quoting Nixon v. Shrink Mo. Gov't PAC, 528 U.S. 377, 387-88 (2000)). Maine advances two arguments to support this contention, which we take in turn.

1. Level of Scrutiny for Restrictions on Contributions

Maine first argues that exacting, rather than strict, scrutiny should apply to the Act's restrictions on contributions. For support, Maine points to this court's precedent applying exacting scrutiny to contribution limits. See id. In response, CMP and Versant acknowledge that exacting scrutiny applies to stand-alone limitations on contributions. See Buckley v. Valeo, 424 U.S. 1, 20-21 (1976). They nevertheless contend that exacting scrutiny is inappropriate here because the Act bans a wide range of spending, including contributions as well as expenditures, restrictions of which are typically subject to strict scrutiny (citing Fortuño, 699 F.3d at 12). CMP additionally argues that the level of scrutiny is irrelevant because the Act would not withstand even a lower level of scrutiny.

At the outset, we note that Maine is correct that exacting scrutiny generally applies to limits on contributions. A less demanding form of scrutiny is appropriate for regulations on contributions because, as the Supreme Court explained, “contributions lie closer to the edges than to the core of political expression.” FEC v. Beaumont, 539 U.S. 146, 161 (2003). Therefore, “a contribution limit involving ‘significant interference’ with associational rights” is not subject to strict scrutiny but instead “passes muster if it satisfies the lesser demand of being ‘closely drawn to match a sufficiently important interest.’ ” Id. at 162 (quoting Nixon, 528 U.S. at 387-88); see also Fortuño, 699 F.3d at 12 (noting that “regulations designed ‘to ensure against the reality or appearance of corruption,’ such as those capping direct contributions to political candidates” are subject to exacting scrutiny (quoting Citizens United, 558 U.S. at 357)); Minn. Chamber of Com. v. Choi, 765 F. Supp. 3d 821, 847-49, 857 (D. Minn. 2025) (applying different levels of scrutiny to a law containing separate restrictions on contributions and expenditures).

CMP and Versant point out that the Act bans a wide variety of political spending: “contribution[s], expenditure[s], independent expenditure[s], electioneering communication[s and] any other donation[s] or disbursement[s] of funds.” Tit. 21-A, § 1064(2). But we agree with CMP that we need not resolve whether strict scrutiny should apply to the Act in its entirety, because the plaintiffs have demonstrated a likelihood of success on the merits even if the contribution ban is evaluated under the somewhat lower standard of exacting scrutiny. We therefore assume without deciding that exacting scrutiny applies to the portion of the Act restricting contributions.⁴ See Beaumont, 539 U.S. at 161-62.

2. Level of Scrutiny for Restrictions on Political Spending by Foreign Citizens

Next, despite acknowledging that restrictions on expenditures generally receive strict scrutiny, Maine argues that the remainder of the Act -- including limits on expenditures -- should be subjected to exacting scrutiny because the Act restricts non-citizens’ participation in our nation's democracy. Maine argues that cases like Citizens United and Fortuño did not concern what level of scrutiny applied to laws targeting foreign government influence in American elections and are therefore inapposite.

It is helpful to begin our discussion by sketching out the overarching legal framework regarding restrictions on corporate spending, including restrictions on foreign corporations. On the one hand, Citizens United makes clear that federal and state governments “may not suppress political speech on the basis of the speaker's corporate identity.” 558 U.S. at 365. In contrast, “foreign organizations operating abroad have no First Amendment rights.” Agency for Int'l Dev. v. All. for Open Soc'y Int'l, Inc., 591 U.S. 430, 436 (2020). This case falls in the middle of the spectrum and calls on us to consider at which point a domestic corporation has sufficient foreign ownership or control that the First Amendment analysis changes.

To make its argument for a lower level of scrutiny, Maine relies heavily on Foley v. Connelie, in which the Supreme Court rejected an equal protection challenge to a state law that limited police jobs to U.S. citizens. 435 U.S. 291, 292-93 (1978). The Court subjected the law to rational basis review, noting that a lower level of scrutiny applied to classifications involving non-citizens where the law prohibited non-citizens “from participation in [a state's] democratic political institutions” or where the state law was “firmly within a State's constitutional prerogatives”; for example, the right to vote, to run for elective office, or to serve on a jury. Id. at 295-96 (quoting Sugarman v. Dougall, 413 U.S. 634, 648 (1973)). Maine seizes on this language to assert that Foley supports a lower level of scrutiny for restrictions on foreign entities’ political speech. Crucially, however, Maine does not argue that the First Amendment's protections do not apply at all to these U.S. corporations, even those with some level of foreign ownership.

In response, Versant and CMP seek to distinguish Foley by noting that it involved an equal protection claim, not the First Amendment, id. at 294, and that restrictions on domestic corporations’ speech trigger strict scrutiny under Citizens United, 558 U.S. at 355, 362. We are not persuaded that Foley applies to this case and accordingly, as we explain below, we apply Citizens United to the restrictions on spending other than contributions.

This is an issue of first impression for this court, and although other courts have faced similar inquiries, we are aware of only one to have decided the issue. In Choi, a federal district court held that a similar law, which would have barred corporations with some foreign ownership from political spending in state and local candidate elections and ballot questions, violated the First Amendment. 765 F. Supp. 3d at 832-33. The district court held that the standard levels of scrutiny under the First Amendment still applied, reasoning that “no case holds that a corporation ceases to be ‘American’ by virtue of any quantum of foreign ownership.” Id. at *12. Other courts facing similar arguments have avoided deciding the applicable level of scrutiny. See Bluman v. FEC, 800 F. Supp. 2d 281, 285-86 (D.D.C. 2011) (noting that determining proper level of scrutiny for federal law barring foreign nationals from making various types of contributions, donations, and expenditures related to candidate elections presented a complex question but concluding that the law withstood even strict scrutiny, thereby avoiding question of applicable level of scrutiny), aff'd, 565 U.S. 1104 (2012) (summarily affirming); see also OPAWL - Bldg. AAPI Feminist Leadership v. Yost, 118 F.4th 770, 772-73, 777 (6th Cir. 2024) (declining to decide appropriate level of scrutiny for Ohio law barring foreign nationals from contributing to candidates in state elections or spending money on ballot initiatives because law withstood all possible levels of scrutiny); cf. id. at 786-87 (Davis, J., dissenting) (agreeing that court did not need to decide applicable level of scrutiny but concluding that was so because law failed both intermediate and strict scrutiny). We find that Foley is inapplicable in this situation, as the Choi court did, and explain our reasoning below.

