VOTING FOR AMERICA, INC.; Brad Richey; Penelope McFadden; Project Vote, Inc., Plaintiffs–Appellees v. John STEEN, in his Official Capacity as Texas Secretary of State, Defendant–Appellant.
-- October 03, 2013
Chad Wilson Dunn, Esq., Brazil & Dunn, Houston, TX, Ryan Morland Malone, David C. Peet Ropes & Gray, L.L.P., Brian Mellor, Project Vote, Washington, DC, for Plaintiffs–Appellees.Jonathan F. Mitchell, Solicitor General, Kathlyn C. Wilson, Esq., Assistant Attorney General, Office of the Attorney General, Austin, TX, for Defendant–Appellant.Lawrence John Joseph, Washington, DC, for Eagle Forum Education and Legal Defense Fund, Amicus Curiae.Farrah Robyn Berse, Paul, Weiss, Rifkind, Wharton & Garrison, L.L .P., Ian Arthur Vandewalker, Esq., Brennan Center for Justice, New York, NY, for Brennan Center for Justice at New York University School of Law, League of Women Voters, Rock the Vote, Amicus Curiae.Susan Lea Hays, Esq., Dallas, TX, for Former Governor Mark White, Former Lieutenant Governor William, Members of The 69th Texas Legislature, Amicus Curiae.
Appellees Voting for America, Inc., Brad Richey, Penelope McFadden, and Project Vote, Inc. (“Appellees”) sued Texas Secretary of State John Steen (“Steen”)1 for declaratory and injunctive relief against several provisions of Texas's law regulating volunteer deputy registrars, Tex. Elec.Code Ann. § 13.031 et seq. (“VDR Law”). The district court granted a preliminary injunction against three provisions for violating the First Amendment and two provisions for violating the National Voter Registration Act (“NVRA”), 42 U.S.C. §§ 1973gg et seq. A motions panel of this court granted a stay pending appeal. Amplifying the motions panel's majority opinion, we conclude that Appellees failed to establish facial unconstitutionality of the challenged provisions. The preliminary injunction is reversed and the case remanded for further proceedings.
Steen administers the VDR Law, which regulates the appointment and activities of volunteer deputy registrars (“VDRs”), individuals trained and empowered to receive and deliver completed voter registration applications. After Texas amended the VDR Law in 2011, Appellees requested Steen's interpretation of several provisions. Unsatisfied with the answers, Appellees filed this lawsuit seeking injunctive relief on the basis that several provisions restrict their ability to conduct voter registration drives in violation of the First Amendment and the NVRA.
In May 2012, Appellees moved for a preliminary injunction. After a hearing, the district court granted the motion in part, enjoining enforcement of the following provisions of the VDR Law:
(1) Texas Election Code § 13.031(d)(3) to the extent it forbids non-Texas residents from serving as VDRs (the “Non–Resident Provision”);
(2) Texas Election Code § 13.038 to the extent it prohibits VDRs appointed in one county from serving in another county (the “County Provision”);
(3) Texas Election Code § 13.008(a)(2) & (3) (the “Compensation Provision”);
(4) Texas Election Code § 13.038 to the extent it prohibits VDRs from photocopying or scanning voter registration applications submitted to the VDR but not yet delivered to the county registrar (so long as no information deemed confidential under § 13.004 is included) (the “Photocopying Provision”); and
(5) Texas Election Code § 13.042 to the extent it prohibits VDRs from sending completed voter registration applications via United States mail (the “Personal Delivery Provision”).2
Steen appealed and moved for a stay pending appeal, which the district court denied. A motions panel of this court granted the stay after hearing oral argument. The U.S. Supreme Court denied Appellees' emergency application to vacate the stay pending appeal. Voting for Am., Inc. v. Andrade, –––U.S. ––––, 133 S.Ct. 99, 183 L.Ed.2d 737 (2012) (Justice Sotomayor would have granted the application in part).
On September 26, 2012, the motions panel issued an unpublished opinion explaining its reasons for granting the stay, along with a dissenting opinion. Voting for Am., Inc. v. Andrade, 488 F. App'x 890 (5th Cir.2012) [hereinafter Andrade II ] (unpublished). Although we are not bound by the ruling of motions panel in the same case, see Mattern v. Eastman Kodak Co., 104 F.3d 702, 704 (5th Cir.1997), we substantially agree with the majority's analysis, and we conclude that Appellees have not established a strong likelihood of prevailing on the merits of their claims.
II. Standard of Review
“To be entitled to a preliminary injunction, the applicants must show (1) a substantial likelihood that they will prevail on the merits; (2) a substantial threat that they will suffer irreparable injury if the injunction is not granted; (3) their substantial injury outweighs the threatened harm to the party to be enjoined; and (4) granting the preliminary injunction will not disserve the public interest.” Tex. Med. Providers Performing Abortion Servs. v. Lakey, 667 F.3d 570, 574 (5th Cir.2012) (internal citation omitted). This court has repeatedly cautioned that “a preliminary injunction is an extraordinary remedy which should not be granted unless the party seeking it has ‘clearly carried the burden of persuasion on all four requirements.’ “ Id. (quoting Lake Charles Diesel, Inc. v. Gen. Motors Corp., 328 F.3d 192, 195–96 (5th Cir.2003)). We review a district court's ultimate issuance of a preliminary injunction for an abuse of discretion. Janvey v. Alguire, 647 F.3d 585, 592 (5th Cir.2011). “As to each element of the district court's preliminary-injunction analysis, however, the district court's findings of fact are subject to a clearly-erroneous standard of review, while conclusions of law are subject to broad review and will be reversed if incorrect.” Dennis Melancon, Inc. v. City of New Orleans, 703 F.3d 262, 267 (5th Cir.2012) (internal quotations omitted).
Initially, we address the constitutional challenges concerning three VDR provisions, and we then move to the alleged conflict between two other provisions and federal law.
A. First Amendment
Principles of judicial restraint must be employed before a federal court may declare a state law unconstitutional. Although their briefing conveys some ambiguity, Appellees essentially assert the facial unconstitutionality of the Non–Resident, County, and Compensation provisions. Courts generally disfavor facial challenges, and for good reason. “[F]acial challenges threaten to short circuit the democratic process by preventing laws embodying the will of the people from being implemented in a manner consistent with the Constitution.” Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 451, 128 S.Ct. 1184, 1191, 170 L.Ed.2d 151 (2008). With the exception of First Amendment cases, a facial challenge will succeed only if the plaintiff establishes that the act is invalid under all of its applications. United States v. Salerno, 481 U.S. 739, 745, 107 S.Ct. 2095, 2100, 95 L.Ed.2d 697 (1987). The standard for facial challenges in First Amendment cases is different, although still daunting. A law implicating the right to expression may be may be invalidated on a facial challenge if “a substantial number of its applications are unconstitutional, judged in relation to the statute's plainly legitimate sweep.” United States v. Stevens, 559 U.S. 460, 130 S.Ct. 1577, 1587, 176 L.Ed.2d 435 (2010) (quoting Washington State Grange, 552 U.S. at 449, n. 6, 128 S.Ct. at 1191 n. 6). Other relevant limits on our authority were well stated by the motions panel majority:
Our task as a federal court is, to the extent possible, to construe the provisions to avoid a constitutional conflict. See, e.g., Ohio v. Akron Ctr. for Reprod. Health, 497 U.S. 502, 514, 110 S.Ct. 2972, 2980, 111 L.Ed.2d 405 (1990). A federal court should not lightly enjoin the enforcement of a state statute. Chisom v. Roemer, 853 F.2d 1186, 1189 (5th Cir.1988). The determination of whether a democratically enacted statute is constitutional on its face requires that “every reasonable construction must be resorted to [ ] in order to save a statute from unconstitutionality.” Nat'l Fed. of Indep. Bus. v. Sebelius, ––– U.S. ––––, 132 S.Ct. 2566, 2594, 183 L.Ed.2d 450 (2012). In this vein, [Steen's] interpretation must be accorded some meaningful weight, as [he] is the official charged with enforcing the statute. See Bellotti v. Baird, 428 U.S. 132, 143, 96 S.Ct. 2857, 2864, 49 L.Ed.2d 844(1976). We defer to [his] interpretation of how the law is to be enforced, so long as it does not conflict with the statutory text. See Hamer v. Musselwhite, 376 F.2d 479, 481 (5th Cir.1967) (deferring to the city's interpretation of an ordinance because city officials are charged with enforcing the statute and are the ones who must apply it).
Andrade II, 488 F. App'x at 895.
