Carey v. Wolnitzek | Cases | Westlaw

Imagine if a State imposed these restrictions on candidates for election to the legislature: (1) They “shall not identify” themselves “as a member of a political party in any form of advertising or when speaking to a gathering”; (2) they “shall not solicit campaign funds”; and (3) they “shall not ... make a statement that a reasonable person would perceive as committing” the candidate to vote “a certain way on a[n] ... issue” likely to come before the legislature. A court faced with a First (and Fourteenth) Amendment challenge to the law would make short work of it. Legislative candidates have a First Amendment right to associate publicly with a political party, see Tashjian v. Republican Party of Conn., 479 U.S. 208, 214, 107 S.Ct. 544, 93 L.Ed.2d 514 (1986), to solicit campaign funds, see Riley v. Nat'l Fed. of the Blind of N.C., 487 U.S. 781, 796, 108 S.Ct. 2667, 101 L.Ed.2d 669 (1988), and to communicate to their constituents how they will vote on the issues of the day, see Brown v. Hartlage, 456 U.S. 45, 55–59, 102 S.Ct. 1523, 71 L.Ed.2d 732 (1982). It is doubtful that a single federal or state court judge in the country would see it differently.

Yet what happens if the same restrictions apply to judicial elections, not legislative elections? Some say the answer is the same. Elections are elections, and the same First Amendment applies to all of them. When the government suppresses election speech based on its content—prohibiting candidates from mentioning a political party with whom they affiliate, barring them from putting their name on a fund-raising letter or telling them what they can and cannot say about their judicial philosophy—the most rigorous form of constitutional second-guessing applies, and no categorical exemption from the First Amendment spares the government from this burden. In modern America, judicial elections are no less relevant to the public policy concerns of the citizenry than legislative elections, and the First Amendment protects electioneering speech in the one context as vigorously as it does in the other. Concerns about impartiality and open-mindedness that might result from unfettered judicial campaigning can be handled after the elections, not before, through the application of case-by-case judicial recusal rules that all States require their judges to follow before they agree to hear a case. Any remaining concerns flow not from the absence of speech restrictions on judicial candidates but from the State's insistence on holding elections for judicial office in the first place. A State cannot simultaneously insist that judges be held accountable to the electorate at regular intervals but deny to sitting judges and candidates alike the communicative tools for explaining how they will be held to account.

Others say it is not that easy. Judges do not represent constituents. They apply the law to the facts one case at a time, and, if they represent anyone or anything, it is the rule of law, which is why they sometimes must rule against the policy preferences of a majority of the voters. The judicial process works only when it is done in a disinterested manner, which is inconsistent with campaigns in which judges commit to rule, or appear to commit to rule, in a certain way in certain cases. It *194 is one thing when a legislator solicits money during a campaign; it is quite another when a judicial candidate, a sitting judge above all, does the same. With a few modest exceptions, see, e.g., Caperton v. A.T. Massey Coal Co., 556 U.S. 868, 129 S.Ct. 2252, 173 L.Ed.2d 1208 (2009); Mich. Court Rule 2.003, judicial-recusal rules are self-enforced and therefore may not provide adequate safeguards against the risks that flow from treating judicial elections like legislative ones. Unlike the other branches of government, the authority of the judiciary turns almost exclusively on its credibility and the respect warranted by its rulings, both of which are likely to be diminished by free-flowing electoral speech that permits the malignant inference that there is such a thing as caucus-bound blue-robed judges and caucus-bound red-robed judges. In some settings, there can be too much of a good thing, and unfettered free speech in judicial elections is one of them.

This is a complicated debate, and today's case requires us to take a side on some of these issues. Most recently in 2005, the Kentucky Supreme Court promulgated a judicial canon along the lines of the hypothetical legislative campaign rules mentioned above. As sitting judges ourselves, we have considerable sympathy for the concerns that prompted the canon, so much so that we embrace a central premise of it: Judicial elections differ from legislative elections, and the Kentucky Supreme Court has a compelling interest in regulating judicial campaign speech to ensure the reality and appearance of an impartial judiciary. Yet because two clauses of the canon overlooked narrower ways of advancing this interest and because, as written, they remain incompatible with the United States Supreme Court's decision in Republican Party of Minnesota v. White, 536 U.S. 765, 122 S.Ct. 2528, 153 L.Ed.2d 694 (2002), we must invalidate them. The third clause is constitutional in the main but contains a material ambiguity, which requires further consideration by the district court. The district court's decision is affirmed in part and vacated in part.

In 1792, Kentucky became the fifteenth State (and the fourth Commonwealth). The original Kentucky Constitution permitted the Governor to appoint judges, see Ky. Const. art. 2, § 8 (1792), but the Commonwealth, in the aftermath of the Age of Jackson, amended its Constitution in 1850 to require its judges to stand for popular election to eight-year terms. See Ky. Const. § 117; Ky. Const. art. 4, §§ 4, 6 (1850). Since 1975, judicial elections in Kentucky have been “nonpartisan,” compare Ky. Const. § 116 (1891) with Ky. Const. § 117 (1976), meaning that political parties have no formal role in any stage of the judicial selection process. Prospective candidates submit petitions for nomination to the Secretary of State. Ky.Rev.Stat. Ann. § 118A.060. The Commonwealth holds a single primary election for each judicial seat with no party identifiers and random ballot positioning. Id. The top two vote-getters in the primary election receive a spot on the general election ballot, which also is held without any party identifier. Id.

In competing for judicial seats, all candidates must abide by the Kentucky Code of Judicial Conduct. Promulgated by the Kentucky Supreme Court, it generally prohibits “a judge or judicial candidate” from “inappropriate political activity.” Rules of Supreme Court of Kentucky 3.130(8.2); 4.300, Canon 5. Sitting judges or judicial candidates violate this admonition, the Code says, if they fail to follow these clauses of Canon 5, among others:

In June 2006, Marcus Carey, then a candidate for a seat on the Kentucky Supreme Court, filed a complaint in federal district court claiming that the party affiliation, solicitation and commits clauses violated his speech and associational rights under the First and Fourteenth Amendments of the U.S. Constitution. The named defendants sit on the Kentucky Judicial Conduct Commission, sit on the Kentucky Inquiry Commission or serve as Bar Counsel.