First, as discussed, Foley involved a Fourteenth Amendment equal protection challenge and not a First Amendment challenge. 435 U.S. at 294; see also Choi, 765 F. Supp. 3d at 849. Second, Foley is factually distinct: it involved a challenge to a state law that limited police jobs to U.S. citizens. 435 U.S. at 292-93. The Court applied rational basis review because states could permissibly exclude non-citizens from participation in certain democratic institutions. Id. at 295-96. The Court upheld the challenged law after “examin[ing] [the] position in question to determine whether it involve[d] discretionary decisionmaking, or execution of policy, which substantially affect[ed] members of the political community.” Id. at 296.

Thus, Foley’s test, formulated in response to a law concerning state employment, does not help guide our decision about the appropriate level of scrutiny here. All of the examples that Foley gives as permissible to reserve to U.S. citizens are roles that individual citizens play in our democracy -- juror, voter, political candidate, public servant -- because these “lie at the heart of our political institutions” and impact “the right to govern.” Id. at 295-97. We do not find this discussion sufficiently applicable to the context of political speech restrictions on a wide range of speakers, including corporations, to persuade us to lower the typical level of scrutiny for the entire Act.

Because the Act applies to domestic actors as well as foreign actors, the First Amendment's protections apply. The Act's restrictions on contributions must withstand exacting scrutiny, see Nixon, 528 U.S. at 387-88, and its remaining burdens on political speech must withstand strict scrutiny, see Citizens United, 558 U.S. at 340.

C. Maine's Proposed State Interests

Below, the district court determined that Maine has a compelling interest in limiting foreign government influence in candidate elections.⁵ Cent. Me. Power Co., 721 F. Supp. 3d at 50-51. Given the “initial stage of the case,” the district court assumed, without deciding, that Maine likely also has a compelling interest in limiting foreign government influence in referenda elections. Id. at 51. However, the district court rejected Maine's argument that it has a compelling interest in limiting the appearance of foreign government influence in both candidate and referenda elections. Id. at 51-52. On appeal, Maine contends that the district court erred in not recognizing all of its proposed compelling interests, while CMP maintains that none of these proposed interests are compelling. Versant also challenges the district court's decision to assume, without deciding, a compelling interest in limiting foreign government influence in referenda elections.

As noted above, the district court applied strict scrutiny to the entirety of the Act, but we think it unnecessary to decide that strict scrutiny should apply to the Act's contribution limits, which would fail even exacting scrutiny. We therefore must consider whether Maine's proposed interests are compelling, in the strict scrutiny analysis, or sufficiently important, in the exacting scrutiny analysis. But we may streamline those analyses because no matter the level of scrutiny, the plaintiffs have established a likelihood of success in establishing that the law is not sufficiently related to any of these proposed state interests. We therefore assume, without deciding, that each of Maine's proposed interests is sufficiently compelling.

D. Tailoring

Having assumed that Maine has the requisite interests in preventing foreign influence or the appearance of foreign influence in its elections, we turn to the question of whether, at this stage of the litigation, Maine is likely to succeed in showing the required fit between the Act's provisions and the state's interests.

For the Act's restrictions on all spending other than contributions, we ask whether Maine is likely to succeed in showing that the Act “serves [the] compelling state interest[s] in a narrowly tailored manner.” We the People PAC v. Bellows, 40 F.4th 1, 19 (1st Cir. 2022). In other words, we first consider the fit between the invoked interest and the Act's restrictions and how the law advances the compelling interest. See id. at 19-20. The restriction on speech must be “actually necessary” to achieve the compelling interest. See United States v. Alvarez, 567 U.S. 709, 725 (2012) (plurality opinion) (quoting Brown v. Ent. Merchs. Ass'n, 564 U.S. 786, 799 (2011)). “There must be a direct causal link between the restriction imposed and the injury to be prevented.” Id. Courts consider the “record evidence or legislative findings” that demonstrate the necessity of curtailing First Amendment rights to fix a problem. FEC v. Ted Cruz for Senate, 596 U.S. 289, 307 (2022) (quoting Colo. Republican Fed. Campaign Comm. v. FEC, 518 U.S. 604, 618 (1996)). Courts also “consider whether the rule is either under- or overinclusive.” Does 1-6 v. Mills, 16 F.4th 20, 33 (1st Cir. 2021). “[U]nderinclusiveness can raise ‘doubts about whether the government is in fact pursuing the interest it invokes, rather than disfavoring a particular speaker or viewpoint.’ ” Williams-Yulee v. Fla. Bar, 575 U.S. 433, 448 (2015) (quoting Brown, 564 U.S. at 802).

Similarly, for the Act's restrictions on contributions, we consider whether Maine is likely to succeed in showing that the Act is “narrowly tailored to serve a sufficiently important governmental interest.” Gaspee Project v. Mederos, 13 F.4th 79, 85 (1st Cir. 2021) (citing Ams. for Prosperity, 594 U.S. at 608). Although the Supreme Court has previously referred to the required relationship as “closely drawn,” see Beaumont, 539 U.S. at 162, the Court has clarified that the required fit amounts to narrow tailoring, see Ams. for Prosperity, 594 U.S. at 608; see also McCutcheon v. FEC, 572 U.S. 185, 197 (2014) (plurality opinion) (describing exacting scrutiny as “a lesser but still ‘rigorous standard of review’ ” (quoting Buckley, 424 U.S. at 29)). Under exacting scrutiny, the restrictions need not “be the least restrictive means of achieving their ends,” but they must be “narrowly tailored to the government's asserted interest.” Ams. for Prosperity, 594 U.S. at 608. This tailoring is required because, “[i]n the First Amendment context, fit matters.” Id. at 609 (quoting McCutcheon, 572 U.S. at 218). Even outside of strict scrutiny, “we still require a fit that is not necessarily perfect, but reasonable; [and] that represents not necessarily the single best disposition but one whose scope is in proportion to the interest served.” Id. (quoting McCutcheon, 572 U.S. at 218). Therefore, although we apply exacting scrutiny to the Act's restrictions on contributions and strict scrutiny to the remainder of the Act's restrictions, our analysis of the fit looks largely the same.

As a reminder, the Act's central prohibition on campaign spending forbids a “foreign government-influenced entity” from making “a contribution, expenditure, independent expenditure, electioneering communication or any other donation or disbursement of funds to influence” a candidate election or a referendum. Tit. 21-A, § 1064(2). The Act provides three definitions for “foreign government-influenced entity”: a foreign government; an entity that is 5% or more owned, directly or indirectly, by a foreign government or a foreign government-owned entity (meaning it is itself more than 50% owned by a foreign government); and an entity in which a foreign government or foreign government-owned entity “[d]irects, dictates, controls or directly or indirectly participates in the decision-making process” regarding the entity's political speech. See id. § 1064(1)(E), (F). We analyze the Act with respect to each of these three definitions.