Under the federalist structure of the United States, the states are responsible for regulating the conduct of their elections. It is well recognized that state regulations will invariably affect “the individual's right to vote and his right to associate with others for political ends.” Anderson v. Celebrezze, 460 U.S. 780, 788, 180 S.Ct. 1564, 1570, 75 L.Ed.2d 547, –––– (1983). Where a state election rule directly restricts or otherwise burdens an individual's First Amendment rights, courts apply a balancing test derived from two Supreme Court decisions, Anderson, supra, and Burdick v. Takushi, 504 U.S. 428, 112 S.Ct. 2059, 119 L.Ed.2d 245 (1992). Using the Anderson/Burdick balancing test, the court “must first consider the character and magnitude of the asserted injury to the rights protected by the First and Fourteenth Amendments that the plaintiff seeks to vindicate.” Anderson, 460 U.S. at 789, 103 S.Ct. at 1570. Next, the court “must identify and evaluate the precise interest put forward by the State as justifications for the burden imposed by its rule.” Id. In passing judgment, the court “must weigh ‘the character and magnitude of the asserted injury to the rights protected by the First and Fourteenth Amendments that the plaintiff seeks to vindicate’ against ‘the precise interests put forward by the State as justifications for the burden imposed by its rule,’ taking into consideration ‘the extent to which those interests make it necessary to burden the plaintiff's rights.’ “ Burdick, 504 U.S. at 434, 112 S.Ct. at 2063 (quoting Anderson, 460 U.S. at 789, 103 S.Ct. at 1570). State rules that impose a severe burden on First Amendment rights must be “narrowly drawn to advance a state interest of compelling importance.” Burdick, 504 U.S. at 434, 112 S.Ct. 2063 (quoting Norman v. Reed, 502 U.S. 279, 289, 112 S.Ct. 698, 116 L.Ed.2d 711 (1992). “Lesser burdens, however, trigger less exacting review, and a State's ‘important regulatory interests' will usually be enough to justify ‘reasonable, nondiscriminatory restrictions.’ “ Timmons v. Twin Cities Area New Party, 520 U.S. 351, 358, 117 S.Ct. 1364, 1370, 137 L.Ed.2d 589 (1997) (internal citations omitted). Thus, Anderson/Burdick provides no “litmus-paper test” that excuses courts reviewing election-related free expression cases from the “hard judgments” common in ordinary litigation. Anderson, 460 U.S. at 789–90.
Here, Appellees face a threshold problem. As the party invoking the First Amendment's protection, they have the burden to prove that it applies. Clark v. Cmty. for Creative Non–Violence, 468 U.S. 288, 293 n. 5, 104 S.Ct. 3065, 3069 n. 5, 82 L.Ed.2d 221 (1984). In Rumsfeld v. Forum for Academic and Institutional Rights, Inc., 547 U.S. 47, 66, 126 S.Ct. 1297, 1310, 164 L.Ed.2d 156 (2006), the Supreme Court reiterated that the First Amendment protects speech as well as certain kinds of conduct. However, the Court went on to underscore that only conduct that is “inherently expressive” is entitled to First Amendment protection. Id. at 66. To determine whether particular conduct possesses sufficient “communicative elements” to be embraced by the First Amendment, courts look to whether the conduct shows an “intent to convey a particular message” and whether “the likelihood was great that the message would be understood by those who viewed it.” Texas v. Johnson, 491 U.S. 397, 404, 109 S.Ct. 2533, 2539, 105 L.Ed.2d 342 (1989) (internal quotation marks and citation omitted). Conduct does not become speech for First Amendment purposes merely because the person engaging in the conduct intends to express an idea. Rumsfeld, 547 U.S. at 66, 126 S.Ct. at 1310.
The district court accepted Appellees' argument that all of the conduct comprising voter registration drives—urging citizens to vote, organizing the drive, conducting voter registration, and verifying that registrations have been accepted by the government—is “expressive conduct” protected by the First Amendment. See Voting for Am., Inc. v. Andrade, 888 F.Supp.2d 816, 839–43 (S.D.Tex.2012) [hereinafter Andrade I ]. The court broadly concluded that these activities implicate core political free speech and association. Wisely declining to apply strict scrutiny, however, the court ruled that burdens imposed by the Non–Resident Provision, the County Provision, and the Compensation Provision were not justified under the Anderson/Burdick balancing test. Id. at 843, 851–52.
Like the motions panel majority, we are unpersuaded that the smorgasbord of activities comprising voter registration drives involves expressive conduct or conduct so inextricably intertwined with speech as to require First Amendment scrutiny. Instead, we analyze the challenged Texas provisions separately because, as will be shown, discrete steps of the voter registration drive are in fact separable and are governed by different legal standards. See Planned Parenthood v. Suehs, 692 F.3d 343, 349 (5th Cir.2012) (reviewing a temporary injunction that impermissibly grouped state regulations on promotion of abortion with the right to affiliate with other pro-choice supporters). Further, this mode of analysis is required by the Supreme Court, which, in the context of election-related burdens on free expression, has long advised against substituting the hard judgments common in ordinary litigation with litmus-paper tests. Anderson, 460 U.S. at 789–90. The Court also has repeatedly explained that non-expressive conduct does not acquire First Amendment protection whenever it is combined with another activity that involves protected speech. See, e.g, Clark, 468 U.S. at 297–98, 104 S.Ct. at 3071 (emphasizing that the activity of camping does not become speech protected by the First Amendment when demonstrators camp as part of a political demonstration); Rumsfeld, 547 U.S. at 66, 126 S.Ct. at 1311 (2006) (“If combining speech and conduct were enough to create expressive conduct, a regulated party could always transform conduct into ‘speech’ simply by talking about it.”); see also United States v. O'Brien, 391 U.S. 367, 376, 88 S.Ct. 1673, 1678, 20 L.Ed.2d 672 (1968) (“We cannot accept the view that an apparently limitless variety of conduct can be labeled ‘speech’ whenever the person engaging in the conduct intends thereby to express an idea.”).
1. Non–Resident and County Provisions
The Non–Resident and County Provisions restrict the appointment of VDRs in two ways: VDRs must be Texas residents; and they may only register voters for counties in which the VDRs have been appointed. A VDR may be trained in one county and may apply by mail for automatic appointment in other counties. County registrars are required to appoint everyone who applies and satisfies the position's minimal criteria. Tex. Elec.Code Ann. § 13.031 (18 years of age, not a felon). The County appointment requirement has existed since 1985. A state residency requirement (for petition circulators) was upheld by the Eighth Circuit. Initiative & Referendum Inst. v. Jaeger, 241 F.3d 614 (8th Cir.2001). As these provisions have been interpreted by Steen, VDRs for each county exclusively are allowed to receive and deliver voter registration forms to the county offices for processing.
Appellees contend that these limits on the scope of VDRs' appointments prevent out-of-state-residents from acting as VDRs, inhibit the efficient functioning of voter registration drives, and truncate their ability to conduct the drives in violation of the First Amendment. We adopt the analysis of the motions panel majority rejecting these propositions, and recapitulate some of it here, together with additional reasons. See Andrade II, 488 F. App'x at 897–900.
The state does not deny that some voter registration activities involve speech—“urging” citizens to register; “distributing” voter registration forms; “helping” voters to fill out their forms; and “asking” for information to verify that registrations were processed successfully. Texas neither regulates nor limits any of this constitutionally protected speech. The district court accepted these representations, and so do we.
To establish their premise that the First Amendment protects what VDRs do—collect, review for completeness, and deliver completed voter registration forms—Appellees must show why such conduct is inherently expressive or why their “speech” is inextricably entwined with such actions. From a practical standpoint, the activities involved in a voter registration drive can be separated in a number of ways. For instance, experienced team leaders from out-of-state could organize the drive and train canvassers, using local citizens as VDRs and canvassers to complete and deliver the registration forms. Appellees in fact testified that they prefer to use local citizens as VDRs or canvassers. Out-of-state or out-of-county canvassers can participate anywhere, in any capacity, except to perform the functions exclusively ascribed to trained volunteer VDRs: collecting, reviewing for completeness, issuing a receipt, and delivering the completed voter registration forms to a county office. See Tex. Elec.Code Ann. §§ 13.038–.040, 13.042. With an appropriate division of labor and organizational forethought, no participant in the drive need suffer a detriment of the ability to urge, advocate, interact, or persuade.
Put otherwise, while voter registration drives involve core protected speech, they are factually distinct from the circulation of petitions addressed by the Supreme Court in Meyer v. Grant, 486 U.S. 414, 108 S.Ct. 1886, 100 L.Ed.2d 425 (1988), and Buckley v. Am. Constitutional Law Fund, 525 U.S. 182, 119 S.Ct. 636, 142 L.Ed.2d 599 (1999). Petitions by themselves are protected speech, and unlike a completed voter registration form, they are the circulator's speech. Assuming a voter registration application is speech, it is the voter's speech indicating his desire to be registered. Soliciting, urging and persuading the citizen to vote are the forms of the canvasser's speech, but only the voter decides to “speak” by registering. Logically, what the VDR does with the voter's form follows the voter's completion of the application but is not itself “speech.” One does not “speak” in this context by handling another person's “speech.” As the state's brief observes, the voter could refuse to return a registration application to the VDR and say, “I'll mail it myself.”