Carey complained that he wanted to disclose his party status, yet he feared the party affiliation clause barred him from doing so. He wanted to ask for campaign contributions by signing fund-raising letters, yet he feared the solicitation clause barred him from doing so. And he wished to respond to a judicial questionnaire distributed by Kentucky Right to Life, raising questions for the candidates about their judicial philosophy and about their positions on specific issues, yet he feared the commits clause barred him from doing so. He asked the court to declare the clauses unconstitutional on their face and to enjoin their enforcement.

In October 2006, roughly one month before the election, the district court preliminarily enjoined enforcement of the party affiliation and the solicitation clauses but dismissed Carey's challenge to the commits clause on ripeness and standing grounds. On November 2, Carey moved to amend his complaint, re-challenging the commits clause, this time detailing the statements he proposed to make in possible violation of the clause. About a week later, Carey lost the election.

In September 2007, the court ruled that Carey's amended challenge to the commits *196 clause was ripe for review and allowed it to proceed along with Carey's challenges to the party affiliation and solicitation clauses. The parties all moved for summary judgment. In ruling on the motions, the district court determined that strict scrutiny applied to all of the challenges. It then invalidated the party affiliation and solicitation clauses on their face but rejected Carey's facial challenge to the commits clause. The state defendants appeal the court's ruling on the party affiliation and solicitation clauses, and Carey appeals the court's ruling on the commits clause.

Before turning to the merits, we must consider two jurisdictional questions implicated by these challenges: Did Carey file his claims too early, making them unripe for judicial review, or too late, making them moot? See Warshak v. United States, 532 F.3d 521, 525 (6th Cir.2008) (en banc).

Ripeness. Designed to ensure that the federal courts resolve “existing, substantial controversies,” Norton v. Ashcroft, 298 F.3d 547, 554 (6th Cir.2002), not disputes “anchored in future events that may not occur as anticipated” or may not occur “at all,” Nat'l Rifle Ass'n of Am. v. Magaw, 132 F.3d 272, 284 (6th Cir.1997), the ripeness doctrine ensures that a dispute is concrete and real before the judicial branch resolves it. Three considerations inform the doctrine: Is the alleged injury likely to occur? Is the factual record sufficiently developed to resolve the question? And what kinds of hardships, if any, will the parties face if the court delays resolution of the question? Warshak, 532 F.3d at 525. In the context of a free-speech overbreadth challenge like this one, a relaxed ripeness standard applies to steer clear of the risk that the law “may cause others not before the court to refrain from constitutionally protected speech or expression.” Broadrick v. Oklahoma, 413 U.S. 601, 612, 93 S.Ct. 2908, 37 L.Ed.2d 830 (1973).

Carey meets these requirements. In future judicial elections, as in prior ones, he claims an interest in engaging in protected speech that implicates, if not violates, each clause. He wants to let voters know his party affiliation. He wants to solicit campaign funds directly, as opposed to indirectly via an election committee. And he wants to answer judicial questionnaires propounded by a local right-to-life organization. These aspects of the canon at least chill, and in some instances prohibit, these forms of communication, and in the course of the November 2006 election, at least until the entry of the October 2006 injunction, Carey censored himself on each topic. All of this establishes a “credible fear of enforcement,” Norton, 298 F.3d at 554, sufficient to overcome any ripeness concerns.

The Kentucky Judicial Conduct Commission persists that the Kentucky Supreme Court and its ethics branch, the Kentucky Judicial Ethics Committee, have yet to apply these clauses to Carey, noting that “an authoritative construction of the canons may significantly alter the constitutional questions.” Commission's Opening Br. at 22. That is all true, but it is a peculiar ground for staying our hand now with respect to all of these challenges, some of which involve clauses with little ambiguity. See City of Lakewood v. Plain Dealer Pub. Co., 486 U.S. 750, 770 n. 11, 108 S.Ct. 2138, 100 L.Ed.2d 771 (1988). This challenge dates from July 2006, and a related challenge, supported by the same counsel, dates from September 2004, see Family Trust Found. of Ky., Inc. v. Wolnitzek, 345 F.Supp.2d 672 (E.D.Ky.2004). The Commission has had ample time to request interpretations or modifications of *197 Canon 5 by the Committee or the Court, yet apparently has not done so. Nor has the Commission, or anyone else in this case, asked the federal courts to certify any questions to the Kentucky Supreme Court. These claims are ripe for review.

Mootness. In one sense, Carey's original challenge seems moot because the November 2006 election has come and gone. Carey filed his original complaint months ahead of the election, and moved to amend it a week before the election, yet here we are more than three years after the election, still considering his claims. Carey, however, retains the right to run for judicial office again, and all candidates for judicial office in Kentucky, whether sitting judges or not, are subject to Canon 5. Under these circumstances, the claims may proceed: The alleged wrongs are “capable of repetition, yet evading review,” saving them from mootness, as the district court correctly held and as the parties do not dispute. See Fed. Election Comm'n v. Wis. Right to Life, Inc., 551 U.S. 449, 462, 127 S.Ct. 2652, 168 L.Ed.2d 329 (2007); Briggs v. Ohio Elections Comm'n, 61 F.3d 487, 492–93 (6th Cir.1995).

Two recent decisions of the Supreme Court—White and Caperton—set the stage for resolving the merits of this dispute. At issue in White was a judicial canon, first promulgated by Minnesota in 1974, providing that “a candidate for a judicial office, including an incumbent judge,” shall not “announce his or her views on disputed legal or political issues.” Minn.Code of Judicial Conduct, Canon 5(A)(3)(d)(i); White, 536 U.S. at 768, 122 S.Ct. 2528. Applying strict scrutiny, the Court rejected Minnesota's contention that the canon preserved judicial “impartiality” in a permissible way.

To the extent the Minnesota announce clause sought to preserve judicial “impartiality” in one sense—a “lack of bias for or against either party to the proceeding,” id. at 775, 122 S.Ct. 2528—the Court accepted the State's interest as a compelling one. Id. at 777 n. 7, 122 S.Ct. 2528. But the clause suffered from a means-end problem because it did “not restrict speech for or against particular parties, but rather speech for or against particular issues.” Id. at 776, 122 S.Ct. 2528. It may be, the Court acknowledged, that, “when a case arises that turns on a legal issue on which the judge (as a candidate) had taken a particular stand, the party taking the opposite stand is likely to lose,” but that is not due to “any bias against that party” for “[a] ny party taking that position is just as likely to lose.” Id. at 776–77, 122 S.Ct. 2528.