1. Prohibition on Spending by a Foreign Government

The first of the three definitions of “foreign government-influenced entity” is simply “[a] foreign government,” id. § 1064(1)(E)(1), as defined in the Act, id. § 1064(1)(D).⁶ The district court held that subsection 2's ban on campaign spending by foreign governments is likely narrowly tailored and thus constitutional. Cent. Me. Power Co., 721 F. Supp. 3d at 52. No party challenges this conclusion, but we note it because it is relevant to comparing the Act's “plainly legitimate sweep” with its unconstitutional applications. Moody, 603 U.S. at 723 (quoting Ams. for Prosperity, 594 U.S. at 615). We therefore move to the next two definitions.

2. Prohibition on Spending by an Entity with 5% or More Foreign Ownership

The second definition is “[a] firm, partnership, corporation, association, organization or other entity with respect to which a foreign government or foreign government-owned entity ․ [h]olds, owns, controls or otherwise has direct or indirect beneficial ownership of 5% or more of the total equity, outstanding voting shares, membership units or other applicable ownership interests.” Id. § 1064(1)(E)(2)(a). The district court held that the Act's restrictions with respect to this 5% ownership threshold are likely not narrowly tailored and therefore unconstitutional, for several reasons. Cent. Me. Power Co., 721 F. Supp. 3d at 52. First, the district court determined that this 5% foreign ownership threshold is overinclusive because it would prohibit a substantial amount of protected speech. Id. at 53. For example, the Act prohibits campaign spending by CMP, a company incorporated in Maine and run by United States citizens. Id. Second, the district court noted that the 5% ownership threshold seems to be arbitrarily chosen. Id. Finally, the court observed that Maine had not yet offered any “evidence that a foreign government or foreign government-influenced entity with less than full ownership of a domestic entity ha[d] exerted influence over that entity's election spending in Maine.” Id. The district court thus concluded that this 5% threshold provision is likely not narrowly tailored to a compelling interest in preventing foreign influence in candidate or referenda elections. Id.

On appeal, Maine argues that the district court erred in underestimating the potential influence that a shareholder who owns 5% or more of a corporation may wield over that corporation's decision-making. That includes the shareholder's ability to sell off all of its stock at once if dissatisfied, which would be a major event for a corporation. Maine argues that a shareholder with well under 50% ownership -- and even under 10% ownership -- can use that influence to accomplish significant changes in a corporation. Because of this possibility, Maine argues, the Act is not overinclusive. Maine also contests the need for evidence showing actual influence, arguing that corporate managers’ fiduciary duties to their shareholders will prompt them to anticipate and infer the interests of their largest shareholders. Maine disputes that the 5% threshold was arbitrarily chosen, pointing to federal laws that require special disclosures for any person who owns 5% or more of the equity of a corporation. See 15 U.S.C. § 78m(d). Finally, Maine points to restrictions in other states and cities that use the 5% ownership threshold to demonstrate that such a threshold is commonly used.

We agree with the district court that this 5% foreign ownership threshold for triggering the Act's prohibition on a wide range of political speech is likely not narrowly tailored to the stated compelling interests in preventing foreign influence or its appearance. Regarding subsection 2's ban on contributions, we similarly think that the 5% foreign ownership threshold is likely not closely drawn to match its sufficiently important state interests.

The law is overinclusive because -- as the district court pointed out -- it silences U.S. corporations that have their own First Amendment rights: CMP and Versant were founded in Maine, have operated exclusively there for over a century, and are entirely run by U.S. citizens. See Mills, 16 F.4th at 33. To illustrate why the law is overinclusive, we will look at the strongest evidence in Maine's favor: the evidence that HQUS, a subsidiary of a foreign government's utility company, was the third highest contributor to political action and ballot question committees in Maine over the last decade. The company's owner, Hydro-Québec, and the province of Québec do not have any First Amendment rights. See Agency for Int'l Dev., 591 U.S. at 436. Subsection 2's ban applies to HQUS because of the 5% definition. But HQUS is 100% indirectly owned by a foreign government, as is Versant. Maine's evidence therefore does not demonstrate why the 5% threshold -- as opposed to 100%, or 50%, or any other number -- is narrowly tailored to its interests in preventing foreign influence in its elections. See Cruz, 596 U.S. at 307 (noting that courts consider the record evidence or legislative findings demonstrating why restricting First Amendment rights is needed). Maine has not shown that the Act's curtailment of First Amendment rights in this way is necessary.

The prohibition is overly broad, silencing U.S. corporations based on the mere possibility that foreign shareholders might try to influence its decisions on political speech, even where those foreign shareholders may be passive owners that exercise no influence or control over the corporation's political spending. See Choi, 765 F. Supp. 3d at 852 (“It is not enough ․ to explain how foreign minority shareholders could exercise influence over corporations.”); see also Cruz, 596 U.S. at 307 (deeming the absence of record evidence “significant”). CMP offers a helpful illustration. CMP is captured by the 5% definition because Qatar's sovereign wealth fund indirectly owns 7 to 11% of CMP. But Maine has shown no evidence that Qatar itself has tried to influence CMP's decisions regarding political speech. True, CMP and its affiliates have spent a lot of money -- nearly $73 million -- in Maine's elections over the past decade. But in that time, CMP faced two ballot initiatives aimed at removing its permit for the CMP Corridor -- after $450 million had already been spent on construction -- in addition to a ballot question that sought to seize all of its assets through eminent domain. The record suggests that CMP's spending was motivated by its desire to protect the company's own interests, rather than the independent interests of Qatar.

With this context, we cannot find that Maine's interest in avoiding the appearance or possibility of Qatar's influence justifies entirely silencing CMP's speech in the face of public referenda that could have such detrimental outcomes to its future as a company. See Citizens United, 558 U.S. at 339 (striking down a restriction on political speech where the “purpose and effect [was] to silence entities whose voices the Government deems to be suspect”).

In the face of these examples, the 5% threshold starts to look either like an end-run around Citizens United, aimed at silencing a large swath of corporations merely because they are corporations, or an effort to shape the ongoing debate in Maine about its two primary utility companies by silencing one side -- the companies themselves. See First Nat'l Bank of Bos. v. Bellotti, 435 U.S. 765, 785 (1978) (striking down a prohibition on political speech where it appeared that the legislature aimed to silence one side of the debate on particular ballot questions). Neither is permissible under the First Amendment.

