with whom ALARCON and FERNANDEZ, Circuit Judges, join, dissenting:
There is no question that political speech is at the core of the First Amendment, and is to be jealously protected. Eu v. San Francisco Democratic Cent. Comm., 489 U.S. 214, 109 S.Ct. 1013, 1020, 103 L.Ed.2d 271 (1989); Buckley v. Valeo, 424 U.S. 1, 14-15, 96 S.Ct. 612, 632, 46 L.Ed.2d 659 (1976). “Debate on the qualifications of candidates [is] integral to the operation of the system of government established by our Constitution.” Id. at 1, 96 S.Ct. at 612. Similarly, there is no question that all persons have the right to associate in political parties for the purpose of espousing their views. Tashjian v. Republican Party, 479 U.S. 208, 214, 107 S.Ct. 544, 548-49, 93 L.Ed.2d 514 (1986). Likewise the state’s power to regulate the time, place and manner of elections does not extinguish the state’s responsibility to be faithful to the First Amendment. Id. at 217, 107 S.Ct. at 550; Eu, 109 S.Ct. at 1019. I also agree that California’s ban on political party endorsements, opposition and support of candidates for nonpartisan elections burdens First Amendment rights, and is subject to strict scrutiny. First Nat’l Bank of Boston v. Bellotti, 435 U.S. 765, 786, 98 S.Ct. 1407, 1421, 55 L.Ed.2d 707 (1978).
However, I disagree that this ease is like any other, including Eu, upon which the majority relies.1 Article II, § 6(b) of the California Constitution, which was designed to keep party politics out of elections for these nonpartisan offices, must therefore be analyzed afresh.
The restriction on party endorsements in non-party elections goes to the heart of the structure of local, educational and judicial governance, which California has determined shall be nonpartisan in form. Cal. Const, art. II, § 6(a); Cal.Elec.Code § 37. No one contests the constitutionality of this structure. Implementing it, the Legislature has enacted a number of statutes that insulate nonpartisan elections from the par*296ty apparatus.2 When the California Supreme Court ruled in 1984 that existing law did not clearly prevent political parties from endorsing and supporting candidates for nonpartisan office,3 the people adopted a constitutional amendment to make clear their intent that no political party or party central committee should endorse, support or oppose a candidate for nonpartisan office.
The question in this case is whether California has a compelling interest in the structure of its nonpartisan government, such that prohibiting political party endorsements, opposition and support of candidates for nonpartisan office is the least restrictive means by which it can be achieved.4 The answer, I submit, requires us to consider the unique nature of this interest in our federal system, and to constrain ourselves in the interests of comity to interpret § 6(b) so as to avoid unconstitutionality, if possible.
California’s interest in this case is of extraordinary importance. It is rooted in the United States Constitution, which assures the citizens of each state the right to determine the structure of their own government so long as it is in a republican form.5 Because of its constitutional underpinning, the state’s interest in this case is particularly compelling. Such an interest requires a different balancing of the burdens on plaintiffs’ First and Fourteenth Amendment rights, the state’s justification for the burdens imposed, and the restrictiveness of the means available to accomplish the goal of preserving the nonpartisan structure of government by keeping nonpartisan offices and nonpartisan elections, nonpartisan, than either the district court or the majority has undertaken.
It is not surprising that the state’s interest in the structure of its government has not been taken into account, because the City argued a different interest (avoiding corruption) in its papers. Based on the argument that was made, the district court granted summary judgment. On one view this was consistent with Celotex Corp. v. Catrett, 477 U.S. 317, 322, 106 S.Ct. 2548, 2552, 91 L.Ed.2d 265 (1986), because the City produced no evidence to show that the state’s interests are endangered by party participation in nonpartisan elections or that § 6(b) is the least restrictive means for serving its interests.6 However, in light of *297the significance of this issue to the people of the State of California, and the overriding importance of the interest at stake, it is appropriate that the parties, and the court, have the benefit of a more searching analysis than has occurred in this case.
When, as here, both the record and the issues are inadequately developed, it is within the province of the district court to decline a summary judgment.7 In this very special situation, I would reverse and remand to permit full consideration of the state interest I view as “super”-compelling.8
I
A
“A Republic, madam, if we can keep it” was Benjamin Franklin’s response when asked what form of government the new Constitution created. A republican government is “a government which derives all its powers directly or indirectly from the great body of the people.” The Federalist No. 39, at 112 (J. Madison)(R. Fairfield ed. 1981). It is “constructed on this principle, that the supreme Power resides in the body of the people.” Chisholm v. Georgia, 2 U.S. (2 Dall.) 419, 457, 1 L.Ed. 440 (1793) (opinion of Wilson, J.). As put by the Court in In re Duncan, 139 U.S. 449, 461, 11 S.Ct. 573, 577, 35 L.Ed. 219 (1891), “the distinguishing feature” of a republican form of government “is the right of the people to choose their own officers for governmental administration, and pass their own laws.”9
Article IV, § 4 of the United States Constitution declares that “[t]he United States shall guarantee to every State in this Union a Republican Form of Government.” As described by one commentator, the guarantee clause
restricts the federal government’s power to interfere with the organizational structure and governmental processes chosen by a state’s residents.... In order to ensure that state and local governments remain responsive to their constituents ... citizens must have the power to choose the governmental forms that work best for them. The guarantee clause, therefore, grants states control over their internal governmental machinery.