To the extent the State meant to advance “impartiality” in another sense—an absence of judicial “preconception in favor of or against a particular legal view ”—that was not a compelling interest. Id. at 777, 122 S.Ct. 2528. “[S]ince avoiding judicial preconceptions on legal issues is neither possible nor desirable, pretending otherwise by attempting to preserve the ‘appearance’ of that type of impartiality can hardly be a compelling state interest either.” Id. at 778, 122 S.Ct. 2528.

And to the extent the State meant to promote “impartiality” in the sense of judicial “open-mindedness”—the “willing[ness] to consider views that oppose [one's] preconceptions”—the Court found it unnecessary to decide whether this “desirable” quality amounted to a compelling interest. Id. at 778, 122 S.Ct. 2528. It held that the clause was so poorly tailored to any interest in open-mindedness that the Minnesota Supreme Court could not have “adopted *198 the announce clause for that purpose.” Id. Judges, both incumbent and prospective, it reasoned, retained so many ways to communicate their views on legal issues other than through election statements that the clause gratuitously limited speech while “leav[ing] appreciable damage to that supposedly vital interest unprohibited.” Id. at 780, 122 S.Ct. 2528.

Caperton dealt with a sitting state supreme court justice whose top campaign donor in the previous election, the head of a mining company, had spent $3 million on his behalf—more than all of his other supporters combined. See Caperton, 129 S.Ct. at 2257. When a high-stakes dispute involving the mining company came before the court, the justice refused to recuse himself from hearing it and ultimately joined the 3–2 majority in ruling for the company. See id. at 2258. The losing party claimed that the justice's participation in the case violated its due process rights. The Supreme Court agreed, holding that, by refusing to disqualify himself, the justice had unconstitutionally deprived the parties of a fair hearing. See id. at 2265. When “there is a serious risk of actual bias,” the Court reasoned, the Constitution requires judges to disqualify themselves, though the Court cautioned that this was “an extraordinary situation,” emphasizing the size of the mining company's support relative to other donors' support, the apparently decisive effect of this support on the justice's election and the close temporal connection between the justice's election and the company's case. Id. at 2263, 2265.

Strict scrutiny applies to all three aspects of this First Amendment challenge. White, for one, suggests as much, even if the decision does not compel that conclusion. In striking down Minnesota's announce clause, the Court said the following about the standard of review:

The state defendants seize on the modest length of the Court's analysis and Minnesota's concession, arguing that we need not apply strict scrutiny here. But White's brevity on this score and Minnesota's concession may suggest something else: that the counter-argument has little to support it. The multi-State amicus brief filed in support of Minnesota did not question the applicability of strict scrutiny in White. See Brief Amicus Curiae of California, et al. in Support of Respondents, Republican Party of Minn. v. Kelly, 536 U.S. 765 (2002). Not one of the Justices, not even one of the four dissenters, objected to the application of strict scrutiny. And if strict scrutiny does not apply to judicial canons like this one and the one at issue in White, it is difficult to understand why the Court exercised its discretion in reviewing White, given that virtually the entire analysis is premised on the applicability of strict scrutiny and given that the outcome of the case under a lower level of scrutiny is far from clear.

Free-speech first principles also suggest that strict scrutiny should apply. The three canons censor speech based on its *199 content in the most basic of ways: They prevent candidates from speaking about some subjects (judicial philosophy, the legal issues of the day, party affiliation) but not others (experience); and they prevent candidates from asking for support in some ways (campaign funds) but not others (a vote, yard signs). The canons refer directly to, and are “justified with[ ] reference to,” the content of candidates' speech, meaning they are not eligible for the relaxed review that content-neutral restrictions receive. Ward v. Rock Against Racism, 491 U.S. 781, 791, 109 S.Ct. 2746, 105 L.Ed.2d 661 (1989).

Content-based restrictions on speech generally face strict scrutiny, see United States v. Playboy Entm't Group, Inc., 529 U.S. 803, 813, 120 S.Ct. 1878, 146 L.Ed.2d 865 (2000), and thus are “presumptively invalid” unless the restriction discriminates on the basis of categorically “proscribable” speech, see R.A.V. v. City of St. Paul, 505 U.S. 377, 382–83, 112 S.Ct. 2538, 120 L.Ed.2d 305 (1992). See also United States v. Stevens, 559U.S. 460, 130 S.Ct. 1577, 1584, 176 L.Ed.2d 435 (2010). None of the categorical carve-outs apply. The canons do not address “fighting words” or incitement, see Chaplinsky v. New Hampshire, 315 U.S. 568, 572, 62 S.Ct. 766, 86 L.Ed. 1031 (1942), defamation, see Beauharnais v. Illinois, 343 U.S. 250, 72 S.Ct. 725, 96 L.Ed. 919 (1952), obscenity, see Roth v. United States, 354 U.S. 476, 77 S.Ct. 1304, 1 L.Ed.2d 1498 (1957), or child pornography, see New York v. Ferber, 458 U.S. 747, 764, 102 S.Ct. 3348, 73 L.Ed.2d 1113 (1982). Far from implicating these exceptions, today's regulations implicate a core area of free-speech protection: elections. See Brown, 456 U.S. at 53–54, 102 S.Ct. 1523; see also Stevens, 130 S.Ct. at 1586 (declining to declare “depictions of animal cruelty” as a “new categor[y] of speech outside the scope of the First Amendment”).

Nor does the nature of the restrictions implicate any of the other areas or types of regulation—time, place and manner restrictions, commercial speech, expressive conduct—in which the Court has applied less-than-rigorous review. The canons instead are of a piece with the kinds of speech regulation—telling candidates what they can and cannot say before an election—that the courts have scrutinized most rigorously. See, e.g., Brown, 456 U.S. at 53–54, 102 S.Ct. 1523; see also Erwin Chemerinsky, Restrictions on the Speech of Judicial Candidates are Unconstitutional, 35 Ind. L. Rev. 735, 740–742 (2002) (explaining that strict scrutiny should apply to First Amendment challenges to judicial canons like these); Mark Spottswood, Comment, Free Speech and Due Process Problems in the Regulation and Financing of Judicial Election Campaigns, 101 Nw. U. L. Rev. 331, 347 (2007) (same).