At oral argument, Maine defended the 5% definition by pointing out that the record showed only two particular companies in Maine with between 5 and 50% foreign ownership. But we are not persuaded that this demonstrates narrow tailoring. Instead, it suggests that the Act was targeted at particular companies. See Williams-Yulee, 575 U.S. at 448 (noting that underinclusiveness may indicate that the government seeks to disfavor a particular speaker); see also Citizens United, 558 U.S. at 340 (laws may violate the First Amendment when they “identif[y] certain preferred speakers”).

Finally, we note that the amount of uncertainty as to which corporations are covered by the law will potentially have a chilling effect. The Act does not set any particular moment in time for determining the level of foreign ownership, which -- for publicly traded corporations -- can fluctuate throughout the course of a day.⁷ As a consequence, U.S. corporations with First Amendment protections will likely choose not to speak at all rather than risk criminal penalties.⁸

We are sympathetic to Maine and amici on the difficulty of ascertaining when foreign shareholders are wielding influence over a domestic corporation's decisions on political speech. That, however, does not alter our conclusion that the Act likely sweeps far too broadly to be narrowly tailored.⁹ See Citizens United, 558 U.S. at 362 (noting in dicta that the provision at issue, which was “not limited to corporations or associations that were created in foreign countries or funded predominately by foreign shareholders,” would still be overbroad even if the government had “a compelling interest in limiting foreign influence over our political process”).

3. Prohibition on Spending by an Entity with Direct or Indirect Foreign Participation in the Decision-making Process

The third and final definition of “foreign government-influenced entity” is:

A firm, partnership, corporation, association, organization or other entity with respect to which a foreign government or foreign government-owned entity ․ [d]irects, dictates, controls or directly or indirectly participates in the decision-making process with regard to the activities of the firm, partnership, corporation, association, organization or other entity to influence the nomination or election of a candidate or the initiation or approval of a referendum, such as decisions concerning the making of contributions, expenditures, independent expenditures, electioneering communications or disbursements.

Tit. 21-A, § 1064(1)(E)(2)(b). We will refer to this as the “actual participation” definition as a shorthand.

The district court held that the Act's restrictions on the entities encompassed by this definition were not narrowly tailored and were likely unconstitutional. Cent. Me. Power Co., 721 F. Supp. 3d at 54-55. The district court noted that this definition seemed, “[a]t first blush,” to be a closer fit to Maine's interest than the previous 5% definition. Id. at 54. But Maine defended this provision by pointing to the Ethics & Election Commission's then-proposed rules, and the district court observed that those proposed rules seemed to broaden the Act by eliminating the statutory requirement that the foreign government or foreign government-owned entity actually “participate[ ]” in the decision-making process. Id. at 54-55. Therefore, in response to Maine's reliance on these proposed rules to defend the statute, the district court found that this category was “overly broad” and “likely to stifle the speech of domestic corporations regardless” of actual foreign influence. Id. at 55. However, the district court noted that its conclusion might change if the Ethics & Election Commission adopted new rules indicating that actual participation was required. Id. at 55 n.21.

Maine argues on appeal that the definition in § 1064(1)(E)(2)(b) mirrors a federal regulation implementing FECA. See 11 C.F.R. § 110.20(i). Maine also focuses, as it did below, on the Ethics & Election Commission's rules and argues that the proposed rules discussed in the district court's decision have since been rewritten to clarify and narrow the definition of “participate.” Finally, Maine asserts that there are no less restrictive means to achieve its compelling interest, as it would not be able to enforce a law that targets only foreign governments and must be able to regulate “the U.S.-based recipients of such influence.”

We start with Maine's argument based on the text of the statute: that the Act's “actual participation” definition was “lifted almost verbatim” from a federal regulation. That federal regulation states:

A foreign national shall not direct, dictate, control, or directly or indirectly participate in the decision-making process of any person, such as a corporation, labor organization, political committee, or political organization with regard to such person's Federal or non-Federal election-related activities, such as decisions concerning the making of contributions, donations, expenditures, or disbursements in connection with elections for any Federal, State, or local office or decisions concerning the administration of a political committee.

11 C.F.R. § 110.20(i). Certainly, the verbs used in the Maine statute are almost the same. See tit. 21-A, § 1064(1)(E)(2)(b) (“[d]irects, dictates, controls or directly or indirectly participates”). But the subjects of the two provisions -- “a foreign national” as opposed to “a foreign government or foreign government-owned entity” -- are not. The federal regulation defines “foreign national” as including:

(1) a government of a foreign country and a foreign political party;

(2) a person outside of the United States, unless it is established that ․ such person is not an individual and is organized under or created by the laws of the United States or of any State or other place subject to the jurisdiction of the United States and has its principal place of business within the United States; and

(3) a partnership, association, corporation, organization, or other combination of persons organized under the laws of or having its principal place of business in a foreign country.

22 U.S.C. § 611(b) (defining “foreign principal”); see 11 C.F.R. § 110.20(a)(3)(i) (defining “[f]oreign national” as equivalent to 22 U.S.C. § 611(b)’s definition of “foreign principal”).

There is some overlap between the Act's definition of a “foreign government or foreign government-owned entity” and the federal regulation's definition of “foreign national.” For example, both use substantially the same definition of “foreign government.” Compare tit. 21-A, § 1064(1)(D), with 22 U.S.C. § 611(e). But the federal provision also encompasses corporations organized or with their principal places of business abroad, many of which have no First Amendment rights. See Agency for Int'l Dev., 591 U.S. at 436. The Maine statute, on the other hand, encompasses entities at least 50% owned by a foreign government -- even if those entities are U.S. corporations, which the federal definition expressly excludes. See 22 U.S.C. § 611(b)(2). The Act thus applies to a broader swath of U.S. corporations than the federal provision and is therefore less tailored. In light of the different scope of the two provisions, we find unpersuasive Maine's argument that is premised on the provisions’ alleged similarity.