The motions panel majority made this point in a slightly different way:
At oral argument, Appellees urged the court to draw a close parallel to the Supreme Court's reasoning in Meyer, 486 U.S. at 422, finding that the circulation of initiative petitions was a matter involving the core political speech rights of the circulators under the First Amendment. The analogy is improper. The circulation and submission of an initiative petition is closely intertwined with the underlying political ideas put forth by the petition. The petition itself is the protected speech. Moreover, the very nature of a petition process requires association between the third-party circulator and the individuals agreeing to sign. In the voter registration context, the underlying expressive conduct (encouraging democratic participation and voting) does not implicate a third-party's right to process the application. Voter registration applications are individual, not associational, and may be successfully submitted without the aid of another. Here, the actual expression is not being limited.
Andrade II, 488 F. App'x at 898 n. 13 (emphasis added).
Buckley and Meyer are further distinguishable because those cases involved laws that specifically regulated the process of advocacy itself, dictating who could speak (only unpaid circulators and registered voters) or how to go about speaking (with name badges and subsequent detailed reports). Thus, the Colorado law had “the inevitable effect of reducing the total quantum of speech,” limiting “the number of voices who will convey [Plaintiffs'] message and the hours they can speak and, therefore, limit[ing] the size of the audience they can reach.” Meyer, 486 U.S. at 422–23, 108 S.Ct. at 1892. The motions panel majority aptly responded to the claim that the Texas provisions are unconstitutional for the same reason:
This argument simply does not apply. No party argues that the law prevents anyone, regardless of domicile or VDR appointment, from passing out registration forms and encouraging others to vote. The size of the audience remains the same, and the “total quantum of speech” is unaffected. While affirming the striking down of some provisions of Colorado law in Buckley, the Court recognized an “important interest” of the state in protecting the integrity of the ballot initiative process there at issue and in deterring fraud. 525 U.S. at 204–05.
Andrade II, 488 F. App'x at 897 n. 12.
The Non–Resident and County provisions do not in any way restrict or regulate who can advocate pro-voter-registration messages, the manner in which they may do so, or any communicative conduct.3 They merely regulate the receipt and delivery of completed voter-registration applications, two non-expressive activities.4 See League of Women Voters of Fla. v. Browning, 575 F.Supp.2d 1298, 1319 (S.D.Fla.2008) (“[T]he collection and handling of voter registration applications is not inherently expressive activity”); Garman v. U.S. Postal Service, 509 F.Supp. 507, 510 (N.D.Ind.1981) (noting that the ministerial task of processing draft registration forms has no speech component that would implicate First Amendment rights).
Similarly, because the canvassers' speech-related activities are distinct from both the collection and delivery of the forms and from the voters' “speech” in registering, the drives themselves cannot be amalgamated into protected “expressive conduct.” In Texas v. Johnson, 491 U.S. at 406, 109 S.Ct. at 2540, for instance, the act of burning the American flag indicated the speaker's contempt for the Republican Party's renomination of Ronald Reagan for President. The conduct was the message. But if actually registering citizens to vote is necessary to “express” the canvasser's belief in the importance of voting, then Appellees essentially seek a “First Amendment right not just to speak out or engage in ‘expressive conduct’ but also to succeed in their ultimate goal regardless of any other considerations.” Andrade II, 488 F. App'x at 898. Only two possibilities flow from this reasoning. As the motions panel majority suggests, throwing voter registration forms in the trash would have to be constitutionally protected expressive conduct. Id. at 898 n. 14.5 But cf. Tex. Elec.Code Ann. § 13.043 (failure of a VDR to deliver completed applications is an offense). Alternatively, and more sensibly, one must concede that supporting voter registration is the canvasser's speech, while actually completing the forms is the voter's speech, and collecting and delivering the forms are merely conduct.
Appellees urged that these limits on who may be VDRs and what they may do will interfere with Appellees' ability to conduct large-scale voter registration drives in Texas. But not every procedural limit on election-related conduct automatically runs afoul of the First Amendment. The challenged law must restrict political discussion or burden the exchange of ideas. See Initiative & Referendum Inst. v. Walker, 450 F.3d 1082, 1100–01 (10th Cir.2006) (en banc) (state constitutional provision requiring a supermajority to pass a wildlife-related ballot initiative does not implicate the First Amendment because it neither regulated the advocacy itself nor limited the “communicative conduct of persons advocating a position”); Dobrovolny v. Moore, 126 F.3d 1111, 1113 (8th Cir.1997) (concluding that a Nebraska law establishing a procedure for calculating the number of signatures required to place an initiative on the ballot does not implicate the First Amendment, although it “may have made it difficult for appellants to plan their initiative campaign and efficiently allocate their resources”); Biddulph v. Mortham, 89 F.3d 1491, 1500–01 (11th Cir.1996) (holding that a law did not violate the First Amendment because it did not burden the exchange of ideas and noting most laws restricting a state's initiative process would not implicate the First Amendment). Whether requiring only VDRs to collect and deliver completed voter registration applications even incidentally affects Appellees' activities is unclear, given the testimony that they prefer to hire local people to handle the actual registrations. That the requirement burdens no one's core political speech is undisputable.
In sum, we agree with the motions panel majority's conclusion that “there is nothing ‘inherently expressive’ about receiving a person's completed application and being charged with getting that application to the proper place.” Andrade II, 488 F. App'x at 898; see also League of Women Voters of Fla., 575 F.Supp.2d at 1319. Because the Non–Resident and County provisions regulate conduct only and do not implicate the First Amendment, rational basis scrutiny is appropriate. See District of Columbia v. Heller, 554 U.S. 570, 628 n. 27, 128 S.Ct. 2817 n. 27, 171 L.Ed.2d 637, –––– n. 27128. Steen justifies the Non–Resident provisions as an important safeguard in preventing fraud. He contends that if out-of-state residents were to act as VDRs, they could be less easily deterred from breaking the law because they could easily remove themselves from the jurisdiction of state investigators and prosecutors. Steen similarly justifies the County Provision as a means to increase accountability of VDRs by facilitating revocation of their appointments in cases of negligence, fraud or violation of registrants' privacy. These are rational bases for the provisions. Cf. Jaeger, 241 F.3d at 616 (holding that “the State has a compelling interest in preventing fraud” and that “[t]he residency requirement allows North Dakota's Secretary of State to protect the petition process from fraud and abuse by ensuring that circulators answer to the Secretary's subpoena power”).
Even assuming arguendo that the Non–Resident and County provisions implicate First Amendment interests, they pass the Anderson/Burdick balancing test. As a preliminary matter, there are fundamental differences in the activities of VDRs and petition circulators. VDRs register fellow citizens to vote. They may “advocate” and “interactively communicate” the importance of registration, but they have a duty to the fellow citizens whose registrations they facilitate. See, e.g., Tex. Elec.Code Ann. § 13.036(a)(2) (registrar may terminate VDR convicted of failing to deliver applications); § 13.036(b) (registrar may terminate VDR for failing to adequately review applications); § 13.044 (offense for a person “purportedly” acting as a VDR without an appointment). They assume a role carefully regulated by the state to serve the citizens who register to vote as well as the public interest in the integrity of the electoral body. Petition circulators, in contrast, are not agents of the state:
Nothing in this opinion should be read to suggest that initiative petition circulators are agents of the state. Although circulators are subject to regulations and are accountable to the State for compliance with legitimate controls [citation omitted], circulators act on behalf of themselves or the proponents of ballot initiatives.
Buckley, 525 U.S. at 192 n. 11, 119 S.Ct. at 642 n. 11. Petition circulators' ultimate responsibility lies in furthering their own or the sponsors' advocacy. To analyze VDRs' responsibilities solely in terms of their advocacy denigrates their statutory duty to register all prospective voters whose applications they collect. See Tex. Elec.Code Ann. § 13.043 (making it a crime to fail to deliver voter applications). It is one thing to say core political speech is involved when the advocates are trying to persuade the voting public to consider supporting an initiative or place a candidate on the ballot; it is quite another for the “advocates” to have free rein in creating the electorate. Thus, the character of any speech limited by the Non–Resident and County provisions is qualitatively different from the political speech restricted by laws that affect petition circulators.