The Commission does not cite a single case, and we have not found one on our own, applying anything less than strict scrutiny to comparable free-speech challenges to judicial election canons. After White, the Eighth Circuit applied strict scrutiny to Minnesota's party affiliation and solicitation clauses. See Republican Party of Minn. v. White, 416 F.3d 738, 749 (8th Cir.2005) (en banc) (“White II ”). After White, the Eleventh Circuit did the same in invalidating Georgia's rules prohibiting judicial candidates from soliciting campaign funds. See Weaver v. Bonner, 309 F.3d 1312, 1319 (11th Cir.2002). After White, the Seventh Circuit applied strict scrutiny to Wisconsin's party affiliation clause and held that Wisconsin's solicitation clause survived both intermediate and strict scrutiny. See Siefert v. Alexander, 608 F.3d 974, 978, 981 (7th Cir.2010). And, before White, the Third Circuit applied *200 strict scrutiny in upholding some judicial speech restrictions. See Stretton v. Disciplinary Bd. of Supreme Court of Pa., 944 F.2d 137, 141–42 (3d Cir.1991).

The Commission urges us to apply a form of intermediate scrutiny, which balances the “competing fundamental rights” of some judicial candidates (who have a right to engage in campaign speech) and some litigants (who have a right to an impartial judiciary). Commission's Opening Br. at 10. But the reality that judicial impartiality is a “vital state interest,” protected by the Due Process Clause, Caperton, 129 S.Ct. at 2266–67, does not require us to dilute the First Amendment. It establishes instead that Kentucky has a compelling interest in preserving the canon, proving that the State can satisfy the first requirement of strict scrutiny, not that, having satisfied this requirement, it may water down the remaining requirements.

The Commission's analogy to Gentile v. State Bar of Nevada, 501 U.S. 1030, 111 S.Ct. 2720, 115 L.Ed.2d 888 (1991), does not hold up. Gentile “balance [d]” litigants' fair trial rights with attorneys' free speech rights in upholding a rule prohibiting attorneys involved in a pending trial from making statements likely to prejudice the proceedings. Id. at 1075, 111 S.Ct. 2720. As these features of the decision suggest, Gentile applies only to speech restrictions imposed on attorneys during a pending case, see id. at 1073 n. 5, 111 S.Ct. 2720, which is one reason—there are others—why a comparable law restricting judges from telling the press about the outcome of a pending case would not be an unconstitutional prior restraint on speech. Today, however, we have a speech restriction aimed not at judges performing court functions but at judges and judicial candidates making campaign statements or solicitations outside of court and outside of the process of deciding cases in their official capacity—all for the purpose of communicating information to voters about whom they should elect. That Gentile upholds a law restricting a lawyer's speech during a trial does not mean that it allows restrictions on lawyers in all settings. Otherwise, a lawyer running to be the Attorney General or Governor of a State could be censored simply because she is an “officer of the court.” That is not the case.

The Commission insists that the solicitation clause is an especially poor candidate for strict scrutiny review, because the Supreme Court applies a “lesser” standard of review to restrictions on political donations. See McConnell v. Fed. Election Comm'n, 540 U.S. 93, 136, 124 S.Ct. 619, 157 L.Ed.2d 491 (2003). But this argument gives analogy a bad name. The solicitation clause does not set a contribution limit, as in McConnell and similar cases. See, e.g., Fed. Election Comm'n v. Nat'l Right to Work Comm., 459 U.S. 197, 208, 103 S.Ct. 552, 74 L.Ed.2d 364 (1982). It flatly prohibits speech, not donations, based on the topic (solicitation of a contribution) and speaker (a judge or judicial candidate)—precisely the kind of content-based regulations that traditionally warrant strict scrutiny.

Because strict scrutiny applies, the Commission faces a daunting gauntlet, as “it is the rare” law that “survives” this kind of review. Burson v. Freeman, 504 U.S. 191, 211, 112 S.Ct. 1846, 119 L.Ed.2d 5 (1992). To survive, the three canons must be “narrowly tailored” to advance a “compelling state interest.” Eu v. San Francisco County Democratic Cent. Comm., 489 U.S. 214, 222, 109 S.Ct. 1013, 103 L.Ed.2d 271 (1989).

At the same time, Carey seeks to invalidate these clauses not just as applied *201 to him but in all of their applications, which is to say on their face. In most constitutional cases, that exceptional remedy requires the claimant to “show one of two things: (1) that there truly are ‘no’ or at least few ‘circumstances' in ‘which the Act would be valid,’ United States v. Salerno, 481 U.S. 739, 745, 107 S.Ct. 2095, 95 L.Ed.2d 697 (1987); see also Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 449–50, 128 S.Ct. 1184, 170 L.Ed.2d 151 (2008); or (2) that a court cannot sever the unconstitutional textual provisions of the law or enjoin its unconstitutional applications.” Connection Distrib. Co. v. Holder, 557 F.3d 321, 335 (6th Cir.2009) (en banc). The courts, however, “rightly lighten this load in the context of free-speech challenges to the facial validity of a law.” Id. In view of the risk that “enforcement of an overbroad law” may “deter[ ] people from engaging in constitutionally protected speech” and may “inhibit[ ] the free exchange of ideas,” the overbreadth doctrine permits courts to invalidate a law on its face “if ‘a substantial number of its applications are unconstitutional, judged in relation to the statute's plainly legitimate sweep.’ ” Stevens, 130 S.Ct. at 1587 (quoting Wash. State Grange, 552 U.S. at 449 n. 6, 128 S.Ct. 1184).

Party affiliation clause. This clause prohibits judges and candidates from disclosing their party affiliation “in any form of advertising, or when speaking to a gathering,” save in answer to a question by a voter in one-on-one or “very small private informal” settings. Rules of Supreme Court of Kentucky 4.300, Canon 5(A)(2); Kentucky Judicial Ethics Opinion JE–105 (2004). The clause advances at least two interests, both sufficiently compelling to satisfy the First Amendment. It furthers the Commonwealth's goal of having a judiciary that is biased neither in fact nor in appearance. See White, 536 U.S. at 775–79, 122 S.Ct. 2528. And it furthers the Commonwealth's interest in diminishing reliance on political parties in judicial selection, a policy grounded in the Kentucky Constitution's requirement that judicial elections be nonpartisan in nature.