In looking at the “actual participation” definition as a whole, we agree with the district court that it appears more tailored than the 5% threshold to Maine's interest in limiting foreign influence or its appearance in state and local elections, because it focuses on conduct. But Maine has made no effort to defend the statute on its own terms, other than its alleged similarity to a federal regulation, which we dismissed above as unpersuasive. We thus remain concerned that the “actual participation” definition applies to too broad a swath of speakers with First Amendment rights to be narrowly tailored. For the same reasons, we think that the “actual participation” definition as applied to the Act's ban on contributions is not narrowly tailored to match a sufficiently important state interest.¹⁰

E. Additional Provisions and Potential Overbreadth of the Injunction

Maine argues that, even if two of the three definitions of “foreign government-influenced entity” have unconstitutional applications, the district court erred in finding the Act facially invalid because the Act's overbreadth is not “substantial ․ relative to its plainly legitimate sweep.” In support of this argument, Maine notes that the court held that the Act was likely constitutional as applied to foreign governments. Maine argues that, if the 5% threshold is too low, the Act is still constitutional as applied to entities with a higher percentage of foreign ownership. In addition, Maine specifically endorses the constitutionality of the disclaimer provision. See tit. 21-A, § 1064(6).

As discussed, most of the applications of the Act's central provision, subsection 2, are likely unconstitutional due to the overly broad definitions of “foreign government-influenced entity.” Regarding the 5% definition, to the extent that Maine implies that the court should have chosen a different threshold of foreign ownership to which the Act should apply, we disagree. We will set aside the issue that we are a federal court reviewing a state law that the state's own court has not yet interpreted, which in itself would be reason to tread cautiously. Cf. United States v. Stevens, 559 U.S. 460, 474 (2010) (“Because [the challenged law] is a federal statute, there is no need to defer to a state court's authority to interpret its own law.”). Courts “will not rewrite a law to conform it to constitutional requirements.” Id. at 481 (cleaned up) (quoting Reno v. Am. C.L. Union, 521 U.S. 844, 884-85 (1997)). “[D]oing so would constitute a ‘serious invasion of the legislative domain’ and sharply diminish [the legislature's] ‘incentive to draft a narrowly tailored law in the first place.’ ” Id. (citation omitted) (first quoting United States v. Nat'l Treasury Emps. Union, 513 U.S. 454, 479 n.26 (1995); and then quoting Osborne v. Ohio, 495 U.S. 103, 121 (1990)).

Moving past subsection 2, most of the substantive provisions of the Act are entwined with that provision. Subsections 3, 4, and 5 prohibit soliciting, accepting, or assisting in transactions that are prohibited under subsection 2, or structuring transactions to evade subsection 2's prohibitions. See tit. 21-A, § 1064(2)-(5). Subsection 9 provides criminal penalties for violating subsections 2 through 5. Id. § 1064(9). Subsection 8 provides civil penalties for violating any of the Act's subsections. Id. § 1064(8). Given how the Act is structured around subsection 2's constitutionally problematic ban (as the definitions currently stand), the overwhelming majority of applications of these other subsections are necessarily unconstitutional as well.

The remaining substantive provisions are subsection 6, which requires a specific disclaimer on any permissible public communications pertaining to a range of political speech by foreign government-influenced entities, and subsection 7, which requires media platforms to conduct due diligence to ensure that they have not published any public communications that violate this Act. Id. § 1064(6), (7). But even if we assume that both subsections 6 and 7 are constitutional, a substantial number of the statute's applications are still likely unconstitutional as compared to the statute's plainly legitimate sweep. See Moody, 603 U.S. at 723.

F. Reserving the Question of Severability

After concluding that two of the three central provisions of the Act are likely to fail under strict scrutiny, the district court held that the Act was likely facially invalid because a substantial number of the Act's applications were unconstitutional as compared to its plainly legitimate sweep. Cent. Me. Power Co., 721 F. Supp. 3d at 55. The district court declined to sever the Act, given the rushed nature of the proceedings, until the parties had a chance to brief the issue of whether particular portions were severable under Maine law. Id.

Maine argues on appeal that the district court abused its discretion by not analyzing severability. Maine further argues that the definitions of “foreign government-influenced entity” are easily severable from the rest of the Act and that the district court should not have enjoined subsection 6's disclaimer requirement without analyzing its constitutionality.

However, while Maine raises colorable arguments concerning the Act's severability under state law, it points to no federal or state authority that required the district court to consider that question at the preliminary injunction stage. We leave the issue of severance for the district court to decide in the first instance.

G. Preemption

Issues of federal preemption are questions of statutory interpretation that we review de novo. See DiFiore v. Am. Airlines, Inc., 646 F.3d 81, 85 (1st Cir. 2011).

The district court concluded that FECA explicitly preempted the Act to the extent that it might be read to apply to federal elections. Cent. Me. Power Co., 721 F. Supp. 3d at 43. Maine agrees that the Act does not apply to foreign spending in elections for federal office, but it argues here, as it did below, that this result should be reached through statutory interpretation rather than preemption. However, Maine only disputes the district court's preemption holding to the extent that it informed the district court's ultimate conclusion, in considering the motions for a preliminary injunction, that the plaintiffs were likely to succeed on the merits of their First Amendment claim.

In response, Versant urges us to affirm the district court's holding that federal law expressly preempts the Act from applying to federal elections, noting that this holding may well have contributed to the district court's conclusion that a substantial number of the statute's applications were likely unconstitutional and therefore the law was likely facially invalid. Versant also worries that a future Ethics & Election Commission will reverse course and apply the Act to federal elections, unconstrained by any limiting statutory language.

We do not read the district court's decision that a substantial number of the statute's applications are likely unconstitutional -- and that plaintiffs were likely to succeed on the merits -- as hinging on its determination that the statute was preempted as to federal candidate elections. The district court determined that subsection 2's ban on campaign spending was likely unconstitutional as it applied to the two broadest of the three statutory definitions because only the “foreign government” definition was likely narrowly tailored. Cent. Me. Power Co., 721 F. Supp. 3d at 52, 55; see tit. 21-A, § 1064(1)(E). As explained, we have determined no error as to these conclusions and, accordingly, it is not necessary to discuss the merits of the preemption determination in affirming the injunction.

V. Conclusion

For the foregoing reasons, we affirm the judgment of the district court.

I write separately for two reasons: first, to state my view that Maine's asserted government interests for its law are inadequate; and second, to identify a possible vagueness problem caused by the law's definition of foreign government.

A. Maine's Interests

I agree with the Court that the plaintiffs will likely prevail on their First Amendment challenges to Maine's law because the law sweeps too broadly into areas of protected speech by American companies. I would, however, reach that conclusion by following a somewhat different path.