The burdens imposed by the Non–Resident and County provisions on canvassers' speech are also minimal. Non–VDRs remain free to organize and run the registration drive, persuade others to register to vote, distribute registration forms, and assist others in filling them out. Unlike the requirements struck down in petition circulator cases, the Texas provisions do not directly reduce the number of voices by preventing out-of-state residents from advocating political or civic messages. See Buckley, 525 U.S. at 193 n. 15, 119 S.Ct. at 643 n. 15 (noting that “Colorado's registration requirement would exclude approximately 964,000 unregistered but voter-eligible residents from circulating petitions”); Nader v. Blackwell, 545 F.3d 459 (6th Cir.2008) (“[I]t is undisputable that [a residency requirement for petition circulators] sharply limited Nader's ability to convey his message to Ohio voters and thereby curtailed Nader's core political speech.”); Krislov v. Rednour, 226 F.3d 851, 860 (7th Cir.2000) (“By preventing the candidates from employing millions of potential advocates to carry their political message to the people of Illinois, the statute places a formidable burden on the candidates' right to disseminate their message.”). Appellees' claim that they may be less successful in achieving the result they advocate or in running a registration drive in the precise way they prefer does not demonstrate that their ability to advocate is significantly burdened by a requirement of Texas residency for VDRs.
Although the restrictions on Appellees' expressive voter registration activities are not severe, the state must justify the Non–Resident and County provisions with “important regulatory interests.” Anderson, 460 U.S. at 788, 103 S.Ct. at 1570. Steen asserts that Texas's interest in preventing voter registration fraud provides adequate justification. Any corruption in voter registration affects a state's paramount obligation to ensure the integrity of the voting process and threatens the public's right to democratic government. See Crawford v. Marion Cnty. Election Bd., 553 U.S. 181, 196, 128 S.Ct. 1610, 1619, 170 L.Ed.2d 574 (2008) (“While the most effective method of preventing election fraud may well be debatable, the propriety of doing so is perfectly clear.”); Storer v. Brown, 415 U.S. 724, 730, 94 S.Ct. 1274, 1297, 39 L.Ed.2d 714 (1974) (noting that a state has the power to engage in “substantial regulation of elections” in order to ensure elections are fair, honest, and orderly).
As the Supreme Court has recognized, the risk posed by fraud during the electoral process is far greater than that in either the initiative or candidate petition process. Meyer, 486 U.S. at 427, 108 S.Ct. at 1895 (“the risk of fraud or corruption, or the appearance thereof, is more remote at the petition stage of an initiative than at the time of balloting”) (citing First Nat'l Bank v. Bellotti, 435 U.S. 765, 790, 98 S.Ct. 1407, 1423, 55 L.Ed.2d 707 (1978) (“the risk of corruption perceived in cases involving candidate elections ․ simply is not present in a popular vote on a public issue.”)). Indeed, Steen points to documented evidence of voter registration fraud committed by canvassers, including those who worked for Appellee Project Vote and its former affiliate ACORN. See, e.g., League of Women Voters of Fla., 575 F.Supp.2d at 1310 (noting that the State of Florida received 13 written complaints in 2004 from “persons who registered to vote with third-party organizations” but who “[a]t the time of voting ․ were advised they were not registered to vote because the forms they had filled out had never been turned in.”); Staff of the H. Comm. on Oversight & Gov't Reform, 111th Cong., Follow the Money: ACORN, SEIU, and their Political Allies 49 (2010) [hereinafter Follow the Money] (noting that a “Project Vote employee was convicted ․ for submitting more than 400 fake voter registration applications.”); Staff of the H. Comm. on Oversight & Gov't Reform, 111th Cong., Is ACORN Internationally Structured as a Criminal Enterprise? 4 (2009) ( “[N]early 70 ACORN employees have been convicted in 12 states for voter registration fraud”).
The district court faulted Steen for not producing evidence of “rampant fraud” by out-of-state VDRs in pre–2011 Texas elections. Andrade I, 888 F.Supp.2d at 845. In Crawford, however, the Supreme Court upheld Indiana's voter ID law under the Anderson/Burdick test despite the fact that were was “no evidence of any [impersonation] fraud actually occurring in Indiana at any time in its history.” Crawford, 553 U.S. at 196, 128 S.Ct. at 1619. Here, as in Crawford, Texas need not show specific local evidence of fraud in order to justify preventive measures. See also Munro v. Socialist Workers Party, 479 U.S. 189, 195–96, 107 S.Ct. 533, 537–38, 93 L.Ed.2d 499 (1986) ( “Legislatures ․ should be permitted to respond to potential deficiencies in the electoral process with foresight rather than reactively․”).6
Texas's chosen means to avert fraudulent voter registrations by requiring state residency and county appointment for VDRs is sufficiently tailored. Compliance with voter registration regulations begins at the county level, where VDRs are approved, trained, and if need be, disciplined. After qualifying in one county, a VDR may perfunctorily qualify to serve in other counties. Multi-county qualification enables oversight of the VDR while eliminating the potential complications of registering voters in metropolitan areas covering several counties. Requiring that VDRs be state residents is obviously important to maintaining any credible possibility of prosecution for registration-related offenses. Election law violations typically carry low penalties and are hard to prosecute against local violators. Requiring the state to authorize itinerant out-of-state VDRs could render enforcement ineffective. Cf. Buckley, 525 U.S. at 196, 119 S.Ct. at 644 (not reaching the issue, but explaining the Tenth Circuit upheld a state residency requirement for petition circulators because it “more precisely achieved” the objective of making them eligible for subpoena and enforcement). We conclude that Texas's interest in preventing voter registration fraud is “sufficiently weighty to justify the limitation[s]” on Appellees' voter registration activities imposed by the Non–Resident and County provisions. Crawford, 553 U.S. at 191, 128 S.Ct. at 1616 (internal quotation and citation omitted); see also League of Women Voters of Fla., 575 F.Supp.2d at 1325 (concluding that Florida's interests in preventing fraud justify restrictions on organizations' voter registration drive activities).
The dissent disagrees with our analysis that disaggregates the advocacy for voter registration from the mere mechanics of registration performed by VDRs. According to the dissent, this mode conflicts with Meyer, where the Court invalidated Colorado's prohibition of paid petition circulators. The dissent quotes Colorado's argument in brief that because petition circulators were the sole parties responsible for the act of validating voters' signatures, their role should not receive First Amendment protection. Brief for Appellants, Meyer, 486 U.S. 414, 108 S.Ct. 1886, 100 L.Ed.2d 425 (No. 87–920), 1987 WL 880992, at *12.1. Judge Davis emphasizes that the Supreme Court rejected Colorado's argument while characterizing all petition-related expressive conduct as core political speech. Post at –––– (quoting Meyer, 486 U.S. at 421–22, 108 S.Ct. at 1892). This reasoning draws the wrong lessons from Meyer.
Although Meyer characterizes the advocacy required in petition circulation as core political speech, Meyer and Buckley also engage in detailed, fact-specific analysis of the impact of specific regulations on the furtherance of petition sponsors' advocacy.7 First, in Meyer, unlike this case, the payment prohibition deprived petition sponsors of a full opportunity to spread their views. Second, both cases recognize, explicitly or implicitly, the propriety of various regulations concerning even this core political speech—minimum age and affidavit of residency for circulators; criminal penalties related to false or purchased signatures; printed warnings on the petitions; sponsor and payment disclosures. Finally, Meyer acknowledges that the risk of improper conduct during petition circulation is less than in petition balloting, and the risk in petition balloting is less than in candidate elections. Meyer, 486 U.S. at 427–28, 108 S.Ct. at 1885, quoting First National Bank of Boston v. Bellotti, 435 U.S. 765, 790, 98 S.Ct. 1407, 1423, 55 L.Ed.2d 707 (1978).
In sum, contrary to the dissent, we find Meyer and Buckley distinguishable factually, indicative of nuanced analysis mandated in this type of case, and highly suggestive that state regulation to counter voter registration fraud should not be hastily overturned. What the dissent ultimately neglects is that controlling cases require the “hard judgments” that are common to ordinary litigation, and not a “litmus-paper test.” Anderson, 460 U.S. at 789, 103 S.Ct. at 1570; Buckley, 525 U.S. at 192, 119 S.Ct. at 642. As earlier explained, the Residence and County Provisions, unlike the rules at issue in Meyer and Buckley, do not infringe in any way on the communicative, interpersonal aspect of voter registration drives. But even if they do infringe on First Amendment rights, the Texas provisions fulfill a vital role in securing the integrity of the voter registration process and are a tailored response to the task. As a result, Appellees' First Amendment challenge cannot prevail.
2. Compensation Provision
The Compensation Provision of the Texas Election Code, § 13.008(a), creates offenses if a person:
(1) compensates another person based on the number of voter registrations that the other person successfully facilitates;
(2) presents another person with a quota of voter registrations to facilitate as a condition of payment or employment;
(3) engages in another practice that causes another person's compensation from or employment status with the person to be dependent on the number of voter registrations that the other person facilitates; or
(4) accepts compensation for an activity described by Subdivision (1), (2) or (3).
Recognizing that these provisions are potentially overbroad, Steen interprets subdivisions (2) and (3) of the Compensation Provision to ban (1) paying canvassers on a per-application basis and (2) conditioning payment or employment solely on the submission of a fixed number of applications.8 More precisely, according to Steen, subdivision (a)(2) applies only when a quota is “present[ed]” to a canvasser, while subdivision (a)(3) applies to any different (“another”) situation when employment is conditioned solely on a quota that has not been presented to the canvasser.