The problem, however, is not the Commonwealth's laudable interests in promulgating this canon; it is the Commonwealth's methods in furthering them. The Court frequently says that censoring speech must be a government's measure of “last—not first—resort” in advancing its policy interests, e.g., Thompson v. W. States Med. Ctr., 535 U.S. 357, 373, 122 S.Ct. 1497, 152 L.Ed.2d 563 (2002), and the narrow-tailoring requirement is proof that the Court means it. If a law does too much, or does too little, to advance the government's objectives, it will fail. See Eu, 489 U.S. at 222, 109 S.Ct. 1013. This canon does both.

The canon's first problem is a White problem—that it suppresses too much speech to advance the government's interest. In invalidating Minnesota's announce clause, White established that a State may not prohibit a judicial candidate from disclosing, say, that “I am for limited government,” “I support a woman's right to choose,” “I prefer tough-on-crime laws,” or, to use an example from White, “I think it is constitutional for the legislature to prohibit same-sex marriage.” 536 U.S. at 779, 122 S.Ct. 2528. The party affiliation clause prohibits all of this, only more so. It prohibits candidates from announcing their position on one issue of potential importance to voters: the party they support. And it prohibits them from announcing their position on many issues of *202 potential importance to voters: the party platform with which they affiliate. A party platform after all is nothing more than an aggregation of political and legal positions, a shorthand way of announcing one's views on many topics of the day. If the single-issue announce canon at play in White prevented candidates from “communicating relevant information to voters” on “matters of current public importance,” and did not narrowly advance the State's interest in a non-partisan judiciary, id. at 781–82, 122 S.Ct. 2528, the same is true of Kentucky's canon, which potentially prevents candidates from announcing their views on many issues at once. See id. at 782, 122 S.Ct. 2528 (“We have never allowed the government to prohibit candidates from communicating relevant information to voters during an election.”).

At the same time, the canon does too little to advance the State's interest in impartiality and the avoidance of partisan influence. Party affiliation, as it turns out, is not a forbidden topic. It is forbidden only when the candidate raises the point. If, by contrast, a voter asks the question in a one-on-one setting or in a small gathering, the candidate is free to say what she wants. That reality undermines the suggestion that a candidate deals a fatal blow to judicial impartiality by revealing her party affiliations. And of course, once that information is disclosed, whether in answer to a question or based on prior publicly known affiliations (including holding other elected offices), nothing in the canon prohibits others, whether newspapers or political parties or interest groups, from disclosing to the world the candidate's party affiliation. “A law cannot be regarded as protecting an interest of the highest order, and thus as justifying a restriction upon truthful speech, when it leaves appreciable damage to that supposedly vital interest unprohibited.” Id. at 780, 122 S.Ct. 2528.

The clause undershoots its target in another respect. Although candidates may not reveal their party affiliation, they may discuss their membership in, affiliation with or support of any other type of organization, including organizations that take positions on judges and judicial philosophy. Although the two major political parties take positions on a wide array of issues, many interest groups advance a narrower set of positions and often do so more vocally, particularly with respect to judges. By identifying themselves with such groups, candidates can communicate more about their political and judicial convictions than they ever could by carrying a party membership card—and, in the process, may do as much to call judicial open-mindedness into question as any party affiliation ever would.

The canon also prohibits only disclosure of a candidate's party membership, not party membership itself. Yet the appearance of judicial closed-mindedness is part and parcel of its reality, not a device designed to disguise reality. If concern over judicial partisanship and the influence of political parties on judging truly underlies the clause, the authorization to belong (secretly) to a political party amounts to a gaping omission. A party's undisclosed potential influence on candidates is far worse than its disclosed influence, as the one allows a full airing of the issue before the voters while the other helps to shield it from public view.

Kentucky responds that the restriction supports the Kentucky Constitution's requirement that judicial elections be nonpartisan—that they operate with no partisan primaries and with no partisan identifiers at the ballot booth. See Ky. Const. § 117; Ky.Rev.Stat. § 118A.060. The point, however, cuts both ways. In one sense, it establishes the bona fides of the Commonwealth's policy. But in another *203 sense, it undermines the Commonwealth's professed need to suppress speech in the process. Carey does not challenge the validity of prohibiting party identifiers on the ballot or the validity of holding non-partisan primaries. He just wishes to communicate about a matter of potential interest to the voters and one that is often already a point of public knowledge—party affiliation—on his own terms.

Most States have not made the choice Kentucky did. Fifteen States choose their Supreme Court justices in contested, “nonpartisan” elections, and only five, including Kentucky, prohibit candidates in those elections from revealing their partisan affiliations. See App'x A. (Three more prohibit candidates from claiming to be “a candidate of a political organization” but do not prohibit revealing membership or affiliation. Id.) And two of these five canons—this one and Wisconsin's—have been invalidated. See Siefert, 608 F.3d at 983. That a majority of the States with nonpartisan Supreme Court elections have opted not to censor their candidates in this way of course does not establish the invalidity of the clause, but it does call into question the necessity of implementing Kentucky's nonpartisan judicial election system in this way and whether it amounts to the “least restrictive means” of protecting the Commonwealth's interests. Playboy Entm't Group, 529 U.S. at 813, 120 S.Ct. 1878.

The Commission says that the clause restricts as little speech as possible while preventing Kentucky's elections from turning into ultra-partisan affairs. It allows judges, the Commission adds, to join political parties, to participate in them and to disclose party affiliation if asked in the proper setting—allowing voters who care about a candidate's partisan affiliation to discover it while preventing widespread advertisement of a candidate's party membership and preventing “judicial races [from] turning into partisan political campaigns.” Commission's Opening Br. at 66. Yet this argument looks at the problem through the wrong end of the telescope: It merely demonstrates that the clause does not restrict as much speech as it might, not that the clause restricts no more speech than is necessary.