I would not assume that Maine's interest in limiting “foreign government influence” or “the appearance of such influence” on political speech by American companies is a compelling or even important government interest. Rather, as I see it, First Amendment principles dictate that the government's only compelling or important interest in this realm is to prevent actual participation by foreign persons and entities in the American political process, i.e., in “activities ‘intimately related to the process of democratic self-government,’ ” Bluman v. FEC, 800 F. Supp. 2d 281, 287 (D.D.C. 2011) (quoting Bernal v. Fainter, 467 U.S. 216, 220 (1984)), aff'd, 565 U.S. 1104 (2012), or the appearance of such participation. Anything less impermissibly interferes with the rights of Americans to engage in political speech. Therefore, I would conclude that Maine's law likely is significantly overbroad because none of the restrictions at issue are commensurate with this far more limited government interest.

The Supreme Court has repeatedly stated that the “ ‘[d]iscussion of public issues and debate on the qualifications of candidates [is] integral to the operation’ of our system of government.” Ariz. Free Enter. Club's Freedom Club PAC v. Bennett, 564 U.S. 721, 734 (2011) (quoting Buckley v. Valeo, 424 U.S. 1, 14 (1976) (per curiam)); see, e.g., Eu v. S.F. Cnty. Democratic Cent. Comm., 489 U.S. 214, 223 (1989); Garrison v. Louisiana, 379 U.S. 64, 74–75 (1964). As such, “the First Amendment ‘has its fullest and most urgent application’ to speech uttered during a campaign for political office.” Eu, 489 U.S. at 223 (quoting Monitor Patriot Co. v. Roy, 401 U.S. 265, 272 (1971)). This principle applies equally to candidate-based and issue-based elections. McIntyre v. Ohio Elections Comm'n, 514 U.S. 334, 347 (1995).

The Supreme Court also has emphasized that “political speech does not lose First Amendment protection ‘simply because its source is a corporation.’ ” Citizens United v. FEC, 558 U.S. 310, 342 (2010) (quoting First Nat'l Bank of Boston v. Bellotti, 435 U.S. 765, 784 (1978)). That is because “[c]orporations ․, like individuals, contribute to the ‘discussion, debate, and the dissemination of information and ideas’ that the First Amendment seeks to foster.” Pac. Gas & Elec. Co. v. Pub. Util. Comm'n of Cal., 475 U.S. 1, 8 (1986) (plurality opinion) (quoting Bellotti, 435 U.S. at 783); see Citizens United, 558 U.S. at 342-43. Thus, the political speech of American corporations should be treated no differently than the speech of American “natural persons.” Citizens United, 558 U.S. at 343 (quoting Bellotti, 435 U.S. at 776).

The Maine law bans, among other things, American corporate political speech that is influenced or appears to be influenced by a “foreign government” or a “foreign government-owned entity.” See Me. Rev. Stat. Ann. tit. 21-A, § 1064(1)(D)-(F), (2) (2024). In other words, the law seeks to limit what American companies may say in Maine political campaigns because foreign sources may supply some of the information that helps to shape an American company's speech choices. That, in my view, presents a serious constitutional problem.

For over a half a century, the Supreme Court has recognized that “the Constitution protects the right to receive information and ideas ․ regardless of their social worth.” Stanley v. Georgia, 394 U.S. 557, 564 (1969). The First Amendment also recognizes that an important “manifestation of the principle of free speech,” “enjoyed [alike] by business corporations generally and by ordinary people,” Hurley v. Irish-Am. Gay, Lesbian & Bisexual Grp. of Bos., 515 U.S. 557, 573-74 (1995), is the “choice[ ] of what to say and what to leave unsaid,” id. at 573 (quoting Pac. Gas & Elec. Co., 475 U.S. at 11).

Maine seems to be concerned that foreign governments will influence the outcome of Maine elections by inducing American companies to spread foreign-sponsored messages to the electorate through an American speaker. But the Maine law does not seek to silence only foreign speech; it also seeks to suppress the speech of American companies that might have been swayed by it. Such targeting of an American speaker's right to engage in core political speech is anathema to the First Amendment: “Those who seek to censor or burden free expression often assert that disfavored speech has adverse effects[, b]ut the ‘fear that people w[ill] make bad decisions’ ” based on arguments and information provided by others generally “cannot justify content-based burdens on speech.” Sorrell v. IMS Health Inc., 564 U.S. 552, 577 (2011) (quoting Thompson v. W. States Med. Ctr., 535 U.S. 357, 374 (2002)).

Every speaker's choice of what to say is influenced by many factors, including information and opinions gleaned from external sources. Generally, we trust counterspeech, not government regulation of the speaker, to dissuade people from adopting bad or false ideas that a speaker may offer. See United States v. Alvarez, 567 U.S. 709, 726-28 (2012) (plurality opinion). Allowing Maine to silence an American speaker because it does not like a source of information which may have influenced that speaker does not square with the basic First Amendment principles recognizing the rights to receive information and to speak one's ideas. See Sorrell, 564 U.S. at 577. I would reject out of hand the interest Maine appears to assert in silencing an American speaker on political matters.

To be clear, that does not mean that Maine is powerless to prevent foreign government speech in its elections. There was much debate in this appeal about whether the Supreme Court's summary affirmance in Bluman binds this Court. I find that debate irrelevant because, even if it does not bind our disposition of this case, Bluman articulates a proper understanding of the contours of the government's permissible interest in restricting foreign participation in American elections.

Bluman recognized that the government “has a compelling interest for purposes of First Amendment analysis in limiting the participation of foreign citizens in activities of American democratic self-government, and in thereby preventing foreign influence over the U.S. political process.” Bluman, 800 F. Supp. 2d at 288. Maine seizes on the phrase “foreign influence over the U.S. political process” as supporting the broad governmental interest that it proposes. But in so doing, Maine isolates that phrase from the rest of the quoted sentence and the case more generally.

Bluman involved a direct restriction on a foreign citizen making a political contribution or independent spending in a political campaign. See id. at 282-83. Bluman held that these acts -- the actual giving to a candidate or independent spending by a foreign citizen -- may be constitutionally proscribed to protect our democratic processes from foreign influence. Id. at 288-89. In other words, Bluman held that the government may close off “activities ‘intimately related to the process of democratic self-government’ ” from foreign participation. Id. at 287 (quoting Bernal, 467 U.S. at 220). Thus, Bluman holds only that the government may forbid foreign persons or entities from actually participating in the American political process. Id.; see Agency for Int'l Dev. v. All. for Open Soc'y Int'l, Inc., 591 U.S. 430, 436 (2020) (recognizing that First Amendment rights do not extend to foreign organizations operating abroad).

But Maine incorrectly reads Bluman to go further. It reads Bluman as authorizing the government to prevent an American company from itself speaking because it consults or has some other contact with a foreign government before it decides what to say. That is a misreading of Bluman. Bluman does not support the regulation of this sort of secondhand foreign influence on the American political process.

In line with Bluman, Citizens United seems to have adopted, in the corporate context, the same interest in preventing actual foreign participation in the American political process, or the appearance thereof. See Citizens United, 558 U.S. at 362. One of the government's arguments in Citizens United against extending First Amendment rights to corporations was that it would allow “foreign individuals or associations [to] influenc[e] our Nation's political process.” See id.

The Supreme Court stated that, even if there were such a compelling interest, it would only extend to “corporations or associations that were created in foreign countries or funded predominately by foreign shareholders.” Id. at 362. The Court's explanation suggests that any government interest in restricting corporate political speech would be limited to situations where the foreign corporations were themselves speaking or where the American company was predominantly funded by foreign shareholders such that these shareholders in effect controlled or appeared to control the company's speech. See id. Maine's asserted interest, however, is far broader insofar as it suggests that a foreign government's influence over an American company, even when small, provides grounds for silencing an American company. Citizens United does not contemplate such an interest. See id.

In sum, this case presents an important question about when the government may prohibit speech in a political campaign by an American corporation. Maine asserts that it has the power to do so whenever it appears that a foreign government might have influenced an American company's speech choice. But absent foreign government control, it is the American company that ultimately decides what to say. That decision by an American speaker is protected by the First Amendment. While I agree with the Court that Maine's law would infringe on the First Amendment even if Maine's asserted interests were compelling or important, I would say now that they are not.

B. Vagueness

I also want to raise a concern about potential vagueness in Maine's law because, even after today's ruling, the case remains in its early stages. My vagueness concern relates to the law's definition of “foreign government.” See tit. 21-A, § 1064(1)(D).

The “foreign government” definition plays a central role in the Maine law's application. The law prohibits, and even criminalizes, otherwise constitutionally protected political speech by American companies that are “foreign government-influenced entities.” See tit. 21-A, § 1064(2). Each method by which an American company becomes a “foreign government-influenced entity” leads back to the law's definition of “foreign government.” See id. § 1064(1)(D)-(F). Thus, under Maine's scheme, an American company must determine whether it has a relevant relationship with a “foreign government.”

The statute defines “foreign government” as follows:

“Foreign government” includes any person or group of persons exercising sovereign de facto or de jure political jurisdiction over any country other than the United States or over any part of such country and includes any subdivision of any such group and any group or agency to which such sovereign de facto or de jure authority or functions are directly or indirectly delegated. “Foreign government” includes any faction or body of insurgents within a country assuming to exercise governmental authority, whether or not such faction or body of insurgents has been recognized by the United States.

Id. § 1064(1)(D).

Maine borrowed this definition from the Foreign Agents Registration Act's definition of “government of a foreign country.” See 22 U.S.C. § 611(e). The Foreign Agents Registration Act, inter alia, prohibits a person from acting as an “agent of a foreign principal” -- which includes acting as an agent of the “government of a foreign country,” id. § 611(b)(1), (c)(1) -- unless the person first files a registration with the Attorney General. Id. § 612(a).

There are a handful of cases holding that the Foreign Agents Registration Act's registration requirement comports with the Constitution. See, e.g., United States v. Peace Info. Ctr., 97 F. Supp. 255, 262 (D.D.C. 1951) (“The statute under consideration neither limits nor interferes with freedom of speech. It does not regulate expression of ideas. Nor does it preclude the making of any utterances. It merely requires persons carrying on certain activities to identify themselves by filing a registration statement.”); Att'y Gen. v. Irish N. Aid Comm., 346 F. Supp. 1384, 1389-90 (S.D.N.Y. 1972) (holding that the registration requirement is constitutional); Att'y Gen. v. Irish N. Aid Comm., 530 F. Supp. 241, 253 (S.D.N.Y. 1981) (same). Because the registration requirement does not implicate constitutional rights and contains a specific-intent mens rea for the criminal penalties arising from a failure to register, see 22 U.S.C. § 618(a), vagueness concerns about identifying a “government of a foreign country,” id. § 611(b)(1), (e), may not be substantial in the registration context. See Screws v. United States, 325 U.S. 91, 102 (1945) (stating that a statute is not likely to be vague when a conviction requires a specific intent to violate the statute); Vill. of Hoffman Ests. v. Flipside, Hoffman Ests., Inc., 455 U.S. 489, 494-95 (1982) (stating that when a law “implicates no constitutionally protected conduct,” a statute is impermissibly vague only if it is “vague in all of its applications”).

But the context into which Maine has imported the “government of a foreign country” definition is quite different. Here, as discussed, the Maine law has obvious and substantial constitutional implications under the First Amendment. According to the law, whether an American company is silenced from engaging in political speech depends on whether that company maintains a covered relationship with a “foreign government.” Tit. 21-A, § 1064(1)(E), (2).

Because of the Maine law's First Amendment implications, it is essential that the definition of “foreign government” be sufficiently clear to provide American companies with adequate notice of when they must desist from otherwise protected speech. That is especially so where a company's wrong assessment of its speech rights exposes it to criminal penalties under a mens rea standard that is less protective than specific intent -- the Maine law imposes a mens rea of “knowing,” which typically indicates a general intent crime. Tit. 21-A, § 1064(9); Bryan v. United States, 524 U.S. 184, 193 (1998) (“[U]nless the text of the statute dictates a different result, the term ‘knowingly’ merely requires proof of knowledge of the facts that constitute the offense.” (footnote omitted)).

The Supreme Court raised similar concerns in Reno v. ACLU, 521 U.S. 844 (1997). Reno recognized that vagueness “raise[s] special First Amendment concerns because of its obvious chilling effect on free speech.” Reno, 521 U.S. at 871-72. These concerns are heightened when criminal penalties are involved because “criminal sanctions may well cause speakers to remain silent rather than communicate even arguably unlawful words.” Id. at 872. “[S]tandards of permissible statutory vagueness are [thus] strict in the area of free expression.” Keyishian v. Bd. of Regents of Univ. of N.Y., 385 U.S. 589, 604 (1967) (quoting NAACP v. Button, 371 U.S. 415, 432 (1963)). “First Amendment freedoms need breathing space to survive, [and therefore] government may regulate in the area only with narrow specificity.” Id. (quoting NAACP, 371 U.S. at 433).