Merritt, The Guarantee Clause and State Autonomy: Federalism for a Third Century, 88 Colum.L.Rev. 1, 41 (1988).10
*298California has chosen to make its judicial, educational and municipal offices nonpartisan. The guarantee clause requires that we “respect the state’s most fundamental structural choices as to how its people are to participate in their own governance.” L. Tribe, American Constitutional Law 398 (2d ed.1988). Put another way, the nonpartisan form of government represents a state interest of constitutional dimension. For that reason, it must be accorded weight of the most compelling sort.
B
The district court recognized only a “legitimate” interest in maintaining the nonpartisan character of local, judicial, and school elections. Against that level of interest the court found that § 6(b) was “not a narrowly drafted enactment designed to advance such interest by the least drastic means.” 11 Judge Zirpoli postulated that the state could safeguard its interests by providing for nonpartisan methods of nomination (which had already been done by statute), and by controlling partisan activities of the candidates, for example, by prohibiting candidates from soliciting or collecting campaign funds from political parties.
There are two problems with this analysis. It misapprehends the nature of the state’s interest, and in so doing, undervalues it. Second, it hypothesizes less restrictive means without considering their relationship to the interest at stake or whether they work.12
The interest at stake is California’s choice of a nonpartisan structure of government for judicial, school and municipal office. I believe it is “super”-compelling. In determining whether the limitations embodied in § 6(b) sufficiently relate to that interest and are narrowly tailored to serve it, we may look to precedent,13 determine that the answer is “inherently persuasive” or “self-evident,” 14 or base a decision on legislative findings or record evidence.15
The majority looks to precedent and finds Eu v. San Francisco Democratic Cent. Comm, persuasive. However, it is not possible to fit this case into that pigeonhole. Eu involved different issues and a different interest.
Eu had to do with the system by which a political party selects its own candidates for partisan office. The purpose of a primary election within this structure is to settle the differences within the party, so that the majority views of the party members are carried forward to the general election. The partisan primary election is designed to ensure that party nominations *299are within party lines, thus preserving the integrity of the party. In Eu we found that regulation of internal party affairs burdens the right of political parties to govern themselves as they see fit, because a partisan primary is by definition theirs.
By the same token, a nonpartisan election is by definition not theirs. The purpose of an election within the nonpartisan structure has nothing to do with settling the internal divisions within a political party. Rather, a nonpartisan structure abandons the political party as a conduit for the electorate’s views. Nonpartisanship envisions “direct representation of citizens rather than indirect representation through parties as intermediaries.” Note, Local Nonpartisan Elections, Political Parties and the First Amendment, 87 Colum.L.Rev. 1677, 1679 (1987).
The interest asserted by the state was quite different as well: protecting political parties from disruption that might result if they were to make preprimary endorsements. “This interference with the parties’ and their members’ First Amendment rights was not justified by a compelling state interest for a State has a legitimate interest ‘in orderly elections, not orderly parties.’ ” Eu, 109 S.Ct. at 1019 (quoting our opinion, 826 F.2d at 831).
The district court, on the other hand, recognized that Eu was not dispositive16 but it proceeded by way of assumption, and was also apparently of the view that the issue was self-evident because it granted summary judgment without a record. Much as I respect Judge Zirpoli, I do not think the result in this case is self-evident. If, however, my only option were to decide what is self-evident, I would say that nonpartisan means no party and there is accordingly no less restrictive means of filling a no-party office, than a no party election. But I believe reasonable judges can reasonably differ on a sensitive issue such as this. It is therefore unsatisfactory that we should indulge our own instincts or experience.
There is a framework within which an objective analysis could be conducted that would be commensurate with the importance of this case. A logical starting point is how the Legislature has filled out the skeleton of the state’s nonpartisan structure for judicial, school and municipal government. A “partisan office” means “an office for which a party may nominate a candidate,” while a nonpartisan office is one whose holders have not been nominated in a partisan primary or identified on the ballot as affiliated with a political party, and who have not been elected on a partisan ballot.17 In 1986 the people of the state amended Section 6 of the California Constitution to reaffirm that “[a]ll judicial, school, county, and city offices shall be nonpartisan” (§ 6(a)) and to provide that “[n]o political party18 or party central committee may endorse, support, or oppose a candidate for nonpartisan office” (§ 6(b)).19 *300This was done to continue a “long-understood ban on partisan electioneering in local and judicial elections”20 which the California Supreme Court had concluded was not enforceable under the Constitution or statutes as then drafted.