Federal courts are required to accept a narrowing construction of a state law in order to preserve its constitutionality. See Frisby v. Schultz, 487 U.S. 474, 483, 108 S.Ct. 2495, 2502, 101 L.Ed.2d 420 (1988) (construing a town ordinance “more narrowly” in part because “[t]his narrow reading is supported by representations of counsel for the town at oral argument”); Bellotti v. Baird, 428 U.S. 132, 143, 96 S.Ct. 2857, 2864, 49 L.Ed.2d 844 (1976) (noting the importance of the interpretation given by the officials charged with enforcing the statute); see also Ohio v. Akron Ctr. for Reprod. Health, 497 U.S. 502, 514, 110 S.Ct. 2972, 2980, 111 L.Ed.2d 405 (1990) (“[W]here fairly possible, courts should construe a statute to avoid a danger of unconstitutionality.” (citation and internal quotation marks omitted)).
The district court refused to accept the narrowing construction, reasoning that it is plainly contradictory to the statutory language. Thus, the court held, if Steen's construction is accepted, subdivision (a)(3) either superfluously overlaps subdivision (a)(2), or on its own, subdivision (a)(3) prevents the employer from making traditional performance-based decisions about the canvassers' effectiveness. We disagree with these conclusions.
With respect to the district court's opinion that Steen's construction would render subsection (a)(3) of the provision “superfluous,” subsection (a)(2) applies when someone “presents another person with a quota of voter registrations to facilitate as a condition of payment or employment.” Tex. Elec.Code Ann. § 13.008(a)(2) (emphasis added). Subsection (a)(3) applies when that fixed quota is used as the sole basis for determining compensation or employment, regardless whether it has been “presented” to the canvasser. Id. § 13.008(a)(3). As such, Steen's interpretation leaves the more comprehensive language of (a)(3) with plenty of work to do. Because the subsection readily lends itself to this interpretation, the district court incorrectly disregarded it. Virginia v. Am. Booksellers Ass'n, Inc., 484 U.S. 383, 397, 108 S.Ct. 636, 646, 98 L.Ed.2d 782 (1988) (maintaining that statues “readily susceptible” to a narrowing construction will survive a First Amendment challenge).
The district court's alternative interpretation is that (a)(3) overbroadly bans performance-based work reviews. Appellees contended that their employment decisions are chilled (and thus the quantum of their speech burdened) because in managing canvassers' work, they must inescapably tie decisions for superior or subpar performance to the canvassers' productivity in terms of voter registrations facilitated. The district court's interpretation is plausible, but it is impermissible pursuant to the standards of deference we have earlier cited.
First, the district court failed to assess whether “a substantial number” of the applications of (a)(2) and (a)(3) are unconstitutional, judged in relation to the provision's “plainly legitimate sweep.” This inquiry is required in a facial challenge on First Amendment grounds under Stevens, 130 S.Ct. at 1587. The existence of a wide swath of constitutional applications of the provision would suffice to prevent a facial remedy. Citizens United v. Fed. Election Comm'n, 558 U.S. 310, 331, 130 S.Ct. 876, 893, 175 L.Ed.2d 753 (distinguishing facial and as-applied challenges by “the breadth of the remedy employed by the Court”); see also United States v. Treasury Employees, 513 U.S. 454, 477–78, 115 S.Ct. 1003, 130, ––– L.Ed.2d ––––, –––– – –––– (1995) (contrasting “a facial challenge” with “a narrower remedy”).9
Second, the narrowing interpretation is not contradictory to the statute, and the court was thus required to accept it for present purposes. Subdivision (a)(2) is premised on “presenting” a quota to the employee, a formulation that implies direct contact in advance to warn the person of a “quid pro quota.” Subdivision (a)(3), making employment decisions “in another practice” “dependent” on the employee's number of voter registrations reasonably lends itself to decisions that are “solely dependent” on that number. Employers are familiar, through Title VII discrimination law, with the difference between “sole” motive and “mixed motive” decision making. Appellees cannot feign their inability to justify reasonable employment decisions on factors other than or in addition to the number of registrations a canvasser produces. Moreover, as a criminal statute susceptible to more than one interpretation, this provision's enforcement would have to be construed with lenity. See Tex. Elec.Code Ann. § 13.008(b). The rule of lenity reinforces the state's proffered narrowing interpretation. Finally, Texas's strong severability statute, which preserves statutes even if in some “applications” they are unconstitutional, clearly applies to the hypothetical situations Appellees invoked. Tex. Gov't Code Ann. § 311.032(c). Severability is a state law issue that binds federal courts. See Leavitt v. Jane L., 518 U.S. 137, 139, 116 S.Ct. 2068, 2069, 135 L.Ed.2d 443 (1996).
As we must accept the state's narrowing construction of the Compensation Provision, we turn to the merits. Because the provision applies to all persons—not just VDRs—and covers any activity that facilitates voter registration, it encompasses activities that involve expression, including voter drives where canvassers seek to persuade eligible voters to register. Further, we assume, without deciding, that prohibiting quota-based pay is a “lesser burden,” not subject to strict scrutiny. Prete v. Bradbury, 438 F.3d 949, 963 (9th Cir.2006) (assessing ban on tying compensation to number of signatures obtained for a petition as a lesser burden); Jaeger, 241 F.3d at 617–18 (declining to apply strict scrutiny to statute banning payment per petition signature procured); see also Person v. N.Y. State Bd. of Elections, 467 F.3d 141, 143 (2d Cir.2006) (finding that statute prohibiting payment of electoral petition signature gatherers on a per-signature basis does not, by itself, violate the First Amendment). Appellees have not made a strong showing that their facial challenge will prevail because the Compensation Provision can apply constitutionally to proscribe quotas or compensation incentives that reward canvassers solely for turning in a very high number of applications, which all parties agree can encourage fraud. Texas has a strong legitimate interest in preventing fraud. Buckley, 525 U.S. at 204–05, 119 S.Ct. at 648. The Compensation Provision was enacted in part to rectify deficiencies in the previous law that created incentives, such as paying canvassers for each application, that resulted in voter registration scandals. See Follow the Money, supra; Tex. Elec. Comm. Rep. on H.B. 239, 82d Leg. (2011) (“In many of the scandals, the convicted individuals specifically cited the compensation or performance quotas as the incentive to fraudulently complete voter registration forms.”).
Nevertheless, the district court faulted the Compensation Provision for being stricter than laws in other states, particularly since there is no indication that Texas is more susceptible to voter registration fraud. Andrade I, 888 F.Supp.2d at 852. This criticism is unwarranted; neither the uniqueness of an election law nor the state-wide prevalence of the type of fraud a law seeks to prevent has any bearing on its constitutionality. See Crawford, 553 U.S. at 195, 128 S.Ct. at 1619 (upholding Indiana's unique voter-ID law as a legitimate means to combat election fraud, despite no evidence of impersonation fraud in Indiana). In addition, the district court's and Appellees' analogies to Meyer, supra, and Citizens for Tax Reform v. Deters, 518 F.3d 375 (6th Cir.2008), are inapposite. Unlike the laws struck down in those cases, the Compensation Provision does not completely ban third-party organizations from compensating canvassers or only permit them to pay hourly wages. To deter fraud, (a)(3) merely prevents organizations from making compensation and employment decisions solely based on a fixed number of applications.10
We conclude that the Compensation Provision, as narrowly construed, does not violate the First Amendment. It is unnecessary to address Steen's abstention argument. Accordingly, Appellees have not demonstrated a likelihood of success on the merits of this claim.
Under the Constitution's Election Clause, Congress may enact laws that preempt state election laws concerning federal elections. See Foster v. Love, 522 U.S. 67, 69, 118 S.Ct. 464, 466, 139 L.Ed.2d 369 (1977). When it does, the federal legislation renders any conflicting state laws inoperative. See Ex parte Siebold, 100 U.S. 371, 384, 25 L.Ed. 717 (1879). “To this end, state election laws cannot ‘directly conflict’ with federal election laws on the subject.” Andrade II, 488 F. App'x at 896 (citing Voting Integrity Project, Inc. v. Bomer, 199 F.3d 773, 775 (5th Cir.2000)). Against this background, we examine whether the challenged provisions of the VDR Law conflict with the NVRA and require preemption.
The NVRA mandates that “[e]ach State shall maintain for at least 2 years and shall make available for public inspection and, where available, photocopying at a reasonable cost, all records concerning the implementation of programs and activities conducted for the purpose of ensuring the accuracy and currency of official lists of eligible voters.” 42 U.S.C. § 1973gg–6(i)(1). The Photocopying Provision states that a VDR “may distribute voter registration application forms throughout the county and receive registration applications submitted to the deputy in person.” Tex. Elec.Code Ann. § 38.038. As interpreted by Steen, this provision limits VDRs' conduct to collecting and delivering completed applications and implicitly precludes photocopying.
