Republican Party v. Suzanne White ( 2001 )

  •                    United States Court of Appeals
                             FOR THE EIGHTH CIRCUIT
                                     No. 99-4021
    Republican Party of Minnesota, an        *
    association; Indian Asian American       *
    Republicans of Minnesota, an             *
    association; Republican Seniors, an      *
    association; Young Republican League     *
    of Minnesota, a Minnesota nonprofit      *
    corporation; Minnesota College           *
    Republicans, an association,             *
           Plaintiffs - Appellants,          *
                                             * Appeals from the United States
    Gregory F. Wersal, individually,         * District Court for the
                                             * District of Minnesota.
           Plaintiff,                        *
    Cheryl L. Wersal, individually; Mark E. *
    Wersal, individually; Corwin C. Hulbert, *
    individually,                            *
           Plaintiffs - Appellants,          *
    Campaign for Justice, an association,    *
           Plaintiff,                        *
    Minnesota African American Republic *
    Council, an association,                 *
           Plaintiff - Appellant,            *
    Muslim Republicans, an association;       *
    Michael Maxim, individually; Kevin J.     *
    Kolosky, individually,                    *
          Plaintiffs,                         *
          v.                                  *
    Verna Kelly, in her capacity as           *
    Chairperson of the Minnesota Board of     *
    Judicial Standards, or her successor,     *
          Defendant,                          *
    Barry M. Lazarus, in his capacity as      *
    Chairperson of the Minnesota Board of     *
    Judicial Standards, or his successor;     *
    Edward J. Cleary, in his capacity as      *
    director of the Minnesota Office of       *
    Lawyers Professional Responsibility, or   *
    his successor; Charles E. Lundberg,       *
    in his capacity as Chair of the           *
    Minnesota Lawyers Professional            *
    Responsibility Board, or his              *
    successor,                                *
          Defendants - Appellees,             *
    Minnesota Civil Liberties Union,          *
          Amicus on Behalf of Appellants,     *
    The Minnesota State Bar Association,      *
          Amicus on Behalf of Appellee.       *
                                     No. 99-4025
    Republican Party of Minnesota, an     *
    association; Indian Asian American    *
    Republicans of Minnesota, as          *
    association; Republican Seniors, an   *
    association; Young Republican League  *
    of Minnesota, a Minnesota nonprofit   *
    corporation; Minnesota College        *
    Republicans, an association; Minnesota*
    African American Republic Council, an *
    association; Cheryl L. Wersal,        *
    individually; Mark E. Wersal,         *
    individually; Corwin C. Hulbert,      *
    individually; Gregory F. Wersal,      *
    individually; Campaign for Justice, an*
    association; Muslim Republicans, an   *
    association,                          *
          Plaintiffs,                     *
    Michael Maxim, individually,          *
          Plaintiff - Appellant,          *
    Kevin J. Kolosky, individually,       *
          Plaintiff,                      *
          v.                              *
    Verna Kelly, in her capacity as       *
    Chairperson of the Minnesota Board of *
    Judicial Standards, or her successor, *
           Defendant,                         *
    Barry M. Lazarus, in his capacity as      *
    Chairperson of the Minnesota Board of *
    Judicial Standards, or his successor;     *
    Edward J. Cleary, in his capacity as      *
    director of the Minnesota Office of       *
    Lawyers Professional Responsibility, or *
    his successor; Edward J. Cleary, in his *
    capacity as Director of the Minnesota     *
    Office of Lawyers Professional            *
    Responsibility, or his successor;         *
    Charles E. Lundberg, in his capacity as *
    Chair of the Minnesota Lawyers            *
    Professional Responsibility Board, or his*
    successor,                                *
           Defendants - Appellees,            *
    The Minnesota State Bar Association, *
           Amicus on Behalf of Appellee.      *
                                       No. 99-4029
    Republican Party of Minnesota, an        *
    association; Indian Asian American       *
    Republicans of Minnesota, an             *
    association; Republican Seniors, an      *
    association; Young Republican League     *
    of Minnesota, a Minnesota nonprofit      *
    corporation; Minnesota College           *
    Republicans, an association,             *
           Plaintiffs,                       *
    Gregory F. Wersal, individually,         *
           Plaintiff - Appellant,            *
    Cheryl L. Wersal, individually; Mark E. *
    Wersal, individually; Corwin C. Hulbert,*
    individually;                            *
           Plaintiffs,                       *
    Campaign for Justice, an association;    *
           Plaintiff - Appellant,            *
    Minnesota African American Republic *
    Council, an association; Muslim          *
    Republicans, an association; Michael     *
    Maxim, individually;                     *
           Plaintiffs,                       *
    Kevin J. Kolosky, individually,          *
           Plaintiff - Appellant,            *
           v.                                *
    Verna Kelly, in her capacity as          *
    Chairperson of the Minnesota Board of *
    Judicial Standards, or her successor;    *
           Defendant,                        *
    Barry M. Lazarus, in his capacity as     *
    Chairperson of the Minnesota Board of*
    Judicial Standards, or his successor;*
    Edward J. Cleary, in his capacity as *
    director of the Minnesota Office of  *
    Lawyers Professional Responsibility, or
    his successor; Charles E. Lundberg, in
    his capacity as Chair of the Minnesota
    Lawyers Professional Responsibility  *
    Board, or his successor,             *
         Defendants - Appellees,         *
    The Minnesota State Bar Association, *
         Amicus on Behalf of Appellants. *
                                 Submitted: May 10, 2000
                                      Filed: April 30, 2001
    Before McMILLIAN, JOHN R. GIBSON, and BEAM, Circuit Judges.
    JOHN R. GIBSON, Circuit Judge.
           The issue before us in this appeal is whether Canon 5 of the Minnesota Code of
    Judicial Conduct, a rule promulgated by the Supreme Court of Minnesota to deal with
    political activity deemed inappropriate to judicial office, violates the First and
    Fourteenth Amendments of the United States Constitution. Canon 5 restricts
    candidates for judicial office from attending and speaking at partisan political
    gatherings; identifying their membership in a political party; seeking, accepting, or
    using a political party endorsement; announcing their views on disputed legal and
    political issues; personally soliciting campaign contributions; or authorizing or
    knowingly permitting others to do these things on the candidates' behalf. The district
    court1 held that Canon 5's provisions, except for the restriction on candidates
    announcing their views on disputed legal and political issues, were narrowly tailored
    to serve a compelling state interest in maintaining the independence and impartiality of
    Minnesota's judiciary, did not offend equal protection, and were not impermissibly
    vague. Republican Party v. Kelly, 
    63 F. Supp. 2d 967
    , 974-83 (D. Minn. 1999). As
    to the "announce" clause, the court concluded that a broad reading of the clause would
    raise constitutional difficulties. The court construed the clause narrowly to uphold its
    constitutionality, however, predicting that the Minnesota Supreme Court would do the
    same. Id. at 983-86. We affirm.
           The Minnesota Constitution provides that judges "shall be elected by the voters
    from the area which they are to serve," and that their term of office shall be six years.
    Minn. Const. art. 6, § 7. In 1912, the Minnesota General Assembly designated judicial
    elections as nonpartisan, meaning that party affiliation is not listed when candidates file
    for office, nor does it appear on the ballot. Act of June 19, 1912, ch. 2, 1912 Minn.
    Laws Spec. Sess. 4-6.
           Ethical codes restricting the campaign conduct of judicial candidates have
    existed in Minnesota since at least 1950, when the Minnesota District Judges
    Association adopted by unanimous vote the ABA's Canons of Judicial Ethics (1924).
    In 1974, the Minnesota Supreme Court promulgated a code of judicial conduct that was
    based in large measure on the ABA's Model Code of Judicial Conduct (1972). Over
    the years, the Minnesota Supreme Court has revised its ethical rules, including those
          The Honorable Michael J. Davis, United States District Judge for the District
    of Minnesota.
    relating to a candidate's ability to attend and speak at political gatherings, to solicit
    campaign funds, and to discuss certain topics.
           Gregory Wersal, a Minneapolis-area attorney and longtime member of the
    Republican Party of Minnesota, ran unsuccessfully for the office of Associate Justice
    of the Minnesota Supreme Court in 1996 and 1998. The year Wersal launched his first
    campaign, the Minnesota Supreme Court revised its code of judicial conduct. The
    court renumbered and reorganized the canons and made several substantive changes,
    bringing the code largely in line with the 1990 version of the ABA's Model Code of
    Judicial Conduct.
           Wersal interpreted one revision to lift a twenty-two-year ban on judicial
    candidates speaking to partisan political gatherings. From 1974 until 1996, Canon 7
    of the Minnesota Code of Judicial Conduct barred candidates and judicial incumbents
    from speaking at partisan political gatherings, but allowed them to accept invitations
    to speak on their own behalf to other groups. Canon 7(A)(2) (1974) (a judicial
    candidate or incumbent "may accept invitations to attend and speak on his own behalf
    at other than partisan political gatherings"). It also prohibited judges from engaging in
    political activity except on behalf of measures to improve the law or the legal system.
    Canon 7(A)(4) (1974). In the 1996 revisions, the Minnesota Supreme Court
    reorganized and revised the subsections of Canon 7 and renumbered it as Canon 5.
    One subsection of the newly revised Canon 5 allowed candidates and judges to speak
    on their own behalf to gatherings generally, while another prohibited candidates and
    incumbents from attending political events. Canon (5)(A)(1)(a) & (d) (1996).
            Consistent with his reading of the canon, Wersal, his wife Cheryl, and members
    of his campaign committee spoke at Republican Party gatherings as part of Wersal's
    1996 bid for office. At these gatherings, they announced that Wersal was a member
    of the Republican Party and that he favored strict construction of the Constitution.
    They distributed campaign literature criticizing several Minnesota Supreme Court
    decisions on issues such as crime, welfare, and abortion as being "marked by their
    disregard for the Legislature and lack of common sense." In addition, the campaign
    committee sought unsuccessfully to obtain an endorsement by the Republican Party.2
           In May 1996, a delegate to a Republican district convention filed an ethical
    complaint against Wersal with the Office of Lawyers Professional Responsibility. The
    Office of Lawyers Professional Responsibility, under the direction of the Minnesota
    Lawyers Professional Responsibility Board (collectively, the Lawyers Board),
    investigates and prosecutes ethical violations of lawyer candidates for judicial office.
    The complainant questioned the propriety of Wersal's attendance at Republican
    gatherings, the campaign committee's solicitation of partisan support, and the campaign
    materials critical of Minnesota Supreme Court decisions. The Director of the Lawyers
    Board dismissed the complaint, concluding that no disciplinary action was warranted
    under Canon 5.
           In the Director's written determination, she first noted that it was unclear whether
    the Minnesota Supreme Court had intended to retain the ban on candidates speaking
    to political gatherings when it revised the code in 1996. Second, she pointed out that
    the 1996 version of Canon 5 restricted only candidates, and not their campaign
    committees, from soliciting publicly stated support. Finally, the Director expressed
            At least in recent times, it has not been the norm in Minnesota for judicial
    candidates to seek political party endorsements or for political parties to consider
    endorsing them. The parties submitted no evidence that other judicial candidate
    campaign committees sought or were considered for such endorsements. In a February
    10, 1998 letter, Wersal wrote a newspaper editor that "the Republican Party had never
    endorsed a judicial candidate before," and a June 20, 1998 Minnesota newspaper
    article characterized the Party's deliberation over whether to endorse Wersal as "a
    historic break with tradition." Party endorsements may not have been uncommon early
    in the twentieth century. See Moon v. Halverson, 
    288 N.W. 579
    , 581-82 (Minn. 1939)
    (Loring, J., concurring) (commenting on recent allegations of party treason directed at
    judges who issued decisions contrary to endorsing party's interests).
    doubts about the applicability of the announce clause to Wersal's campaign statements.
    Citing several decisions from other jurisdictions in which similar language was either
    struck down or interpreted narrowly, she also questioned whether the clause was
           Ethical complaints are deemed confidential in Minnesota, but notification of the
    Director's initial determination whether a complaint should be dismissed summarily or
    investigated further is provided to the complainant and to the respondent lawyer.
    Sometime after Wersal received this notification, he withdrew his candidacy for the
    1996 race, fearing that further ethical complaints would jeopardize his ability to
    practice law. In January of the following year, Wersal announced his candidacy for a
    supreme court seat opening up in 1998, and he and his campaign committee began
    campaigning as they had in the 1996 race, including the campaign committee's pursuing
    the Republican Party endorsement.
           The Minnesota Board on Judicial Standards (the Judicial Board), the body
    charged with enforcing ethical codes against judges, petitioned the Minnesota Supreme
    Court in September 1997 to amend Canon 5, primarily to clarify the nonpartisan nature
    of judicial elections. The Judicial Board, which had become aware that Wersal's
    campaign committee was soliciting a political party endorsement, urged the court to
    add language that would limit the ability of candidates to identify themselves as
    members of a political organization and that would prevent campaign committees from
    seeking endorsements from political organizations. The Judicial Board also
    recommended that the court clarify that judicial candidates could not speak to political
    gatherings. Following a hearing, the court adopted each of these recommendations.
    Order Amending Canon 5 of the Code of Judicial Conduct, No. C7-81-300 (Dec. 23,
    1997). The changes took effect January 1, 1998.
          In February 1998, Wersal sought an advisory opinion from the Lawyers Board.
    He asked whether the Board would prosecute him for ethical violations if he spoke at
    political party gatherings or sought a Republican Party endorsement. He also inquired
    whether the Board would enforce the provision of Canon 5 restricting candidates from
    announcing their views on disputed issues. The Director answered by stating that
    Wersal would be subject to discipline regarding the first two actions he proposed, but
    that the Board could not advise him on the latter question since Wersal had not
    provided the Board with information about any particular statements he wished to make
    that might have been a view on a disputed legal or political issue. The Director also
    stated that the Board continued to have "significant doubts as to whether or not [the
    announce clause] would survive a facial challenge to its constitutionality" and that it
    would not enforce the provision unless the speech at issue violated other portions of the
    judicial ethics code.
           A few days after he received this advisory opinion, Wersal filed this complaint
    under 42 U.S.C. § 1983 (1994), seeking declaratory and injunctive relief from the
    provisions of Canon 5. Joined as plaintiffs were the Republican Party of Minnesota
    (the Party), its affiliated organizations,3 and several other individuals and organizations
    interested in Wersal's candidacy.4 The complaint alleged that Canon 5 violated the free
            These affiliated organizations included the Indian Asian American Republicans,
    the Republican Seniors, the Young Republicans League of Minnesota, and the
    Minnesota College Republicans. The Minnesota African American Republican Council
    and the Muslim Republicans, also affiliated organizations, were later added as
            Other plaintiffs included Wersal's campaign committee, and Party members
    Cheryl Wersal, Mark Wersal, and Corwin Hulbert. Michael Maxim, another member
    of the Minnesota Republican Party, and Kevin Kolosky, also a candidate for judicial
    office in 1996 and 1998, later joined as plaintiffs. Because the Party, its affiliated
    organizations, and its members Cheryl Wersal, Mark Wersal, and Corwin Hulbert have
    filed a single brief, we refer only to the Party when addressing their arguments.
    Likewise, we refer only to Gregory Wersal when addressing his contentions, those of
    Kolosky, and the campaign committee, who also have submitted a single brief.
    speech and association guarantees of the First Amendment and the Equal Protection
    Clause of the Fourteenth Amendment.
           Wersal and the other plaintiffs moved for a temporary restraining order and/or
    preliminary injunction to enjoin the Lawyers Board and Judicial Board from enforcing
    Canon 5 so that Wersal would be free to participate in Republican precinct caucuses
    scheduled for early March 1998. The district court denied the motion,5 and we
    subsequently affirmed. Republican Party v. Kelly, 
    966 F. Supp. 875
     (D. Minn. 1998),
    aff'd, No. 98-1625 (8th Cir. Nov. 2, 1998).
          While appeal of the district court's order was pending, Wersal continued his
    campaign, but was forced to cancel the numerous speaking engagements he had
    scheduled at the precinct caucuses and other Republican events. He wrote at least five
    persons who he knew intended to advocate his candidacy at these events and asked
    them not to do so because of his concern that someone might infer that he had
    "knowingly permitted" them to do so in violation of Canon 5. On advice of counsel,
    he declined to answer questions posed by members of the press and public when he
    thought answering might impermissibly disclose his views on disputed issues. His wife
    Cheryl and his brother Mark, a member of his campaign committee, refrained from
    taking part in certain campaign activities on Wersal's behalf for fear that their action
    would subject Wersal to discipline.
         At the state Republican convention held in June, Michael Maxim, a delegate
    who was later joined as a plaintiff in this case, moved that the Party endorse Wersal.
            The district court found that the plaintiffs established a likelihood of success on
    their claim that the announce clause was unconstitutional, but concluded that the State's
    interest in maintaining a nonpartisan judicial election weighed against enjoining
    enforcement of all of Canon 5. Republican Party v. Kelly, 
    966 F. Supp. 875
    , 879-80
    (D. Minn. 1998).
    After considerable debate, the motion failed by a vote of 344 to 390. No other
    candidates for judicial office were nominated for endorsement. Because of the
    Republican Party's policy of not supporting candidates it had not formally endorsed, the
    Party did nothing to further Wersal's bid for office. His campaign ended when he came
    in third in the primary.
           The parties filed cross motions for summary judgment. The district court found
    in favor of defendants on all claims. Republican Party v. Kelly, 
    63 F. Supp. 2d 967
    Minn. 1999). The court concluded that the State had compelling interests in
    maintaining the actual and apparent integrity and independence of the judiciary and that
    the restrictions on candidates' political activity and fund solicitation were narrowly
    tailored to serve those interests. Id. at 974-80. It also upheld the provisions against
    vagueness and equal protection challenges. Id. at 980-82. In its analysis of the
    announce clause, the court determined that the critical issue was whether the provision
    was narrowly tailored to serve the State's interest in maintaining the integrity and
    independence of the judiciary. The district court construed the clause to reach only the
    discussion of issues likely to come before the court, having considered that the Judicial
    Board had argued for a narrow interpretation of the clause and that the Minnesota
    Supreme Court, when possible, construes laws to prohibit their application to
    constitutionally protected expression. Id. at 985-86. The court then concluded that the
    provision did not offend the First Amendment. Id. at 986.
          Wersal and those joined with him appeal,6 challenging the determinations made
    by the district court.
           We review the district court's grant of summary judgment de novo, applying the
    same standard the district court applied to the motion. Essco Geometric v. Harvard
    46 F.3d 718
    , 729 (8th Cir. 1995). We will affirm if we conclude there are no
    genuine issues of material fact and the Lawyers and Judicial Boards are entitled to
    judgment as a matter of law. Fed. R. Civ. P. 56(c); Celotex Corp. v. Catrett, 
    477 U.S. 317
    , 322-23 (1986).