Next, it makes sense to assess the burden § 6(b) imposes on political parties.21 It does not prevent individuals or groups of individuals (such as the California Democratic Council or “Democrats for Bradley”) from making endorsements. Nor does it preclude political parties from speaking out on issues. For example, the Republican Party could support the building of a jail in Fresno and the Democratic Party could oppose it, even though candidate A for mayor was for the building of the jail and candidate B was against it. Narrowly construed, § 6(b) does not inhibit a political party from calling attention to a candidate’s record and experience and qualifications. Further, the identification of candidates with particular parties plays much less of a role in the nurturing of an informed electorate for local, school and judicial offices, than for national (or state) elections.22 To the extent that the natural *301focus of a ballot-qualified party is on issues of broad application and on partisan offices such as the Presidency, the Congress, or the Governorship, party labels provide scarcely any “shorthand designation of the views of party candidates on matters of public concern” in connection with nonpartisan, specialized local elections.23 This is especially so since candidates for nonpartisan office neither run as a party member, nor are labelled as such on the ballot.24
These factors should then be considered in light of the “super”-compelling importance of the state’s interest in its chosen form of government, to determine if § 6(b) is narrowly tailored, whether there are less restrictive means that serve that interest, and whether alternatives work. The majority concludes that “the mere fact that § 6(b) targets the collective rather than the individual voices of party members does not suffice to render it ‘precisely drawn.’ ”25 I agree.26 But the fact that it targets the collective voice only with respect to endorsements for nonpartisan offices may render it drawn as precisely as it can be,27 for to preclude party endorsements in nonparty elections is the flip side of a candidate’s running for nonpartisan office without party identification. It also conforms exactly to the California Election Code’s definition of partisan office as one “for which a party may nominate a candidate,” 28 and conversely to that of a nonpartisan office, for which a party may not nominate a candidate.
The “less drastic” possibilities posited by the district court (and adopted by the majority) may be links in the chain, but cannot serve the state’s structural interest if there is any missing link. For example, Judge Zirpoli suggested that nonpartisan methods of nomination would be less restrictive. Clearly these serve the applicable interest. However, if nonpartisan methods of nomination were not supplemented by prohibiting party endorsements, the track to nonpartisan office would be virtually indistinguishable from the track to partisan office. Candidates are nominated for nonpartisan office by petition which may not disclose their party affiliation; those for partisan offices are nominated through partisan primaries. There would thus be no material difference between a nonpartisan method of nomination, with party endorsements as Judge Zirpoli suggests, and a partisan method of election.29 Therefore it does not appear that the means of relying entirely *302on nonpartisan nomination methods, without the limitation in § 6(b), would work.30 The district court’s other suggestion, prohibiting candidates from soliciting or collecting campaign funds from political parties, might serve the state’s interest in corruption, but would only incidentally advance its interest in structure.31
Because the district court was considering corruption, it measured neither alternative against a structural interest that is “super”-compelling. While these or other means might work, there is no basis in the record for reaching a reasoned conclusion that they will or won’t.
Instead, the final step in the analysis ought to be whether the state’s interest in the nonpartisan structure of government, constitutionally recognized in the guarantee clause, warrants a burden of the sort imposed on the First and Fourteenth Amendment rights of a political party. Both the Congress and the Supreme Court have acknowledged that there is a difference of constitutional dimension between political activity by political parties and political activities of other organizations or individual political actors. In Civil Serv. Comm’n v. Letter Carriers, 413 U.S. 548, 565, 93 S.Ct. 2880, 2890, 37 L.Ed.2d 796 (1973), the Supreme Court upheld the Hatch Act, which prohibited federal employees from involvement in political activity, against a First Amendment challenge. In so doing, it emphasized the right of government employees to be involved with everything except political parties.
Conversely there is nothing about the state’s interest in the nonpartisan form of its local, school and judicial government that implicates the rights of anyone but a ballot-qualified political party from endorsing, opposing or supporting candidates for nonpartisan office. The same goes for the rights of a political party to espouse views on controversial issues, even when those views correspond to views held by a particular candidate.32
It was clear to the Court in Letter Carriers that the danger to proper government *303functioning presented by political party involvement was different in kind from the danger presented by the involvement of other organizations, ideas, considerations, or even by political actors of an independent stripe. In other words, in the political context nonpartisanship does not mean that an officeholder adheres to no causes or factions, or even that the officeholder is not committed to the tenets or goals of a political party. It does mean that there should not be activity that gives (or seems to give) a party any power whatever over the holder of a nonpartisan office.
Accordingly, there is no reason inhering in the state’s interest in the structure of its government that inhibits California from justifying its restriction on endorsements in this case.
II
Because the state’s interest in the structure of nonpartisan government is so compelling, I do not believe it is necessary to reach the corruption issue. The majority does, but its approach is flawed in a way that affects its analysis of the state’s justification for § 6(b). The majority focuses solely on financial corruption of the sort addressed in Buckley, FEC v. National Conservative Political Action Comm., 470 U.S. 480, 106 S.Ct. 1459, 84 L.Ed.2d 455 (1985), and Austin v. Michigan Chamber of Commerce, — U.S. -, 110 S.Ct. 1391, 108 L.Ed.2d 652 (1990), and purports to distinguish the kind of corruption “decried” by California in this case.33 However it is clear that corruption can occur from political as well as financial debt. See, e.g., Letter Carriers, 413 U.S. at 565, 93 S.Ct. at 2890 (partisan political activities by government employees poses danger to “fair and effective government”); Bellotti, 435 U.S. at 788 n. 26, 98 S.Ct. at 1422 n. 26 (importance of government interest in preventing corruption of elected representatives through creation of political debts “has never been doubted”); Buckley, 424 U.S. at 26-27, 96 S.Ct. at 638. Letter Carriers is closely on point, implicitly recognizing that partisan politics may have a corrupting influence on a system designed to be nonpartisan.34 The interest California actually asserted in this case is not that endorsement by a political party corrupts the voter by influencing him or her to cast a ballot one way or the other;35 rather, it is that endorsement corrupts the nonpartisan candidate and officeholder in the performance of his duties.36 Along *304with the state’s interest in preventing corruption of that sort, “[preservation of the individual citizen’s confidence in government is equally important.” Bellotti, 435 U.S. at 789, 98 S.Ct. at 1422.
Even if a limitation on speech does keep information from the voters, Buckley teaches that it is permissible if there is a compelling state interest balanced against a marginal burden on First Amendment rights. Applying Buckley to § 6(b) suggests that endorsements are more analogous to large campaign contributions than to independent expenditures; unlike independent campaign expenditures, endorsements are prearranged and coordinated.37 Unlike a contribution, however, endorsement itself creates the corruption, not the amount of endorsement. The state therefore has a justification for banning endorsements that neither the district court nor the majority recognizes.
The majority also relies on Austin in concluding that the kind of corruption at issue here is different from that reached elsewhere. While the relevance of Austin to this case is by no means clear, it is not unambiguous. However tempting it may be to whitewash its holding or pick and choose what it says, the Court can fairly be read to suggest that the state’s interest in avoiding corruption is reasonably furthered by restricting speech, in the form of expenditures, by organizations which have amassed political “war chests” with the aid of legal advantages conferred by the state. As ballot-qualified parties have that status by virtue of statute in California, it could be said that they, too, have amassed their clout with the aid of legal advantages conferred by the state such that, in the context of a nonpartisan election, their “war chest” would have a correspondingly corrosive effect.38
Ill
We have an obligation to construe any law, if fairly possible, to avoid raising doubts of its constitutionality. New York v. Ferber, 458 U.S. 747, 769-70, 102 S.Ct. 3348, 3361, 73 L.Ed.2d 1113 (1982); Letter Carriers, 413 U.S. at 571, 93 S.Ct. at 2892. However, neither the district court nor the majority tries to do so.
The issue in this case is whether the City of San Francisco erred in not permitting party endorsements to appear on the official San Francisco Voter Pamphlet, in reliance on § 6(b). It is not necessary to construe § 6(b) as the majority does, as “impospng] a total ban on any partisan gesture of support for or opposition to a candidate.”39
Not even plaintiffs assert that the ban encompasses “any partisan gesture of support.” Plainly it does not: parties are free to speak out on issues and relate those issues to a candidate; partisan sub-groups are free to do anything they want in as partisan a way as they choose.
More significantly, § 6(b) can be construed to prohibit only endorsements, and not to extend to communications about the *305candidates’ qualifications, position on the issues, experience or platform. It can also be construed as applying only to public endorsements, or endorsements on the official voter information pamphlet. In any event it could be construed not to apply to recommendations made to members. See, e.g., United States v. CIO, 335 U.S. 106, 68 S.Ct. 1349, 92 L.Ed. 1849 (1948) (declining to construe § 313 of the Corrupt Practices Act to prohibit corporations and unions from expressing views on candidates directly to members). To the extent § 6(b) provides that a political party may not “support or oppose” a candidate for nonpartisan office, it may be too vague or too encompassing to be enforced. If excision of those terms will narrow the statute so that it can be constitutional, the necessary surgery should be performed.40 Brockett v. Spokane Arcades, Inc., 472 U.S. 491, 502, 105 S.Ct. 2794, 2801, 86 L.Ed.2d 394 (1985).
IV
It is particularly troubling in this ease that there is virtually no record.41 There is, for example, no evidence showing whether the relative voice of political parties has been unduly significant or influential in nonpartisan elections where endorsements have occurred. Nor is there any evidence bearing on feasibility of alternate means to aid the state’s interest. The absence of a record leads inexorably to judges judging on their own instinct or experience.
This case presents an issue of fundamental importance to the State of California, and concerns an interest of the people that lies within the heartland of the Constitution. That being so, we should not conclude that California’s concern is not compelling enough, without searching inquiry and an adequate record.
I would, therefore, reverse and remand.42
. Eu had to do with whether the state could interfere with the internal processes of a party by preventing its governing body from endorsing, supporting or opposing a candidate for nomination to partisan offices in the party primary. It has nothing to do with the issue in this case, which concerns the nonpartisan structure of the judicial, municipal and educational units of government and the role of a partisan political party in nonpartisan elections for nonpartisan offices.
As Judge Norris wrote for our court in Eu, "[i]t is the right of the party, not the State, to decide how the party shall be governed." 826 F.2d 814, 827 (9th Cir.1987). The Supreme Court recognized the same point: ‘‘[a] primary is not hostile to intraparty feuds; rather it is an ideal forum in which to resolve them.” 109 S.Ct. at 1022. Based on this principle it followed that the state may not enact laws "‘to prevent the parties from taking internal steps affecting their own process for the selection of candidates.’" Id. (quoting Tashjian, 479 U.S. at 224, 107 S.Ct. at 553-54) (emphasis added).