We do not doubt one of the premises of the canon—that party affiliation may not be a reliable indicator of the qualities that make a good judge. Yet “[i]t is simply not the function of government to select which issues are worth discussing or debating in the course of a political campaign,” White, 536 U.S. at 782, 122 S.Ct. 2528 (quotation omitted), and it is difficult to see how Kentucky's speech restriction does not do just that. Informational bans premised on the fear that voters cannot handle the disclosure have a long history of being legislatively tried and judicially struck, whether in the election setting or elsewhere. See, e.g., Brown, 456 U.S. at 60, 102 S.Ct. 1523 (“The State's fear that voters might make an ill-advised choice does not provide the State with a compelling justification for limiting speech.”); Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, Inc., 425 U.S. 748, 773, 96 S.Ct. 1817, 48 L.Ed.2d 346 (1976). Voters often resort to a variety of proxies in selecting judges and other office holders, some good, some bad. And while political identification may be an unhelpful way to pick judges, it assuredly beats other grounds, such as the all-too-familiar formula of running candidates with familiar or popular last names. In that respect, this informational ban increases the likelihood that one of the least relevant grounds for judicial selection—the fortuity of one's surname—is all that the voters will have to go on. As the district court correctly concluded, *204 this clause violates the First Amendment on its face.

Solicitation clause. Kentucky prohibits judicial candidates from “solicit[ing] campaign funds,” a restriction that extends to all fundraising by the candidate, including in-person solicitations, group solicitations, telephone calls and letters. Rules of Supreme Court of Kentucky 4.300, Canon 5(B)(2). The clause permits the candidate to establish a committee that may solicit campaign donations, and it permits the committee to disclose to the candidate the names of people who donated to the campaign and those who declined. See id. Kentucky says that the clause satisfies the First Amendment, but we, like the district court, conclude that it does not.

As with the party affiliation clause, we do not doubt the bona fides of the solicitation clause: that it serves Kentucky's compelling interest in an impartial judiciary. The same goes for its interest in preserving the appearance and reality of a non-corrupt judiciary, an objective often served by fundraising limitations. See Fed. Election Comm'n v. Nat'l Conservative Political Action Comm., 470 U.S. 480, 496–97, 105 S.Ct. 1459, 84 L.Ed.2d 455 (1985). Litigants have a due process right to a trial before a judge with no “direct, personal, substantial pecuniary interest” in the outcome, Tumey v. Ohio, 273 U.S. 510, 523, 47 S.Ct. 437, 71 L.Ed. 749 (1927), and the legitimacy of the judiciary rests on delivering on that promise and in furthering the public's trust in the integrity of its judges, see Mistretta v. United States, 488 U.S. 361, 407, 109 S.Ct. 647, 102 L.Ed.2d 714 (1989). See generally Caperton, 129 S.Ct. 2252.

Preserving these interests, we also acknowledge, grows more complicated when a State exercises its sovereign right to select judges through popular elections. Judicial elections, like most elections, require money—often a lot of it. Kentucky's 2006 Supreme Court election, which featured four contested races and one uncontested race, saw ten candidates raise a total of $2,119,871, of which the candidates spent $772,563 on 2,357 television commercials. Brennan Center for Justice at NYU School of Law, The New Politics of Judicial Elections 2006 at 3, 16. “Unless the pool of judicial candidates is limited to those wealthy enough to independently fund their campaigns, a limitation unrelated to judicial skill, the cost of campaigning requires judicial candidates to engage in fundraising.” White, 536 U.S. at 789–90, 122 S.Ct. 2528 (O'Connor, J., concurring). Complicating matters further, the general public often, though not invariably, pays less attention to judicial elections than other elections, forcing judicial candidates to focus their fundraising efforts on the segment of the population most likely to have an interest in judicial races: the bar. “This leads to the unseemly situation in which judges preside over cases in which the parties are represented by counsel who have contributed in varying amounts to the judicial campaigns.” Stretton, 944 F.2d at 145.

That the clause advances important government interests, however, does not establish that it does so narrowly. Prohibiting candidates from asking for money suppresses speech in the most conspicuous of ways and, in the process, favors some candidates over others—incumbent judges (who benefit from their current status) over non-judicial candidates, the well-to-do (who may not need to raise any money at all) over lower-income candidates, and the well-connected (who have an army of potential fundraisers) over outsiders. For these reasons, it is tempting to say that any limitation on *205 a candidate's right to ask for a campaign contribution is one limitation too many. But there are at least two areas covered by the clause that test such an interpretation—face-to-face solicitations, particularly by sitting judges, and solicitations of individuals with cases pending in front of the court. Yet we need not decide the validity of such restrictions today because Kentucky goes well beyond them.

Besides covering in-person solicitations and those directed at individuals with pending cases, the canon prohibits a range of other solicitations, including speeches to large groups and signed mass mailings. Such indirect methods of solicitation present little or no risk of undue pressure or the appearance of a quid pro quo. No one could reasonably believe that a failure to respond to a signed mass mailing asking for donations would result in unfair treatment in future dealings with the judge. Nor would a speech requesting donations from a large gathering have a “coercive effect” on reasonable attendees. Commission's Opening Br. at 55; compare Ohralik v. Ohio State Bar Ass'n, 436 U.S. 447, 465–66, 98 S.Ct. 1912, 56 L.Ed.2d 444 (1978) (State may regulate lawyers' in-person for-profit solicitation of clients because of “intrusive[ness]” of “persuasion under circumstances conducive to uninformed acquiescence”) with In re Primus, 436 U.S. 412, 435–36, 98 S.Ct. 1893, 56 L.Ed.2d 417 (1978) (regulation of lawyer's written, not-for-profit solicitation merited heightened scrutiny because it did not “afford any significant opportunity for overreaching or coercion”) and Shapero v. Ky. Bar Ass'n, 486 U.S. 466, 475, 108 S.Ct. 1916, 100 L.Ed.2d 475 (1988) ( “Targeted, direct-mail solicitation is distinguishable from the in-person solicitation” because there is no “badgering advocate breathing down [a potential client's] neck,” asking for “an immediate yes-or-no answer.”).

At the same time, the clause does too little to protect the Commonwealth's interests. Although the candidate himself may not solicit donations, his campaign committee may. And nothing prevents a committee member from soliciting donations in person. That leaves a rule preventing a candidate from sending a signed mass mailing to every voter in the district but permitting the candidate's best friend to ask for a donation directly from an attorney who frequently practices before the court. Are not the risks of coercion and undue appearance far less with the first (prohibited) solicitation than the second (permitted) one?

Although the clause prevents judicial candidates from saying “please, give me a donation,” it does not prevent them from saying “thank you” for a donation given. The clause bars any solicitation, whether in a large group or small one, whether by letter or one on one, but it does not bar the candidate from learning how individuals responded to the committee's solicitations. That omission suggests that the only interest at play is the impolitic interpersonal dynamics of a candidate's request for money, not the more corrosive reality of who gives and how much. If the purported risk addressed by the clause is that the judge or candidate will treat donors and non-donors differently, it is knowing who contributed and who balked that makes the difference, not who asked for the contribution. If Kentucky fears that judges will allow campaign donations to affect their rulings, it must believe that “[s]uccessful candidates will feel beholden to the people who helped them get elected regardless of who did the soliciting of support.” Weaver, 309 F.3d at 1323.

Two other circuits have considered the validity of similar canons and have come to similar conclusions. In Weaver, the Eleventh Circuit considered a Georgia rule providing *206 that judicial candidates “shall not themselves solicit campaign funds.” 309 F.3d at 1315. Relying on many of the same means-end problems identified here, the court concluded that the canon was “not narrowly tailored to serve Georgia's compelling interest in judicial impartiality.” Id. at 1322. In White II, on remand after the Supreme Court invalidated Minnesota's “announce” clause, the Eighth Circuit invalidated a canon that prohibited judicial candidates from “personally solicit[ing] or accept[ing] campaign contributions.” 416 F.3d at 745. In Siefert, 608 F.3d 974, the Seventh Circuit upheld Wisconsin's prohibition on judges' “personal[ ] solicit[ation]” of campaign contributions. See 608 F.3d at 990. But, in doing so, it focused on the problems associated with “direct” solicitation and did not consider the validity of applying the canon to mass mailings and group solicitations—the most troubling scenarios here. See id. at 989–90.

The state defendants push back, arguing that, even though a candidate may discover her donors' identities from the campaign committee, the solicitation clause makes “favoritism” toward contributors “more difficult.” Commission's Opening Br. at 57. After all, they reason, when a candidate asks for a donation in person, she immediately will find out whether the donor gives and how much. Id. That may or may not be true. But even if we grant the Commonwealth's premise—that in-person solicitations always lead to more immediate information about donations or rejections—that suggests only that the solicitation clause may be constitutional in some settings. It does not resolve the clause's considerable overbreadth: its application to mass-mailing solicitations or speeches to a large audience.

But the solicitation clause must be constitutional, the state defendants add, because most other States with judicial elections also prevent candidates from soliciting funds. See Commission's Opening Br. at 53. The argument is not as helpful as they suggest. By our count, twenty-two States currently elect judges to their highest courts in contested elections. (States with retention elections are less relevant because by definition they do not involve two candidates competing for the same seat.) Of these twenty-two States, thirteen, including Kentucky, prohibit candidates from soliciting campaign contributions. See App'x B. (Two more have hortatory canons telling candidates they “should not” or are “strongly discouraged” from personally soliciting. Id.) Yet this bare majority is no more dispositive here than it was in White, where twenty-six States had some form of announce clause. See White, 536 U.S. at 786, 122 S.Ct. 2528. No less importantly, we do not decide today whether a State could enact a narrowly tailored solicitation clause—say, one focused on one-on-one solicitations or solicitations from individuals with cases pending before the court—only that this clause does not do so narrowly.

The Commonwealth to its credit wishes to avoid cases like Simes v. Ark. Judicial Discipline & Disability Comm'n, 368 Ark. 577, 247 S.W.3d 876 (2007), where a judge “made direct, personal solicitations” to attorneys who “had cases currently pending in the judge's court.” Id. at 880. But Kentucky's clause goes well beyond these sorts of solicitations. Kentucky has chosen to elect its judges in competitive elections and must abide by some of the risks that go with that decision. See White, 536 U.S. at 792, 122 S.Ct. 2528 (O'Connor, J., concurring). While we do not question Kentucky's right to select judges through popular elections, the Commonwealth cannot exempt itself from the demands of the First Amendment in the process. See *207 id. at 788, 122 S.Ct. 2528 (majority); Geary v. Renne, 911 F.2d 280, 294 (9th Cir.1990) (en banc) (Reinhardt, J., concurring) (“The State ... cannot have it both ways. If it wants to elect its judges, it cannot deprive its citizens of a full and robust election debate.”), vacated on other grounds by Renne v. Geary, 501 U.S. 312, 111 S.Ct. 2331, 115 L.Ed.2d 288 (1991). The solicitation clause is overbroad and thus invalid on its face.

The commits clause. In prohibiting judicial candidates from “intentionally or recklessly mak[ing] a statement that a reasonable person would perceive as committing a judge or candidate to rule a certain way in a case, controversy, or issue that is likely to come before the court,” Canon 5(B)(1)(c), the commits clause covers a range of campaign statements. Some of those restrictions are legitimate. Others may not be. And there is a “vast middle ground of uncertainty” between the two. Outlaw v. Airtech Air Conditioning & Heating, Inc., 412 F.3d 156, 161 (D.C.Cir.2005) (Roberts, J.).

In what seems to be its core sense, the clause, found in one form or another in 39 States, see App'x C, runs the gauntlet of strict scrutiny. By preventing candidates from making “statement[s]” that “commit[ ]” them “to rule a certain way in a case [or] controversy,” the clause secures a basic objective of the judiciary, one so basic that due process requires it: that litigants have a right to air their disputes before judges who have not committed to rule against them before the opening brief is read. See Caperton, 129 S.Ct. at 2266–67; Bracy v. Gramley, 520 U.S. 899, 904–05, 117 S.Ct. 1793, 138 L.Ed.2d 97 (1997). Whatever else a fair adjudication requires, it demands that judges decide cases based on the law and facts before them, not based on “express ... commitments that they may have made to their campaign supporters.” Buckley v. Ill. Judicial Inquiry Bd., 997 F.2d 224, 227 (7th Cir.1993). No one, Carey included, disputes that the Commonwealth has a compelling interest in “prohibit[ing] candidates from promising to rule a certain way on cases.” Carey's Opening Br. at 14–15.

