OPINION

PAEZ, Circuit Judge:

A state sets itself on a collision course with the First Amendment when it chooses to popularly elect its judges but restricts a candidate's campaign speech. The conflict arises from the fundamental tension between the ideal of apolitical judicial independence and the critical nature of unfettered speech in the electoral political process. Here we must decide whether several provisions in the Arizona Code of Judicial Conduct restricting judicial candidate speech run afoul of First Amendment protections. Because we are concerned with content-based restrictions on electioneering-related speech, those protections are at their apex. Arizona, like every other state, has a compelling interest in the reality and appearance of an impartial judiciary, but speech restrictions must be narrowly tailored to serve that interest. We hold that several provisions of the Arizona Code of Judicial Conduct unconstitutionally restrict the speech of non judge candidates because the restrictions are not sufficiently narrowly tailored to survive strict scrutiny. Accordingly, we reverse the district court's grant of summary judgment in favor of Defendants.

I.

Arizona counties with fewer than 250,000 people popularly elect local judicial officers. See Ariz. Const. art. VI, §§ 12, 40.1 The Arizona Code of Judicial Conduct2 (the "Code") regulates the conduct of judges campaigning for retention and judicial candidates campaigning for office. The Code provides for discipline if a candidate is elected as a judge, but lawyers who are unsuccessful in their candidacy may also be subject to discipline under the Arizona Rules of Professional Conduct.3 See Ariz.Rev.Stat. Ann. § 17A, Sup.Ct. Rules, Rule 42, Rules of Prof. Conduct, ER 8.2 (2003).

Plaintiff Randolph Wolfson was an unsuccessful candidate for judicial office in Mohave County, Arizona in 2006 and 2008. Wolfson I, 616 F.3d at 1052-53. He intends to run in a future election. Id. at 1054-55. As a candidate, Wolfson wished to conduct a number of activities he believed to be prohibited by the Code, but refrained from doing so, fearing professional discipline.4 He brought this action challenging the facial and as-applied constitutionality of certain provisions of the Code, seeking declaratory and injunctive relief. Defending this appeal are the members of the Arizona Commission on Judicial Conduct (the "Commission") and Arizona Chief Bar Counsel ("State Bar Counsel"), collectively the "Arizona defendants."5

Wolfson challenges five clauses of Rule 4.1 of the Code (the "Rules"):

Ariz.Rev.Stat. Ann. § 17A, Sup.Ct. Rules, Rule 81, Code of Jud. Conduct (2009).

This is the second time that this case is before us. We previously held in Wolfson I that Wolfson's challenges to these clauses (hereinafter the "solicitation" clause (6) and "political activities" clauses, (2)-(5)) were justiciable and remanded them to the district court to consider them on the merits. Wolfson I, 616 F.3d at 1054-62, 1066-67. With respect to his challenge to a now-defunct "pledges and promises" clause, we held that Wolfson lacked standing to challenge it insofar as it applied to the speech of judges. Id. at 1064. "Wolfson cannot assert the constitutional rights of judges when he is not, and may never be, a member of that group." Id.

On remand, ruling on cross-motions for summary judgment, the district court applied a balancing test articulated by the Seventh Circuit in Siefert v. Alexander, 608 F.3d 974 (7th Cir.2010), and Bauer v. Shepard, 620 F.3d 704 (7th Cir.2010), and upheld the constitutionality of the five challenged Code provisions. Wolfson II, 822 F.Supp.2d at 929-30. The balancing test from Siefert/Bauer "derives from the line of Supreme Court cases upholding the limited power of governments to restrict their employees' political speech in order to promote the efficiency and integrity of government services." Id. at 929. The district court held that this standard "strikes an appropriate balance between the weaker First Amendment rights at stake and the stronger State interests in regulating the way it chooses its judges," apparently because the speech at issue was not "core speech" deserving of strict scrutiny but "behavior short of true speech." Id. at 929-30.

The district court proceeded to balance the interests of the state against the interests of a judicial candidate. With respect to the political activities restrictions (the campaigning and endorsement clauses), the district court held that "[e]ndorsements, making speeches, and soliciting funds on behalf of other candidates is not... core political speech." Id. at 931. The district court distinguished between announcing one's own political views or qualifications — speech protected by Republican Party of Minn. v. White, 536 U.S. 765, 788, 122 S.Ct. 2528, 153 L.Ed.2d 694 (2002) (White I) — and the type of speech prohibited by the Rules, which only "advance[s] other candidates' political aspirations, or ... garner[s] votes by way of political coattails." Wolfson II, 822 F.Supp.2d at 931-32. Moreover, although the district court recognized that its review was "limited to the constitutionality of the Rules as applied to judicial candidates who are not also sitting judges," id. at 928, it nonetheless

Id. at 932. The district court assumed the constitutional validity of the Rules restricting political activities as applied to sitting judges, holding that "the Pickering line of cases [upholding the government's power to restrict employees' political speech to promote efficiency and integrity of government services] remains relevant to restrictions on the speech of sitting judges." Id. The court concluded that Rules 4.1(A)(2)-(5) appropriately balanced the state's interest in "protecting the due process rights of litigants and ensuring the real and perceived impartiality of the judiciary" against a candidate's interest in "participating in the political campaigns of other candidates" and upheld the political activities clauses as constitutional. Id.

As for the solicitation clause (Rule 4.1(A)(6)) prohibiting a judicial candidate from "personally solicit[ing] or accept[ing] campaign contributions other than through a campaign committee," the district court held that it was constitutional as applied to non judge candidates because it struck "a constitutional balance" between the state's interest in the appearance and actuality of an impartial judiciary and a candidate's need for funds. Id. at 931. The district court found that all forms of personal solicitation, whether in-person or via signed mass mailings, created "the same risk of coercion and bias." Id. Wolfson timely appealed.

II.

A.

We review de novo an order granting summary judgment on the constitutionality of a statute. See Nunez by Nunez v. City of San Diego, 114 F.3d 935, 940 (9th Cir.1997).

B.