Appellees contend that because completed voter registration applications in the possession of VDRs are “public records,” the restriction against photocopying them violates the NVRA. However, Appellees disregard a crucial distinction: the NVRA only pertains to records “maintain[ed]” by the State, while the Photocopying Provision only applies to voter registration applications in the hands of VDRs, before they are officially received or maintained by the State. For this reason, the district court misplaced reliance on Project Vote/Voting for America, Inc. v. Long, 682 F.3d 331 (4th Cir.2012), a case that specifically addressed the denial of access to voter registration applications in the government's long-term possession, rather than those in the hands of VDRs. The question here is not whether such applications will be made available for photocopying but how. Thus, we disagree with the district court's reasoning that the applications received and delivered by VDRs are within the “constructive possession” of the state. For one thing, this conclusion is not supported by any statutory text and is contrary to state law prohibiting VDRs from “maintaining” the applications. Tex. Elec.Code Ann. § 13.038 (deputizing VDRs to receive and deliver voter registration applications, not to “maintain” them for the state); § 13.042(b) (requiring VDRs to deliver the voter registration applications to the county within five days). Moreover, allowing VDRs indiscriminately to photocopy registration applications places at risk the private information, e.g., social security numbers, they contain, because Steen and counties have limited means to enforce privacy protections against temporary volunteers. Because the NVRA and Texas law do not conflict, Appellees cannot prevail in this preemption claim.
Personal Delivery Provision
The NVRA requires states to “accept and use” a federal voter registration application sent through the United States mail. 42 U.S.C. § 1973gg–4(a)(1). The Personal Delivery Provision mandates that VDRs deliver completed voter registration applications to the county registrar in person. Tex. Elec.Code Ann. § 13.042(a). State law neither prevents prospective voters from mailing in their voter registration applications nor prohibits counties from accepting those applications. Significantly, county registrars must accept every application received by mail, even those sent by VDRs in violation of the Personal Delivery Provision. See Tex. Elec.Code Ann. §§ 13.071–.072. These facts differentiate Texas law from the Supreme Court's recent decision in Arizona v. InterTribal of Arizona, ––– U.S. ––––, 133 S.Ct. 2247, 186 L.Ed.2d 239 (2013), in which the Court overturned, as inconsistent with the NVRA, the imposition of additional conditions by the state on its acceptance of voter registrations. Also distinguishable is Charles H. Wesley Educ. Found., Inc., v. Cox, 408 F.3d 1349, 1354–55 (11th Cir.2005), where the Eleventh Circuit rejected Georgia's argument that the state could refuse to accept voter registration applications mailed by third-party organizations that did not meet additional state requirements. Moreover, as the court emphasized, the NVRA “simply requires that valid registration forms delivered by mail and postmarked in time be processed.” Id. at 1355. Texas law ensures this requirement is met. Because the laws do not conflict, Appellees have failed to demonstrate a substantial likelihood that the Personal Delivery Provision is preempted by the NVRA.
Appellees have not made a clear showing that they are likely to succeed in demonstrating that the challenged provisions of the VDR Law violate their First Amendment rights or are preempted by the NVRA. Consequently, it is unnecessary to address the remaining elements required for preliminary injunctive relief. La Union Del Pueblo Entero v. Fed. Emergency Mgmt. Agency, 608 F.3d 217, 225 (5th Cir.2010). The district court's preliminary injunction is REVERSED and the case REMANDED for further proceedings consistent with this opinion.
I respectfully dissent from the majority's decision reversing the preliminary injunction entered by the district court barring the Texas Secretary of State from enforcing five provisions of the Texas Election Code. The enjoined provisions are described as follows:
(1) Texas Election Code § 13.031(d)(3) to the extent it forbids non-Texas residents from serving as volunteer deputy registrars (“VDRs”) (the “Non–Resident Provision”);
(2) Texas Election Code § 13.038 to the extent it prohibits VDRs appointed in one county from serving in another county (the “County Provision”);
(3) Texas Election Code § 13.008(a)(2) & (3) (the “Compensation Provision”);
(4) Texas Election Code § 13.038 to the extent it prohibits VDRs from photocopying or scanning voter registration applications submitted to the VDR but not yet delivered to the county registrar (so long as no information deemed confidential under § 13.004 is included) (the “Photocopying Provision”); and
(5) Texas Election Code § 13.042 to the extent it prohibits VDRs from sending completed voter registration applications via United States mail (the “Personal Delivery Provision”).
The district court, following a lengthy evidentiary hearing and extensive briefing, held that the Non–Resident Provision, the County Provision, and the Compensation Provision unconstitutionally interfere with the plaintiffs' rights under the First Amendment and the Photocopying Provision and the Personal Delivery Provision are preempted by the National Voter Registration Act (“NVRA”), 42 U.S.C. §§ 1973gg et seq.
In its opinion, the majority concludes that the plaintiffs' activities affected by the Non–Resident Provision and the County Provision are not protected speech and, after accepting an interpretation of the Compensation Provision that is textually unsupportable, finds that the provision does not interfere with the plaintiffs' First Amendment rights. The majority also disagrees with the district court's conclusion that the Photocopying Provision and the Personal Delivery Provision are preempted by the NVRA.
My difference with the majority, in general, is two twofold: first, in my view the majority takes an unsupportably restrictive view of the scope of plaintiffs' activity to register voters that is protected by the First Amendment. The majority also ignores clear conflicts between two provisions of the Texas Election Code and the NVRA. In short, I would affirm the district court's thorough, well-reasoned opinion.
My disagreement with my colleagues' treatment of this case rests in large part on how the majority slices and dices the activities involved in the plaintiffs' voter registration drives instead of considering those activities in the aggregate. Ordinarily, when Project Vote and Voting for America set up a voter registration drive, they cooperate with local organizations in the targeted area, while maintaining control of the process. They hire canvassers from the local community, but also rely on experienced canvassers and organizers from out of state to manage the drive, train employees, and demonstrate proper techniques for voter registration. Once trained, the canvassers are deployed throughout the targeted community to attempt to persuade eligible voters to register to vote. Canvassers raise issues of local importance and stress that voting is a forum for voters to voice their views. If the canvasser is successful, he provides a blank application to the applicant and assists in completing the form. The canvasser then collects the completed form and returns it to the organization, where it is reviewed for completeness and signs of fraud. The non-confidential portions of the application are then scanned or photocopied for tracking purposes and then delivered to the appropriate registrar. The plaintiffs use the photocopy to follow up with the registrar to verify that the registration application has been processed and the applicant added to the voting rolls. If not, the plaintiffs follow up as needed to correct the situation. At election time, the plaintiffs urge the newly registered voters to actually vote and may provide transportation assistance to the voting area.
The Secretary and the majority concede that “some voter registration activities involve speech—‘urging’ citizens to register; ‘distributing’ voter registration forms; ‘helping’ voters fill out their forms; and ‘asking’ for information to verify that registrations were processed successfully.” However, the majority draws a line between these portions of the voter registration drive they must concede are protected and all the other activity which they dismiss as outside the protection of the First Amendment and simply the administrative process of collecting and handling voter registration forms. In the words of the majority, “[o]ne does not ‘speak’ in this context by handling another person's ‘speech,’ “ i.e., the voter registration application, which is the voter's “speech.” In the majority's view, all of the plaintiffs' activities that occur after the voter completes the registration application are not speech. This would include processing the application and checking it for errors, submitting it to the appropriate registrar, following up to ensure that the application was processed and the applicant added to the voting rolls, and encouraging the voter to participate in subsequent elections.
Two Supreme Court cases applying First Amendment protections to laws regulating parties engaged in similar activities reject the majority's line-drawing. In Buckley v. American Constitutional Law Foundation, Inc., 525 U.S. 182, 119 S.Ct. 636, 142 L.Ed.2d 599 (1999), the Court struck down Colorado's requirements that those collecting signatures for a ballot initiative wear a badge and be registered Colorado voters. In Meyer v. Grant, 486 U.S. 414, 108 S.Ct. 1886, 100 L.Ed.2d 425 (1988), the Court struck down Colorado's ban on paying petition circulators. As the district court pointed out, it could have been said in Buckley and Meyer “that the invalidated regulations did not prevent anyone from actually speaking to another citizen in an attempt to persuade her to sign a petition; rather, the regulations just governed the act of collecting the signatures.” The Supreme Court rejected that distinction.
In Meyer, Colorado argued to the Supreme Court that the statute banning paid petition circulators (much as Texas does in this case) did not implicate speech. Colorado argued:
[T]he petition circulator [is] the person with the public duty to determine the validity of the signatures of the persons who sign the petitions․ The verification of signatures does not constitute speech, and the prohibition against payment of petition circulators constitutes nothing more than the prohibition against payment for the act of verifying signatures. The fact that a person voluntarily links his conduct with a speech component does not transform the conduct into speech.