I am concerned that Maine's law fails to regulate with the required “narrow specificity.” Keyishian, 385 U.S. at 604 (quoting NAACP, 371 U.S. at 433). The definition of “foreign government,” the fulcrum on which the law pivots, is exceedingly broad. It covers “de facto ․ political jurisdiction” exercised by a “group” or “any subdivision of any such group” over “any part of [any] country” other than the United States. Tit. 21-A, § 1064(1)(D). It also reaches “any faction or body of insurgents within a country assuming to exercise governmental authority, whether or not such faction or body of insurgents has been recognized by the United States.” Id.

We live in a complex world. Are the Houthis a “foreign government” in Yemen under Maine's foreign government definition? How about MS-13 in El Salvador? Boko Haram in Nigeria? Or even kibbutzim in Israel? The hard calls are everywhere and endless.

That Maine requires each company to monitor what groups or people may be purchasing its shares is difficult enough. But the law also requires each company to make granular judgments about the power that each “group,” “subdivision of ․ such group,” or “body of insurgents” has within any part of any country at any time. Id. It would be a tall task for our State Department to make these determinations. It seems to me it would be almost impossible for a business or media group confidently to make such judgments in constantly changing political environments.

As I see it, there is a likelihood that the “foreign government” definition, the linchpin provision of Maine's law, is sufficiently vague that people “of common intelligence must necessarily guess at its meaning and differ as to its application.” Baggett v. Bullitt, 377 U.S. 360, 367 (1964) (emphasis added). If I am right, a company otherwise wishing to participate in a Maine election would likely abstain from political speech entirely -- especially given the criminal penalties that may attach from an inaccurate evaluation of the political situation in a faraway place at any given time. See tit. 21-A, § 1064(9). It is precisely to avoid such chilling of speech that the Supreme Court has closely policed statutory vagueness in areas implicating free expression. See NAACP, 371 U.S. at 432-33. As this case returns to the district court, I urge consideration of this potential vagueness problem.

FOOTNOTES

1.   A second section of the initiative, aimed at promoting an anticorruption amendment to the U.S. Constitution that would limit spending in state and federal elections, was not challenged in this case and is not at issue.

2.   Iberdrola now owns 100% of Avangrid.

3.   Maine uses the term “ ‘closely drawn’ scrutiny” (quoting Nixon v. Shrink Mo. Gov't PAC, 528 U.S. 377, 387 (2000)), but we use the term “exacting scrutiny” for the same concept. See Daggett v. Comm'n on Governmental Ethics & Election Pracs., 205 F.3d 445, 454 (1st Cir. 2000) (discussing the Supreme Court's reference to “exacting scrutiny”); see also Fortuño, 699 F.3d at 11 (describing Daggett as “applying exacting scrutiny to limits on direct contributions”).

4.   CMP argues that we would have to overturn Fortuño to apply anything other than strict scrutiny to any aspect of the Act. But Fortuño does not control here. The law in Fortuño created detailed requirements with which corporations and unions had to comply to make campaign contributions or political expenditures. 699 F.3d at 5. There, we applied strict scrutiny to the entire law because it “impose[d] substantial burdens on the very process through which a [speaker] determine[d] whether and how to exercise its free speech rights,” “reach[ing] deep into the mechanics of an organization's own self-governance” and “regulat[ing] the if and how of a[n organization]’s political speech.” Id. at 12 (emphasis added). In contrast, the Act contains a more straightforward restriction on contributions and does not regulate the process by which decisions about speech are made.

5.   The district court relied on Bluman in making this determination, and the parties now dispute whether that case is controlling here. Cent. Me. Power Co., 721 F. Supp. 3d at 50-51 (citing Bluman, 800 F. Supp. 2d at 282-83, 285, 288, 292 n.4). The Supreme Court summarily affirmed Bluman and, accordingly, there is some question as to whether its conclusion is binding or merely persuasive. See Anderson v. Celebrezze, 460 U.S. 780, 784 n.5 (1983) (holding that “the precedential effect of a summary affirmance extends no further than ‘the precise issues presented and necessarily decided by those actions.’ ” (quoting Ill. State Bd. of Elections v. Socialist Workers Party, 440 U.S. 173, 182-83 (1979))); Auburn Police Union v. Carpenter, 8 F.3d 886, 894 (1st Cir. 1993) (cautioning that “summary dispositions ‘should not be understood as breaking new ground but as applying principles established by prior decisions to the particular facts involved’ ” (quoting Mandel v. Bradley, 432 U.S. 173, 176 (1977))). However, we need not reach this dispute because we assume that all of Maine's proposed interests are compelling.

6.   The Act defines “foreign government” as including “any person or group of persons exercising sovereign de facto or de jure political jurisdiction over any country other than the United States or over any part of such country.” Tit. 21-A, § 1064(1)(D). The term also includes “any subdivision of any such group and any group or agency to which such sovereign de facto or de jure authority or functions are directly or indirectly delegated” and “any faction or body of insurgents within a country assuming to exercise governmental authority, whether or not such faction or body of insurgents has been recognized by the United States.” Id.

7.   Indeed, CMP's ownership changed while this appeal was pending.

8.   Maine cited similar provisions from other states that restrict political speech based on foreign ownership of corporations, but these are not persuasive. The St. Petersburg law has been preempted by the Florida legislature. Fla. Stat. § 106.08(11) (2023). The Minnesota law, which applied to companies with only 1% foreign ownership, Minn. Stat. § 211B.15(d), was struck down as a violation of the First Amendment. Minn. Chamber of Com. v. Choi, 765 F. Supp. 3d 821, 858 (D. Minn. 2025). While Alaska has a similar 5% threshold, Alaska Stat. § 15.13.068(e)(5)(A) (2018), it does not appear to have faced a constitutional challenge. Neither has Seattle's ordinance, which has a lower threshold. See Seattle, Wash., Mun. Code §§ 2.04.010, .370, .400 (2025).

9.   We also agree with the district court that the federal securities law that Maine cites does not provide a persuasive analogy here, as that law requires a particular disclosure at the 5% ownership threshold but does not indicate that 5% is necessarily a proxy for control. See Cent. Me. Power Co., 721 F. Supp. 3d at 53; 15 U.S.C. § 78m(d)(1)-(3).

10.   Maine argues that the Ethics & Election Commission's new rules clarify that the statutory term “directly or indirectly participate” is sufficiently narrow and that these new rules render the “actual participation” definition constitutional. These rules were not before the district court, see Cent. Me. Power Co., 721 F. Supp. 3d at 55 n.21, and we thus conclude that the district court acted within its discretion in issuing its injunction based on an earlier iteration of them.

MONTECALVO, Circuit Judge.