. For example, Cal.Elec.Code § 6401.5 provides that declarations of candidacy and other nomination papers for nonpartisan office may not refer to party affiliation; the name of the party to which a nonpartisan candidate belongs may not appear on the ballot (§ 10200.5); a voter may cast a ballot for a candidate for such an office without regard to party affiliation (§ 10214); and partisan and nonpartisan offices are listed in separate columns of the ballot (§ 10214). Political parties are also prohibited from nominating a candidate for a nonpartisan office (§ 36). These provisions have the effect of creating a two-track process of candidate selection and election, paralleling the structural difference between nonpartisan and partisan offices.
For a comprehensive and detailed history of these statutes and the nonpartisan structure in California see Unger v. Superior Court, 37 Cal.3d 612, 209 Cal.Rptr. 474, 692 P.2d 238 (1984), and the original panel opinion by Judge Trott in Geary v. Renne, 880 F.2d 1062 (9th Cir.1989).
. Unger, 37 Cal.3d 612, 209 Cal.Rptr. 474, 692 P.2d 238.
. Although freedom to discuss matters of public concern is fundamental to our national heritage and the system of government ordained by our Constitution, the protection is not absolute and restraints are permitted for appropriate reasons. See, e.g., Austin v. Michigan Chamber of Commerce, — U.S.-, 110 S.Ct. 1391, 108 L.Ed.2d 652 (1990) (expenditure limitation); Buckley v. Valeo, 424 U.S. 1, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976) (contribution ceilings); Storer v. Brown, 415 U.S. 724, 94 S.Ct. 1274, 39 L.Ed.2d 714 (1974) (orderly elections); Munro v. Socialist Workers Party, 479 U.S. 189, 107 S.Ct. 533, 93 L.Ed.2d 499 (1986); Jenness v. Fortson, 403 U.S. 431, 91 S.Ct. 1970, 29 L.Ed.2d 554 (1971) (ballot access).
. U.S. Const, art. IV, § 4.
. The government bears the burden of showing a subordinating interest that is compelling, and of proving that a regulation is a precisely drawn means of serving its interests. Consolidated Edison Co. v. Public Serv. Comm'n, 447 U.S. 530, 100 S.Ct. 2326, 65 L.Ed.2d 319 (1980); San Francisco County Democratic Cent. Com. v. Eu, 826 F.2d at 834. At the same time, the district court did not consider the "legislative findings” set forth in the California Ballot Pamphlet for the June 3, 1986 primary election (at which the voters enacted the amendment to Section 6) in determining that there were no triable issues of fact.
. In re Rigden, 795 F.2d 727, 731-32 (9th Cir.1986).
. The state’s interest in the structure of its government is "super" compelling because of its constitutional significance. The guarantee clause commits the United States, including the courts of the United States, to protect the structure chosen by the citizens of each state. In this sense, structure is more compelling than corruption, for example, which is an evil that the state has a "compelling” interest in avoiding but which lacks the constitutional underpinnings of the state’s form of government. Having recognized this, the state's interest in this case is to be weighed in accordance with accepted First Amendment jurisprudence.
. This power reposing in the citizens of each state is also basic to our federal system. "The federal Constitution forms a happy combination [between a large republic and a small one]; the great and aggregate interests being referred to the national, the local and particular to the State legislatures." The Federalist No. 10, at 22 (J. Madison)(R. Fairfield ed. 1981). See also, Oregon v. Mitchell, 400 U.S. 112, 125, 91 S.Ct. 260, 265, 27 L.Ed.2d 272 (1970)(Black, J.)("[n]o function is more essential to the separate and independent existence of the States and their governments than the power to determine within the limits of the Constitution ... the nature of their own machinery for filling local public offices’’); FERC v. Mississippi, 456 U.S. 742, 790 n. 28, 102 S.Ct. 2126, 2153 n. 28, 72 L.Ed.2d 532 (1982) (O’Conner, J., concurring in part and dissenting in part)(quoting I. Silone, The School for Dictators 119 (W. Weaver trans.): "[a] regime of freedom should receive its lifeblood from the self-government of local institutions. When democracy, driven by some of its baser tendencies, suppresses such autonomies, it is only devouring itself.... [I]f the central government’s representative runs the city and the province, ... you can no longer speak of democracy”).
.Merritt makes the further point that "the tenth amendment underscores the need to enforce the federalism principle found in the guarantee clause. The structure of the Constitution, which clearly contemplates a federal system, likewise affirms the importance of interpreting the guarantee clause as a restraint on federal power.” Id. at 2 n. 8.
. Geary v. Retine, No. CV-87-4724, at 4 (Order filed April 27, 1987).
. "It is not helpful to ponder what is 'less restrictive' without thinking about the functions the rules serve and how these functions would be affected by ‘less restrictive alternatives.’ Every alternative is more restrictive than some other one." Stevenson v. State Bd. of Elections, 794 F.2d 1176, 1180 (7th Cir.1986) (Easterbrook, J., concurring)(emphasis in original).