Nor does Carey dispute that the clause narrowly advances this interest—if, that is, the clause is confined to campaign “commitments” with respect to “cases” or “controversies.” So limited, the clause targets the kinds of interests White suggests the States may protect, as judicial commitments with respect to cases and controversies implicate not just a lack of open mindedness about the law but a lack of impartiality in its most essential sense—a commitment to rule for one party over another. See White, 536 U.S. at 775–76, 122 S.Ct. 2528 (“[I]mpartiality” in “the traditional sense” means “apply[ing] the law to [one party] in the same way [one] applies it to any other party.”). The First Amendment permits a State to limit speech when the Due Process Clause demands nothing less.

But the canon does not stop there. It also prevents candidates from making commitments about “issues.” A commitment to rule a certain way on “issues likely to come before the court” covers a raft of electioneering stands, and it unmoors the prohibition from “cases” or “controversies” and the party-specific connotations that come with those terms. As White reminds us, “there is almost no legal or political issue that is unlikely to come before a judge of an American court, state or federal, of general jurisdiction.” White, 536 U.S. at 772, 122 S.Ct. 2528 (emphasis added and quotation omitted). To make matters worse, the commentary to the clause says that it covers the “appearance” of *208 making “issue”-related commitments. Canon 5(B)(1)(c), Cmt.

Think back to White. How is a judicial candidate's “announcement” of a position on a legal “issue” during an election campaign not likely to create the “appearance” that the candidate has “committed” to “rule a certain way” on the “issue”? And, if that is so, how can this aspect of the canon survive White, given that it seems to ban what White permits? These are good questions, but they prompt an even more basic one: what exactly does the “issues” prohibition cover?

The clause contains a serious level-of-generality problem. At the broadest level of meaning, it would seem to cover issue-related promises like these: “I commit to follow stare decisis ”; “I commit to follow an originalist theory of constitutional interpretation” or for that matter “a living constitutionalist theory”; “I commit to a purposive method of statutory interpretation” or for that matter a “textual” one; “I commit to use (or not to use) legislative history”; or “I commit to be a rule-of-law judge.” One might reasonably say that the clause covers all of these statements, as they all relate to “issues” likely to come before a court and they all create an “appearance” of commitment. Yet if that is what the clause means, it is hard to square with the Constitution. A restriction on such promises does nothing to prevent the kind of “impartiality” that the States have an interest in securing—defined as bias (or the appearance of bias) toward particular parties or cases. See White, 536 U.S. at 776–77, 122 S.Ct. 2528.

In a narrower sense, however, the “issues” prohibition may serve that interest. In White itself, the Court contemplated that a State could prevent a candidate from highlighting an “unbroken record of affirming convictions for rape” because such statements would “exhibit a bias against parties,” namely against these types of criminal defendants and in favor of the prosecutor in these types of appeals. 536 U.S. at 777 n. 7, 122 S.Ct. 2528; id. at 800–01, 122 S.Ct. 2528 (Stevens, J., dissenting). An interpretation of the clause confined to these kinds of statements thus might advance a compelling state interest and do so narrowly.

In a facial challenge like this one, the ultimate question is one of overbreadth: Does the law “prohibit[ ] a substantial amount of protected speech both in an absolute sense and relative to [the canons'] plainly legitimate sweep”? Connection Distrib. Co., 557 F.3d at 336 (internal quotations omitted). To determine the extent of a law's illegitimate reach, one needs to know what it means, as “it is impossible to determine whether a statute reaches too far without first knowing what the statute covers.” United States v. Williams, 553 U.S. 285, 293, 128 S.Ct. 1830, 170 L.Ed.2d 650 (2008).

That inquiry has not happened here—at least with respect to the “issues” prohibition. In upholding this clause, the district court focused on its application to “cases” and “controversies” and, to that extent, we agree with its analysis for the reasons noted. But the district court did not explore the clause's applicability to “issues,” the array of settings in which that part of the clause and commentary may apply and the tension of several of them with White. At oral argument, we asked the parties about the point. Carey agreed that the commits clause would satisfy the First Amendment if the clause did not contain an “issues” component to it, and saw the addition of the “issues” language (together with the commentary) as having two impermissible effects: chilling candidates' free-speech rights to discuss their legal philosophies freely, and effectively sidestepping White by prohibiting candidates *209 from announcing their positions on legal issues. The state defendants suggested that a narrowing construction of the “issues” clause could save it.

Under these circumstances, discretion, to say nothing of respect for a co-equal sovereign, is the better part of valor. At this point it is not clear what the Commonwealth's position on the term is, and the district court has not yet explored these issues. If we remand this aspect of the case to the district court, the court will have that chance. So too will the parties—particularly the state defendants, who retain considerable authority over shaping the clause and the commentary that goes with it. The state defendants may be able to obtain authority to remove the “issues” language; they may be able to identify an acceptable narrowing construction of the “issues” language along with a modification to the commentary; or they may suggest certification to the Kentucky Supreme Court. Any of these options may spare the federal courts the task of resolving a difficult constitutional question, and at a minimum they will give the Commonwealth a first shot at addressing the question.

* * *

There is room for debate about whether the election of state court judges is a good idea or a bad one. Yet there is no room for debate that, if a State opts to select its judges through popular elections, it must comply with the First Amendment in doing so. In this case, we have upheld some components of Kentucky's Code of Judicial Conduct, invalidated others and sought clarification of still one other provision. Through it all, no one should lose sight of the reality that a judicial candidate's right to engage in certain types of speech says nothing about the desirability of that speech. The First Amendment protects the meek and brazen, the “offensive” and agreeable. Texas v. Johnson, 491 U.S. 397, 414, 109 S.Ct. 2533, 105 L.Ed.2d 342 (1989). Today's case is about the meaning of the First Amendment, not about the virtues of some types of judicial campaign speech relative to others.

For these reasons, we affirm the district court's judgment as to the party affiliation and solicitation clauses and vacate its judgment as to the commits clause and remand the case for further consideration of the meaning and validity of that clause.