Wolfson seeks to invalidate the challenged Rules on their face, including as to sitting judges campaigning for retention or reelection. In Wolfson I, however, we held that "Wolfson cannot assert the constitutional rights of judges when he is not, and may never be, a member of that group." 616 F.3d at 1064. Nonetheless, although we reject the Arizona defendants' argument, which the district court adopted, that the balancing test applicable to government employee speech cases also applies to sitting judges and thus fairly extends to non judge candidates campaigning for office, we must establish the scope of our review of the challenged Rules.

We decline to adopt the district court's approach because such reasoning requires a series of unnecessary constitutional decisions.7 Rather, our analysis of the challenged Rules is based on Wolfson's status as a non judge candidate. While the Rules apply to judges whether or not a judge is actively campaigning for retention or reelection, they only apply to non judge candidates during an election campaign for judicial office.8 There is a meaningful distinction in how the Rules actually apply to judges versus non judge candidates that may warrant distinct levels of scrutiny. Regulated non judge speech only takes place during a campaign. As noted above, political speech is subject to the highest degree of First Amendment protection. Because Wolfson's desired speech would only take place in the context of a political campaign for judicial office, we do not decide whether the restrictions as applied to judges — whether campaigning or not — fit into the "narrow class of speech restrictions" that may be constitutionally permissible if "based on an interest in allowing governmental entities to perform their functions." Citizens United v. Fed. Election Comm'n, 558 U.S. 310, 341, 130 S.Ct. 876, 175 L.Ed.2d 753 (2010).

We are not persuaded that "fundamental fairness," see Wolfson II, 822 F.Supp.2d at 929, warrants making an advisory decision about the constitutional speech rights of judges who are not presently before us and whose rights Wolfson cannot assert, Wolfson I, 616 F.3d at 1064. Under strict scrutiny, see Part III.A, the proponents of a speech regulation must establish a compelling state interest served by the regulation. Neither the Commission nor the State Bar Counsel has argued that Arizona has a compelling state interest in applying the same election regulations to incumbent sitting judges as to candidates who are not sitting judges — only that such an equal application is principled, logical, and fair.

Our decision to limit our review to non judge candidates is ultimately based on judicial restraint. We need not decide today what restrictions on judges' speech are constitutionally justified by the interest in allowing the judiciary to function optimally, nor are we squarely presented with that question. We neither "`anticipate a question of constitutional law in advance of the necessity of deciding it' nor `formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied.'" Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 450, 128 S.Ct. 1184, 170 L.Ed.2d 151 (2008) (quoting Ashwander v. TVA, 297 U.S. 288, 347, 56 S.Ct. 466, 80 L.Ed. 688 (1936) (Brandeis, J., concurring)). The only constitutional question we address is whether the challenged Rules violate the First Amendment rights of non judge candidates.

III.

A.

Strict scrutiny applies to this First Amendment challenge. The regulations in question are content- and speaker-based restrictions on political speech, which receives the most stringent First Amendment protection. Republican Party of Minn. v. White, 416 F.3d 738, 748-49 (8th Cir.2005) (White II); see also Eu v. San Francisco Cnty. Democratic Cent. Comm., 489 U.S. 214, 223, 109 S.Ct. 1013, 103 L.Ed.2d 271 (1989) ("[T]he First Amendment has its fullest and most urgent application to speech uttered during a campaign for political office." (internal quotation marks omitted)). We recently applied strict scrutiny to another state statute regulating judicial elections because it was, "on its face, a content-based restriction on political speech and association [which] thereby threaten[ed] to abridge a fundamental right." Sanders Cnty. Republican Cent. Comm. v. Bullock, 698 F.3d 741, 746 (9th Cir.2012) (holding unconstitutional a ban on political party endorsement of judicial candidates).

Content-based restrictions on speech receive strict scrutiny. See United States v. Playboy Entm't Grp., Inc., 529 U.S. 803, 813, 120 S.Ct. 1878, 146 L.Ed.2d 865 (2000). Here, the Rules at issue

Carey v. Wolnitzek, 614 F.3d 189, 198-99 (6th Cir.2010). The canons do not address any of the "categorical carve-outs" of proscribable speech. See id. at 199. Nor are they the types of regulations to which the Supreme Court has applied a less rigorous standard of review, such as time, place and manner restrictions, commercial speech, or expressive conduct. Id.

Every sister circuit except the Seventh that has considered similar regulations since White I has applied strict scrutiny as the standard of review. See Wersal v. Sexton, 674 F.3d 1010, 1019 (8th Cir.2012) (en banc), cert. denied, ___ U.S. ___, 133 S.Ct. 209, 184 L.Ed.2d 40 (2012); Carey, 614 F.3d at 198-99; White II, 416 F.3d at 749, 764-65; Weaver v. Bonner, 309 F.3d 1312, 1319 (11th Cir.2002). We are not persuaded by the Seventh Circuit's approach, which the Arizona defendants urge us to adopt by asking us to affirm the district court.

The Seventh Circuit treated the solicitation ban in Siefert as a "campaign finance regulation" and applied the "closely drawn scrutiny" framework of Buckley v. Valeo, 608 F.3d at 988 (citing 424 U.S. 1, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976) (per curiam)). The court treated the solicitation ban like a restriction on a campaign contribution — though by default, because the solicitation ban was not an expenditure restriction. Id. Contrary to the Arizona defendants' argument, the solicitation clause at issue here is not a restriction on a campaign contribution within the meaning of Buckley, 424 U.S. at 26-27, 96 S.Ct. 612. Arizona's solicitation ban does nothing at all to limit contributions to a judicial candidate's campaign — either in amount or from certain persons or groups. Contribution restrictions, like those at issue in Buckley, restrict the speech of potential contributors. 424 U.S. at 21-22, 96 S.Ct. 612. The Rule at issue here restricts only the solicitation for the contributions — the speech of the candidate.9 Indeed, Buckley says nothing at all about solicitation, other than to note that candidates will ask for contributions. Buckley's framework is inapposite here.10

Considering a rule prohibiting a judge or judicial candidate from making endorsements or speaking on behalf of a partisan candidate or platform, the Seventh Circuit applied "a balancing approach" derived from a line of cases determining the speech rights of government employees. Siefert, 608 F.3d at 983-87. As noted in Part II.B, here we consider only the speech rights of Wolfson as a private citizen and judicial candidate — not yet, and perhaps never, a government employee. "[Wolfson] [i]s not a sitting judge but a challenger; he ha[s] not voluntarily entered into an employment relationship with the State or surrendered any First Amendment rights. His speech may not be controlled or abridged in this manner." See White I, 536 U.S at 796, 122 S.Ct. 2528 (Kennedy, J., concurring). For the reasons discussed above, we decline to extend the rationale from the employee-speech cases to apply a lower level of scrutiny to the restrictions on Wolfson's First Amendment rights during a judicial campaign.

The Seventh Circuit also reasoned that a balancing approach was appropriate because endorsements are "a different form of speech" outside of "core" political speech thus having "limited communicative value," and when judges make endorsements they are "speaking as judges, and trading on the prestige of their office to advance other political ends." Siefert, 608 F.3d at 983, 984, 986.11 We do not hold the same view of endorsements by non-judge candidates. In Sanders County, we held that endorsements of judicial candidates are no different from other types of political speech: "Thus, political speech — including the endorsement of candidates for office — is at the core of speech protected by the First Amendment." 698 F.3d at 745. Similarly, endorsements by candidates for office is also political speech protected by the First Amendment. Moreover, endorsements made by a non-judge candidate cannot trade on the prestige of an office that candidate does not yet hold.

We share the Seventh Circuit's concerns about protecting litigants' due process rights, which we recognize as a compelling state interest. That court reasoned that because "restrictions on judicial speech may, in some circumstances, be required by the Due Process Clause," states could regulate even political speech by judges if the regulations served the state's interest in protecting litigants' constitutional right to due process. Siefert, 608 F.3d at 984. We agree that due process concerns are paramount, but this concern does not justify a categorically lower level of constitutional scrutiny for political speech by judicial candidates. Applying strict scrutiny, we can adequately assess whether regulations on a judicial candidate's political speech are narrowly tailored to serve the state's compelling interest in protecting litigants' due process rights. Narrow tailoring is most appropriate. Although we could scarcely imagine a more compelling state interest, we also recognize that "due process" concerns arise not in the ether, but "only ... in the context of judicial proceedings." See Michelle T. Friedland, Disqualification or Suppression: Due Process and the Response to Judicial Campaign Speech, 104 Colum. L.Rev. 563, 613 (2004).12 We are mindful of the fact that we should endeavor to protect litigants from even the "potential for due process violations" or the "probability of unfairness." See White I, 536 U.S. at 815-16, 122 S.Ct. 2528 (Ginsburg, J., dissenting) (emphasis added) (internal quotation marks omitted). The potential for and probability of a problem that in actuality arises only in real cases does not, however, translate into a generalized concern about the appearance or reality of an impartial judiciary warranting a lower level of scrutiny. Indeed, the Eighth Circuit identified the flaw in this argument.

Weaver, 309 F.3d at 1320; accord White I, 536 U.S. at 792, 122 S.Ct. 2528 (O'Connor, J., concurring) ("If the State has a problem with judicial impartiality, it is largely one the State brought upon itself by continuing the practice of popularly electing judges.").13 Moreover, there is an equally compelling state interest in the free flow of information during a political campaign. "Deciding the relevance of candidate speech is the right of the voters, not the State." White I, 536 U.S. at 794, 122 S.Ct. 2528 (Kennedy, J., concurring). Whether and to what extent a judicial candidate chooses to engage in activities such as endorsing and making speeches on behalf of other candidates, fundraising for or taking part in other political campaigns, or asking for contributions is information that the electorate can use to decide whether he or she is qualified to hold judicial office. "The vast majority of states have judicial elections because of a belief that judges as government officials should be accountable to their constituents. By making this choice, the states, by definition, are turning judges into politicians." Erwin Chemerinsky, Restrictions on the Speech of Judicial Candidates Are Unconstitutional, 35 Ind. L.Rev. 735, 736 (2002). Along with knowing a candidate's views on legal or political issues, voters have a right to know how political their potential judge might be.14 To the extent states wish to avoid a politicized judiciary, they can choose to do so by not electing judges.

B.

Under strict scrutiny, the Arizona defendants have the burden to prove that the challenged Rules further a compelling interest and are narrowly tailored to achieve that interest. Citizens United, 558 U.S. at 340, 130 S.Ct. 876. First we consider Arizona's state interests. Then, we analyze whether the solicitation clause (Rule 4.1(A)(6)) and the political activities clauses (Rules 4.1(A)(2)-(5)) are narrowly tailored to serve those interests.

1.

Every court to consider the issue has affirmed that states have a compelling interest in the appearance and actuality of an impartial judiciary. See, e.g., White I, 536 U.S. at 775-76, 122 S.Ct. 2528. The meaning of "impartiality" is lack of bias for or against either party to a case. Id. at 775, 122 S.Ct. 2528. This definition accords with the idea that due process violations arise only in case-specific contexts. The Supreme Court has also recognized that states have a compelling interest in preventing corruption or the appearance of corruption through campaign finance regulations. Buckley, 424 U.S. at 26-27, 96 S.Ct. 612; see also Citizens United, 558 U.S. at 357, 130 S.Ct. 876. Thus, we recognize that Arizona has a compelling interest in an uncorrupt judiciary that appears to be and is impartial to the parties who appear before its judges.

The Arizona defendants also argue for two other compelling interests that we do not find persuasive. First, the Commission defendants argue that "the State has a compelling interest in preventing candidates (who will after all be the next judges if and when elected) from trampling on the interests of impartiality and public confidence." This argument is, essentially, that states have a compelling interest in regulating candidates' speech; we do not find an interest in regulating speech per se to be compelling. We do agree, however, that states have a compelling interest in maintaining public confidence in the judiciary. In a similar vein, State Bar Counsel argues that Arizona has a compelling interest in avoiding "judicial campaign abuses that threaten to imperil public confidence in the fairness and integrity of the nation's elected judges." But, as explained above, any imperilment of public confidence has its roots in the very nature of judicial elections, and not in the speech of candidates who must participate in those elections to become judges. See White I, 536 U.S. at 792, 122 S.Ct. 2528 (O'Connor, J., concurring).15 If a judicial candidate wishes to engage in politicking to achieve a seat on the bench, keeping the public ignorant of that fact may conceal valuable information about how well that candidate may uphold the office of an ideally impartial, apolitical adjudicator.

Second, the Commission defendants argue that Arizona has a compelling interest in "preventing judges and judicial candidates from using the prestige of their office or potential office for purposes not related to their judicial duties." We are not persuaded by this argument as applied to non judge candidates, who cannot abuse the prestige of an office they do not yet and may never hold.

2.

The solicitation clause prohibits a judicial candidate from "personally solicit[ing] or accept[ing] campaign contributions other than through a campaign committee authorized by Rule 4.4." Rule 4.1(A)(6).16 The Code defines "personally solicit" as "a direct request made by a judge or a judicial candidate for financial support or in-kind services, whether made by letter, telephone, or any other means of communication." Ariz.Rev.Stat. Ann. § 17A, Sup.Ct. Rules, Rule 81, Code of Jud. Conduct, "Terminology" (2009). We hold that Rule 4.1(A)(6) is unconstitutional as applied to non judge judicial candidates because it restricts speech that presents little to no risk of corruption or bias towards future litigants and is not narrowly tailored to serve those state interests.

Arizona's sweeping definition of "personally solicit" encompasses methods not likely to impinge on even the appearance of impartiality. The Sixth Circuit recently invalidated a similar clause in Kentucky that also extended beyond one-on-one, in-person solicitations to group solicitations, telephone calls, and letters. Carey, 614 F.3d at 204. We agree with our sister court's cogent analysis of this issue. "[I]ndirect methods of solicitation [such as speeches to large groups and signed mass mailings] present little or no risk of undue pressure or the appearance of a quid pro quo." Id. at 205. The clauses are also underinclusive: a personal solicitation by a campaign committee member who may be the candidate's best friend or close professional associate (such as a law practice partner) is likely to have a greater risk for "coercion and undue appearance" than a signed mass mailing or request during a speech to a large group. Id. Moreover, the Code does not prohibit a candidate's campaign committee from disclosing to the candidate the names of contributors and solicited non-contributors.

Id.17 The lack of narrow tailoring is obvious here: if impartiality or absence of corruption is the concern, what is the point of prohibiting judges from personally asking for solicitations or signing letters, if they are free to know who contributes and who balks at their committee's request? Wersal teaches that the in-person "`ask' is precisely the speech [a state] must regulate to maintain its interest in impartiality and the appearance of impartiality" because of the greater risk of a quid pro quo. 674 F.3d at 1029-31. Indeed, we agree with State Bar Counsel's argument that "the very act of asking for money, personally, creates the impression that judge (and justice) may be for sale." But the clause here sweeps more broadly. It is not necessary "to 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." Carey, 614 F.3d at 206 (emphasis in original).18 The solicitation clause is invalid as applied to non judge candidates.

3.

We analyze Rules 4.1(A)(2)-(5) as the "political activities" clauses. Judicial candidates are prohibited from speechifying for another candidate or organization, endorsing or opposing another candidate, fundraising for another candidate or organization, or actively taking part in any political campaign other than his or her own. These clauses are also not sufficiently narrowly tailored to serve the state's interest in an impartial judiciary, and are thus unconstitutional restrictions on political speech of non judge candidates for judicial office.

Rules 4.1(A)(2)-(4) — prohibiting speechifying, endorsements, and fundraising — present the closest question. There is an argument that these rules are sufficiently narrowly tailored to be constitutional because they curtail speech that evidences bias towards a particular (potential) party within the scope of White I: the candidate or political organization endorsed or spoken of favorably by the judicial candidate. A plurality of the Eighth Circuit, sitting en banc, upheld a nearly identical Minnesota prohibition on a judge or judicial candidate endorsing "another candidate for public office" because such an endorsement "creates a risk of partiality towards the endorsed party and his or her supporters." Wersal, 674 F.3d at 1024, 1025. The plurality concluded that the clause was narrowly tailored to serve the state's compelling interest in the appearance and reality of an impartial judiciary. Id. at 1028.19

Nonetheless, we hold that these regulations are underinclusive because they only address speech that occurs beginning the day after a non judge candidate has filed his intention to run for judicial office.20 The day before a private citizen becomes a judicial candidate, he or she could have been a major fundraiser or campaign manager for another elected official, or may have donated large sums of money to another's political campaign, or may have himself been an elected politician. The Supreme Court confronted a similar underinclusive issue in White I. There, in explaining why the "announce clause" was underinclusive, the Court said

White I, 536 U.S. at 779-80, 122 S.Ct. 2528. Here too, Rules 4.1(A)(2)-(4) are "woefully underinclusive" because they only address speech made after a candidate has filed his intention to enter the race. Id. at 780, 122 S.Ct. 2528. Contrary to the dissent, we fail to see why this same concern does not apply here.

Moreover, the Arizona defendants have failed to show why the less restrictive remedy of recusal of a successful candidate from any case in which he or she was involved in a party's political campaign or gave an endorsement is an unworkable alternative. "[B]ecause restricting speech should be the government's tool of last resort, the availability of obvious less-restrictive alternatives renders a speech restriction overinclusive." Valle Del Sol Inc. v. Whiting, 709 F.3d 808, 826 (9th Cir.2013). Here, it seems that if a candidate indeed becomes a judge, a less restrictive means of addressing the state's concerns would be to require recusal in cases where the new judge's bias against or in favor of a party is clear.21 Unlike the dissent and the plurality of the Eighth Circuit in Wersal, we decline to address hypothetical situations involving potential frequent litigants and single-judge counties. See Dissent at 1168; Wersal, 674 F.3d at 1027-28 (posing the hypothetical that "candidates and judges would be free to endorse individuals who would become frequent litigants in future cases, such as county sheriffs and prosecutors"). The Arizona defendants have not offered any evidence nor argued that these concerns exist, cf. Siefert, 608 F.3d at 987, though they bear the burden of demonstrating that the Rules survive strict scrutiny. We decline to speculate on whether such a problem would exist in the Arizona judicial elections affected by these Rules.

We hold Rule 4.1(A)(5), which prohibits a judicial candidate from "actively tak[ing] part in any political campaign other than his or her own campaign for election, reelection, or retention in office" to be unconstitutional because it is overbroad. By its terms, it is not limited to restrictions on participation in political campaigns on behalf of persons who may become parties to a suit, but may also include political campaigns on ballot propositions and other issues, including political campaigns for ballot propositions that present no risk of impartiality towards future parties. Thus, Rule 4.1(A)(5) unconstitutionally prohibits protected speech about legal issues. White I, 536 U.S. at 776-78, 122 S.Ct. 2528.

IV.

For these reasons, we reverse the district court's grant of summary judgment to the Arizona defendants. We hold that strict scrutiny applies and that the challenged portions of the Arizona Code of Judicial conduct unconstitutionally restrict the speech of non judge judicial candidates. We remand the case for further proceedings consistent with this opinion.

REVERSED and REMANDED.

BERZON, Circuit Judge, concurring:

Sitting for judicial election while judging cases, Justice Otto Kaus famously quipped, is like "brushing your teeth in the bathroom and trying not to notice the crocodile in the bathtub." Joseph R. Grodin, In Pursuit of Justice: Reflections of a State Supreme Court Justice 177 (1989) (quoting Kaus). Kaus would know. He sat on the California Supreme Court from 1981 to 1985, Gerald T. McLaughlin, Memorial Dedication to Otto Kaus, 30 Loy. L.A. L.Rev. 923, 923 (1997), having narrowly won a retention election in 1982 and retiring from the court soon before the 1986 vote that would unseat three of his former colleagues, Stephen R. Barnett, Otto and the Court, 30 Loy. L.A. L.Rev. 943, 947 & n.19 (1997).1

Kaus' point about the psychology of judging applies outside the context of judicial elections, for the temptation to engage in overt political behavior affects judges generally. And so I write separately to identify, and hopefully to tame, the "crocodile" stalking today's majority opinion: the prospect that the principles we apply now will be used in future litigation to challenge the constitutionality of restrictions on the political behavior of sitting judges. The opinion studiously — and designedly — does not address that issue. But it is worth explaining why, in my view, the considerations pertinent to evaluating the complex of constitutional issues raised by such restrictions are quite different than those the majority opinion applies today.

I.

Today's opinion addresses the constitutionality of certain provisions of the Arizona Code of Judicial Conduct ("Code") only as they apply to judicial candidates who, like Wolfson, have not yet ascended to the bench. It does not decide those provisions' constitutionality as they apply to elected judges who, like Kaus, have already taken their oaths of office. Still less does it decide the constitutionality of restrictions on the political activity of judges who, like us on the federal bench, "hold their Offices during good Behaviour," U.S. Const. art. III, § 1, and never sit for election. In the name of prudence and constitutional avoidance, the majority's opinion rightly reserves judgment on the constitutionality of restricting the speech of sitting judges, an issue neither properly before us nor necessary to the resolution of this case.

I emphasize the limited scope of today's decision for fear that future litigants might otherwise seek to obscure it, despite the repeated admonishments in the opinion. Of the five Code provisions we strike today, only one — the solicitation ban — directly relates to a judicial candidate's own campaign for office.2 The remainder prohibit a would-be judge's efforts to advance the political fortunes of other candidates or causes, through speeches, endorsements, fundraising, financial support, or other campaign assistance.3 As these proscriptions bear little direct relation to judicial candidates' personal political fortunes, a casual reader might be forgiven for assuming that they are just as constitutionally offensive as applied outside the election context, to sitting judges, whether or not they reached the bench via election.

In my view, that is not so, for at least two reasons: The analytic framework applicable to political restrictions on sitting judges may well differ from the one we apply today. And the compelling state interest that could well justify such restrictions differs from the one emphasized in the majority opinion. I address each difference in turn.

II.

In applying strict scrutiny to a judicial candidate who is not now a judge, today's majority opinion rightly rejects the Seventh Circuit's approach, which applies to political restrictions on elected sitting judges a balancing test derived from the Supreme Court's cases on public employee speech. Bauer v. Shepard, 620 F.3d 704 (7th Cir.2010); Siefert v. Alexander, 608 F.3d 974 (7th Cir.2010). Although such a tempered standard has no application to a candidate who has not yet taken his oath of judicial office, whether it would be appropriately applied to political restrictions governing sitting judges is quite a different manner.

The Constitution permits the government to prohibit its employees from speaking about matters of public concern where the government's interest "in promoting the efficiency of the public services it performs through its employees" outweighs the First Amendment interest in speech. Pickering v. Bd. of Educ. of Twp. High Sch. Dist. 205, Will Cnty., Ill., 391 U.S. 563, 568, 88 S.Ct. 1731, 20 L.Ed.2d 811 (1968). The Pickering balancing test seeks "both to promote the individual and societal interests that are served when employees speak as citizens on matters of public concern and to respect the needs of government employers attempting to perform their important public functions." Garcetti v. Ceballos, 547 U.S. 410, 420, 126 S.Ct. 1951, 164 L.Ed.2d 689 (2006). And that test recognizes that "there are certain governmental functions that cannot operate without some restrictions on particular kinds of speech." Citizens United v. Fed. Election Comm'n, 558 U.S. 310, 341, 130 S.Ct. 876, 175 L.Ed.2d 753 (2010).

Republican Party of Minnesota v. White, 536 U.S. 765, 122 S.Ct. 2528, 153 L.Ed.2d 694 (2002), did not decide whether the public employee speech cases would justify restrictions on judges' active support for political causes or the candidacies of others. Justice Kennedy, who was a member of the five justice majority, wrote a separate concurrence, explaining this limitation: "Whether the rationale of Pickering [, 391 U.S. 563, 88 S.Ct. 1731], and Connick v. Myers, 461 U.S. 138, 103 S.Ct. 1684, 75 L.Ed.2d 708 (1983), could be extended to allow a general speech restriction on sitting judges — regardless of whether they are campaigning — in order to promote the efficient administration of justice, is not an issue raised here." White, 536 U.S. at 796, 122 S.Ct. 2528 (Kennedy, J., concurring).

In Siefert, 608 F.3d at 985, the Seventh Circuit extended the public employee speech cases to a provision of the Wisconsin Code of Judicial Conduct prohibiting an elected sitting judge from "[p]ublicly endors[ing] or speak[ing] on behalf of [a political party's] candidates or platforms," id. at 978-79. It reasoned that the government's authority as an employer, "its duty to promote the efficiency of the public services it performs," and the imperative that "the work of the judiciary conform[] with the due process requirements of the Constitution" justified a less rigorous balancing test for restrictions on elected sitting judges' participation in the political campaigns or candidacies of others. Id. at 985. In a subsequent decision, the Seventh Circuit extended this balancing test to provisions of the Indiana Code of Judicial Conduct prohibiting elected judges from leading or holding office in political organizations or making speeches on behalf of such organizations. Bauer, 620 F.3d at 710-11.

The core rationale of the public employee speech cases, on which Siefert and Bauer relied, does not apply to the case presently before us. Wolfson has never been an employee of Arizona, let alone a judge. Indeed, he may never become one. While the public employee speech cases do not rest solely on the now-antiquated principle that the government can condition employment on the waiver of First Amendment rights, see Myers, 461 U.S. at 143-44, 103 S.Ct. 1684, the nature of government employment is a necessary component of their reasoning. Pickering recognized as much, commenting that "it cannot be gainsaid that the State has interests as an employer in regulating the speech of its employees that differ significantly from those it possesses in connection with regulation of the speech of the citizenry in general." 391 U.S. at 568, 88 S.Ct. 1731. The public employee speech cases thus recognize the "crucial difference, with respect to constitutional analysis, between the government exercising `the power to regulate or license, as lawmaker,' and the government acting `as proprietor, to manage [its] internal operation.'" Engquist v. Or. Dep't of Agric., 553 U.S. 591, 598, 128 S.Ct. 2146, 170 L.Ed.2d 975 (2008) (alteration in original) (quoting Cafeteria & Rest. Workers v. McElroy, 367 U.S. 886, 896, 81 S.Ct. 1743, 6 L.Ed.2d 1230 (1961)). Critically, the balancing test the Pickering line of cases articulates does not apply to governmental restrictions on the speech of those, like judicial candidates, not employed by the government. We could not abandon that determinative distinction without dangerously expanding the scope of constitutionally permissible regulation of speech.

But our refusal to apply to a judicial candidate not yet a state employee a balancing test derived from the public employee speech cases says nothing whatever about the applicability of such a test to individuals who have already taken their oaths of judicial office and already receive wages from the state. That question remains unanswered. Resolving the First Amendment challenge of a sitting judge to similar restrictions on his speech will require answering it. And, without prejudging whether we should adopt the Siefert analysis for restrictions on political activity by sitting judges on behalf of political causes or the candidacies of others, I suggest that the analogy to the Pickering line of cases has much to commend it.

III.

Even if we determined that restrictions on the political activity of sitting judges were subject to strict scrutiny, the state interest supporting such a restriction would be far stronger than the one we hold inadequate to justify the restrictions on judicial candidate Wolfson's speech today.

The Supreme Court has recognized as a "vital state interest" the interest in maintaining those "safeguard[s] against judicial campaign abuses that threaten to imperil public confidence in the fairness and integrity of the nation's elected judges." Caperton v. A.T. Massey Coal Co., 556 U.S. 868, 889, 129 S.Ct. 2252, 173 L.Ed.2d 1208 (2009) (emphasis added) (internal quotation marks and citation omitted). Preserving public confidence includes maintaining the perception of judicial propriety. In other words, "`justice must satisfy the appearance of justice.'" In re Murchison, 349 U.S. 133, 136, 75 S.Ct. 623, 99 L.Ed. 942 (1955) (quoting Offutt v. United States, 348 U.S. 11, 14, 75 S.Ct. 11, 99 L.Ed. 11 (1954)). "[T]he appearance of evenhanded justice ... is at the core of due process." Mayberry v. Pennsylvania, 400 U.S. 455, 469, 91 S.Ct. 499, 27 L.Ed.2d 532 (1971) (Harlan, J., concurring).

The majority opinion, taking its cue from Supreme Court cases on judicial elections, focuses its strict scrutiny analysis on the interest in preserving the actuality and appearance of judicial impartiality. The case law's emphasis on impartiality derives from the obligations imposed by the due process clause, particularly "the proposition that an impartial judge is essential to due process." White, 536 U.S. at 776, 122 S.Ct. 2528. This compelling interest in preserving the appearance of impartiality is both weighty and narrow: weighty, because it rises to the level of a constitutional obligation, requiring a judge to recuse himself from a particular case in the name of due process, Caperton, 556 U.S. at 886-87, 129 S.Ct. 2252; and narrow, because it refers only to "lack of bias for or against either party to the proceeding," White, 536 U.S. at 775-76, 122 S.Ct. 2528 (emphasis in original). Given this narrow focus on the parties appearing before a judge in an actual proceeding, the less-restrictive remedy of mandatory recusal is available to a state seeking to protect, as it must, the due process rights of litigants appearing in its courts.

But I would define the state's interest in preserving public confidence in its judiciary more broadly, as reaching beyond the process due specific litigants in particular cases. Maintaining public trust in the judiciary as an institution driven by legal principles rather than political concerns is a structural imperative. The rule of law depends upon it.

The fundamental importance of this structural imperative has been recognized from the founding of the nation. As Alexander Hamilton emphasized in The Federalist No. 78, the courts possess "neither FORCE nor WILL, but merely judgment...." Id. at 433 (Clinton Rossiter ed., 1961). Deprived of those alternative sources of power, the authority of the judiciary instead "lies ... in its legitimacy, a product of substance and perception that shows itself in the people's acceptance of the Judiciary as fit to determine what the... law means and to declare what it demands." Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 865, 112 S.Ct. 2791, 120 L.Ed.2d 674 (1992); see also White, 536 U.S. at 793, 122 S.Ct. 2528 (Kennedy, J., concurring) ("The power and the prerogative of a court ... rest, in the end, upon the respect accorded to its judgments."). It is the courts' perceived legitimacy as institutions grounded in established legal principles, not partisanship, "that leads decisions to be obeyed and averts vigilantism and civil strife." Bauer, 620 F.3d at 712. Loss of judicial legitimacy thus corrodes the rule of law, "sap[ping] the foundations of public and private confidence, and ... introduc[ing] in its stead universal distrust and distress." The Federalist No. 78, at 438. In this sense, "[t]he rule of law, which is a foundation of freedom, presupposes a functioning judiciary respected for its independence, its professional attainments, and the absolute probity of its judges." NY State Bd. of Elections v. Lopez Torres, 552 U.S. 196, 212, 128 S.Ct. 791, 169 L.Ed.2d 665 (2008) (Kennedy, J., concurring).

This nation's political history demonstrates the disastrous effects of the perceived politicization of the courts. Charges that King George "ha[d] obstructed the Administration of Justice" and "ha[d] made judges dependent on his Will alone ...." were among the founding generation's justifications for the 1776 revolution. The Declaration of Independence para. 11 (U.S.1776). Similar concerns apply outside the context of a monarchy: Where the judiciary is drawn into the political intrigues of its coordinate branches, the public might well "fear that the pestilential breath of faction may poison the fountains of justice. The habit of being continually marshaled on opposite sides will be too apt to stifle the voice both of law and of equity." The Federalist No. 81, at 452 (Alexander Hamilton) (Clinton Rossiter ed., 1961).4 And where the politicization of the judiciary brings it into alliance with the politicians who staff the other two branches of government, the public may no longer consider "the courts of justice ... as the bulwark of a limited Constitution against legislative encroachments," The Federalist No. 78, at 437, or executive excesses. In short, when sitting judges support the campaigns of nonjudicial candidates — via endorsements, speeches, money, or other means — the public may begin to see them not as neutral arbiters of a limited system of governance, but as participants in the larger game of politics.5

The defendants here express precisely this concern — that if sitting judges may support the campaigns of others, the public will perceive them as masters of the political game, powerbrokers "trading on the prestige of their office to advance other political ends...." Siefert, 608 F.3d at 984; see also Model Code of Judicial Conduct R. 4.1, cmt.4 (2011) (justifying prohibitions on endorsements and speeches on behalf of other candidates as "prevent[ing sitting judges] from abusing the prestige of judicial office to advance the interests of others"). The opposite fear is equally justified: Today's powerbroker is tomorrow's pawn, as the political winds shift and the next election cycle approaches. The endorsing judge entwines his fate with whomever he endorses and earns the enmity of his favored politician's opponents. "This kind of personal affiliation between a member of the judiciary and a member of the political branches raises the specter — readily perceived by the general public — that the judge's future rulings will be influenced by this political dependency." Wersal v. Sexton, 674 F.3d 1010, 1034 (8th Cir.2012) (Loken, J., concurring in the judgment) (emphasis in original).

In his concurrence in Wersal, Judge Loken concluded that there is a "compelling state interest ... in protecting the political independence of its judiciary." Id. at 1033. I have no reason at this juncture to come to rest on that question. Instead, I emphasize that, at the very least, there is a powerful state interest in preventing sitting judges from playing the part of political powerbroker and creating the publicly visible interdependence that corrodes confidence in judicial autonomy. Assessing whether that interest qualifies as "compelling," in the lexicon of First Amendment doctrine, awaits a properly presented case — particularly as the issue will never arise if we first determine that the Pickering balancing test, rather than strict scrutiny, applies to speech restrictions on sitting judges.

Almost certainly, a state does not forfeit this powerful interest in judicial autonomy by selecting its judges via popular election. It was in the context of a state prohibition against judicial candidates expressing their personal views on disputed legal and political issues during their own campaigns that the Supreme Court has explained that "`the greater power to dispense with elections altogether does not include the lesser power to conduct elections under conditions of state-imposed voter ignorance. If the State chooses to tap the energy and the legitimizing power of the democratic process, it must accord the participants in that process ... the First Amendment rights that attach to their roles.'" White, 536 U.S. at 788, 122 S.Ct. 2528 (alteration in original) (quoting Renne v. Geary, 501 U.S. 312, 349, 111 S.Ct. 2331, 115 L.Ed.2d 288 (1991) (Marshall, J., dissenting)). But that observation does not seem to extend to prohibitions on campaigning on behalf of issue elections or for nonjudicial candidates. The Supreme Court's case law on the political behavior of government employees has "carefully distinguishe[d] between [proscribable] partisan political activities and mere expressions of views," which are constitutionally protected. Biller v. U.S. Merit Sys. Prot. Bd., 863 F.2d 1079, 1089 (2d Cir.1988) (citing U.S. Civil Serv. Comm'n v. Nat'l Ass'n of Letter Carriers, AFL-CIO, 413 U.S. 548, 554-56, 93 S.Ct. 2880, 37 L.Ed.2d 796 (1973), and United Pub. Workers of Am. v. Mitchell, 330 U.S. 75, 98-99, 67 S.Ct. 556, 91 L.Ed. 754 (1947)); Siefert, 608 F.3d at 984; see also Citizens United, 558 U.S. at 341, 130 S.Ct. 876 (citing Letter Carriers in support of the proposition that the Supreme Court has often "upheld a narrow class of speech restrictions that operate to the disadvantage of certain persons, ... based on an interest in allowing governmental entities to perform their functions").6 Indeed, prohibitions on supporting the campaigns of others complement, rather than contradict, the decision to select judges via popular election: By adopting such restrictions alongside judicial elections, states harness the "legitimizing power of the democratic process" while avoiding worrisome interdependence between judges and politicians from the remaining two branches.

Nor should we forget that our own federal scheme supplements its structural protections for judicial autonomy with direct prohibitions on politicking. Structurally, our Constitution endows judges with life tenure and prohibits the diminution of their salaries. U.S. Const. art. III, § 1. Such protections seek to encourage "that independent spirit in the judges which must be essential to the faithful performance of so arduous a duty," The Federalist No. 78, at 437, and help "preserve[] the independence of the Federal Judiciary," White, 536 U.S. at 795, 122 S.Ct. 2528 (Kennedy, J., concurring). In addition to those structural safeguards the federal judiciary has adopted a code of ethics that regulates directly the behavior of federal judges, including restrictions on supporting the political causes and candidacies of others.7 Our ethical code is independent of the structural safeguards that insulate us from the political branches, and it performs a slightly different function. I see no reason why a state cannot adopt the one without the other, except with regard to a judicial candidate's personal campaign for judicial office in states where judicial elections are held.

Critically, the state interest in preserving an autonomous judiciary is powerful only insofar as it applies to sitting judges; it has no application to judicial candidates who, like Wolfson, have not yet reached the bench. The spectacle of sitting judges aiding partisan allies in their political struggles corrodes the public repute of the judiciary in a way that the participation of a mere candidate never can. Indeed, the interest in an independent judiciary does not come into existence until a judge assumes office; the politicking of lay people cannot damage the reputation of a body whose ranks they have not yet joined. Individuals who run for judicial office may themselves be officers of political parties or holders of nonjudicial political office when they decide to run for a judgeship. That politicians can become judges is no secret. But that is different from allowing judges to remain or become politicians while still on the bench. Moreover, as the majority opinion explains, a layman who has not yet assumed office has no prestige derived from the office he has not yet attained to lend his political brethren. Essentially, ascending to the bench is like taking the veil, and that veil does not descend until the oath of office is sworn.

Meanwhile, to the extent White sought to preserve voters' access to "relevant information" and to prevent "state-imposed voter ignorance" about the candidates sitting for election, 536 U.S. at 782, 788, 122 S.Ct. 2528 (internal quotation marks omitted), such concerns are weaker for already seated judges. Such judges already possess a record of decisions that interested voters can analyze to inform themselves about the desirability of competing judicial candidates; under White, they are free to campaign for their own reelection by drawing attention to their records on the bench. By contrast, lay people, like Wolfson, who have not yet sat on the bench lack any such judicial record, making their campaign speech — including endorsements — relatively more valuable for what it reveals about how they might perform in office.

* * *

In sum, the principles applicable to the constitutionality of political restrictions on sitting judges diverge dramatically from those we apply to today's challenge to restrictions on a judicial candidate not now a judge. The standard of review may well differ. And the powerful interests supporting such restrictions differ, too. I need not address, as the issue is not before us, whether the particular restrictions we review today would be constitutional as applied to sitting judges. But I am quite sure that the analysis required to resolve that question will receive scant support from our decision in this case.

TALLMAN, Circuit Judge, dissenting in part:

I agree with the majority that strict scrutiny — not Seifert — is the appropriate standard. I agree that we should limit our decision to non-incumbent judicial candidates. And I agree that Rules 4.1(a)(5) (campaigning for others) and 4.1(a)(6) (personal solicitation) are unconstitutional as applied to those candidates. I concur in the majority opinion only on those points. I part company with my colleagues as to Rules 4.1(a)(2) (giving speeches on behalf of others), (3) (endorsing others), and (4) (soliciting money for others). These three rules are constitutional because they are narrowly tailored to serve the state's compelling interest in maintaining judicial impartiality and its appearance — the hallmark of government's third branch.

My colleagues acknowledge that these three rules "present the closest question," and that the Eighth Circuit upheld similar ones. Wersal, 674 F.3d at 1024-25. Nonetheless, the majority concludes that they are not narrowly tailored for two reasons: timing and recusal. The timing argument is that the rules are underinclusive because "they only address speech that occurs beginning the day after a non judge candidate has filed his intention to run for judicial office." The recusal argument is that the rules are more restrictive than recusal, i.e., requiring judges who have campaigned for others to recuse themselves when those others show up as litigants. I dissent because I do not find these reasons persuasive.

The majority's timing argument is clever but impractical. Its breadth alone suggests this. The argument would cut down any restriction (a) that is subject to strict scrutiny and (b) that starts to apply to people only after some triggering event. If the restriction's enactment counts as a triggering event, and I don't see why it wouldn't, then strict scrutiny would always be fatal. That cannot be the law.

Moreover, the argument doesn't actually answer the question, which is whether there are less restrictive ways to preserve judicial impartiality and its appearance. Having no rules is, of course, less restrictive. But it isn't an alternative means of furthering the interest at stake here. Any actual alternative will suffer from the timing problem the majority identifies. So the timing argument tells us nothing about which alternative is the least restrictive; it only identifies a problem that all conceivable alternatives share.

The majority's recusal argument, like the timing argument, is too impractical in my view. In Arizona, only very small counties elect judges. And some small counties may well have only one superior court judge. If that one judge campaigns for someone who is then elected sheriff or district attorney, an outside judge would be necessary in every criminal case and in all civil cases involving the county where the district attorney is its lawyer. Constant recusal is no solution.

That's what the Eighth Circuit held in Wersal, after it considered this obvious problem. 674 F.3d at 1027-28. The majority, on the other hand, recognizes the problem, but then sidesteps it, claiming that the state failed to raise it and that dealing with it would require us to speculate. I disagree. There's no need to speculate about something so self-evident. And it's hard to fault the state for failing to dwell on the obvious.

In sum, I don't buy the timing or recusal arguments. And without them, there's nothing that prevents us from declaring that these three rules are the least restrictive means at Arizona's disposal for furthering their compelling interest in maintaining judicial impartiality and its appearance. Simply affixing the label of strict scrutiny and then declaring that unspecified less restrictive means are required gives no guidance as to what rules pass constitutional muster. And it encourages an elective free-for-all that undermines respect for the third branch of government. Because my colleagues disagree, I respectfully dissent.