            For the first time on appeal, the Lawyers and Judicial Boards contend that the
    plaintiffs who are not candidates for judicial office and cannot be sanctioned by the
    Boards lack standing to challenge Canon 5. Citing Laird v. Tatum, 
    408 U.S. 1
    the Boards argue that the plaintiffs' voluntary decisions not to advocate Wersal's
    candidacy do not constitute legally cognizable injuries. However, the plaintiffs have
    alleged more than the subjective "chilling" of their free speech rights held insufficient
    by the Supreme Court in Tatum, 408 U.S. at 13-14. Wersal and the other plaintiffs
    have stated they were unable to associate with each other at political party gatherings
    and alleged they would have done so absent the ethical restrictions. Deprivation of the
    right to associate with others politically is a cognizable "injury in fact." See Lerman
    v. Board of Elections, 
    232 F.3d 135
    , 143 (2d Cir. 2000), petition for cert. filed (U.S.
    Feb. 28, 2001) (No. 00-1360); Krislov v. Rednour, 
    226 F.3d 851
    , 858 (7th Cir. 2000),
    cert. denied, 
    121 S. Ct. 1085
     (2001) (No. 00-923). Because the plaintiffs' associational
    injuries are fairly traceable to Canon 5 and the injunctive relief they request will redress
    them, the plaintiffs have standing to assert their claims. See Lujan v. Defenders of
    504 U.S. 555
    , 560-61 (1992) (Article III standing requires plaintiffs allege
    "injury in fact" fairly traceable to defendant's conduct likely to be redressed by
    favorable court decision); National Solid Waste Management Ass'n v. Williams, 
    146 F.3d 595
    , 598 (8th Cir. 1998).
           The plaintiffs contend that Canon 5 contravenes their First Amendment rights
    to freedom of speech and of association, made applicable to the States by the Due
    Process Clause of the Fourteenth Amendment. Tashjian v. Republican Party, 
    479 U.S. 208
    , 214 (1986).
           Freedom of speech reaches its high-water mark in the context of political
    expression. Debate about the qualification of candidates for public office is at the core
    of our First Amendment freedoms, Eu v. San Francisco County Democratic Cent.
    489 U.S. 214
    , 222-23 (1989), valuable not only as a personal liberty, but also
    because of the role it plays in the proper functioning of our entire democratic form of
    government. Burson v. Freeman, 
    504 U.S. 191
    , 196 (1992); Brown v. Hartlage, 
    456 U.S. 45
    , 52-53 (1982); Buckley v. Valeo, 
    424 U.S. 1
    , 14-15 (1976) (per curiam). The
    closely related right of association is also particularly acute in the context of elections.
    Eu, 489 U.S. at 224-25 (ban on endorsement in primary election acts at "crucial
    juncture" to prevent parties from promoting candidates who can translate shared ideas
    into action); Buckley, 424 U.S. at 15.
           But even among restrictions that touch upon the core First Amendment rights to
    political speech and association, some restrictions are more burdensome than others,
    depending on the kind of activity that is restrained and on the nature of the restraint.
    See Timmons v. Twin Cities Area New Party, 
    520 U.S. 351
    , 358 (1997) (standard of
    review depends on character and magnitude of burden on associational rights). In
    Buckley v. Valeo, both campaign contribution limits and expenditure limits affected
    core First Amendment expression and association interests, 424 U.S. at 14, yet the
    contribution limits placed a lesser burden on those interests than the expenditure limits
    did, id. at 19-23. Consequently, contribution limits have been reviewed under a less
    demanding standard. Nixon v. Shrink Mo. Gov’t PAC, 
    120 S. Ct. 897
    , 903-04 (2000).
           In this case, the restriction applies to conduct of candidates for judicial office.7
    There are important differences between judicial office, on the one hand, and legislative
    or executive office, on the other, that affect the nature of the candidate's interest in
    certain kinds of policy debate.
           The functioning of the judicial system differs markedly from those of the
           executive and legislative. In those areas, the public has the right to know
           the details of the programs that candidates propose to enact into law and
           administer. Pledges to follow certain paths are not only expected, but are
           desirable so that voters may make a choice between proposed agendas
           that affect the public. By contrast, the judicial system is based on the
           concept of individualized decisions on challenged conduct and
           interpretations of law enacted by the other branches of government.
    Stretton v. Disciplinary Bd., 
    944 F.2d 137
    , 142 (3d Cir. 1991). Accord Buckley v.
    Illinois Judicial Inquiry Bd., 
    997 F.2d 224
    , 228 (7th Cir. 1993) ("Judges remain
    different from legislators and executive officials . . . in ways that bear on the strength
    of the state's interest in restricting their freedom of speech."); In re Chmura, 
    608 N.W.2d 31
    , 39-40 (Mich.) ("[T]he differences between judges and other government
    officials bear on the strength of the state's interest in restricting political speech."), cert.
    121 S. Ct. 77
     (2000). Whereas affiliation with a partisan program is thus at the
    heart of executive and legislative campaigns, a State may conclude that it has no role
    in judicial campaigns because of the neutral, decision-making nature of the judicial
    function. "Because the judicial office is different in key respects from other offices, the
    state may regulate its judges with the differences in mind . . . . [The judicial candidate]
    cannot, consistent with the proper exercise of his judicial powers, bind himself to
    decide particular cases in order to achieve a given programmatic result." Morial v.
    Judiciary Comm'n, 
    565 F.2d 295
    , 305 (5th Cir. 1977). The judicial candidate simply
    does not have a First Amendment right to promise to abuse his office. See Brown, 456
           We discuss the claim that Canon 5 also reaches the conduct of persons who are
    not judicial candidates infra at 36-38.
    U.S. at 54-56 (some kinds of campaign promises "may be declared illegal without
    constitutional difficulty"); In re Kaiser, 
    759 P.2d 392
    , 400 (Wash. 1988) (holding
    statements of party affiliation do not refer to subject relevant to judicial qualification
    and therefore are not protected by First Amendment). Recognizing that the judiciary
    has a different job to do, Minnesota has provided that its judicial offices are
    nonpartisan. Minn. Stat. § 204B.06, subd. 6 (1998); Peterson v. Stafford, 
    490 N.W.2d 418
    , 420 (Minn. 1992). Thus, restrictions on Minnesota judicial candidates' speech are
    entirely different from limitations on the speech of candidates for partisan office, such
    as those in Eu, 489 U.S. at 217 (striking statute making it misdemeanor for primary
    candidates to claim endorsement by party).
           Another important aspect of the restraint in this case is that it does not
    discriminate in favor of one viewpoint or other. Minnesota's restraint of judicial
    candidates' First Amendment rights is a straightforward restriction of expression, rather
    than being incidental to some other type of regulation, such as a content-neutral time,
    place, or manner restriction, and in one sense, it is not content-neutral because it
    restricts speech on the basis of subject-matter. In Burson v. Freeman, 
    504 U.S. 191
    (1992), the Supreme Court considered a state law establishing a 100-foot "campaign-
    free" zone around polling places. Justice Blackmun rejected the assertion that the law
    was content-neutral: "Whether individuals may exercise their free speech rights near
    polling places depends entirely on whether their speech is related to a political
    campaign." Id. at 197. Similarly, in this case, the bans on accepting political party
    endorsements, attending political party gatherings, identifying oneself as a party
    member, announcing views on disputed legal and political issues, and soliciting funds
    depend entirely on the subject matter of the speech. See Hill v. Colorado, 
    120 S. Ct. 2480
    , 2500 (2000) (Souter, J., concurring) (government held to very exacting and rarely
    satisfied standard when it disfavors discussion of particular subjects or articulation of
    particular viewpoints).
           However, the restriction in this case avoids discriminating against a particular
    viewpoint, which is the most serious threat to First Amendment rights. See R.A.V. v.
    City of St. Paul, 
    505 U.S. 377
    , 387-90 (1992) (treating "content discrimination" not as
    an evil in itself, but as evidence that "official suppression of ideas is afoot"). In United
    States Civil Service Commission v. National Association of Letter Carriers, 
    413 U.S. 548
     (1973), the Supreme Court entertained a challenge to the Hatch Act, which
    prohibited federal employees from taking an active part in political management or
    political campaigns. The Court prefaced its analysis with the observation that the
    challenged restrictions did not discriminate on the basis of viewpoint:
           The restrictions so far imposed on federal employees are not aimed at
           particular parties, groups, or points of view, but apply equally to all
           partisan activities of the type described. They discriminate against no
           racial, ethnic, or religious minorities. Nor do they seek to control political
           opinions or beliefs, or to interfere with or influence anyone's vote at the
    Id. at 564. Accord Morial, 565 F.2d at 301-02 (requirement that sitting judge resign
    before running for a non-judicial office did not penalize belief in any particular idea and
    therefore did not touch core First Amendment values). Thus, although Canon 5 does
    burden First Amendment rights, the burden is less onerous than it might otherwise be
    because Canon 5 does not discriminate on the basis of viewpoint and because it
    governs only judicial elections.
           Because the restrictions on the employees' rights in Letter Carriers were similar
    to those of Canon 5, the defendants suggest we look to Letter Carriers and cases
    following it, which judged restrictions on speech of government employees under a
    balancing test less rigorous than strict First Amendment scrutiny. See Letter Carriers,
    413 U.S. at 564. Wersal argues that we cannot apply the Letter Carriers balancing test
    to a restriction on his speech because he is not a government employee. But even
    though he is not presently a government employee, Canon 5 applies to him only insofar
    as he seeks to become one. The State can reasonably conclude that Wersal's actions
    as a candidate could affect his actions as a judge if he is elected, see Buckley v. Valeo,
    424 U.S. at 26-27 (discussing danger of corruption from contributions to "current and
    potential office holders"), and so the reasoning of the balancing test cases fairly applies
    to him. However, the restrictions in this case are different from the Hatch Act
    provisions challenged in Letter Carriers because, while the Hatch Act restrained
    political activity of government employees, Canon 5 restrains the activity of candidates
    engaged in an election contest. The burden on the plaintiff in either case may be
    comparable, but the public's interest in free speech is greater where the person subject
    to restrictions is a candidate for public office, about whom the public is obliged to
    inform itself. Therefore, we will invoke strict scrutiny and examine the restrictions at
    issue to determine whether they are narrowly tailored to serve a compelling state
    interest.8 See Stretton, 944 F.2d at 141-42; see generally California Democratic Party
    v. Jones, 
    120 S. Ct. 2402
    , 2412 (2000); Burson, 504 U.S. at 198; Brown, 456 U.S. at
            The governmental interests put forth to justify Canon 5 are undeniably
    compelling. The Boards contend that the restrictions are necessary to guarantee the
    independence of the Minnesota judiciary, which in turn is crucial to preserve the justice
    of its courts of law and its citizens' faith in those courts. There is simply no question
    but that a judge's ability to apply the law neutrally is a compelling governmental interest
    of the highest order.
           The effects of Canon 5 on speech and associational rights, respectively, are so
    closely intertwined that, having decided to apply strict scrutiny because of the effects
    on speech, we need not consider whether the effects on associational rights alone
    would have called for a less rigorous standard.
           Judges should decide cases in accordance with law rather than with any
           express or implied commitments that they may have made to their
           campaign supporters or to others . . . . Justice under law is as
           fundamental a part of the Western political tradition as democratic self-
           government and is historically more deeply rooted, having been
           essentially uncontested within the mainstream of the tradition since at
           least Cicero's time . . . .
    Buckley v. Illinois Judicial Inquiry Bd., 
    997 F.2d 224
    , 227 (7th Cir. 1993). Accord
    Stretton v. Disciplinary Bd., 
    944 F.2d 137
    , 142 (3d Cir. 1991) ("There can be no
    question . . . that a state has a compelling interest in the integrity of its judiciary . . . .
    If judicial candidates during a campaign prejudge cases that later come before them,
    the concept of impartial justice becomes a mockery."); Morial v. Judiciary Comm'n,
    565 F.2d 295
    , 302 (5th Cir. 1977) ("Ours is an era in which members of the judiciary
    often are called upon to adjudicate cases squarely presenting hotly contested social or
    political issues. The state's interest in ensuring that judges be and appear to be neither
    antagonistic nor beholden to any interest, party, or person is entitled to the greatest
           The plaintiffs contend that Minnesota has no interest in the independence of its
    judiciary because it has chosen to make its judges stand for election. Wersal says:
    "Contrary to popular understanding—even among lawyers—Minnesota does not have
    an 'independent' judiciary. Minnesota has an elected judiciary that its founding citizens
    specifically adopted in the Minnesota Constitution to insure its judiciary was not
    'independent' from the voters." Wersal argues that when the citizens at the Democratic9
    Minnesota constitutional convention in 1857 debated whether the judiciary should be
           In Minnesota there were two constitutional conventions, one Democratic and
    one Republican. The two drafts resulting from those conventions were consolidated
    and then submitted for ratification. Peterson v. Stafford, 
    490 N.W.2d 418
    , 420 n.10
    (Minn. 1992).
    appointed or elected, they considered the then-recent Dred Scott10 decision. They
    decided to choose judges by election and to give them finite terms, rather than life
    tenure, in order to retain the ability to oust unsatisfactory judges.
           The Minnesota Supreme Court has considered the broad history of Minnesota
    judicial elections and has concluded that the State has historically pursued the ideal of
    an independent judiciary. In Peterson v. Stafford, 
    490 N.W.2d 418
     (Minn. 1992), the
    Minnesota Supreme Court considered an equal protection challenge to Minn. Stat. §
    204B.36, subds. 4 and 5 (1990), which provided that an incumbent judge should be
    designated as such on the ballot. In explaining the reason for the designation, the court
    identified the underpinnings of the Minnesota selection process as the search for an
    independent judiciary:
                 The methods by which the federal system and other states initially
          select and then elect or retain judges are varied, yet the explicit or implicit
          goal of the constitutional provisions and enabling legislation is the same:
          to create and maintain an independent judiciary as free from political,
          economic and social pressure as possible so judges can decide cases
          without those influences[11] . . . . While the framers of our state
          constitution have developed a system of selection and election quite
          different from that federal scheme, they too designed a plan to recognize
          the uniqueness and independence of the state judiciary.
    Id. at 420. Peterson traced the history of the Minnesota judicial selection process
    beginning with the adoption of the Minnesota Constitution in 1857, which provided that
    judges of the Supreme Court should be elected for seven-year terms, Minn. Const. art.
               Dred Scott v. Sandford, 60 U.S. (19 How.) 393 (1856).
           The Minnesota Supreme Court has thus with great precision and specificity
    answered the dissent's rhetorical question, "Independent from what?" Infra at 59.
    6, § 3 (1857), so that judicial elections did not coincide with elections for other offices.
    490 N.W.2d at 420. Moreover, the 1857 Constitution provided that judges must be
    "learned in the law." Minn. Const. art. 6, § 5 (1857). "Implicit in this requirement is
    recognition that those elected as judges will be subject to the restrictive canons of
    conduct governing the profession of law." 490 N.W.2d at 422. The term length was
    reduced to six years in 1883, which made more apparent the "difficulties associated
    with partisan judicial elections." Id. at 420. In 1912, the legislature enacted
    nonpartisan ballot legislation, Act of June 19, 1912, ch. 2, 1912 Minn. Laws, Spec.
    Sess. 4-6, which made the State's judicial offices nonpartisan. See id. at § 182. The
    court characterized this reform as an attempt to ensure the "judicial impartiality
    required to decide cases free from political maneuvering." 490 N.W.2d at 422.
            Further, the Minnesota Constitution provided that whenever there was a judicial
    vacancy, the governor should appoint the successor, who would serve until his
    successor was elected.12 Minn. Const. art. 6, § 10 (1857). In 1972, at the time when
    the State was studying the possible advantages of a merit selection plan, the time frame
    for this successor election was extended from the next general election occurring more
    than thirty days after the vacancy to the next general election occurring more than one
    year following the vacancy, in order to allow the electorate enough time to assess the
    new judge's competence. Minn. Const. art. 6, § 8 (amended 1972); see Peterson, 490
    N.W.2d at 422-23.
         In the context of this history, Peterson held that the ballot designation of the
    incumbent was rational:
            Since 1991 a judicial selection commission, which is composed of members
    appointed by the governor and by the justices of the supreme court, evaluates the merits
    of applicants for mid-term vacancies and recommends to the governor three to five
    nominees for each open position. Minn. Stat. §§ 480B.01, subds. 2, 8-11 (1998).
                 It seems clear that Minnesota has adopted its own middle-of-the-
          road approach to judicial selection. The open election process has been
          retained, but with a quasi-retention feature which simply informs the voter
          who the incumbent candidate is and who the challenger is. This
          arrangement acts as a check on the gubernatorial appointment process by
          keeping the ultimate choice with the voters while, at the same time,
          recognizing the unique independent nature of the judicial function.
    490 N.W.2d at 425. Thus, the Minnesota Supreme Court has considered in depth the
    history and structure of Minnesota judicial elections procedures and has concluded that
    they are designed in large part to protect the independence of the State’s judiciary.13
           Other courts considering the same question have held that the decision to elect
    judges cannot be regarded as abandonment of a State’s interest in an independent
    judiciary. See Buckley, 997 F.2d at 227 ("Judges remain different from legislators and
             The dissent concludes that Peterson's use of the phrase "independent 'as
    possible,'" infra at 67, must be viewed, not in light of Peterson's detailed historical
    review, but rather in light of the dissent's independent review of Minnesota history. See
    infra at 59-66. The deference we must give to state court opinions on matters of state
    law does not allow us to supplant Peterson's considered views in favor of our own de
    novo interpretation, nor may the dissent. The dissent also states that Canon 5
    "subverts" Minnesota's policy, infra at 66, and seemingly presumes that Canon 5 is not
    part of state law because it emanates from the state supreme court, rather than from the
    state legislature. No one, including the dissent, has questioned the authority of the
    Minnesota Supreme Court to promulgate ethical rules for incumbent judges and judicial
    candidates. The same federalism principles which require us to defer to state courts'
    interpretations of state law and to recognize that state laws embody the will of a State
    also dictate that we recognize Canon 5 as a regulation that expresses Minnesota's will
    as a sovereign entity. See Bush v. Gore, 
    121 S. Ct. 525
    , 534 (2000) (Rehnquist, C.J.,
    concurring) ("in ordinary cases, the distribution of powers among the branches of a
    State's government raises no questions of federal constitutional law, subject to the
    requirement that the government be republican in character").
    executive officials, even when all are elected . . . ."); Stretton, 944 F.2d at 142 ("The
    fact that a state chooses to select its judges by popular election . . . does not signify the
    abandonment of the ideal of an impartial judiciary carrying out its duties fairly and
    thoroughly."); In re Chmura, 
    608 N.W.2d 31
    , 39-40 (Mich.) ("By providing for the
    election of judges, the people of Michigan have not transformed judges into legislators
    or executives . . . ."), cert. denied, 
    121 S. Ct. 77
            The governmental interest in an independent and impartial judiciary is matched
    by its equally important interest in preserving public confidence in that independence
    and impartiality. See Cox v. Louisiana, 
    379 U.S. 559
    , 565 (1965) ("A State may also
    properly protect the judicial process from being misjudged in the minds of the public.");
    Suster v. Marshall, 
    149 F.3d 523
    , 532 (6th Cir. 1988) (State's interest in preventing
    judicial corruption or appearance of corruption compelling); cf. Reeder v. Kansas City
    Bd. of Police Comm'rs, 
    733 F.2d 543
    , 547 (8th Cir. 1984) ("It is proper for a state to
    insist that the police be, and appear to be, above reproach, like Caesar's wife."). Letter
    Carriers stressed that the appearance of corruption resulting from partisan activities of
    government employees could affect the entire government: "[I]t is not only important
    that the Government and its employees in fact avoid practicing political justice, but it
    is also critical that they appear to the public to be avoiding it, if confidence in the
    system of representative Government is not to be eroded to a disastrous extent." 413
    U.S. at 565.
           Finally, a State has an interest in protecting its judges from pressure to
    participate in partisan activities, if it reasonably concludes that this protection is
    necessary to retain a high caliber of judges. "There could hardly be a higher
    governmental interest than a State's interest in the quality of its judiciary." Landmark
    Communications, Inc. v. Virginia, 
    435 U.S. 829
    , 848 (1978) (Stewart, J., concurring
    in judgment). Letter Carriers applied an analogous principle when it justified the Hatch
    Act partly on the ground that it meant to protect government employees from political
    pressure from their superiors. 413 U.S. at 566-67.
          We therefore conclude that the Lawyers and Judicial Boards have shown
    compelling governmental interests14 to justify Canon 5, and we turn to the necessity and
    narrow tailoring of the various restrictions found within it.
          We first address Wersal's challenge to the three provisions of Canon 5(A) that
    prohibit him from attending political party gatherings; from seeking, accepting, or using
    endorsements from a party; and from identifying himself as a member of the party,
             In arguing that the state interests we have identified are not compelling, the
    dissent relies on Landmark Communications, Inc. v. Virginia, 
    435 U.S. 829
     (1978), in
    which the State sought to govern not the conduct of judges, but of the press in reporting
    on judicial disciplinary proceedings. The dissent's discussion of Landmark
    Communications suggests that the compelling interest we rely on is protection of judges
    from criticism. Infra at 79. We rely on the State's interest in its judges' integrity and
    independence, which requires keeping judges from doing (or appearing to do) things
    that compromise their neutrality, rather than keeping others from talking or writing
    about them. Minnesota seeks not to "shield[] judges from published criticism," infra
    at 80 (quoting Bridges v. California, 
    314 U.S. 252
    , 270-71 (1941)), but to prevent them
    from deserving such criticism. Moreover, the Supreme Court in Landmark
    Communications did not reject the legitimacy of the State's proffered interests in the
    reputation of its judiciary and the integrity of its disciplinary proceedings, but held that
    the State had not shown that these interests would be "seriously undermined" unless the
    State could use criminal sanctions against the defendant newspaper. Id. at 841, 845.
    except as necessary to vote. Minn. Code of Jud. Conduct Canon 5(A)(1)(a) & (d)
           We must determine whether these restraints protect Minnesota's interests in the
    integrity and quality of its judiciary. United States v. Playboy Entertainment Group,
    120 S. Ct. 1878
    , 1891 (2000). The burden of proof rests with the State. Id. at
          A threshold question is what sort of evidence the Boards must provide
    substantiating the threat to the governmental interest. In Nixon v. Shrink Missouri
    Government PAC, 
    120 S. Ct. 897
     (2000), the Supreme Court declined to answer this
    question definitively, id. at 907, but gave this guidance: "The quantum of empirical
    evidence needed to satisfy heightened judicial scrutiny of legislative judgments will
    vary up or down with the novelty and plausibility of the justification raised." Id. at 906.
    The asserted danger to the governmental interest in that case, danger of corruption
            Canons 5(A)(1)(a) and (d) provide:
          Each justice of the supreme court and each court of appeals and district
          court judge is deemed to hold a separate nonpartisan office. MS 204B.06
          (1) Except as authorized in Section 5B(1), a judge or a candidate for
          election to judicial office shall not:
                 (a) act as a leader or hold any office in a political
                 organization; identify themselves as members of a political
                 organization, except as necessary to vote in an election.
                 (d) attend political gatherings; or seek, accept or use
                 endorsements from a political organization . . . .
    attendant on large campaign contributions, was neither novel nor implausible, so that
    the State's burden was amply satisfied by its production of a state senator's affidavit and
    newspaper articles showing cases of apparent corruption. Id. at 907-08. The Court
    also assigned some evidentiary value to the passage of a campaign contribution
    limitation by a large margin at a referendum, which demonstrated public concern about
    corruption.16 Id. at 908. The Court quoted City of Renton v. Playtime Theatres, Inc.,
    475 U.S. 41
    , 51-52 (1986), in which the city's evidence was said to be adequate, "so
    long as whatever evidence the city relies upon is reasonably believed to be relevant to
    the problem that the city addresses." Id. at 907 n.6. Shrink Missouri also noted the
    lack of evidence tending to disprove the State's theory. Id. at 908.
           The idea that judicial integrity is threatened by judges deploying political
    organizations in connection with campaigns for judicial office is neither novel nor
    implausible. In Letter Carriers, the Supreme Court recognized that partisanship of
    governmental officials created a risk of corruption that justified the restraint of those
    officials' partisan activities. Although the Hatch Act applied to employees of the
    executive branch, the Court's reasoning could as well have been written about judges
    and in fact applies with even greater urgency to them:
                 It seems fundamental in the first place that employees in the
          Executive Branch of the government, or those working for any of its
          agencies, should administer the law in accordance with the will of
          Congress, rather than in accordance with their own or the will of a
          political party. They are expected to enforce the law and execute the
          programs of the Government without bias or favoritism for or against any
          political party or group or the members thereof. A major thesis of the
             We held the voter initiative imposing contribution limits invalid under the First
    Amendment in Carver v. Nixon, 
    72 F.3d 633
    , 645 (8th Cir. 1995); consequently, its
    restrictions were not in force when Shrink Missouri was decided.
          Hatch Act is that to serve this great end of Government—the impartial
          execution of the laws—it is essential that federal employees, for example,
          not take formal positions in political parties, not undertake to play
          substantial roles in partisan political campaigns, and not run for office on
          partisan political tickets. Forbidding activities like these will reduce the
          hazards to fair and effective government.
    United States Civil Serv. Comm'n v. National Ass'n of Letter Carriers, 
    413 U.S. 548
    563-64 (1973). Accord LaMontagne v. St. Louis Dev. Corp., 
    172 F.3d 555
    , 557 (8th
    Cir. 1999) (city employees), cert. denied, 
    120 S. Ct. 176
     (1999); Reeder v. Kansas City
    Bd. of Police Comm'rs, 
    733 F.2d 533
    , 547 (8th Cir. 1984) (police officers); Otten v.
    655 F.2d 142
    , 144 (8th Cir. 1981) (same).
           The restriction here was adopted by the Minnesota Supreme Court, which has
    consistently over time perceived partisanship as posing a particular threat to judicial
    integrity. Canon 7(A) of Minnesota's 1974 Code of Judicial Conduct also restricted a
    judicial candidate's political activities broadly, specifically permitting a candidate for
    elective judicial office to appear at "other than partisan political gatherings during the
    year in which he is a candidate." Minn. Code of Judicial Conduct Canon 7(A)(2)
    (1974) (emphasis added). The comments of the advisory committee accompanying the
    1995 Minnesota Code of Judicial Conduct state: "Deference to the judgments and
    rulings of courts depends upon public confidence in the integrity and independence of
    judges. The integrity and independence of judges depends in turn upon their acting
    without fear or favor." Canon 1(A) cmt. In proposing to the court the 1997
    amendments to clarify the prohibitions on partisan conduct the Judicial Board stated:
    "The Judicial Standards Board members believe the proposed language changes will
    make it clear that judicial elections are to be nonpartisan. Minnesota has a long
    tradition of nonpartisan judicial elections, and these changes will assure a strong,
    independent judiciary."
           Long before the present canons were adopted, justices of the Minnesota Supreme
    Court had also expressed the idea that merely avoiding party designations on the ballot
    was insufficient to protect the Minnesota judiciary from the dangers of partisan
    involvement. In Moon v. Halverson, 
    288 N.W. 579
     (Minn. 1939), a candidate for
    registrar of deeds, an office designated by statute as nonpartisan, had solicited and
    received the endorsement of a political party. The court held that the Minnesota
    nonpartisan statute prescribed only that the candidate's party affiliation could not
    appear on the ballot and that the candidate could not declare his party when filing for
    office. The statute did not forbid party endorsement of candidates for nonpartisan
    office. 288 N.W. at 581. Justice Loring, joined by Justice Olson, concurred specially
    to add that this statute was not sufficient to protect the State's judiciary from the
    dangers of partisan pressure, which he felt had been at work:
          When candidates for such offices were placed on a non-partisan ballot it
          was, it seems to me, the purpose of the legislature to lift the judgeships
          above sordid political influence and to free the candidates from obligation
          to a political party so that if elected they might render judicial instead of
          partisan political decisions on matters where party programs, party
          interests or even prominent party leaders might be involved. The abuse
          and accusations of party treason which have been heaped upon some
          judges in the recent past because of decisions thought to be contrary to
          the interests of an indorsing party ought to be evidence enough of the
          impropriety of party indorsements and of their purpose to induce partisan
          political rather than impartial judicial decisions.
    Id. at 581-82.
           More recently, the court in Peterson v. Stafford, 
    490 N.W.2d 418
    , 425 (Minn.
    1992), stated that the nature of judicial office requires the "holder studiously to avoid
    partisan politics, refrain from all discussions of public issues and restrict one's
    membership and participation in organization to those primarily of a professional
           Additionally, the Boards have adduced evidence of the necessity of reform
    measures in the form of affidavits by a former governor, Arne H. Carlson, and a former
    Chief Justice of the Minnesota Supreme Court, A. M. (Sandy) Keith. Chief Justice
    Keith was a member of the Minnesota Supreme Court when the Court considered and
    approved the amendments to Canon 5. He testified that, based on his experience, if the
    Minnesota Code were changed to permit partisan activity, judges would be under
    pressure to "decide cases in ways that would impress the judge's supporters favorably,"
    and eventually, partisanship would damage the public's confidence in the judiciary.
    Governor Carlson testified: "For the public to read newspaper headlines that a political
    party has endorsed and will work to elect a particular candidate would greatly harm the
    public's confidence in the independence of the judiciary." Both Chief Justice Keith and
    Governor Carlson testified that they believed that allowing judicial elections to become
    partisan contests would discourage many qualified candidates from seeking election.
           The Boards have also produced a news account of the dangers partisan elections
    have posed in Texas, which included interviews with witnesses who believed that
    judges' decisions were affected by the need for political and financial support and that
    the public perceived "that justice is for sale in this state." Editorials from Texas
    newspapers reported a public perception that rulings were influenced by campaign
    obligations. The testimony at the 1997 hearing before the Minnesota Supreme Court
    on the proposal to amend Canon 5 was replete with references to the undesirable
    situation existing in other States with partisan judicial elections, particularly Texas.
    Amendment to Canon 5 of the Code of Judicial Conduct: Hearing Before the Minnesota
    Supreme Court, No. C7-81-300, at 32, 36, 50, 65-68 (1997).
           In contrast to the evidence amassed by the Boards, the plaintiffs have not
    adduced evidence tending to disprove the threat to the integrity or reputation of the
    judiciary from involvement with partisan politics. See Shrink Mo., 120 S. Ct. at 908.
    This is an issue where "a long history, a substantial consensus, and simple common
    sense," Burson, 504 U.S. at 211, combine to show that regulation is necessary to
    protect the institution of the judiciary from the dangers of partisanship and corruption.
    The record suffices to support the Minnesota Supreme Court's assessment that the
    threat of actual and apparent corruption is real and reform measures are necessary. Cf.
    Eu v. San Francisco Democratic Cent. Comm., 
    489 U.S. 214
    , 229 (1989) (striking ban
    on party endorsement of candidates in party primaries where there was no evidence that
    ban served purpose of preventing fraud and corruption). Additionally, there is record
    evidence showing that without protection from partisan pressures, the State will be less
    able to recruit judicial candidates of the highest caliber.
            The dissent contends that Canon 5 is not necessary because it does not achieve
    its stated end. Infra at 82-83. The argument is not precisely that Canon 5 does not
    accomplish its end as far as it goes, but that it is underinclusive—in other words, that
    the measure does not eliminate all the conduct posing the threat the government claims
    to be addressing because a candidate may attend meetings of and accept endorsements
    from politically active groups other than political parties. The dissent contends that in
    cases burdening a fundamental right "failure to solve the entire asserted problem is
    fatal." Infra at 83 n.44 (citing 44 Liquormart, Inc. v. Rhode Island, 
    517 U.S. 484
    (1996), and City of Cincinnati v. Discovery Network, Inc., 
    507 U.S. 410
    , 418 (1993)).
    Wersal makes a similar argument, relying on Carey v. Brown, 
    447 U.S. 455
    and R.A.V. v. City of St. Paul, 
    505 U.S. 377
     (1992). These cases do not stand for such
    a broad rule. Generally, underinclusiveness of a speech regulation is not an
    independent ground for invalidating the regulation, but rather points to the possibility
    that the government is discriminating on the basis of content or that the government's
    asserted interest is not truly pressing. See City of Ladue v. Gilleo, 
    512 U.S. 43
    , 51-53
    (1994); Glickman v. Wileman Bros. & Elliott, Inc., 
    521 U.S. 457
    , 493 (1997) (Souter,
    J., dissenting). When underinclusiveness of a regulation betrays content-discrimination,
    the standard of review must be raised to strict scrutiny. For instance, R.A.V.
    invalidated an ordinance prohibiting fighting words directed at some groups, but not
    others. The Supreme Court held that the principle for distinguishing between fighting
    words that were prohibited by the ordinance and those that were not had nothing to do
    with the rationale for permitting regulation of fighting words in the first place, but
    instead amounted to content discrimination. Therefore, the Court could not apply the
    relaxed standard appropriate for review of regulations of fighting words. 505 U.S. at
    386. The ordinance failed under strict scrutiny review. Id. at 395-96. In this case, we
    have already established that Canon 5 discriminates on the basis of content (though not
    on the basis of viewpoint) and that we must conduct strict scrutiny. Supra at 17-19.
            In cases in which underinclusiveness of a regulation suggests that the regulation
    is not truly necessary to further a compelling governmental interest, the government's
    obligation is to "establish the empirical reality of the problems it purports to be
    addressing." Glickman, 521 U.S. at 443 (Souter, J., dissenting). As part of our strict
    scrutiny, we have determined that the State established that a candidate's use of
    political parties in judicial campaigns poses a real threat to the State's compelling
    interests. Supra at 26-31. Thus, we have already undertaken the heightened scrutiny
    that underinclusiveness of a regulation would suggest is necessary.
           But even if the cases cited by the dissent and Wersal indicated that
    underinclusiveness poses a First Amendment problem in its own right, these cases
    concern only arbitrary underinclusiveness, which arises when the principle for
    distinguishing regulated conduct from non-regulated conduct does not promote the
    government's asserted purpose. In Carey, 447 U.S. at 457, 464, the challenged
    regulation prohibited residential picketing other than labor picketing in order to protect
    residential privacy. In Discovery Network, 507 U.S. at 415, the city banned use of
    newsracks to distribute commercial handbills, but not newspapers, as a way of
    preventing newsracks from becoming a safety hazard and eyesore. The types of
    picketing and newsracks, respectively, that were regulated posed no greater threat to
    the asserted governmental interest than those that were exempted from regulation, and
    the regulations were thus invalid. See Carey, 447 U.S. at 465; Discovery Network, 507
    U.S. at 424-28. In 44 Liquormart a ban on advertising alcoholic beverages was invalid
    where any connection between the advertising ban and the State's desired result of
    reducing alcohol consumption would have been "purely fortuitous." 517 U.S. at 506-07
    (Opinion of Stevens, J.).
           In contrast to these cases, when underinclusiveness results from a choice to
    address a greater threat before a lesser, it does not run afoul of the First Amendment.
    See, e.g., Cornerstone Bible Church v. City of Hastings, 
    948 F.2d 464
    , 470-71 (8th Cir.
    1991); Stretton v. Disciplinary Bd., 
    944 F.2d 137
    , 146 (3d Cir. 1991) ("[W]e cannot
    say that the state may not draw a line at the point where the coercive effect [of judicial
    campaign fund-raising], or its appearance, is at its most intense—personal solicitation
    by the candidate."). Indeed, a categorical rule against underinclusiveness for its own
    sake would coerce governments to regulate speech more broadly than they consider
    necessary, which is hardly the usual goal of First Amendment jurisprudence.17 Here,
    the State has shown that a candidate's use of political parties in judicial campaigns
    poses a greater threat to the compelling state interests than involvement of other kinds
    of groups. See infra at 39-40. Therefore, the State did not violate the First Amendment
    in so limiting Canon 5.
             It is significant that the dissent in this case first argues that the Minnesota
    Supreme Court has violated the Constitution in enacting Canon 5, and then makes the
    antithetical assertion that it has not gone far enough.
           As the final step of our review of the three restrictions on Wersal's political
    activity, we ask whether they are narrowly drawn to address the compelling state
    interests. See California Democratic Party v. Jones, 
    120 S. Ct. 2402
    , 2412, 2414
    (2000); Brown v. Hartlage, 
    456 U.S. 45
    , 54 (1982) (restriction must operate without
    "unnecessarily circumscribing protected expression").
           The Minnesota Supreme Court has attempted to prevent judicial candidates from
    incurring, or seeming to incur, debts to political parties that could compromise their
    independence, while allowing them alternative means of communicating subjects of
    valid interest to voters. Candidates may, for instance, speak to gatherings other than
    political organization gatherings; appear in newspaper, television, or other media
    advertisements supporting their candidacy; and distribute pamphlets and other
    promotional literature. Canon 5(B)(1). Furthermore, candidates may establish
    committees to campaign for them in various ways, including obtaining public
    statements of support (other than from political organizations). Canon 5(B)(2).
           Wersal argues that since various other States and the drafters of the ABA Model
    Code of Judicial Conduct have not found it necessary to forbid attendance at party
    gatherings, acceptance of party endorsements, and identification of party affiliation,
    Minnesota's restrictions must be broader than necessary. But Wersal relies on the laws
    of States that have in fact limited political activity by measures similar to those before
    us.18 There is also the authority of Letter Carriers and the other cases in which courts
            E.g., Fla. Code of Jud. Conduct Canon 7(C)(3) (imposing various limits on
    speaking at political gatherings, including prohibition on announcing party affiliation);
    Ky. Sup. Ct. R. 4.300, Canon 5(A)(2) (allowing candidate to state party affiliation only
    in response to direct question); Okla. Stat. Ann. tit. 20, § 1404(B)(6) (West 1991 &
    held it was proper to limit partisan political activity of executive employees. See supra
    at 24-25.
           The Party contends that the prohibition on disclosing party affiliation is not
    narrowly tailored because it prohibits disclosing past party affiliation as well as present.
    The Judicial Board states that it "simply prevents the judge from affirmatively stating
    his current political affiliation." However, the report submitted to the Minnesota
    Supreme Court at the time the party affiliation language was adopted states that it
    would prohibit candidates from identifying themselves "as past or present members of
    a political organization." The language was adopted as proposed. The Minnesota
    Supreme Court apparently concluded that the public would infer that identification of
    past membership was tantamount to identification of present membership. See In re
    759 P.2d 392
    , 395 (Wash. 1988) (en banc) (concluding judge violated
    prohibition on stating party affiliation by stating past affiliation). It was therefore
    reasonable to extend the prohibition on stating party affiliation to include past affiliation
    as well, lest the measure be rendered ineffective.
           The Party further contends that Canon 5 is not narrowly tailored because Canon
    5(A)(3) applies to a candidate's supporters as well as to the candidate. Specifically,
    Maxim argues that he had to avoid "seeking out any relationship with Gregory Wersal"
    in order to avoid coming within the scope of persons who were prohibited by Canon
    5(A)(3) from doing for Wersal what he could not do for himself—i.e., attending a party
    Supp.) (making it a ground for removal of judge from office to make party affiliation
    publicly known in connection with campaign for office); Or. Code of Jud. Conduct JR
    4-102, 4-104 and Or. Rev. Stat. § 249.015 (1999) (candidate shall not publicly identify
    candidate's political party membership except by registering to vote); S.D. Code of Jud.
    Conduct Canon 5(C)(1)(a)(ii) (permitting candidate to identify party affiliation only to
    vote); Wash. Code of Jud. Conduct Canon 7(A)(1)(e) (same).
    gathering or seeking a party endorsement. Similarly, the members of Wersal's
    campaign committee contend that the restriction is overbroad because the committee
    was prohibited from seeking a party endorsement of Wersal. This argument has two
    parts: first, that the regulation is overinclusive because it affects third parties who have
    a relationship with Wersal; and second, that the regulation is vague because third
    parties cannot predict under what circumstances their actions will be imputed to
           Two district court cases have anticipated the overinclusiveness issue. In
    Concerned Democrats v. Reno, 
    458 F. Supp. 60
    , 65 (S.D. Fla. 1978), and California
    Democratic Party v. Lungren, 
    919 F. Supp. 1397
     (N.D. Cal. 1996), courts held that
    state laws prohibiting political party endorsements of judicial candidates or of all
    nonpartisan candidates, respectively, were not narrowly tailored because they restricted
    the political parties' behavior when it was only necessary to restrict the candidate's
    behavior. Lungren assumed, arguendo, that the State had a compelling interest in
    preventing nonpartisan office holders from engaging in partisan activity, but held that
    the law "misse[d] the mark" in addressing that interest:
           The evil sought to be combatted is the unseemly partisanship of
           nonpartisan officeholders once they are in office, not the partisanship of
           political parties (which, of course, is the very nature of political parties).
           Section 6(b) purports to prevent officeholders from being "beholden" to
           political parties by imposing a ban on the parties' speech about candidates
           for office, rather than (as the Supreme Court approved in Letter Carriers)
           a ban on the partisan political conduct of the officeholders themselves.
           The distinction is crucial. The government has an interest in the manner
           in which its elected officials conduct themselves while in office. The
           government does not and cannot have a legitimate interest in silencing the
           speech of third parties about the qualifications and political views of
           candidates for those offices.
    919 F. Supp. at 1402.
           These district court cases convincingly reason that because the State's compelling
    interest is in the rectitude of the candidate, a narrowly tailored restriction will regulate
    expressions by the candidate, not third parties. However, the language of Canon
    5(A)(3) is so limited. It does not purport to govern what other persons do, but what the
    candidate authorizes or knowingly permits them to do. The canon therefore does not
    affect persons not subject to the candidate's authority and control. See Johnson v.
    116 N.W.2d 673
    , 679 (Minn. 1962) ("knowingly permit" extends only to
    persons having right of authoritative control over actor who are actually aware of his
    or her activity). Canon 5(A)(3) simply prohibits the candidate from accomplishing by
    the acts of an agent what the candidate is forbidden to personally. The Director of the
    Lawyers Board, the body charged with enforcing Canon 5(A)(3) in Wersal's case,
    confirms that there is no threat of disciplinary action without the candidate himself
    authorizing or permitting the action of the supporter. Insofar as Canon 5(A)(3)(a)
    refers to members of the candidate's family, it does not limit their activities, but rather
    affects the candidate's actions towards them: the candidate shall "encourage" them to
    adhere to the standards applicable to the candidate. As for persons who are within the
    candidate's authority and control as members of the candidate committees authorized
    by Canon 5(B)(2), they have voluntarily assumed that relation, and therefore may
    subject themselves to certain restraints which are necessary to make effective the
    restrictions on the candidate. Cf. Letter Carriers, 413 U.S. at 581-82 (Appendix to the
    Court's opinion) (United States Civil Serv. Comm'n Form No. 1236, which Congress
    intended to serve as its definition of the general proscription against partisan activities,
    id. at 572-74, prohibits "activity by indirection" when, by collusion or coercion,
    employee causes another to do what employee may not do directly). The limitation of
    Canon 5's reach to the candidate and to persons within the candidate's control thus
    distinguishes this case from Eu, 454 U.S. at 221, in which there was an "outright ban"
    on political parties endorsing candidates.
            The Party argues that a candidate's supporters cannot tell under what
    circumstances their actions will be imputed to the candidate. A restriction on speech
    must not be so vague that people of ordinary intelligence cannot tell what it prohibits.
    Broadrick v. Oklahoma, 
    413 U.S. 601
    , 607 (1973). But see Berger v. Supreme Court,
    No. 87-3935, 
    1988 WL 114792
    , at *3 (6th Cir. Oct. 31, 1988) (notice requirements
    less stringent in enforcing non-criminal code of judicial conduct than if criminal
    penalties at stake). Canon 5(A)(3)(c) limits the scope of its language by adding the
    requirement that the candidate's permission be "knowing." This requirement prevents
    the sort of violation by mistake the non-candidate plaintiffs claim has chilled their
    speech. See Chulchian v. City of Indianapolis, 
    633 F.2d 27
    , 31 (7th Cir. 1980)
    (ordinance prohibiting person from "permitting" conduct includes requirement that
    person know of conduct and therefore is not unduly vague). Moreover, both Canon
    5(A)(3)(a), dealing with family members, and Canon 5(A)(3)(c), dealing with
    authorizing or knowingly permitting others to act, include the concept of the candidate's
    right to control the speaker. Johnson, 116 N.W.2d at 679. Canon 5(A)(3) is therefore
    not unduly vague.
            The Party contends that the restrictions on a candidate's partisan activity violate
    its rights to equal protection because Canon 5 restricts association with political parties,
    but not other organizations that could affect a judge's independence or the public's
    perception of that independence. We have already determined that the restraints on
    Wersal's speech and association are necessary to serve a compelling state interest. To
    conclude that the same restraints violate the Equal Protection Clause, we would have
    to determine that Canon 5 burdens the rights of political party members more than
    others and that "such differential treatment is not justified." California Med. Ass'n v.
    453 U.S. 182
    , 200 (1981). In Broadrick, in rejecting a First Amendment
    overbreadth challenge to Oklahoma's restrictions on partisan activities by government
    employees, the Supreme Court also rejected an equal protection challenge to the same
          Appellants also claim that § 818 violated the Equal Protection Clause of
          the Fourteenth Amendment by singling out classified service employees
          for restrictions on partisan political expression while leaving unclassified
          personnel free from such restrictions . . . . [T]he legislature must have
          some leeway in determining which of its employment positions require
          restrictions on partisan political activities and which may be left
          unregulated. And a State can hardly be faulted for attempting to limit the
          positions upon which such restrictions are placed.
    413 U.S. at 607 n.5 (citation omitted).
            Our discussion regarding the State's interest in the independence of its judiciary
    demonstrates how Minnesota has historically viewed partisanship as a particular threat
    to the integrity of its courts, supra at 28-31, and we need not reiterate this discussion.
    Restrictions reaching only partisan, rather than nonpartisan, political activity have been
    upheld in other cases. See Letter Carriers, 413 U.S. at 562; Bauers v. Cornett, 
    865 F.2d 1517
    , 1524-25 (8th Cir. 1989) (application of statutes to nonpartisan activity
    would have raised a different constitutional question than application to partisan
    activity). Political parties specialize in the business of electing candidates and have a
    powerful machinery for achieving that end, including large membership and fund-
    raising organizations. Those parties are simply in a better position than other
    organizations to hold a candidate in thrall. Moreover, because political parties have
    comprehensive platforms, obligation to a party has a great likelihood of compromising
    a judge's independence on a wide array of issues. Finally, legislatures are bodies in
    which, for the most part, the members owe allegiance to a political party, not only for
    financial support and endorsement in their campaigns for office, but also for political
    support within the legislative process itself. No single legislator has the power to enact
    laws. Therefore, the sharing of common partisan affiliation plays an integral role in
    enactment of legislation. If the judiciary is then expected to review such legislation
    neutrally, a State may conclude that it is crucial that the judges not be beholden to a
    party responsible for enactment of the legislation, or to one that opposed it.
            It is also relevant that a judge's participation in associations other than political
    parties is regulated by Canon 4, which imposes broad requirements that judges avoid
    involvement that could cast doubt on their impartiality or interfere with proper
    performance of their judicial duties. At the Minnesota Supreme Court's 1997 hearing
    on amending Canon 5, DePaul Willette, Executive Secretary of the Judicial Board,
    testified that the danger of judicial candidates affiliating with single-issue interest
    groups was adequately addressed by the provision of Canon 5 prohibiting
    announcement of the candidate's views on disputed legal or political issues, Canon
    5(A)(3)(d)(i). Amendment of Canon 5 of the Code of Judicial Conduct: Hearing
    before the Minnesota Supreme Court, No. C7-81-300, at 12-13 (1997).
          We conclude that the State has justified its differential treatment imposing
    greater restrictions on a judicial candidate's partisan political activities than on
    association with other kinds of organizations.
          We turn next to Wersal's claim that Canon 5's "announce" clause, which
    prohibits judicial candidates from announcing their "views on disputed legal or political
    issues,"19 violates his free speech rights because it is not narrowly tailored to achieve
    any compelling state interest.20 We have already determined that the interests proffered
    by the State to justify this restriction are compelling governmental interests of the
    highest order, so we proceed to the next step of our inquiry and address whether the
    restriction is necessary to further these interests. See United States v. Playboy
    Entertainment Group, Inc., 
    120 S. Ct. 1878
    , 1891 (2000); see also Brown v. Hartlage,
    456 U.S. 45
    , 53-54 (1982).
           As we have discussed, it is consistent with the essential nature of campaigns for
    legislative and executive offices for candidates to detail and make promises about the
    programs that they intend to enact into law and to administer. For judicial officers,
                Canon 5(A)(3)(d)(i) provides that a judge or a candidate for judicial office shall
           (i) make pledges or promises of conduct in office other than the faithful
           and impartial performance of the duties of the office; announce his or her
           views on disputed legal or political issues; or misrepresent his or her
           identity, qualifications, present position or other fact, or those of the
             Wersal also challenges the announce clause as being vague. However, his
    argument is based solely on the clause's relationship to Canon 5(A)(3), which prevents
    him from "knowingly permit[ting]" others to announce his views on his behalf. We have
    already held supra at 38 that the "knowingly permit" language is not unconstitutionally
    vague, and we therefore reject this claim. Furthermore, our clarification of the
    "knowingly permit" language is dispositive of the Party's claim that the announce clause
    is unconstitutional as applied to third parties. Candidates are subject to discipline only
    when third parties act as agents of the candidate or have voluntarily assumed a
    relationship with the candidate as a member of the candidate's campaign committee.
    The Code does not regulate the independent activities of third parties.
    however, a State may determine that this mode of campaigning, insofar as it relates to
    how judges will decide cases, is fundamentally at odds with the judges' obligation to
    render impartial decisions based on the law and facts. At the time of the campaign, the
    candidate simply cannot predict what the facts or arguments in a particular case may
    be, the precise way in which legal issues will present themselves, or other crucial
    factors that need be considered before a court issues a final decision. See Berger v.
    Supreme Court, No. 87-3935, 
    1988 WL 114792
    , at *3 (6th Cir. Oct. 31, 1988) ("[T]he
    very purpose of the judicial function makes inappropriate the same kind of
    particularized pledges and predetermined commitments that mark campaigns for
    legislative and executive office.") (internal quotations omitted); Buckley v. Illinois
    Judicial Inquiry Bd., 
    997 F.2d 224
    , 227 (7th Cir. 1993) (free discussion of judicial
    candidate's views conflicts with the "historically more deeply rooted" concept of justice
    under the law); Stretton v. Disciplinary Bd., 
    944 F.2d 137
    , 142 (3d Cir. 1991)
    (prejudging cases at campaign stage makes mockery of the concept of an impartial
    judiciary and undermines public confidence in rule of law).
          Canon 5's announce clause restrains candidates from making statements in their
    campaigns about their views on disputed legal and political issues, and thus prevents
    candidates from implying how they would decide cases that might come before them
    as judge. Wersal maintains that this restriction is not necessary because Canon 5
    already protects the State's interests through bans on candidates making pledges or
    promises of conduct in office other than the faithful performance of their duties and on
    candidates manifesting bias or prejudice inappropriate to the judicial office. See Canon
    5(A)(3)(d)(i) & (ii).
           To be sure, the pledges and promises provision of Canon 5(A)(3)(d)(i) addresses
    the type of campaign conduct that most blatantly subverts the judicial office—pledges
    by candidates to make specific decisions on the bench. However, it does not reach the
    full range of campaign activity that can undermine the State's interests in an
    independent and impartial judiciary. It would not, for example, reach declarations by
    candidates that legislation relating to hot-button social issues is or is not constitutional.
    It also would not apply to candidates who publicized their opinions about how unsettled
    legal issues should be resolved. Both instances raise the specter that the candidates are
    declaring how they would decide questions that might come before them as judges in
    order to gain support for their candidacies. See Laird v. Tatum, 
    409 U.S. 824
    , 836 n.5
    (1972) (Rehnquist, J.) (Mem. on Motion for Recusal) ("In terms of propriety, rather
    than disqualification, I would distinguish quite sharply between a public statement
    made prior to nomination for the bench, on the one hand, and a public statement made
    by a nominee to the bench. For the latter to express any but the most general
    observation about the law would suggest that, in order to obtain favorable consideration
    of his nomination, he deliberately was announcing in advance, without the benefit of
    judicial oath, briefs, or argument, how he would decide a particular question that might
    come before him as a judge.").
           When a candidate is later called upon as a judge to preside over cases involving
    disputed issues about which he or she has made campaign announcements, the judge
    is placed in an awkward, if not impossible, position. Should the judge properly
    determine, for example, that a law violates the Constitution, having already expressed
    this opinion during the campaign, the judge risks appearing as though he or she
    prejudged the case rather than gave it due consideration in light of the law, arguments,
    and facts. This apparent rigidity can undermine the faith of the litigants and public in
    the judge's decision and in the State's judicial system generally. See Stretton, 944 F.2d
    at 142. On the other hand, should the judge reach a conclusion that departs from the
    opinion expressed during the campaign, the judge risks being assailed as a dissembler.
    Thus, the judge may hesitate to decide the case in a way that might lose votes at the
    next election. Certainly, many judges have the fortitude to resist such pressures, but
    we do not doubt that the potential of supporter abandonment at the next election can
    weigh heavily on judges who know they were elected based on representations they
    made during the last campaign. Cf. The Federalist No. 78, at 471 (Alexander
    Hamilton) (Clinton Rossiter ed. 1961) (arguing against periodic popular election of
    judges because it gives them "too great a disposition to consult popularity" when
    rendering decisions).
           Wersal does not explain why he believes the canon restricting candidates from
    manifesting bias or prejudice inappropriate to judicial office fully protects the State's
    interests left unshielded by the ban on candidates making improper pledges and
    promises, and we conclude it does not. Canon 5(A)(3)(d)(ii) was added to the Code
    when the Minnesota Supreme Court updated its ethical rules and brought them closer
    to the 1990 version of the ABA Model Code, some twenty-two years after the
    announce clause was promulgated; and we do not read the two provisions to address
    the same conduct. The manifestation of bias provision, fairly read, is directed at words
    and conduct that display personal animus against persons or groups, such as women or
    minorities. Even if it could be construed to cover announcements about how a
    candidate would decide cases if elected, settled rules of statutory interpretation require
    that we decline to read it in this way. See Radzanower v. Touche Ross & Co., 
    426 U.S. 148
    , 153 (1976) (subsequently enacted law cast in general terms should not be
    interpreted to supplant one dealing with specific subject unless absolutely necessary to
    give effect to the language). We therefore reject any implication that the adoption of
    the manifestation of bias provision rendered the announce clause unnecessary.
            Wersal also challenges the sufficiency of the evidence the Boards have put forth
    to justify this speech restriction. He argues that the State did not fulfill its evidentiary
    obligation as a matter of law because the Boards did not prove there was any concrete
    evidence showing the harmfulness of candidates announcing their views in the public
    record before the Minnesota Supreme Court at the time the court adopted the announce
    clause. This argument misapprehends the Boards' evidentiary burden in this case.
           In Nixon v. Shrink Missouri Government PAC, 
    120 S. Ct. 897
    , 907-08 (2000),
    for example, the Court stated that to sustain a law setting contribution limits, it was
    enough for the State to adduce evidence substantiating legislative concerns about the
    actual or perceived corruption associated with large contributions. As we have noted
    supra, the corroborating evidence included a state senator's sworn statement that "large
    contributions have 'the real capacity to buy votes,'" contemporaneous newspaper
    accounts tending to show public officials were actually corrupted by contributions, and
    the recent overwhelming passage of a statewide initiative indicating that the majority
    of voters had concluded that contribution limits were necessary. Id. None of these
    items were shown to be part of the public record before the legislature at the time the
    legislature enacted the law.
           Burson v. Freeman, 
    504 U.S. 191
     (1992), provides an example of the type of
    corroborating evidence that can support a speech restriction subject to strict scrutiny.
    In Burson, the Court entertained a challenge to a law establishing a 100-foot
    "campaign-free" zone around polling places that was designed to protect against voter
    intimidation and election fraud. As evidence showing that the law was necessary, the
    Court looked to nineteenth century ballot reforms and traced the development of
    restrictions on election-day electioneering. Id. at 202-05. It also noted that all fifty
    States limit access to areas in or around polling places. Id. at 206. Taken together, this
    evidence revealed a "widespread and time-tested consensus" about the problems of
    voter intimidation and voter fraud which corroborated the State's conclusion that some
    sort of "campaign-free" zone around voting compartments was necessary. Id.
           In this case, the Boards have met their evidentiary burden by offering evidence
    of widespread and longstanding consensus among members of the bench and bar about
    the necessity of restrictions on campaign speech that conveys a judicial candidate's
    propensity to decide cases in a particular way. Three-quarters of a century ago, the
    ABA promulgated the nation's earliest formalized standards of professional conduct for
    judges, the Canons of Judicial Ethics. Canon 30 provided that a candidate for judicial
    office "should not announce in advance his conclusions of law on disputed issues to
    secure class support." Canon 30 (1924). Developed and drafted by a committee
    chaired by Chief Justice William Howard Taft and later approved by delegates to the
    ABA's 1924 annual meeting, see Jeffery M. Shaman et al., Judicial Conduct and Ethics
    §1.02 (3d ed. 2000), the canon reflects the views of members of the legal profession
    as to what ethical restrictions are needed to protect the integrity of the judiciary. Many
    state courts reached similar conclusions and adopted the ABA Canons for their own
    jurisdictions, id. at §1.02 n.14, including the Minnesota District Judges Association,
    whose members voted to do so unanimously at a conference in 1950.
           In the early 1970s, the ABA committee charged with drafting the Model Code
    of Judicial Conduct included a restriction on candidate speech similar to the restriction
    contained in 1924 code. See ABA Model Code of Jud. Conduct Canon 7(B)(1)(c)
    (1972). The ABA's 1972 Model Canon 7 is the pattern on which Minnesota's announce
    clause is based, and a majority of States that have an elected judiciary promulgated
    ethical rules prohibiting candidates from announcing their views on disputed legal and
    political issues. See Patrick M. McFadden, Electing Justice: The Law and Ethics of
    Judicial Election Campaigns 85 (1990). In 1990, when the ABA revised its Model
    Code, it adopted a different formulation of the candidate speech restriction, but one that
    continues to restrict more speech than the pledges-and-promises provision. Model
    Canon 5 restrains candidates from making "statements that commit or appear to commit
    the candidate with respect to cases, controversies or issues that are likely to come
    before the court." ABA Model Code of Jud. Conduct Canon 5(A)(3)(d)(ii) (1990). A
    number of States have revised their rules in accordance with the 1990 "commitment"
    canon.21 Today, most States with an elected judiciary have campaign speech
    restrictions patterned after either the 1972 or 1990 ABA model canons.22
           Minnesota Supreme Court decisions such as Moon v. Halverson, 
    288 N.W. 579
    (Minn. 1939), and Peterson v. Stafford, 
    490 N.W.2d 418
     (Minn. 1992), also provide
    insight into factors that would have informed the court's decision to adopt the announce
    clause. As we have discussed, in both cases Minnesota justices commented on
    historical problems resulting from close proximity between judicial elections and
             See, e.g., Ariz. Sup. Ct. R. 81, Canon 5(B)(1)(d)(ii); Ark. Code of Jud.
    Conduct Canon 5A(3)(d)(ii); Cal. Code of Jud. Ethics Canon 5(B); Fla. Code of Jud.
    Conduct Canon 7(A)(3)(d)(ii); Ga. Code of Jud. Conduct Canon 7(B)(1)(c); Ill. Sup.
    Ct. R. 67, Canon 7(A)(3)(d)(i); Kan. Sup. Ct. R. 601A, Canon 5(A)(3)(d)(ii); Ky. Sup.
    Ct. R. 4.300, Canon 5(B)(1)(c); La. Code of Jud. Conduct Canon 7(B)(1)(d)(ii); N.Y.
    Code of Jud. Conduct Canon 5(A)(4)(d)(ii); Ohio Code of Jud. Conduct Canon
    7(B)(2)(d); S.D. Stat., ch. 16-2, app., Canon 5(A)(3)(d)(ii); Tenn. Sup. Ct. R. 10,
    Canon 5(A)(3)(d)(ii); Wash. Code of Jud. Conduct Canon 7(B)(1)(c)(ii).
            A few States proscribe only improper campaign pledges and promises. See,
    e.g., Or. Code of Jud. Conduct JR 4-102(B); Utah Code of Jud. Conduct Canon
    5(C)(1). Other States impose different, yet significant speech restrictions. See, e.g.,
    Tex. Code of Jud. Conduct Canon 5(2)(i) (no "indicat[ing] an opinion on any issue that
    may be subject to judicial interpretation by the [judge/candidate if elected into office],
    except that discussion of an individual's judicial philosophy is appropriate if conducted
    in a manner which does not suggest to a reasonable person a probable decision on any
    particular case"); Wis. Sup. Ct. R. 60.06(3) (prohibiting "suggestions of conduct in
    office which appeal to the cupidity or partisanship of the electing or appointing
    power"); see also Ala. Canons of Jud. Ethics Canon 7(B)(1)(c) (candidate "shall not
    announce in advance the candidate's conclusions of law on pending litigation"); Colo.
    Code of Jud. Conduct Canon 7(B)(1)(c) (candidate may not "announce how the judge
    would rule on any case or issue that might come before the judge").
    partisan politics. It is reasonable to infer that the prohibition on candidates announcing
    their views on disputed issues was intended in part to prevent judicial campaigns from
    becoming routine political contests, thereby jeopardizing the independence and integrity
    of the State's judiciary. This inference is corroborated by evidence showing that the
    Minnesota Bar Association, District Judges Association, and the Conference of Chief
    Judges recently recommended against the adoption of the less-restrictive 1990 ABA
    commitment canon because of concerns that liberalizing Canon 5's speech restrictions
    would politicize judicial elections.
          The plaintiffs have offered no contradictory evidence which might increase the
    need for a more extensive showing by the State. See Shrink Mo.,120 S. Ct. at 906.
    Wersal argues that the Lawyers Board's refusal to enforce the announce clause calls
    into question the Boards' attempt to demonstrate that the clause serves compelling state
    interests. That the Director of the Lawyers Board stated she would not prosecute
    violations under the clause as a matter of prosecutorial discretion until it was upheld
    by a court is not sufficient to increase the evidentiary burden. The Lawyers Board's
    determination was based on the likelihood of a constitutional challenge to the announce
    clause and the limited resources available to the Board for defending such a suit; it has
    little bearing on whether certain campaign announcements can impair or be perceived
    to impair judicial independence and impartiality.
          The Lawyers and Judicial Boards have shown that this restriction furthers
    compelling governmental interests in the independence and actual and perceived
    impartiality of the judiciary.
          Having concluded the restriction furthers compelling interests, we must examine
    whether it is narrowly tailored. See California Democratic Party v. Jones, 
    120 S. Ct. 2402
    , 2414 (2000); Brown, 456 U.S. at 54.
            The district court construed the announce clause to apply only to discussion of
    a candidate's predisposition on issues likely to come before the candidate if elected into
    office. Republican Party v. Kelly, 
    63 F. Supp. 2d 967
    , 986 (D. Minn. 1999). None of
    the plaintiffs claimed in their opening briefs on appeal that the district court erred by
    interpreting the language as it did. Wersal attempts to raise the issue in his reply brief.
    It is well established that issues not argued in an opening brief cannot be raised for the
    first time in a reply brief,23 and we will therefore analyze the announce clause as it has
    been construed by the district court. We should add, however, that if the issue were
    properly before us, we would not find error in the district court's interpretation. The
    longstanding principle that courts should construe laws to sustain their constitutionality,
    see, e.g., Edward J. DeBartolo Corp. v. Florida Gulf Coast Bldg. & Constr. Trades
    485 U.S. 568
    , 575 (1988), the Minnesota Supreme Court's adherence to this
    interpretive principle, see In re R.A.V., 
    464 N.W.2d 507
    , 509 (Minn. 1991), rev'd on
    other grounds by R.A.V. v. City of St. Paul, 
    505 U.S. 377
     (1992), and the Judicial
    Board's endorsement of the narrowed construction form a solid foundation for the
              See, e.g., United States v. Vincent, 
    167 F.3d 428
    , 432 (8th Cir. 1999), cert.
    120 S. Ct. 124
     (1999); South Dakota Mining Ass'n v. Lawrence County, 
    155 F.3d 1005
    , 1011 (8th Cir. 1998); United States v. Davis, 
    52 F.3d 781
    , 783 (8th Cir.
    1995); French v. Beard, 
    993 F.2d 160
    , 161 (8th Cir. 1993); see also United States v.
    70 F.3d 1507
    , 1549 n.18 (8th Cir. 1995) ("Absent some reason for failing to
    [raise issue in opening brief], we will not consider an issue first raised in a reply
    district court's conclusion. Furthermore, the interpretation accords with Stretton, where
    the Third Circuit interpreted identical language similarly. 944 F.2d at 144.24
           As construed by the district court, the restriction prohibits candidates only from
    publicly making known how they would decide issues likely to come before them as
    judges. A similar construction was determined to be narrowly drawn in Stretton, 944
    F.2d at 144, and we are persuaded that it is narrowly drawn here as well. These sorts
    of campaign announcements are (or can appear to be) calculated to show that the
    candidate will decide cases in a certain way if elected into office, and the implication
    that the candidate will carry through with his campaign announcements can haunt the
    candidate on the bench to the detriment of the State's judicial system. See id. But see
    Buckley, 997 F.2d at 229 (questioning whether limiting construction would significantly
    circumscribe scope of announce clause, giving example of civil war in Yugoslavia as
    likely to come before courts as persecution defense to deportation).
           Wersal has stated that the restriction prohibits candidates from discussing
    virtually every topic related to their campaigns, and the dissent maintains that it
            In claiming error, Wersal relies heavily on Buckley v. Illinois Judicial Inquiry
    997 F.2d 224
    , 229-30 (7th Cir. 1993) where the Seventh Circuit declined to read
    an analogous canon narrowly. Having carefully studied both Stretton v. Disciplinary
    944 F.2d 137
     (3d Cir. 1991), and Buckley, we believe that Stretton is more
    persuasive. Buckley gives little heed to "our task [as courts] to construe [laws] so as
    to comport with constitutional limitations" if consistent with the will of the lawmaker.
    United States Civil Serv. Comm'n v. National Ass'n of Letter Carriers, 
    413 U.S. 548
    571 (1973). Buckley also distinguished Stretton in part due to factors not present in
    this case or in Stretton: a disciplinary body's determination that a judge violated the
    clause when he commented truthfully on his past record and counsel's convoluted
    arguments that the announce clause contained an implied "right of reply." 997 F.2d at
    effectively bans all campaigning. But nothing in Canon 5(A)(3)(d)(i) restricts
    candidates from discussing or publicizing information about their character, fitness,
    integrity, background (with the exception of their political affiliation), education, legal
    experience, work habits, and abilities, which are subjects the Minnesota General
    Assembly has determined to be highly relevant to a candidate's qualification for office.
    See Minn. Stat. § 480B.01, subd. 8 (listing criteria by which commission on judicial
    selection shall evaluate persons applying for judicial vacancies to be filled by
    gubernatorial appointment). The Minnesota Supreme Court has also indicated that
    candidates may discuss and state their views on how they would handle administrative
    duties if elected. See Bundlie v. Christensen, 
    276 N.W.2d 69
    , 72 (Minn. 1979) (no
    ethical violation under Canon 7(B), the precursor to Canon 5(A), when candidate made
    cutting costs of courtroom administration part of campaign platform); cf. Ackerson v.
    Kentucky Judicial Ret. & Removal Comm'n, 
    776 F. Supp. 309
    , 314 (W.D. Ky. 1991)
    (restraint on campaign speech about court administration would violate First
           We further believe the Minnesota Supreme Court would conclude that general
    discussions of case law or a candidate's judicial philosophy do not fall within the scope
    of the announce clause. In 1990 the Judicial Board issued an advisory opinion, which
    the Board has since made known to judicial candidates by letter, stating that Canon 5
    does not prohibit candidates from discussing appellate court decisions. See Minnesota
    Bd. on Judicial Standards, Informal Op. 10/10/1990. The Judicial Board has also
    approved extensive lists of sample questions that interested voters could pose to
    judicial candidates consistent with Canon 5. These questions are wide-ranging and
    include such topics as a candidate's judicial philosophy, issues relating to the
    administration of justice in criminal, juvenile, and domestic violence cases, and the
    candidate's perception of a judge's role in the judicial system.
           While it is true that some judicial ethics advisory bodies in other States have
    interpreted language identical to the announce clause to reach topics such as these, see,
    e.g., Berger v. Supreme Court, 
    598 F. Supp. 69
    , 74 (S.D. Ohio 1984) (disciplinary
    counsel's position that truthful criticisms of judicial administration and incumbents
    prohibited); see generally, Patrick M. McFadden, Electing Justice: The Law and Ethics
    of Judicial Election Campaigns 86-87 (1990) (collecting advisory opinions stating that
    discussion of court administration, court rules, criminal sentencing, judicial philosophy,
    published court decisions, etc. were improper), we believe the Minnesota Supreme
    Court would view the matter consistently with the Judicial Board. The Judicial Board
    has been intimately involved with the process of amending the Code over the years, and
    its views are likely to reflect those of the Minnesota Supreme Court. As the Third
    Circuit in Stretton put it, "[w]e would be naive not to recognize that the [Judicial
    Board's] position is, at the very least, a straw in the wind indicating the direction that
    [the supreme] court will go." Stretton, 944 F.2d at 143. Furthermore, some of the
    advisory opinions of other States' ethics boards conflict with the Minnesota Supreme
    Court's decision in Bundlie on whether the announce clause prohibits announcements
    about judicial administration.
           Wersal cites several decisions in which courts have struck down similar canons
    as overly broad restrictions on speech. E.g., Buckley, 997 F.2d at 228-29; Beshear v.
    863 F. Supp. 913
    , 917-18 (E.D. Ark. 1994); ACLU v. Florida Bar, 
    744 F. Supp. 1094
    , 1097 (N.D. Fla. 1990); J.C.J.D. v. R.J.C.R., 
    803 S.W.2d 953
    , 956-57 (Ky.
    1990). However, these cases proceed on the assumption that judicial candidates are
    left with little to discuss and to put before the voters other than their "name, rank, and
    serial number," Buckley, 997 F.2d at 227, "biographical data," ACLU v. Florida Bar,
    774 F. Supp. at 1098, or "professional history," J.C.J.D., 803 S.W.2d at 956, and are
    therefore distinguishable.
           The announce clause, as construed by the district court, is narrowly tailored to
    further compelling governmental interests.
          Wersal contends that Canon 5(B)(2),25 which prohibits judicial candidates from
    personally soliciting campaign funds, violates the First Amendment because it is not
    narrowly tailored.
           We have no difficulty in concluding that Canon 5(B)(2) is necessary to serve
    compelling state interests. Indeed, Wersal does not dispute these points. Raising
    campaign funds, a practical requirement for judicial candidates who face popular
    election, has been identified as presenting "the greatest of all conflicts" between
    necessity and judicial impartiality. See E. Wayne Thode, Reporter's Notes to Code of
            In full, Canon 5(B)(2) provides:
          A candidate shall not personally solicit or accept campaign contributions
          or solicit publicly stated support. A candidate may, however, establish
          committees to conduct campaigns for the candidate through media
          advertisements, brochures, mailings, candidate forums and other means
          not prohibited by law. Such committees may solicit and accept campaign
          contributions, manage the expenditure of funds for the candidate's
          campaign and obtain public statements of support for his or her
          candidacy. Such committees are not prohibited from soliciting and
          accepting campaign contributions and public support from lawyers, but
          shall not seek, accept or use political organization endorsements. Such
          committees shall not disclose to the candidate the identity of campaign
          contributors nor shall the committee disclose to the candidate the identity
          of those who were solicited for contribution or stated public support and
          refused such solicitation. A candidate shall not use or permit the use of
          campaign contributions for the private benefit of the candidate or others.
    Judicial Conduct 98 (ABA 1973). Judges, more than officeholders in other branches
    of government, risk the appearance that those who contribute to their campaigns can
    impermissibly influence governmental processes. When judges obtain funds from a
    group that has an interest in the outcome of litigation, such as the plaintiffs' or
    defendants' bar, judges can appear beholden to that group for their accession to office,
    creating the expectation that the judges will favor their benefactors accordingly. See
    In re Fadeley, 
    802 P.2d 31
    , 40 (Or. 1990) ("A judge's direct request for campaign
    contributions offers a quid pro quo or, at least, can be perceived by the public to do
    so."); cf. Nixon v. Shrink Mo. Gov't PAC, 
    120 S. Ct. 897
    , 905-06 (2000) (public
    perception that large donors can "call the tune" for candidates can harm democratic
    processes); Buckley v. Valeo, 
    424 U.S. 1
    , 27 (1976) (per curiam). Even if judges
    receive contributions from a broad cross-section of persons and interests, the
    appearance of impropriety hangs over them if they adjudicate cases in which a litigant
    or counsel has contributed, or refused to contribute, to their campaign. See Stretton,
    944 F.2d at 145 (noting the unseemliness of judges presiding over cases in which
    lawyers have helped fund judicial campaign chests).
           Recognizing the unique difficulties presented by judicial campaign fundraising,
    Canon 5(B)(2) seeks to insulate judicial candidates from the solicitation and receipt of
    funds while leaving open ample alternative means for candidates to raise the resources
    necessary to run their campaigns. To this end, Canon 5(B)(2) provides that candidates
    may establish campaign committees to conduct fundraising on their behalf. The Code
    does not limit the amount individuals can contribute to the candidate's campaign fund
    or the amount the campaign committee may raise. Canon 5(B)(2) imposes only two
    restrictions related to campaign finance on the committee: its members cannot disclose
    to the candidate the identity of contributors or those who were solicited to contribute
    but declined, and they cannot permit the campaign funds to be used for private benefit.
            Wersal's objection to this fundraising arrangement is limited: He does not
    challenge the campaign committee setup or the restrictions placed on the committee's
    members. Rather, he suggests that a narrowly tailored rule would allow candidates to
    solicit funds from large groups and to send out letters over their signatures requesting
    money. He argues that the State's concern over the corrupting influence of money on
    judicial candidates is fully redressed by the provision that prevents the campaign
    committee from revealing the identity of campaign contributors.
            Wersal is correct that the knowledge-screening provision of Canon 5(B)(2)
    lessens several of the problems attendant on judicial campaign fundraising; we reject
    his conclusion, however, that it obviates the need for limitations on candidates
    soliciting funds from groups personally or by letter. When a judicial candidate seeks
    funds directly from those who stand to gain from the candidate's decisions in office, the
    State may conclude that his activity is incongruous with the decorum of judicial office
    whether or not the candidate is shielded from knowing who ultimately contributed.
    Particularly when candidates target groups who have pecuniary or ideological interests
    in litigation likely to come before the court, the mere act of solicitation can contribute
    to the appearance, accurate or not, that "justice is for sale" and the expectation of
    impermissible favoritism. See Fadeley, 802 P.2d at 40.
           Wersal makes a secondary argument that when Canon 5(B)(2) deprives him of
    the opportunity to solicit funds from large groups or by mail it prevents him from
    running an effective campaign. The Supreme Court has said that restrictions on
    campaign financing must leave candidates with the ability to accumulate sufficient
    resources for effective advocacy. Shrink Mo., 120 S. Ct. at 908-09; Buckley, 424 U.S.
    at 21. Citing the impracticability of his being accompanied by a campaign committee
    member who could "utter the magic words" to solicit funds, Wersal argues his
    statewide campaign efforts were frustrated. However, no provision in the Code
    specifies that members of a campaign committee must request contributions in person.
    Canon 5(B)(2) expressly permits committees to conduct campaigns through mailings,
    brochures, and advertisements. Wersal has not alleged that these alternatives are
    ineffective fundraising mechanisms. Accordingly, he has not demonstrated that Canon
    5's restrictions stifled his campaign efforts. See also Shrink Mo., 120 S. Ct. at 909 ("a
    showing of one affected individual does not point up a system of suppressed political
    advocacy that would be unconstitutional under Buckley").
            In a separate argument, the Party asserts that Canon 5(B)(2) is not narrowly
    tailored as applied to third parties. Much of this argument is based on a misreading of
    a Code section we have already clarified. Canon 5(A)(3)(c), which states that
    candidates "shall not authorize or knowingly permit any other person to do for the
    candidate what the candidate is prohibited from doing," is violated only when persons
    act as agents of the candidate. As a result, no ethical violations occur if third parties
    solicit or expend funds independently of Wersal and his campaign committee.
           The district court did not err in determining that this restriction is narrowly
    tailored to further compelling state interests.
          The judgment of the district court is affirmed.
    BEAM, Circuit Judge, dissenting.
           At the outset, I admit that, insofar as I have been able to discern, Minnesota has
    a judicial system peopled with individuals of intellect, integrity and character, fully
    capable of making difficult decisions, consistent with the interests of the people of the
    state. However, such a condition is beside the point because many other states have
    achieved the same essential goals without trenching upon clearly established
    constitutional rights. The court today holds that laws prohibiting candidates for public
    office from voicing their views as to issues pertinent to their conduct in office, or their
    opponent's conduct in office, or from appearing or speaking to like-minded citizens, are
    not only permissible but are apparently necessary to a well-ordered democracy.
    Because in my view such a result flatly contradicts the edicts of the First Amendment,
    I respectfully dissent.
           The court and I part ways in three critical respects. First, the court misconstrues
    Minnesota law, reading it with such latitude as to contradict one-hundred fifty years of
    development in Minnesota's judicial selection processes. Second, the court supplants
    constitutionally guaranteed rights with its own notions of preferred judicial policy.
    Finally, the court countenances restrictions on fundamental, protected activity that are
    neither necessary nor narrowly tailored. In the final analysis, the court sustains a set
    of restrictions which probably have no practical effects other than to quash election-
    related speech and association, and thus undermine a democratic process–precisely the
    fear that prompted the drafting of the First Amendment.
           Prior to even contemplating the suppression of otherwise protected First
    Amendment activity, a state must assert a compelling interest. Presumably, that interest
    must reflect a state's actual and lawfully adopted policy. The court accepts as
    "undeniably compelling" appellees assertion that in effecting Canon 5, the Minnesota
    Supreme Court furthered Minnesota's interest in maintaining the "independence" of its
    judiciary. Ante at 19. Certainly, all states share such an interest generally–such is the
    preferred state of the American judiciary.26 But different sovereigns give different
    meanings to the term "independence." I believe the court has misconstrued Minnesota
    precedent, and in doing so has used it to justify a policy prescription that lacks support
    in Minnesota's constitutionally established public regimen.
            The court, without further analysis, simply accepts appellees' claim that
    Minnesota has "historically pursued the ideal of an independent judiciary." Ante at 21.
    It relies heavily on the Minnesota Supreme Court's observation in Peterson v. Stafford,
    490 N.W.2d 418
     (Minn. 1992), that while the federal and its own system differ, they
    both "create and maintain an independent judiciary as free from political, economic and
    social pressure as possible." Id. at 420. In this, the court finds license to sustain as
    consistent with Minnesota public policy anything which it considers to enhance judicial
             The court distorts my views on judicial independence, implying, ante at 20, that
    I would sacrifice such independence, and, presumably, its alleged by-product, "justice
    under law," on an altar of "democratic self-government." (I assume the court is
    referring to my concern for the democratic election process.) The court's statements
    are seriously at variance with my views and my arguments in dissent. This case is not
    about rules under which a neutral judiciary dispenses justice under canons of ethics
    designed to support judicial independence, once the selection process is over. I wholly
    agree that such a system should exist and be designed to enhance, through reasonable
    state regulation, the ideals of judicial independence. This case is about the point at
    which free speech and association collide directly with an attempt by the Minnesota
    Supreme Court to manage (and I believe over manage) the election process in the name
    of judicial independence of some definition.
           My dissent simply and appropriately recognizes and discusses these competing
    interests and reaches a result in line with the Constitution and Supreme Court
    precedent. It is the court that insists that "one of the principles should give way
    completely to the other," that is, judicial independence, whatever that proves to be
    under the court's analysis, should trump the First Amendment when the election of a
    judge is at stake. There is absolutely no support in federal law for the court's position.
    independence. But this latter term is subject to broad and varied interpretations, and
    merely begs the question "independent from what?" While appointed judges are
    "independent" from the electorate, elected judges, once elected, are hardly "dependent"
    on the electors, and are "independent" from other elected officials. However, we are
    not called upon to sustain our own notions of "independence," but Minnesota's. We
    must therefore read the Minnesota Supreme Court's statement in Peterson in light of
    that state's public policy as prescribed by its citizens. As it turns out, Minnesotans have
    consistently rejected all attempts to narrow or curtail their elective franchise when it
    comes to selecting judges. Rather, they have repeatedly strengthened popular control
    over their judiciary.
            The debate whether to appoint, retain or directly elect judges predates
    Minnesota's founding. The federal drafters debated it in Philadelphia. See The Anti-
    Federalist Papers & the Constitutional Convention Debates 120-27 (Ralph Ketcham
    ed., Mentor 1986). Alexander Hamilton defended their subsequent choice in print,
    arguing that elected judges would suffer "too great a disposition to consult popularity
    to justify a reliance that nothing would be consulted but the Constitution and the laws."
    The Federalist No. 78, at 471 (Clinton Rossiter ed. 1961). Anti-Federalists hotly
    contested that position. Brutus counseled trepidation at the advent of a judiciary
    "altogether unprecedented in a free country" which was to be "rendered totally
    independent, both of the people and the legislature." Brutus XI, Letter of January 31,
    1788, in Ketcham, ante, at 293. See also, George Mason, Speech in Opposition to the
    Constitution, and Patrick Henry, Speech to the Virginia Ratifying Convention, both in
    Ketcham, ante, at 174, 212 (arguing that the federal judiciary made the courts too far
    removed from the citizenry).
          The federal founders ultimately decided upon lifetime appointments, subject to
    good behavior, placing us (ostensibly) beyond the cavil of partisan activity and populist
    suasion. U.S. Const. art. III. As federal judges, we understandably have a bias for this
    method. Yet, we should not confuse the rationale behind our federal system with some
    universal judicial raison d'etre, for different sovereigns value different policies. Some
    elect only a few officers and rely largely on appointments, while others insist on
    electing virtually every office-holder. Since Hamilton and Brutus debated, the
    arguments on each side have been well known, and were so at Minnesota's founding.
           Minnesota's first judiciary arrived under the auspices of the United States with
    the Northwest Ordinance of 1787, which provided the Northwest Territories with
    congressionally-appointed judges and governor-appointed local magistrates. Northwest
    Ordinance §§ 4, 7 (1787); see generally, Hiram F. Stevens, History of the Bench & Bar
    of Minn. (1904). In 1849, Congress created the Territory of Minnesota and provided
    a similarly appointed judiciary. Act of Congress, March 3, 1849 §§ 2, 9, 11, 9 Stat.
    403. Minnesotans did not warm to these arrangements. As one state founder noted,
    "it has been the complaint of the Territory . . . that we have to submit to have our
    Judges sent to us and have no voice in the selection . . . and [the people] have looked
    forward with hope to the time when they could elect their own men." The Debates &
    Proceedings of the Minn. Constitutional Convention 495 (1857) ("Democratic Debates
    & Proceedings"). That chance came in 1857 when Congress authorized Minnesotans
    to draw up a constitution. Act of Congress, Feb. 26, 1857, 11 Stat. 166.
           Minnesota chose an interesting time to seek statehood. With the Civil War three
    years off, and the Supreme Court's decision in Dred Scott v. Sandford, 
    60 U.S. 393
    (1856), one year past, any prospective new state necessarily stood in the eye of a
    hurricane.27 Against this backdrop, on June 1, 1857, Minnesotans elected delegates to
             Abolitionist sentiment ran high in Minnesota and, while not reaching the levels
    in other states, did result in some violence. Interestingly, among the slaves brought to
    Minnesota during the pre-civil war period was Dred Scott, who from 1835-40 lived
    a constitutional convention. One historian gave the following description. "The feeling
    between and within political parties was bitter. Republicans were not just Republicans
    but 'Black Republicans' and 'Nigger Lovers.' The Democrats were divided into
    'Regular' and 'Moccasin' Democrats, but were united for convention purposes." Julius
    E. Haycraft, Territorial Existence & Constitutional Statehood of Minnesota (1946),
    republished in 1 Minn. Stat. Ann. 145, 151; see generally, William Anderson, A
    History of the Constitution of Minnesota (1921). Moreover, "[s]o intense was the
    feeling that the Republicans and Democrats would not meet in the same convention.
    The result was two conventions; delegates of each party held separate sessions in
    different rooms in the old capitol." Haycraft, ante, at 151. The two bodies met
    separately, largely without consultation. Even after a conference committee produced
    a compromise document, each convention refused to sign a document soiled by the
    other's signatures, so each separately signed largely identical documents. Id. at 151-52.
           The records of these proceedings demonstrate that Minnesota's founders
    grappled with precisely the same contours as their federal forebears. The Republicans
    steadfastly insisted on an elected judiciary, but did debate the form of those elections.
    Debates & Proceedings of the Constitutional Convention for the Territory of Minn.
    334-35, 402 (1858) ("Republican Debates & Proceedings"). The question of politics
    arose when a Mr. Galbraith proposed an amendment requiring that judicial elections
    never be held on the same day as an election for any other office. He reasoned that
    "the excitements of mere political and party issues should be kept as far as possible
    from the election of the judges–that the people should elect the Judges, upon their
    merits as judges–that they should be elected with the view of making them as
    independent of political parties as possible." Id. at 406. In opposition, a Mr. Lowe
    argued, "I do not believe it will be possible to separate the election of judges from
    with his "owner," Surgeon J. Emerson, at Fort Snelling. Stevens, ante, at 30-36.
    political considerations." Id. The convention rejected Galbraith's amendment, and
    opted for concurrent judicial elections.
           The Democratic convention proved far more contentious. After a committee
    failed to agree on a judicial selection method, the issue came to the convention floor.
    Democratic Debates & Proceedings, ante, at 493-509. Those favoring the Hamiltonian
    appointment model did so for two reasons, fear of populism and fear of political
          First, they feared the populism that might result from an elected judiciary. One
    vigorous opponent, a Mr. Setzer, argued:
          The great object of an appointed Judiciary, is to secure stability upon the
          part of the government, by having a power within the State conservative
          enough to restrain the waves of popular excitement, when they sweep
          over us as they have done in different States for years past. . . . Judges
          represent no constituency and are elected by no constituency. They
          represent nothing except the abstract ideas of equity and justice.
    Id. at 495-96. The "abstract ideas of equity and justice" heralded were those set out
    in Dred Scott–the culmination of the grand experiment that seventy years earlier Brutus
    had observed was "altogether unprecedented in a free country." And the "waves of
    popular excitement" that so rankled delegates were the popular sentiments of radical
    abolitionist fervor. He continued, "[w]e see it in Wisconsin, in Iowa, and in all those
    States where popular excitement in reference to negro-worship and disunion has had
    its effect upon an elective Judiciary." Id.
           Minnesota's founding debate did not turn entirely on slavery. Those opposed to
    an elected judiciary also raised precisely the same, eminently sensible concerns
    troubling the court today. Another delegate, a Mr. Meeker, argued:
          I contend that the Judges who are elected, are elected by parties, and are
          the mere fuglemen of caucuses. The best trickster or the best manager of
          caucuses is just as likely to be the nominee of a party as the most learned
          man in the nation. . . . [T]he greatest curse that could befal [sic] any
          people would be the establishment of a political court . . . . [A]nd our
          Judges . . . if they are to be elected, must necessarily be essentially
          political Judges; they cannot be anything else.
    Id. at 500. They feared the instability such a system would create. "Whichever party
    is in the ascendancy will change the system of jurisprudence to its own standard, and
    there will be no security, no stability in anything." Id. at 501. Accordingly, a Mr.
    Sherburne recommended adopting the venerable appointment mechanisms common
    along the east coast. Id. at 497.
          The prevailing delegates, those favoring judicial elections, met these arguments
    head-on. Said a Mr. Emmett:
          We hear a great deal of talk about an independent Judiciary. The phrase
          is in everybody's mouth. What does it mean? Independent of whom?
          Independent of what? Independent of the people . . . ? I say then that in
          order to correct the errors of Judges–and it may be important to correct
          them,–the office should be made elective.
    Id. at 503. And as regards the fear that politics would infect an elected system, he
    continued, "if the people are incapable of selecting their Judges, they are also incapable
    of selecting the man who is to appoint the Judges . . . . The governor always selects
    men belonging to his own political party, while the people often select them regardless
    of parties." Id. A Mr. Curtis responded to Mr. Sherburne: "[I]s a sufficient argument
    in favor of an appointed Judiciary, that it is old? Is it all that can be said in its favor,
    that it has grown hoary by age and usurpation–because it is all covered over from one
    end to the other by corruption and fraud?" Id. at 498.
           As these exchanges demonstrate, the arguments surrounding election and
    appointment were well known and well met. Despite the concerns raised then and
    marshaled again by the court today, Minnesotans of both parties opted for an elected
    judiciary. Ultimately, they more feared the potential politics of an appointed bench and
    saw popular election as the proper remedy. Given the political climate, they desired
    more control over their judiciary. That was the policy adopted by the people of
           Minnesota's electorate has never since intimated a change in this policy. Indeed,
    it has frequently reaffirmed judicial accountability to the electorate.28 In 1912, for
    instance, the legislature required judicial candidates to run on a non-partisan ballot.
    Peterson, 490 N.W.2d at 420. That statute in no way restricted a candidate's political
    activities or ability to educate the voters as to his merits, opinions or values. Moon v.
    288 N.W. 579
    , 580-81 (Minn. 1939). The reform's objective was to ensure
    that candidates were elected on their own merits. Id. at 580. A merit-based election
    necessitates the flow of information regarding candidates to the electorate.
             Minnesota courts have consistently struck down rules depriving judicial offices
    of their elective nature. See State ex rel La Jesse v. Meisinger, 
    103 N.W.2d 864
    , 866
    (Minn. 1960); State ex rel Smallwood v. Windom, 
    155 N.W. 629
    , 633 (Minn. 1915);
    State ex rel Rosckes v. Dreger, 
    106 N.W. 904
    , 906 (Minn. 1906); State ex rel Jordan
    v. Bailey, 
    33 N.W. 778
    , 779 (Minn. 1887) .
            Minnesota has four times rejected the so-called Missouri Plan, whereby the
    governor could fill vacancies by non-partisan appointment, and such appointees would
    later be subject only to a retention vote. Maynard E. Pirsig, The Proposed Amendment
    of the Judiciary Article of the Minn. Constitution, 
    40 Minn. L
    . Rev. 815, 815-19 (1955-
    56); Peterson, 490 N.W.2d at 421 n.14. Again, voters retained their system of direct
    electoral control.
           In 1948, a commission recommended that candidates run against a specific
    incumbent rather than against the field. Peterson, 490 N.W.2d at 421 n.14. It argued,
    "[w]here several incumbents are running for re-election, each should stand or fall on
    the basis of his own record. The present system does not permit this since opposing
    candidates run against the field." Report of the Constitutional Comm'n of Minn. 43-44
    (Oct. 1, 1948). Minnesota law was amended accordingly. See 14A Minn. Stat. Ann.
    § 204B.06; Gustafson v. Holm, 
    44 N.W.2d 443
     (Minn. 1950) (sustaining requirement
    that judicial candidates declare which specific seat is sought). Such seat-specific
    elections increase the electorate's focus on a specific judge's conduct.
           Thus, since first permitted to select its own judiciary, Minnesota has consistently
    favored electorally-responsive judges. Its citizens have considered and rejected the
    court's policy concerns,29 and have rejected appointment and retention systems that
            Indeed, the Minnesota Supreme Court specifically refused to resurrect the
    policy debates undergone at the constitutional convention.
          There are many arguments, on grounds of both principle and expediency,
          against the election of judges, either in special cases or generally. Such
          arguments do not enter into our consideration. The fundamental law of
          this state is, and always has been, that the selection of judges must be
          submitted to the electors.
    would have curtailed or eliminated popular control. Minnesota has repeatedly affirmed
    its citizens' right to elect their judges, and has bolstered that franchise with laws
    enhancing merit-based elections and furthering the flow of information regarding the
    candidates to the electorate.30 Canon 5 subverts this policy. The suppression of
    election-related speech and association does not enhance the franchise. In fact, it
    militates against the public's ability to elect judges based on their merits by denying
    citizens a reasonable vehicle for self-information. The court assures that candidates
    may still discuss their "character, fitness, integrity, background (with the exception of
    their political affiliation), education, legal experience, work habits and abilities." Ante
    at 51. However, these are simply not all of the qualities that piqued the interest of
    Minnesota's founders, nor those which fully inform voters who are charged with the
    selection of Minnesota's judiciary.
    La Jesse, 103 N.W.2d at 866.
             The court attempts to divorce my review of Minnesota judiciary history from
    that undertaken by the Minnesota Supreme Court in Peterson, and paints the latter as
    justifying its conclusion today. Ante at 23 n.13. But the review undertaken in Peterson
    supported a regulation wholly different from that before us today. In Peterson, the
    Minnesota Supreme Court reconciled Minnesota history with a measure requiring the
    judicial election ballot to designate incumbent jurists as such. 490 N.W.2d at 424. The
    Minnesota Court found that initiative consistent with Minnesota history and sustained
    it because it enhanced the flow of information to Minnesota voters, which stands in
    stark contrast to the measure before us today which suppresses that same flow of
    information. Rather than address what the Minnesota Supreme Court meant by
    "independent" in the proper context, the court simply transports the Minnesota court's
    purposely cabined conclusions into today's opinion. But as my review, and that
    undertaken in Peterson indicate, in the context of Minnesota public policy, an
    "independent judiciary" means judges elected by well-informed, independent voters.
    In undermining that process by stifling the free flow of information, the court merely
    exacerbates the partial transfer of the Minnesota judicial selection machinery into the
    hands of other narrow political influences as noted elsewhere in this dissent.
           The question here is whether Minnesota has expressed a sufficiently fundamental
    policy interest in "independent" judicial elections to even warrant our proceeding to the
    next stage of our constitutional inquiry. Certainly, in Peterson, the Minnesota Supreme
    Court instructed us that Minnesota has an interest in making its judiciary as
    independent "as possible." 490 N.W.2d at 420. But as the Minnesota Supreme Court
    cannot trump its citizens' prescribed policy, "as possible" cannot be read to include
    restrictions which curtail the elective franchise. Because Canon 5 does precisely that,
    I believe the court has clearly misapplied Minnesota law.
           The preceding discussion may prove largely academic, because the restrictions
    at issue in this case were not enacted by the Minnesota legislature but rather were
    created by the Minnesota Supreme Court itself. While that court has not yet ruled that
    its own regulations comport with fundamental state policy, it could just as easily do so.
    And as a federal court, we defer to state courts on questions of state law. Indeed, we
    wade into such questions only in the most dire of circumstances, for instance where
    state law is being intentionally tortured to discriminate against a protected minority,
    see, e.g., Bouie v. City of Columbia, 
    378 U.S. 347
     (1964), or where an error of state
    law works a singular national impact, see, e.g., Bush v. Gore, 
    121 S. Ct. 525
    But while we ought not inject ourselves into the state law question, we are charged with
    enforcing federal law. Federal law, and in particular the Constitution, operate to
    protect the rights of citizens against encroachment by both federal and state
    governments. In my view, even accepting arguendo that the people of Minnesota did
    actually intend to immunize their judicial candidates from some of the rigors of an
    election campaign, Canon 5 still flatly contradicts controlling precedent and violates the
           The court repeatedly emphasizes the importance of judicial independence, a
    proposition with which I heartily agree. Judicial independence is of the utmost
    importance to a fair and well-functioning judicial system. Moreover, I agree that a state
    has an interest not only in warding off an actual breakdown of that independence, but
    also the appearance of such. Every court to address this or a similar question has
    recognized the importance of this principle. Moreover, all have recognized, again
    rightly so, that by virtue of their function, judges fundamentally differ from other
    elected officials. The court cites extensively to these authorities. Yet the court omits
    noting that almost all of these courts have then further recognized that such policy
    notions cannot trump constitutionally-enshrined rights.31 The Constitution makes strict
    demands. Often times, important and justifiable public policy goals must bow before
    its restraints. As judges, we must put aside our inclinations as jurists, lawyers,
    residents, citizens, and parents, and enforce the Constitution's restraints. The
    Constitution protects the most basic elements of our democratic processes, in particular
    the rights of candidates and citizens to express and receive views pertinent to public
    office, and to associate in the election context with others of their own choosing.
             See e.g., Suster v. Marshall, 
    149 F.3d 523
     (6th Cir. 1998) (striking down
    restrictions on judicial campaign spending); Buckley v. Illinois Judicial Inquiry Bd.,
    997 F.2d 224
     (7th Cir. 1993) (striking down "announce" clause); Stretton v.
    Disciplinary Bd. of the Supreme Court of Pa., 
    944 F.2d 137
     (3d Cir. 1991) (narrowing
    construction of "announce clause"); Weaver v. Bonner, 
    114 F. Supp. 2d 1337
    Ga. 2000) (striking down canon restricting advertising and election speech); Butler v.
    Alabama Judicial Inquiry Comm'n, 
    111 F. Supp. 2d 1224
     (M.D. Ala. 2000) (restraining
    enforcement of campaign advertisement and conduct restrictions); ACLU of Florida v.
    The Florida Bar, 
    744 F. Supp. 1094
     (N.D. Fla 1990) (enjoining "announce" canon); In
    re Chmura, 
    608 N.W.2d 31
     (Mich.) (narrowing canon to prohibit only knowingly false
    statements), cert. denied, 
    121 S. Ct. 77
     (2000); J.C.J.D. v. R.J.C.R., 
    803 S.W.2d 953
    (Ky. 1991) (striking down "announce" and other campaign restrictions).
    Because Canon 5 sharply curtails, if not outright bans these activities, regardless of
    how laudable its purpose, it must be struck down.
           Because election speech and association lie at the heart of the First Amendment,
    any restriction thereof must survive strict scrutiny. Eu v. San Francisco County
    Democratic Cent. Comm., 
    489 U.S. 214
    , 223 (1989). A restriction on otherwise
    protected activity must be necessary and narrowly tailored to serve a compelling state
    interest. Id. at 222. "[I]t is the rare case in which . . . a law survives strict scrutiny,"
    Burson v. Freeman, 
    504 U.S. 191
    , 211 (1992), yet the court today, with little
    reservation, affirms a set of restrictions more sweeping in scope and effect than ever
    adopted elsewhere. I discuss in turn, the First Amendment rights implicated by the
    court's decision, the court's purported interests and the need for and tailoring of the
    challenged restrictions.
            The Supreme Court has consistently afforded election speech and association the
    highest protection. "The free exchange of ideas provides special vitality to the process
    traditionally at the heart of American constitutional democracy–the political campaign."
    Brown v. Hartlage, 
    456 U.S. 45
    , 53 (1982). Few, if any, state interests prove
    sufficiently compelling to trump these activities. While "a state has a legitimate interest
    in upholding the integrity of their electoral process . . . , when a State seeks to uphold
    that interest by restricting speech, the limitations on state authority imposed by the First
    Amendment are manifestly implicated." Id. at 52. Canon 5 runs roughshod over the
    clearly established speech and associational rights of candidates, political parties and
    the voting public. That alone should foreclose any thought of Canon 5's
           Candidates for public office enjoy both speech and associational rights.
           The candidate, no less than any other person, has a First Amendment right
           to engage in the discussion of public issues and vigorously and tirelessly
           to advocate his own election and the election of other candidates. Indeed,
           it is of particular importance that candidates have the unfettered
           opportunity to make their views known so that the electorate may
           intelligently evaluate the candidates' personal qualities and their positions
           on vital public issues before choosing among them on election day.
    Buckley v. Valeo, 
    424 U.S. 1
    , 52-53 (1976) (emphasis added). To that end, the First
    Amendment protects the right to campaign and restricts states to preventing only actual
    fraud and corruption.32 Brown, 456 U.S. at 55.
           Candidates similarly enjoy protected association rights. These interpose
    themselves "at the crucial juncture at which the appeal to common principles may be
    translated into concerted action, and hence to political power in the community."
    Tashjian v. Republican Party of Conn., 
    479 U.S. 208
    , 216 (1986). A candidate's
    association with a political party, organization or another candidate provides a basic
    means of election communication. Associational rights work hand in hand with a
    candidate's speech rights, for without the former, the latter may prove meaningless.
    "'The right to associate with the political party of one's choice is an integral part of this
             The court cites Burson where the Supreme Court sustained the prohibition of
    "the solicitation of votes and the display or distribution of campaign materials within
    100 feet of the entrance to a polling place." 504 U.S. at 193. While not a traditional
    "time, place, and manner" restriction in that it was content-specific, the rule curtailed
    speech only within that 100-foot bubble. Id. at 197. Against a vast history of election
    fraud and intimidation at polling places, the Court permitted only an extremely narrow,
    location-specific restriction on campaign speech. Burson hardly supports the absolute
    bans imposed by Canon 5.
    basic constitutional freedom.'" Id. at 214 (quoting Kusper v. Pontikes, 
    414 U.S. 51
    , 57
           The First Amendment protects not only speakers' rights, but also those of
    listeners. Virginia State Bd. of Pharmacy v. Virginia Citizens Consumer Council, Inc.,
    425 U.S. 748
     (1976). A campaign implicates not only the candidate's right to self-
    expression, but also voters' rights to receive the candidate's message. A state may not
    "hamstring[] voters seeking to inform themselves about the candidates and the
    campaign issues." Eu, 489 U.S. at 223. Nor may a state limit public debate as "[i]t is
    simply not the function of government to 'select which issues are worth discussing or
    debating.'" Brown, 456 U.S. at 60 (quoting Police Dep't of Chicago v. Mosley, 
    408 U.S. 92
    , 96 (1972)); accord United States v. Playboy Entm't Group, Inc., 
    120 S. Ct. 1878
     (2000). In short, "[a] state's claim that it is enhancing the ability of its citizenry
    to make wise decisions by restricting the flow of information to them must be viewed
    with some skepticism." Eu, 489 U.S. at 228 (quotation omitted).
           Voters' associational rights parallel their right to receive information. "To the
    extent that party labels provide a shorthand designation of the views of party candidates
    on matters of public concern, the identification of candidates with particular parties
    plays a role in the process by which voters inform themselves for the exercise of the
    franchise." Tashjian, 479 U.S. at 220.
           A party's speech and association also receive protection, particularly in the
    election context. Eu, 489 U.S. at 224; First Nat'l Bank of Boston v. Bellotti, 
    435 U.S. 765
     (1978). In endorsing or inviting a candidate to speak, a party "reflects its members'
    views about the philosophical and governmental matters that bind them together [and]
    seeks to convince others to join those members in a practical democratic task," an
    election. Colorado Republican Fed. Campaign Comm. v. FEC, 
    518 U.S. 604
    , 615
    (1996) (plurality opinion of Justice Breyer). Limitations on a party's ability to endorse
    a candidate "hamper[] the ability of a party to spread its message." Eu, 489 U.S. at
    223. "[A] political party has a right to . . . select a standard bearer who best represents
    the party's ideologies and preferences." Id. at 224 (citations and quotations omitted).
    The suppression of a party's right to associate with candidates of its choice, renders it
    powerless.33 "The Party's attempt to broaden the base of public participation in and
    support for its activities is conduct undeniably central to the exercise of the right of
    association." Tashjian, 479 U.S. at 214.
           Canon 5 trenches upon these rights in a variety of ways. The "announce" clause,
    for instance, effectively bans campaigning itself. The court disagrees, saying "[t]he
    judicial candidate simply does not have a First Amendment right to promise to abuse
    his office." Ante at 16. But the court proves too much. The Supreme Court
    established in Brown that no candidate for any office possesses such a right. "[T]he
    state may ban such illegal agreements without trenching on any right . . . protected by
    the First Amendment." 456 U.S. at 55. Beyond prohibiting fraud and corruption, and
    the appearance of the same, a state may prevent judicial candidates from prejudging or
    pledging outcomes in specific cases or even in particular types of cases. But beyond
    such, a state cannot further unreasonably curtail election-related speech.
             The court relies in part on Timmons v. Twin Cities Area New Party, 
    520 U.S. 351
     (1997), in which the Court sustained a ban on fusion candidates–those appearing
    on the ballot as representing more than one party. That law governed form of the ballot
    but did not restrict parties from endorsing the candidate of their choice. This prevented
    small parties from bootstrapping their way into public election funds by endorsing a
    major party candidate, and then claiming the resulting votes. As it did not affect
    candidate or party speech and associational rights, the Court sustained the rule. That
    case does not support the banning of all election-related discussion and association.
           The announce clause, however, bars discussion of all disputed legal and political
    issues. The court's narrowing construction to "issues likely to come before the
    candidate," ante at 49, is meaningless. See Buckley v. Illinois Judicial Inquiry Bd., 
    997 F.2d 224
    , 229 (7th Cir. 1993) (recognizing the true scope of such a "limitation").34
    Even could such a category be properly cabined, it would still roam far beyond the
    narrow category of speech left unprotected by the Court in Brown. The court assures
    us, for instance, that a candidate may discuss his "judicial philosophy." Ante at 51. I
    cannot fathom "disputed legal issues" more likely to come before a court than the
    proper role of stare decisis, narrow or strict construction, original intent and substantive
    due process. Yet are these not captured by the term "judicial philosophy?" The term
    is as sweeping as the court's allegedly narrow construction.35
                Construing similar language, similarly curtailed, Judge Posner noted:
           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. The
           civil war in Yugoslavia? But we have cases in which Yugoslavs resist
           deportation to that nation on the ground that they face persecution . . . ;
           and some years ago the Illinois courts were embroiled in a custody fight
           involving a child who didn't want to return to the then Soviet Union with
           his Soviet parents.
    Buckley, 997 F.2d at 229.
             To truly grasp the impact of today's opinion, one must consider the plight of the
    judicial candidate who in 2002 must decipher Minnesota ethical obligations. Between
    them, Canon 5, the district court's opinion and the court's opinion today supply three
    different versions of acceptable speech. Moreover, all three remain contingent on the
    Minnesota Supreme Court's agreeing with the court's construction, for if it does not, the
    judicial candidate may still face sanctions. The threat of chilled speech is too great to
           Along with the announce clause, the restrictions on attending and speaking at
    party gatherings, and the bans on seeking or using endorsements, and on self-
    identifying as a party member similarly tread upon protected First Amendment rights.
    Obviously, by banning such activities, Canon 5 directly impacts the rights of judicial
    candidates. The court insists that this impact goes no further, that Canon 5 affects only
    the rights of candidates, and not those of any third parties. Ante at 36-38. But this
    cannot be, as laws that affect candidates necessarily affect voters. Bullock v. Carter,
    405 U.S. 134
    , 143 (1972). The court strives mightily to justify the impact wrought on
    candidates' rights, arguing for instance that judicial candidates have less interest in
    certain types of policy debate than do other types of candidates. Ante at 16. Such
    arguments do little to justify collateral effects on the rights of others.36
           Canon 5 affects third parties in a variety of ways. For instance, the announce
    clause curtails voters' rights to hear a candidate's views on those issues most pertinent
    to future conduct in office. By preventing candidates from seeking endorsements,
    Canon 5 potentially prevents voters and parties from gauging a candidate's attitude
    towards a party, something central to many voters' decision-making processes. Despite
    the court's protestations, it also de facto curtails a party's ability to endorse the
    candidate of its choosing. Parties are unlikely to endorse their preferred candidate so
    long as they fear handing an opponent a potential election issue–specifically, the
    specter of an ethics violation–even if none exists. This in turn curtails voters' rights to
    hear parties' views as to candidates. Additionally, by preventing candidates from
    attending or speaking at party gatherings, Canon 5 trenches upon parties' rights to
             Of course, the court belies its own contention with the argument, addressed
    infra, that political parties work a unique effect upon the electoral scheme. Ante at 33-
    34, 40-41. How the court reconciles its concern with party involvement with its
    assurances that the rights of candidates alone are affected escapes me.
    express themselves in the manner of their own choosing, specifically by presenting the
    candidate of their choice.
           Our First Amendment rulings usually parse the hypothetical. We worry about
    "potential" chilled speech, and the supposed ill effects that might flow from quashed
    association. Presented only with the text of Canon 5, one might well hypothesize that
    restricting candidates' abilities to express their views and to associate with political
    parties would lead to two results: higher rates of incumbent retention; and
    disproportionate campaign fund-raising. This hypothesis follows from a rudimentary
    understanding of public campaigns and the judiciary.
           Campaigns for public office, judicial, legislative or executive, necessarily favor
    incumbents. Incumbents have better name recognition, better resources, are usually
    better staffed, and barring some recent disaster, are afforded the presumptions of
    experience and competence. Of the three branches of government, the judiciary seems
    to least capture the public's attention. Judicial incumbents, therefore, may enjoy these
    benefits to an even greater extent.
           Unlike most of our First Amendment cases, however, this case need not stick to
    the hypothetical. Canon 5 has been in effect in Minnesota for some time, in varying
    forms. We therefore can look to recent elections to see whether these supposed harms
    have come to pass. Minnesota is no exception to the rule that the public pays the least
    attention to the judiciary. Indeed, one Minnesota Supreme Court Justice has observed
    of judicial elections that "it's pretty darn hard to get people to care." Amendment of
    Canon 5 of the Code of Judicial Conduct: Hearing Before the Minn. Supreme Court 44
    (1997). Election statistics bear this out. In 1996 one-half million voters cast a ballot
    in the Presidential race yet did not vote in the Supreme Court race. League of Women
    Voters of Minnesota, Choosing Minnesota's Judges: An Examination of the Present
    System and Alternative Proposals 6 (1998). Moreover, 86% of responding voters said
    they needed more information about judicial elections and 77% observed they got less
    information about judicial than other elections. Id.
           Fund-raising and other election data demonstrate disturbing distortions in a
    supposedly democratic process.37 During the 2000 election cycle, of those judges up
    for election, only five of sixty-seven trial court benches were contested and only one
    of four appeals court judges faced a challenger. In 1998, only ten of eighty-nine trial
    court benches were contested.38 Judicial fund-raising reflects the same bias. In the
    2000 election, the four incumbent Supreme Court justices seeking re-election raised a
    combined $505,070.63. Of their opponents, only two raised sufficient funds to warrant
    disclosure. These two together raised only $23,582.67. Candidate Filings, Minnesota
    Campaign Finance and Public Disclosure Board (2000). A stable judiciary is, on
    balance, a laudable pursuit and perhaps Minnesotans have no cause to find fault with
    any of their judges. But it is just as likely that, at least in part, these imbalances flow
            Prior to every Supreme Court election, the Minnesota State Bar Association
    holds a plebiscite between the candidates among its members. The results display in
    dramatic fashion the extent to which an incumbent may count on support from groups
    most vested in the judiciary. Since 1974, incumbents have out-polled their challengers
    by the following margins: 1974: 91% to 9% and 85% to 15%; 1978: 77% to 23% and
    95% to 5%; 1982: 95% to 5%; 1992: 61% to 39% and 85% to 15%; 1996: 75% to 25%
    and 90% to 10%. Minnesota State Bar Ass'n, Judicial Elections Task Force Report &
    Recommendations 32 Appendix D (1997). While this case has been pending, the 2000
    election cycle has come and gone. I occasionally use the most recent data, which post-
    dates the record on appeal.
             These numbers reflect a consistent trend. Since 1980, the number of contested
    trial benches of those facing election were as follows: 2 of 25; 0 of 25; 5 of 56; 9 of 54;
    2 of 67; 4 of 93; 5 of 82; 6 of 42; 14 of 111. Minnesota State Bar Ass'n, Judicial
    Elections Task Force Report & Recommendations 7 (1997).
    from a lack of information resulting from the suppression of election-related speech and
    association and from other impediments inherent in Canon 5.
           Not surprisingly then, in recent history prior to the 1998 election which spawned
    this suit, only two incumbent judges had lost elections. See Mary Ellen Egan, Judging
    for Himself, City Pages 4-5 (June 17, 1998). The record details appellant Wersal's
    campaign difficulties, hamstrung by both fund-raising and speech limitations. Indeed,
    an ethics complaint forced him out of the 1996 campaign for fear of losing his law
    license and being unable to support his family. Another Supreme Court candidate,
    Michael Demoss, faced a similar problem after making comments about his views on
    abortion. With challengers unable to create traction by discussing actual issues, or to
    raise funds, the results in judicial elections are hardly surprising.
           Minnesota's judicial elections thus reveal the precise results one might suspect
    Canon 5 would produce. Moreover, its chilling effects on electoral activity are
    precisely those which spurred the First Amendment. It might be naive to argue that
    incumbents must never be permitted to use their offices to bolster their own
    positions–such conduct seems inimical to politics generally. But American tradition
    has long recognized that the best defense against such conduct is the right of a people
    to freely associate, assemble and speak. It may be, as a matter of policy, that an
    appointed judiciary best protects our liberties. I certainly think so, and the court
    obviously agrees. But Minnesota has made a different choice, and having chosen an
    elective selection method it cannot then turn around and quash candidates'
    constitutionally guaranteed rights. The upshot is this: when a state opts to hold an
    election, it must commit itself to a complete election, replete with free speech and
    association. Certainly, a state may regulate corruption as in Brown, the integrity of the
    voting process as in Burson and the form of its ballot as in Timmons v. Twin Cities
    Area New Party, 
    520 U.S. 351
     (1997). But the state may not unduly curtail speech and
    association, for they are too central to First Amendment interests in the election
           The court does not discuss the foregoing authorities. Rather, it tries to sidestep
    this entire doctrinal inconvenience with a simple argument–that judges are different.
    It surmises that given the nature of their office, judges, even when elected, warrant
    different treatment under the First Amendment.39 The court identifies three interests,
    all under the rubric of enhancing judicial independence, which it deems sufficiently
    compelling to warrant the suppression of fundamental rights: safeguarding a judge's
    ability to apply the law neutrally; maintaining the public confidence in judicial
    impartiality; and insulating judges from political pressure. Ante at 19, 24. Many courts
    have noted, and I agree, that judges differ from other officials. As a matter of
    constitutional law, however, the judiciary has no more or less right than any other
    coordinate branch to insulate itself from the rigors of public debate, particularly in the
    election context. The court embarks on a path long since foreclosed, as the Supreme
    Court has directly and pointedly refused to permit such specialized treatment, even
    given the court's stated concerns.
             The court dallies briefly with an analogy between judges and civil servants,
    drawing heavily on United States Civil Service Commission v. National Association
    of Letter Carriers, 
    413 U.S. 548
     (1973). There, the Court found compelling the
    interests of forging a civil service capable of enforcing laws without regard to party
    allegiances, and in avoiding the appearance of bias among federal employees in order
    to maintain the public confidence. Id. at 565. The analogy fails. Offices are made
    elective rather than appointive in order to introduce a degree of public control. Given
    the discretion that judges wield, that voters would desire such control is hardly
    surprising. The state's interests in these two areas are quite divergent.
           The judiciary may not rely on the "maintenance of confidence in the judicial
    system," essentially the interest advanced by the court today, as a justification for the
    suppression of protected expression. Landmark Communications, Inc. v. Virginia, 
    435 U.S. 829
    , 833 (1978). In Landmark Communications, Virginia attempted to rely on
    this interest to justify a statute making it a misdemeanor to report on confidential
    judicial disciplinary proceedings.40 Id. Recognizing that "the law gives judges as
    persons, or courts as institutions . . . no greater immunity from criticism than other
    persons or institutions," the Court ruled "neither the Commonwealth's interest in
    protecting the reputation of its judges, nor its interest in maintaining the institutional
    integrity of its courts is sufficient to justify the subsequent punishment of speech at
    issue here." Id. at 839, 841 (quotation omitted, emphasis supplied). "[I]njury to
    official reputation is an insufficient reason for repressing speech that would otherwise
    be free [and] the institutional reputation of the courts, is entitled to no greater weight
    in the constitutional scales." Id. at 841-42 (quotation omitted, emphasis supplied).
           Landmark Communications extended a well-established body of precedent
    forbidding judicial efforts at self-isolation from public debate. Potential "disrespect for
    the judiciary" and worries of the "disorderly and unfair administration of justice" cannot
    compel the suppression of protected expression. Bridges v. California, 
    314 U.S. 252
              The court may seek to cabin Landmark Communications as merely preventing
    the criminalization of speech whereas this case regards an administrative or
    professional sanction. The statute in Landmark Communications made the challenged
    conduct a misdemeanor and subjected the paper to a $500 fine. While criminal in form,
    the actual harm in that case was much less severe than the threat judicial candidates
    face in Minnesota. Failure to abide by ethical canons can subject a lawyer or a judge
    to sanction, removal or disbarment. In terms of gravity, the restrictions at issue here
    pose much the greater danger. See In Re Primus, 
    436 U.S. 412
     (1978) (holding that
    a state's interest in effective judicial administration must bend before an attorney's First
    Amendment rights).
    270 (1941). In Bridges, the Court prevented a trial court from silencing newspaper
    editorial criticism of its conduct during ongoing legal proceedings, because:
           The assumption that respect for the judiciary can be won by shielding
           judges from published criticism wrongly appraises the character of
           American public opinion . . . . [A]n enforced silence, however limited,
           solely in the name of preserving the dignity of the bench, would probably
           engender resentment, suspicion, and contempt, much more than it would
           enhance respect.
    Id. at 270-71.
           This argument grows all the stronger in the election context. In Wood v.
    370 U.S. 375
    , 379-80 (1962), the Court reversed a contempt order issued
    against a sheriff who told a grand jury that he opposed its having been empaneled. The
    state argued such conduct constituted a "clear and present danger" to the administration
    of justice. Id. at 387. However, both the sheriff and the judges involved were elected
    officials charged with public responsibilities. Id. at 390, 395. The sheriff, therefore,
    "had the right to enter the field of political controversy, particularly where his political
    life was at stake. The role that elected officials play in our society makes it all the more
    imperative that they be allowed freely to express themselves on matters of current
    public importance." Id. at 394-95 (citation omitted); see also, Craig v. Harney, 
    331 U.S. 367
    , 370, 376-77 (1947) ("Judges are supposed to be men of fortitude, able to
    thrive in a hardy climate. . . . Judges who stand for reelection run on their records.
    That may be a rugged environment. Criticism is expected."); Pennekamp v. Florida,
    328 U.S. 331
    , 336-40 (1946) (rejecting a "threat to the impartial and orderly
    administration of justice" as justification for a contempt order stemming from a political
    cartoon); accord Richmond Newspapers, Inc. v. Virginia, 
    448 U.S. 555
     (1980) (holding
    that without good cause, criminal trials must be open to the public); Nebraska Press
    Ass'n v. Stuart, 
    427 U.S. 539
     (1976) (affirming principle of judicial transparency).
           The First Amendment therefore forecloses the use of criminal or contempt
    sanctions to suppress criticism of the judiciary. The use of punishing civil and
    professional sanctions to accomplish the same end is equally offensive. I see little
    difference between truthful criticisms leveled at a judge by a newspaper editor or
    elected official on one hand, and a judicial candidate on the other.41 I similarly see little
    difference between the use of the contempt power, and the use of a Supreme Court's
    ability to regulate bench and bar, to silence judicial criticism. Judges, elected or
    otherwise, are thought to be citizens of honor, integrity and fortitude. By holding office
    they subject themselves to public criticism.42 Monitor Patriot Co. v. Roy, 
    401 U.S. 265
    , 274 (1971). And as elected officers, this holds doubly so for Minnesota's
    judiciary. A judicial candidate has a right to criticize the opinions, philosophies, rulings
    and public statements of the incumbent, who has a concomitant right to respond in kind.
    In sustaining Canon 5, the court permits precisely what these precedents say may not
             The court argues that Minnesota "seeks not to shield judges from published
    criticism, but to prevent them from deserving such criticism." Ante at 25 n.14 (citations
    and quotations omitted). But the court's cure kills the patient. The worry, presumably,
    is that candidates will do or say things that pledge or seem to pledge them to certain
    outcomes in return for support, thus threatening judicial impartiality. The cure is to
    ensure that candidates can do nothing at all, thus destroying the election Canon 5 set
    out to save. Whether the court likes it or not, facing public criticism, whatever the
    source, comes with being a public official. Moreover, it is most certainly part of being
    a public candidate. This regulation flies in the face of both.
             Couching these prohibitions as necessary professional regulations falls similarly
    short, for the Supreme Court has repeatedly favored First Amendment expression over
    professional ethical regulations. See Ohralik v. Ohio State Bar Ass'n, 
    436 U.S. 447
    (1978); In re Primus, 
    436 U.S. 412
    be done, except that rather than only protecting judges from the speech of others, they
    are apparently designed, according to the court, to protect incumbent judges and
    judicial candidates from the effects of their own speech. Such paternalistic notions
    have long been suspect, and are patently unconstitutional. Even granting the court's
    contention that Canon 5 will prevent some undesirable conduct, the vast majority of the
    speech and association suppressed will be legitimate, wholesome, democratic activity.
    Contrary to the court's assumption, judges, in this context, are not different. In the eyes
    of the First Amendment, they are the same.43
           I turn finally to the more mundane application of the Supreme Court's strict
    scrutiny jurisprudence. Even accepting that a restriction may serve a compelling state
    interest, in order to warrant the suppression of otherwise protected speech and
    association, it must be necessary to achieve its purpose. Burson, 504 U.S. at 198.
    Despite the court's best efforts, Canon 5 fails this test.
           Because a regulation that does not achieve its stated end cannot be necessary,
    the touchstone of necessity is effectiveness. Eu, 489 U.S. at 228-29; Tashjian, 
    479 U.S. 214
    . For instance, in Eu, the state argued that its ban on political party
    endorsements of primary candidates served the compelling interest of protecting
             To bolster its "judges are different" argument, the court relies in part on Cox
    v. Louisiana, 
    379 U.S. 559
     (1965), where the Court reviewed a statute prohibiting the
    picketing of a courthouse with the intent to influence a judge, juror or court officer in
    the course of his duties. Id. at 560. The Court reversed the conviction on the grounds
    that an official had misrepresented to the defendant that his protesting was legal. Id.
    at 569-70. Despite this holding, the court engaged in a lengthy dictum discussing the
    statute itself. The statute, the Court observed, regulated only conduct, and did not
    infringe upon speech or assembly at all. Id. at 563.
    primary voters from confusion and undue influence. Eu, 489 U.S. at 228. The Court
    doubted whether the restriction could achieve its stated rationale because of all the
    organizations with an interest in primary elections, the ban silenced only political
    parties. Id. at 228 n.18. Many other equally or more vocal groups remained free to
    endorse whomsoever they wished. Moreover, "the growing number of endorsements
    by political organizations using the labels 'Democratic' or 'Republican' has likely misled
    voters into believing that the official governing bodies were supporting the candidates."
    Id. As the restriction did not effect its goal, it was not necessary.44
           Canon 5 suffers from precisely this same flaw. It bars judicial candidates from
    attending or speaking at "political organization gatherings" or from seeking, accepting
    or using endorsements from a "political organization," and defines a "political
    organization" as a "political party organization." Canon 5 purports to safeguard a
    judge's ability to apply the law neutrally, to maintain the public confidence in judicial
    impartiality and to insulate judges from public pressure. Given its restricted
    application, however, Canon 5 simply cannot achieve these goals.
           The court worries that by appearing at, speaking to or being endorsed by a
    political party, a candidate will appear "in hoc" thereto. As a general proposition, this
             Ordinarily lawmakers may address problems piecemeal rather than having to
    propose all-encompassing solutions. Indeed, permitting "over-inclusiveness" and
    "under-inclusiveness" underpins basic rationality review. See Williamson v. Lee
    Optical of Okla., Inc., 
    348 U.S. 483
     (1955). Such is not the case regarding
    fundamental rights, however, where we enforce a much higher standard. Playboy, 120
    S. Ct. at 1888. Here, failure to solve the entire asserted problem is fatal. See, e.g., 44
    517 U.S. 484
     (Opinion of Stevens, J.); Discovery Network, 507 U.S. at
    418. Strict scrutiny imposes a much heavier burden on Canon 5 than on the "time,
    place and manner" restrictions reviewed in Cornerstone Bible Church v. City of
    948 F.2d 464
    , 470-71 (8th Cir. 1991).
    is simply wrong. The Supreme Court recognized in Eu that one cannot conclude from
    such conduct that a candidate will necessarily tow a party line. 489 U.S. at 225 n.15.
    But even accepting the court's premise, Canon 5 falls far short of its goals because it
    fails to include candidate appearances before, speeches to, or endorsements by other
    politically active organizations. Candidates remain free to consort with organizations
    such as the National Organization of Women, the NAACP, the Christian Coalition, the
    National Association of Manufacturers, the AFL-CIO or the NRA. Many of these
    organizations are implicitly, and some explicitly, aligned with one or another of the
    major parties.45 Candidates remain free to seek and use the endorsements of the
    insurance lobby, the Minnesota Trial Lawyers Association, or any other political or
    quasi-political organization not organized as a "political party." If the court's
    "appearance" concerns are valid, this ought to cause it grave anxiety.
          The court argues that political parties are more likely to command a candidate's
    acquiescence to party positions on the bench. Ante at 33-34, 40. Were this true, the
    court's argument would still fail because one does not satisfy the necessity requirement
    merely by solving for the greatest influence, but rather by addressing all significant
    influences. The court's assertion, however, is untrue. Of all the politically-oriented
    organizations around, the two major parties are both "big-tent"
    organizations–ideologically consistent only in their aggregate distinctions one from
    another. Within each party a variety of opinions on most issues may be found. A
    candidate's appearance at a party function hardly indicates endorsement of its views.
    (I am certain that Republicans everywhere will be either delighted or dismayed to know
    that Ralph Nader's appearance at their national convention last fall signified his support
    for their platform. See Eun-Kyung Kim, Nader Addresses Youth Convention, AP
           Indeed, statistics from the 2000 election indicate that NAACP members voted
    for democratic candidates at a higher rate than did members of the democratic party.
    Online, WL 24553132 (August 2, 2000).) A speech at a party function supports
    nothing beyond the candidate's own words, and an endorsement suggests only that the
    endorser finds a candidate more palatable than any other presently available.
           On the other hand, quasi-political organizations, particularly those that focus on
    a single or only a few issues, are much more likely to demand and command a
    candidate's attention and support. The Federalist Society, for instance, is an
    organization dedicated to strict construction. Were it to endorse a candidate, it stands
    to reason that the candidate would be seen as a "strict constructionist." Abortion issue
    groups are among those most likely to endorse only candidates who will be thought to
    adhere strictly to the relevant "party line." Given their focus and intensity, such groups
    are much more capable than political parties of bringing their members' wrath to bear
    on elected officials including judges who stray from their point of view.
           The record reflects the increasing role in and influence over judicial election
    politics exerted by these groups. Among those groups active in recent Minnesota
    judicial elections were state and local bar associations, the League of Women Voters,
    an organization called "People for Responsible Government," Minnesota Women
    Lawyers, Lavender Magazine and the Minnesota Family Council. Such groups can
    easily bring pressure to bear in judicial elections.46 Two 1996 examples from judicial
    elections in other states bear this out. In Nebraska, Supreme Court Judge David
    Lanphier was defeated in a retention vote after authoring an opinion striking down
             During the 2000 campaign, two Supreme Court justices received between eight
    and fifteen percent of their total contributions from one plaintiff's law firm, which had
    previously been accused of attempting to buy elections with tobacco settlement money.
    Candidate Filings, Minnesota Board of Campaign Finance & Public Disclosure (2000).
    Canon 5 does not solve electoral problems, but merely selectively sweeps them under
    the rug.
    Nebraska's term limits law. His defeat, largely over the opinion, appears to have been
    engineered by a single-issue organization named "Citizens for Responsible
    Government." Similarly, Justice Penny White of the Tennessee Supreme Court was
    hounded out of office after voting with the majority to remand a death penalty case.
    The influence of such groups featured prominently in the testimony when the Minnesota
    Supreme Court held a hearing on whether to expand Canon 5 to exclude all such
    groups, rather than just political parties.
           The court leaves single-issue political parties such as the Twin-Cities Area New
    Party, which brought the suit in Timmons, entirely out in the cold. Many smaller
    parties including the Right to Life Party, the Natural Law Party, and until this recent
    election cycle, the Green Party, tend to focus on only a few key issues. These groups
    are distinguished from advocacy groups only by having exercised their right to organize
    as a political party. Under the court's opinion, having done so will cost them their
    rights to engage in election-related speech and association.
            If the Minnesota Supreme Court is truly worried about the appearance of or
    actual judicial bias, it ought to be concerned with more than political parties. By
    ignoring other political groups, Canon 5 leaves unregulated sources of funding,
    endorsements and pressures, which, if the court's concerns about judicial partiality are
    justified, do much more to undermine public perception than political parties. In fact,
    barring political party involvement will most likely make the system less tenable. By
    removing the only organizations that endorse candidates across a spectrum of issues,
    voters are left with only the shrill voice of narrow advocacy coming from special
    interest groups. Because the canons do not achieve their stated goals, they cannot be
    considered necessary, and therefore cannot pass strict scrutiny.
            Not only must a restriction on speech be necessary, it must also be narrowly
    tailored to use the least restrictive means in achieving its stated goal. Burson, 504 U.S.
    at 198. In doing so, a restriction must suppress as little protected activity as possible.
    "If a less restrictive alternative would serve the Government's purpose, the [state] must
    use that alternative." Playboy, 120 S. Ct. at 1886. Canon 5 runs rampant through acres
    of protected speech, and in doing so eschews many less restrictive options.
           If the court's goals do accurately reflect the actual interests of the people of
    Minnesota, they could be most efficiently and effectively achieved by amending the
    Minnesota Constitution to allow for the appointment, rather than the election of judges.
    This would achieve the court's interests in one stroke, without requiring the suppression
    of a single protected word. In undertaking its "narrowly tailored" analysis, the court
    does not discuss this option. Of course, doing so would reopen for the court the sticky
    problem of Minnesota's having repeatedly expressed a preference for elected judges.
           Compliance with the First Amendment does not require so radical a change, for
    other states have successfully adopted a variety of mechanisms to insulate their judges
    from the pressures of elected office without impermissibly trenching on protected
    constitutional rights. Various states, including Nebraska and Tennessee, have adopted
    the Missouri Plan, discussed in Section I, whereby initially-appointed judges stand in
    periodic retention elections. Other states have used a more piecemeal approach,
    remedying specific problems. Some, for instance, alleviate the financial pressures
    faced by judges for whom the bench provides a livelihood by lengthening terms of
    office or by providing generous pensions in order to ease the transition back to private
    practice. Of course, Minnesota's failure to adopt any such scheme does not license the
    Minnesota Supreme Court to impose on judges and their putative supporters ethical
    canons abusive to their First Amendment rights.
           In sum, the court starts with a series of premises, many of which I accept. The
    Supreme Court of Minnesota has every right to regulate the state's bench and bar. It
    may protect its judiciary against both actual and the appearance of partiality. It may
    adopt measures tempered to do so in the ordinary course of its judiciary's conduct.
    However, that principle of judicial independence, as has been recognized by almost
    every court to take up the question, must give way before rights which, unlike the
    court's policy, have been constitutionally secured. Judges, whatever their differences
    from other officials, have no more right than others to avoid the rigors of public debate
    and public elections. No matter what the wisdom of placing the judiciary beyond the
    rigors of such activities, and no matter what selection method we as federal judges may
    prefer, the people of Minnesota have adopted an elective judiciary–a system which
    under our Constitution, must come replete with speech and associational rights.
           Because the court finds compelling an interest that is not shared by the people
    of Minnesota, trumps well established speech and associational rights with values not
    adopted by the citizens of the state, permits the erection of an unconstitutional self-
    insulating barrier between the judiciary and the public, and permits regulations that are
    neither necessary nor narrowly tailored, I must respectfully dissent.
          A true copy.