Brief for Appellants, Meyer, 486 U.S. 414, 108 S.Ct. 1886, 100 L.Ed.2d 425 (No. 87–920), 1987 WL 880992, at *12. The Supreme Court rejected this argument and concluded that petition circulation involves “interactive communication concerning political change that is appropriately described as ‘core political speech.’ “ Meyer, 486 U.S. at 421–22. Importantly, the Court did not isolate and limit the scope of its definition of core political speech to the verbal exchange between the petition circulator and the person whose signature was being solicited. Instead, it considered the solicitation activity in the aggregate as core speech. Specifically, the court in Meyer concluded that Colorado's ban on paid petition circulators impermissibly implicated the First Amendment by restricting political expression. The Meyer court explained that this ban restricted core speech because it limited the voices available to convey the message and therefore reduced the size of the audience that the canvassers could reach, making it less likely the campaign would be successful.
Similarly, in Buckley, the Supreme Court rejected Colorado's argument that collection of signatures was a ministerial act performed on behalf of the state and therefore state regulations requiring canvassers to be Colorado registered voters and wear badges did not implicate the First Amendment. As Justice Thomas explained in his concurrence,
Even where a State's law does not directly regulate core political speech, we have applied strict scrutiny ․ because we have determined that initiative petition circulation of necessity involves both the expression of a desire for political change and a discussion of the merits of the proposed change.
Buckley, 525 U.S. at 207 (Thomas, J., concurring). Thus, even though the Colorado regulation in Buckley did not directly regulate speech, the requirements that solicitors wear a badge and be Colorado registered voters were offensive in the same way the ban on paid circulators was in Meyer. Both regulations reduced the voices available to convey political messages. Id. at 210.
The Non–Resident Provision, the County Provision, and the Compensation Provision all limit who can participate in voter registration drives. As explained in more detail below, these provisions, like those in Buckley and Meyer, limit the number of voices available to convey the plaintiffs' message. The majority proposes methods by which the plaintiffs could organize their voter registration drives to reduce the effect of these regulations. But judges are not experts in conducting voter registration drives and the Supreme Court made it clear in Meyer that “[t]he First Amendment protects appellees' right not only to advocate their cause but also to select what they believe to be the most effective means for so doing.” 486 U.S. at 424.
In Meyer and Buckley, as in this case, the plaintiffs were advocates seeking to collect signatures (and verify them) as part of core democratic activity. More particularly, the activity in Meyer and Buckley was to persuade as many voters as possible to approve a referendum initiative to be placed on the ballot. This required circulators to collect signatures on a petition, verify the signatures, and deliver the petition to the appropriate state official. In our case the activity is to persuade as many citizens as possible to participate in the democratic process. The first step is registering voters. This requires persuading and assisting a citizen to complete a voter application and ensure the delivery of the application to the appropriate registrar. Once the voter is registered, the plaintiffs urge registered voters to cast their votes in scheduled elections. The speech rights of the plaintiffs in this case seeking to influence citizens to participate in the democratic process are as strong or stronger than the plaintiffs' rights in Meyer and Buckley. The First Amendment rights of the canvassers in Meyer and Buckley and the canvassers in this case are indistinguishable: The right to conduct their campaign without unjustified burdens “limits the number of voices who will convey [plaintiffs'] message.” Meyer, 486 U.S. at 422; see also Citizens United v. Fed. Election Comm'n, 558 U.S. 310, 130 S.Ct. 876, 175 L.Ed.2d 753 (2010) (holding that federal election laws prohibiting corporations and unions from using general treasury funds to make independent expenditures for speech defined as “electioneering communication” or for speech expressly advocating the election or defeat of a candidate is unconstitutional suppression of political speech).
The Supreme Court has a long line of cases holding that restrictions on expressive conduct, other than pure speech, may implicate the First Amendment. See, e.g., Texas v. Johnson, 491 U.S. 397, 404–06, 109 S.Ct. 2533, 105 L.Ed.2d 342 (1989) (flag burning); Tinker v. Des Moines Indep. Cmty. Sch. Dist., 393 U.S. 503, 505, 89 S.Ct. 733, 21 L.Ed.2d 731 (1969) (arm bands to protest the Vietnam War); Brown v. Louisiana, 383 U.S. 131, 141–42, 86 S.Ct. 719, 15 L.Ed.2d 637 (1966) (sit-ins to protest segregation). The plaintiffs' conduct in this case is inherently expressive of their message of increasing citizen participation in the democratic process. Rumsfield v. Forum for Academic & Institutional Rights, Inc., 547 U.S. 47, 66, 126 S.Ct. 1297, 164 L.Ed.2d 156 (2006).
Also, the majority's effort to slice the various portions of the registration effort into protected and unprotected activity ignores the fact that plaintiffs' registration activity implicates not only their speech rights, it also implicates their freedom of association. Both are equally protected by the First Amendment. The plaintiffs' activities do not cease when the voter registration application is complete. The plaintiffs receive and submit the application for processing and follow up with the registrar to ensure that the registration application resulted in a registered voter. They also follow up with voters to encourage them to vote. The freedom of the plaintiffs to associate with others for the advancement of common beliefs is protected by the First and Fourteenth Amendments. NAACP v. Alabama ex rel. Patterson, 357 U.S. 449, 460, 78 S.Ct. 1163, 2 L.Ed.2d 1488 (1958); Kusper v. Pontikes, 414 U.S. 51, 56–57, 94 S.Ct. 303, 38 L.Ed.2d 260 (1973).
As the Secretary concedes, nearly every federal court addressing this issue has found that the expressive conduct of actually registering voters, to the extent separable from the speech involved in persuading voters to register, is protected expressive conduct of those conducting the voter registration drive. See, e.g., Am. Ass'n of People with Disabilities v. Herrera, 690 F.Supp.2d 1183, 1200 (D.N.M.2010) (“In short, to participate in voter registration is to take a position and express a point of view in the ongoing debate whether to engage or to disengage from the political process. The Court concludes that the act of voter registration is expressive conduct worthy of First–Amendment protection.”); Project Vote v. Blackwell, 455 F.Supp.2d 694, 702 (N.D.Ohio 2006) (“[T]he Court is satisfied that participation in voter registration implicates a number of both expressive and associational rights which are protected by the First Amendment. These rights belong to—and may be invoked by—not just the voters seeking to register, but by third parties who encourage participation in the political process through increasing voter registration rolls.”).
If you start from what I believe is the correct baseline—that under the Supreme Court's precedent, the plaintiffs' entire voter registration activity is protected core political speech—then “the question is not whether Plaintiffs' conduct comes within the protections of the First Amendment, but whether Defendants have regulated such conduct in a permissible way.” League of Women Voters v. Cobb, 447 F.Supp.2d 1314, 1334 (S.D.Fla.2006). Assuming without deciding that the district court correctly applied the Anderson–Burdick balancing test to this question,1 I am satisfied that the district court correctly weighed the effect of each regulation on the plaintiffs' First Amendment rights against the justifications raised by the state to conclude that the plaintiffs were entitled to a preliminary injunction barring enforcement of the above listed provisions. I consider briefly below the individual provisions of the Texas Election Code plaintiffs challenge in this case.
A. The Non–Resident Provision
The Non–Resident Provision, one of the provisions under review in the Texas Election Code, provides that only a Texas resident may be appointed as a VDR. See Tex. Elec.Code Ann. §§ 13.031(d)(3), 11.002(a)(5). Because only VDRs can handle or submit a registration application to the registrar, this regulation in effect makes it a crime for a non-resident to handle or submit a registration application. The district court accepted the plaintiffs' testimony that this prevented the out-of-state campaign organizer from using its managers effectively in troubleshooting and identifying problems and generally organizing the campaign. Out-of-state managers cannot train, lead, demonstrate best practices, or perform quality control without ever touching an application-the registration drives' central tool for engaging voters. The only state interest raised to justify the prohibition is the assertion that Texas residents will take better care of their fellow citizens' registration forms than non-residents would. The Secretary provides no authority or analysis to support this bold assertion. I agree with the district court that any state interest identified is not strong and the regulation is not a narrowly tailored attempt to curtail any fraud that might be associated with out-of-state canvassers. This restrictions flies directly into the teeth of Buckley in which the Supreme Court struck down Colorado's attempt to restrict petition circulators to registered voters of the state. The restriction in this case, as in Buckley, limits the number of voices available to convey plaintiffs' message and, contrary to the conclusion of the majority, the plaintiffs have clearly demonstrated a likelihood of success on the merits of their challenge to this provision.