. See, e.g., Buckley, 424 U.S. at 17-19, 96 S.Ct. at 633-35; Bellotti, 435 U.S. at 790, 98 S.Ct. at 1424; EEC v. National Right to Work Comm., 459 U.S. 197, 205-07, 103 S.Ct. 552, 558-59, 74 L.Ed.2d 364 (1982); Anderson v. Celebrezze, 460 U.S. 780, 103 S.Ct. 1564, 75 L.Ed.2d 547 (1983) (no precedent on point).
. See, e.g., Bellotti, 435 U.S. at 790, 793, 98 S.Ct. at 1424, 1425.
. Compare Bellotti, 435 U.S. 765, 98 S.Ct. 1407 (without showing that relative voice of corporation has been overwhelming or even significant in influencing referenda, or that there has been any threat to the confidence of the citizenry in government, argument that participation would exert undue influence is not persuasive) and Eu, 109 S.Ct. at 1019 (state had failed to submit " ‘a shred of evidence’ ” that regulations of party internal affairs helped minimize party factionalism) (quoting Ninth Circuit opinion, 826 F.2d at 833, in turn quoting Civ. No. C-83-5599 (N.D. Cal. May 3, 1984) with Jenness, 403 U.S. 431, 91 S.Ct. 1970); American Party of Texas v. White, 415 U.S. 767, 782, 94 S.Ct. 1296, 1306, 39 L.Ed.2d 744 (1974); Storer v. Brown, 415 U.S. at 736, 94 S.Ct. at 1282; Munro v. Socialist Workers Party, 479 U.S. at 194-95, 107 S.Ct. at 537-38 (states are not required to make a particularized showing of the existence of voter confusion, ballot overcrowding, or the presence of frivolous candidacies prior to imposition of reasonable restrictions on ballot access; otherwise, state’s political system would have to sustain some level of damage before legislature could take corrective action).
. Order Denying Defendants’ Motion to Vacate Judgement, Civ. No. C-87-4724 (N.D.Cal. June 9, 1988), 1-2.
. See n. 2, supra. These statutes could be taken to define what a "nonpartisan" election means. On the other hand, they could reasonably be characterized as merely the procedures for a nonpartisan election that the state has chosen so far. As such they would neither be definitive, nor indicative of how much baggage a nonpartisan election can carry and still be nonpartisan. There is no basis in the record in this case to make this kind of determination, one way or the other.
. Because "political party” is elsewhere defined to refer to ballot-qualified organizations, the parties affected are the Democratic Party (Cal. Elec.Code §§ 8500-8945), the Republican Party (§§ 9000-9510), the American Independent Party (§§ 9600-9745), and the Peace and Freedom Party (§§ 9750-9855).
.While it is irrelevant for constitutional purposes that the voters rather than a legislative body enacted § 6(b), Citizens Against Rent Control v. Berkeley, 454 U.S. 290, 295, 102 S.Ct. 434, 436, 70 L.Ed.2d 492 (1981), in this particular case it is not without significance that the decision to limit party endorsements in nonpartisan elections was made by the very people who are the members and the constituents of the parties whose speech was limited. Members of the state central committees are indirectly elected by the people. See Cal.EIec.Code § 8660. Thus § 6(b) was not the product of a representative body acting out of paternalistic or preservationist instincts, but of the people themselves, burdening their own instrumentalities. See Bellotti, 435 U.S. at 791 n. 31, 98 S.Ct. at 1424 n. 31 *300C‘[g]°vernment is forbidden to assume the task of ultimate judgment, lest the people lose their ability to govern themselves”); cf. Buckley, 424 U.S. at 93-97, 96 S.Ct. at 670-72 (state may not "unnecessarily burden either a minority party’s or an individual candidate's ... interest in the continued availability of political opportunity”) (citations omitted); American Party of Texas v. White, 415 U.S. at 780-81, 94 S.Ct. at 1305-06 (state must afford minority political parties a real and essentially equal opportunity for ballot qualification).
.Argument in Favor of Proposition 49 (to amend Section 6), California Ballot Pamphlet, Primary Election, June 3, 1986, p. 26. See Lee, The Politics of Nonpartisanship 102-104 (1960), and discussion in Unger, 37 Cal.3d at 616, 209 Cal.Rptr. 474, 692 P.2d 238.
According to Lee’s survey, there was “public or openly visible activity” on the part of county committees, local political clubs, party officials or representatives in local city, county or school elections in only 25% of the counties surveyed. Lee at 102-103, Table 32. Whether there was any involvement in judicial elections even in those counties is unclear.
Regardless, it is incorrect to infer from Lee, as the majority appears to do, that political parties endorsed candidates for nonpartisan office in 25% of the counties prior to 1986. Majority opn., supra, at 282 ("Nevertheless, because the legal status of such endorsements was unclear, political parties did not endorse candidates in 75% of California counties"). Table 32, from which that statistic is drawn, concerns "public or openly visible activity," not endorsements. Table 33, which shows the type of political activity reported on the survey, indicates that public endorsement of candidates occurred in 4 counties (about 10%) — although there is no way of telling from the survey whether the occurrence was singular, sporadic or repetitive.
Referring to the Lee survey, or relying on it to suppose that party endorsements caused no harm to the nonpartisan structure before § 6(b) was adopted in 1986, illustrates why it is risky to make decisions based on assumptions without an evidentiary basis. While Lee is no doubt an able scholar, neither the methodology nor significance of the results he reports has been tested by cross-examination.
. When deciding whether laws that impinge on associational freedoms violate the First Amendment:
[A] court ... must first consider the character and magnitude of the asserted injury to the rights protected by the First and Fourteenth Amendments that the plaintiff seeks to vindicate. It then must identify and evaluate the precise interests put forward by the State, as justifications for the burden imposed by its rule. In passing judgment, the court must not only determine the legitimacy and strength of each of those interests: it also must consider the extent to which those interests make it necessary to burden the plaintiff’s rights.
Tashfian, 479 U.S. at 214, 107 S.Ct. at 548-49 (quoting Anderson v. Celebrezze, 460 U.S. at 789, 103 S.Ct. at 1573). While the restrictions embodied in § 6(b) implicate both First and Fourteenth Amendment rights, Eu, 109 S.Ct. at 1020, the magnitude of the injury has to be measured against both the strength of the interest, and the alternatives available. The district court did not touch these bases because of its assumption that the state’s interest was corruption and that alternative means would be less restrictive.
. States have a greater interest in regulating statewide or local elections than national elections. Anderson, 460 U.S. at 795, 103 S.Ct. at 1573. Indeed, one purpose of nonpartisan elections is to insulate state and local elections from the national issues that are the natural focus of national political parties. See Note, Local Nonpartisan Elections, Political Parties and the First Amendment, 87 Colum.L.Rev. at 1679-81.
It is no answer to suggest, as did the dissent in the original panel opinion, that ”[t]here may be no Republican, or Democratic, or Libertarian way to build a city jail, but there may very well be Republican, Democratic or Libertarian positions on whether a jail should be built and who should be taxed for it.” 880 F.2d at 1084. *301There is nothing about § 6(b) which would impede the Republican Party, or the Democratic Party, or the Libertarian Party, or the Peace and Freedom Party from expressing their views on such matters.
. See Tashjian, 479 U.S. at 220, 107 S.Ct. at 551-52 (“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").
. Indeed Lee’s survey includes an appraisal which shows that party activity helped in six cities where both Republican and Democrats were active, and made no difference in three; when only Democrats were active, it helped in two, made no difference in one, and hurt in two. Lee, The Politics of Nonpartisanship 104 (Table 34).
. Majority opn., supra, at 284-85.
. As Bellotti says, "[t]he inherent worth of the speech in terms of its capacity for informing the public does not depend upon the identity of its source, whether corporation, association, union, or individual." 435 U.S. at 777, 98 S.Ct. at 1416.
. The ban on support and opposition is more problematic. See discussion, infra at 304-05, and n. 40.
. Cal.Elec.Code § 36.
. In other words, if Diane Feinstein, John Van de Kamp, and Pete Wilson were running for mayor, and if the Democratic Party endorsed Van de 'Kamp and the Republican Party endorsed Wilson, for all practical purposes the process would parallel the partisan track. The only difference would be that in the case of the nonpartisan track, the decision about which candidate would carry the party banner would be made by the party, not by the people as is their prerogative in the partisan track. The irony is, this would allow political parties even greater power in the nonpartisan track than in the partisan. That not only runs counter to what the voters intended when they amended Section 6 of the California Constitution, but would effectively destroy the nonpartisan system — and that runs counter to the intent of Article IV, § 4 of the United States Constitution.
. That certainly was the judgment of supporters of the amendment to Section 6. See, e.g., an editorial in The Sacramento Bee, To Protect Non-Partisan Elections, which opined that "[t]he principal danger is that the decision [Un-ger ] opens the way ... to school boards whose members make curricular decisions according to the political orientation of the parties that support them and, worse, look toward advancement by the party as a reward for faithful service in 'non-partisan' positions.” August 13, 1985, at B6, col. 1 (final ed.) (editorial), quoted in 87 Colum.LRev. at 1681. See also California Ballot Pamphlet, p. 26 ("To assure that our courts will not be manipulated by political bosses, your yes vote on Proposition 49 [to amend Section 6] is absolutely necessary.”).
. A Columbia Law Review Note proposes other means which the author suggests are also less restrictive: one would be to increase party competition, thereby rendering candidates less vulnerable to manipulation because they would have alternative means of support; another, to appoint administrative local officials rather than elect them. 87 Colum.LRev. at 1698-99. The former relates primarily to the state's interest in corruption but would turn its interest in structure on its head; the latter would too, in that presumably those administrative personnel would be appointed by persons elected in partisan elections. Neither means comports with a nonpartisan structure, and that is a choice both the guarantee clause and the Tenth Amendment leave for the people to make.
.The state could not, for example, justify a ban on endorsements by special interest groups on the ground that government should be "non-special interest." The state’s interest in the structure of government goes to such things as whether it is to be partisan or nonpartisan, town meeting or city council, appointive or elective; the state could not rely on an interest stemming from the guarantee clause to justify a government based on religious beliefs or race or ideas. Such things have nothing to do with structure.
By the same token, if the state’s asserted interest is corruption, its justification for banning endorsements by political parties is weakened because the corrupting effect of endorsements or support by a special interest group may be essentially the same. However, clout of a special interest group has no relevance to the state’s interest in the structure of government. Special interest groups have no access to the ballot and their membership may cut across party lines. Their raison d’etre is to influence voters and officeholders issue by issue, sometimes in accord with the views of a political party, sometimes not. They are quintessentially non-party organizations. By contrast, political parties have ballot access, are comprised of persons of the same political persuasion, and exist to win elections and run governments. See Moore v. Panish, 32 Cal.3d 535, 186 Cal.Rptr. 475, 652 P.2d 32 (1982). They are quintessentially the opposite of non-party or nonpartisan.
. Majority opn., supra, at 283.
. The Hatch Act does not prohibit political activity in connection with nonpartisan elections or with questions "not specifically identified with a National or State political party." 5 U.S.C. § 7326. Thus, it specifically addresses the corrupting influence that political party organizations have on the administration of the law in a nonpartisan context.
. That would be indefensible. Bellotti, 435 U.S. 765, 98 S.Ct. 1407; J. Ely, Democracy and Distrust 111 (1980) (state cannot regulate political speech because of “a fear of how people will react to what the speaker is saying”). It would also be inconsistent with the intent of the people when they adopted § 6(b). For example, the proponents argued:
To assure that our courts will not be manipulated by political bosses, your yes vote on Proposition 49 is absolutely necessary.
WHO WOULD TRUST THE FAIRNESS OF TRIALS TO JUDGES WHO WERE CHOSEN— NOT BECAUSE THEY ARE IMPARTIAL— BUT BECAUSE THEY OWE ALLEGIANCE TO THE POLITICAL PARTIES WHICH GOT THEM ELECTED?
WHO WANTS TO RELY ON THE DECISIONS OF JUDGES WHO ARE CHOSEN— NOT BECAUSE THEY ARE WISE OR BECAUSE THEY KNOW THE LAW — BUT BECAUSE THEY HAVE PROMISED TO TOE THE PARTY LINE?
Californians do not want their judges to become beholden to political parties.
UNLESS YOU VOTE YES ON PROPOSITION 49, JUDGES MAY WELL BE INDEBTED TO PARTY BOSSES TO WIN ELECTIONS. THEIR JOBS WILL DEPEND ON IT.
Local officeholders support this amendment and are equally concerned that partisan electioneering will harm decision-making at the local level. They are concerned that the more they have to rely on money, help, and endorsements from political machines, the more they will owe the political machines.
Local officeholders do not want to have to check with the party bosses before they make decisions important to their constituents!
California Ballot Pamphlet at 26.
.Everyone knows, for example, that Mayor Bradley is a Democrat. That knowledge — or the communication of information that the Democratic Party or some group identifying itself *304with Democratic causes favors his election as Mayor — would not likely spoil anyone but the candidate. As the Mayor’s brief in this case points out, the problem is not “the impact that endorsements may have on voters’ choices but rather [their] indirect impact on elected officials’ independence from partisan political pressures.”
.Declaration of Richard Mountjoy, E.R. 72. (Although I have some concern about citing the Mountjoy Declaration because it arguably was not part of the summary judgment record, it was apparently reviewed in connection with the motion to vacate. In any event what he says is a matter of common sense: it is unlikely that a party endorsement would appear on the ballot information pamphlet unless the candidate had agreed to it.) Opposing or supporting a candidate, on the other hand, is closer to a restriction on independent expenditures, which Buckley declined to approve.
. See also California Medical Ass’n v. FEC, 453 U.S. 182, 201, 101 S.Ct. 2712, 2724, 69 L.Ed.2d 567 (1981). In upholding limitations on union and corporate campaign activity, the Supreme Court stated that the "differing structures and purposes” of entities “may require different forms of regulation in order to protect the integrity of the electoral process.” The Court further distinguished restrictions placed on individuals and unincorporated associations, on the one hand, and unions and corporations on the other.
. Majority opn., supra, at 285.
. To eliminate the ban on opposition and support would also decrease the appearance of prearrangement and coordination that Buckley found to be indicators of corruption. To the same extent, the risk of a direct quid pro quo is lessened. Limiting § 6(b) to endorsements would also obviate the problem of how to enforce a restraint on "opposing" or "supporting” on the margin, for instance, whether Republican Party opposition to Mayor Bradley’s campaign for Governor (a partisan office) would violate its obligation not to "oppose” him in his capacity as a candidate for, or officeholder of, the mayoralty (a nonpartisan office).
. The case comes to us in a curious posture. Relating to the issues presented in this case and in connection with the motion for summary judgment, the plaintiffs produced the affidavits of Arlo Hale Smith and Terence Falkner. These declarations went to standing, not to substantive issues. The city presented no evidence in opposition. Subsequently, the City submitted several declarations in connection with its motion to vacate, as did plaintiffs in opposition. In denying the motion to vacate, the district judge indicated that he had reviewed the motion but was not persuaded to change his decision. In that sense, the declarations are a part of the record. However it is unclear that they form any part of the basis of the district’s order granting partial summary judgment.
. See Mandel v. Bradley, 432 U.S. 173, 177-78, 97 S.Ct. 2238, 2241, 53 L.Ed.2d 199 (1977) (remanded "to permit further findings with respect to the extent of the burden imposed” by state electoral regulations).