B. The County Provision
Texas Election Code § 13.038 provides that a VDR may distribute and accept applications for voter registration throughout the county in which he is qualified. As interpreted by the state of Texas, a VDR must be appointed in every county in which an applicant resides so that a VDR who is appointed in County A yet submits an application for a citizen who resides in County B is subject to criminal prosecution.2 The district court found that the plaintiffs had demonstrated that the County Provision imposes heavy time and administrative burdens on their organizations. These rules force the organizations to have their canvassers and managerial staff appointed as VDRs in multiple counties. This is especially burdensome in the larger metropolitan areas where voters may reside in one of several area counties. As pointed out by the district court, a VDR active in the City of Dallas would need to be appointed in five different counties in order to accept applications in all parts of the city. Texas has 254 counties which magnifies the burden of this limitation. A VDR will not always know in which county a potential voter resides simply by the fact that he is present at a registration drive rally. Thus, the VDR in this situation risks criminal sanctions for accepting a voter registration application from a resident of a county in which the VDR has not been appointed. It is obvious how this rule would chill the plaintiffs' registration activities.
The Secretary provided little justification for this rule. State, not county, laws govern voter registration, so there are no county-specific issues relevant to VDR appointment. Although the Secretary argued that the County Provision helps prevent fraud by making local county registrars more aware of VDR activity in their counties, the registrars are made aware of the identity of the VDRs when they submit voter registration applications. See Tex. Elec.Code Ann. § 13.040 (VDRs must provide a receipt to the registrar with submitted voter registration applications that identifies the county of their appointment). The district court correctly found that this provision violated plaintiffs' First Amendment rights.
C. The Compensation Provision
This provision subjects a campaign organizer, as employer, to criminal prosecution for compensating employees assisting in the campaign in a manner prohibited in this section. Tex. Elec.Code Ann. § 13.008. The Compensation Provision has three subparts: (1) a ban on “compensat[ing] another person based on the number of voter registrations that the other person successfully facilitates,” Id. § 13.008(a)(1); (2) a ban on “present[ing] another person with a quota of voter registrations to facilitate as a condition of payment or employment,” Id. § 13.008(a)(2); and (3) a ban on “engag[ing] in any other practice that causes another person's compensation from or employment status with the person to be dependent on the number of voter registrations that the person facilitates.” Id. § 13.008(a)(3). The plaintiffs do not challenge subpart (1). They contend, however, that the remaining provisions severely burden their ability to conduct registration drives by preventing them from rewarding or sanctioning employees based on performance. They submit that the provisions expose them to criminal sanctions if they either: (1) terminate or discipline a canvasser who performs poorly; or (2) reward high performers by promoting them or increasing their pay.
The Secretary interprets the provisions as only imposing criminal sanctions on an employer who pays canvassers on a per-application basis or conditions payment or employment solely on a preset quota. However, I agree with the district court that this interpretation is inconsistent with the plain language of the statute.3
Subparts (2) and (3) prohibit an employer from conditioning employment on the number of applications collected or basing the employee's compensation on applications collected. These restrictions effectively subject the plaintiffs to criminal sanctions for engaging in many common hiring and firing decisions. The Secretary argues that neither provision precludes general consideration of an employee's productivity. However, when the employee's job is to gather voter registration applications, the number of applications he obtains or facilitates is clearly an important measure of his productivity—a measure the Compensation Provision bars the plaintiffs from using as a basis for employment and compensation decisions.
That these provisions hamper the voter registration activities of the plaintiffs is obvious. As the Sixth Circuit explained in striking down an Ohio compensation prohibition that banned “pay[ing] any other person for collecting signatures on election-related petitions or for registering voters except on the basis of time worked”:
[W]hen petitioner's means are limited to volunteers and to paid hourly workers who cannot be rewarded for being productive and arguably cannot be punished for being unproductive, they carry a significant burden in exercising their right to core political speech.
Citizens for Tax Reform v. Deters, 518 F.3d 375, 385–87 (6th Cir.2008).
Again, the Secretary's justification for this provision is the need to combat voter registration fraud. The ban on paying compensation directly for each application obtained serves this interest and the plaintiffs do not challenge that provision. However, the remaining provisions further the state interest minimally, if at all, and burden speech and association by banning commonly accepted employment practices such as performance evaluations, performance-based pay, and the requirement of performance as a condition of employment. Violation of these provisions subject the employer to criminal prosecution and place an undue burden on the plaintiffs' First Amendment rights. In my judgment, the district court correctly enjoined enforcement of subparts (2) and (3) of this provision.
Preemption by the NVRA
The district court found that the remaining two provisions-the Photocopying Provision and the Personal Delivery Provision-were preempted by the NVRA. The majority concedes that state election law may not “directly conflict” with federal election laws on the subject. Voting Integrity Project, Inc. v. Bomer, 199 F.3d 773, 775 (5th Cir.2000). The recent Supreme Court decision in Arizona v. Inter Tribal Council of Arizona, ––– U.S. ––––, 133 S.Ct. 2247, 186 L.Ed.2d 239 (2013), held that the state enjoys no presumption against preemption in election clause cases. Id. at 2256–57. Instead, we are instructed to interpret “Election Clause legislation ․ to mean what it says .” Id.
A. The Photocopying Provision
With respect to the Photocopying Provision, plaintiffs challenge the Secretary's interpretation of § 13.038. The Secretary argues that the Texas Election Code prohibits VDRs from photocopying registration applications because state law does not explicitly authorize this activity and because these documents are considered confidential under § 13.004. The plaintiffs contend that the NVRA's public disclosure provision4 preempts the prohibition on photocopying.
I agree with the district court that voter registration applications are “records concerning the implementation of programs and activities conducted for the purpose of ensuring the accuracy and currency of official lists of eligible voters” under the relevant provision of the NVRA. 42 U.S.C. § 1973gg–6(I); Project Vote/Voting for Am., Inc. v. Long, 682 F.3d 331, 340 (4th Cir.2012). VDRs are deputized to act in the place of the county registrar when they distribute and receive voter registration applications. Tex. Elec.Code Ann. § 13. 031. Section 1973gg–6(I) does not require that the records be in the hands of the state. It merely requires the state to maintain the records and make them available for public inspection. The myriad of regulations governing VDRs are based on the assumption that the state has the ability to protect that application by regulating how it is handled until it is in the hands of the local registrar. In doing so, the state (through the VDR) “maintains” the completed voter registration applications until they are submitted to the local registrar. Thus, a completed application in the hands of a VDR is a record that must be made available for photocopying under the NVRA. I agree with the district court that it would be an absurd result to forbid private parties from copying applications they are authorized to receive on behalf of the state before they are submitted to the state, when the NVRA requires the state to allow them to make a copy once the record has been submitted.
The privacy concerns raised by the Secretary are also answered by the NVRA. As stated in Long, “[i]t is not the province of this court ․ to strike the proper balance between transparency and voter privacy․ Congress has already answered the question by enacting [section 1973gg–6(I) ], which plainly requires disclosure of complete voter registration applications.” 682 F.3d at 339. This conclusion is buttressed by the Supreme Court's holding in Arizona that no presumption against preemption applies in these election clause cases.
B. The Personal Delivery Provision
The Texas Election Code states that “[a] volunteer deputy registrar shall deliver in person, or by personal delivery through another designated volunteer deputy, to the registrar each completed voter registration application submitted to the deputy.” Tex. Elec.Code Ann. § 13.042(a). This section of the election code prohibits VDRs from using U.S. mail to deliver the applications. Any VDR who violates this ban is subject to criminal prosecution. Id. § 13.043.
Several provisions of the NVRA require states to allow voter registration by mail. 42 U.S.C. § 1973gg–4(a)(1) (“Each State shall accept and use the mail voter registration application form prescribed ․ pursuant to section 1973gg–7(a)(2).”); 42 U.S.C.1973gg–2(a)(2) (“[N]otwithstanding any other Federal or state law, in addition to any other method of voter registration provided for under State law, each State shall establish procedures to register to vote in Elections for Federal office ․ by mail application pursuant to section 1973gg–4․”).
I agree with the district court that the Texas Election Code provision presents a clear and direct conflict with the NVRA. The NVRA makes no distinction between applications submitted directly by a voter and those submitted by a third-party. The prospect of a criminal sanction effectively prevents the plaintiffs from using the mails to deliver applications to the registrar.
Because of the patent conflict, I would find that the plaintiffs have shown that they have a substantial likelihood of success on the merits on this issue and that the district court correctly enjoined its enforcement.
For the above reasons and those stated by the district court in its careful opinion, I am convinced that the plaintiffs have demonstrated a substantial likelihood of success on the merits of each provision of the Texas Election Code the district court enjoined the state from enforcing. No serious argument is advanced that plaintiffs failed to establish the remaining factors for issuance of the preliminary injunction—irreparable injury, balancing of harms, and the public interest. For reasons advanced by the district court, I conclude that plaintiffs clearly established these factors. I would AFFIRM the district court's preliminary injunction and therefore DISSENT from the majority's contrary conclusion.
EDITH H. JONES, Circuit Judge: