I agree with the majority that the City and County of San Francisco’s (San Francisco) ban on write-in voting in its *184December 1999 runoff election is constitutional. I arrive at that result through a different route. The majority would overrule Canaan v. Abdelnour (1985) 40 Cal.3d 703 [221 Cal.Rptr. 468, 710 P.2d 268, 69 A.L.R.4th 915] (Canaan) “to the extent, but only to the extent, it is inconsistent with the views expressed” in the majority opinion. (Maj. opn., ante, at p. 183.) The extent to which Canaan is overruled by the majority is unclear. I disagree that Canaan should be overruled and would instead distinguish it from the present case, as explained below.
As the majority recognizes, the United States Supreme Court’s holding that a complete ban on write-in voting does not contravene the First Amendment of the United States Constitution (see Burdick v. Takushi (1992) 504 U.S. 428 [112 S.Ct. 2059, 119 L.Ed.2d 245] (Burdick)) does not compel the conclusion that such a ban, or even a more limited ban, would pass muster under the California Constitution’s free speech clause, article I, section 2, subdivision (a). The meaning of this clause is derived not only from the distinctive text and history of the clause, or from our present discernment of its purposes and contours, but also from previous judicial interpretations of the clause. Thus, our natural starting point for determining whether a write-in ban in runoff elections violates the free speech clause is Canaan, which considered a similar, although not an identical, question.
The Canaan court adopted into California law the federal constitutional balancing test for evaluating the constitutionality of election law articulated in Anderson v. Celebrezze (1983) 460 U.S. 780 [103 S.Ct. 1564, 75 L.Ed.2d 547] (Anderson). “ ‘Constitutional challenges to specific provisions of a State’s election laws . . . cannot be resolved by any “litmus-paper test” that will separate valid from invalid restrictions. [Citation.] Instead, a court must resolve such a challenge by an analytical process that parallels its work in ordinary litigation. It 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 plaintiffs rights. Only after weighing all these factors is the reviewing court in a position to decide whether the challenged provision is unconstitutional. [Citations.] The results of this evaluation will not be automatic; as we have recognized, there is “no substitute for the hard judgments that must be made.” [Citation.]”’ (Canaan, supra, 40 Cal.3d at p. 712, fn. omitted, quoting Anderson, supra, 460 U.S. at pp. 789-790 [103 S.Ct. at p. 1570].) Applying this test, the Canaan court then invalidated a scheme prohibiting *185write-in voting for the San Diego mayoral contest in a general election occurring five months after the primary election.
The majority does not disagree with the test, only its application. In their view, the Canaan court gives too much importance to the “expressive function” of voting and not enough importance to what the Burdick court identified as “ ‘[T]he function of the election process . . . “to winnow out and finally reject all but the chosen candidates.” ’ ” (Maj. opn., ante, at p. 181, quoting Burdick, supra, 504 U.S. at p. 438 [112 S.Ct. at p. 2066].) Because the Canaan court overvalued this expressive function, the majority concludes, it correspondingly undervalued a local government’s legitimate interest in having a winning candidate chosen by a majority of the voters. This analysis has an appealing simplicity. But as majority itself seems to recognize at points, it is overly simple.
Although the Canaan court, to be sure, articulated an interest in the “expressive function” of write-in voting, it addressed another concern more fundamental to the basic purpose of the electoral process. In Canaan, the petitioner sought to advance his write-in candidacy for the general mayoral election after one of the two candidates who had qualified for that election, the incumbent mayor, was indicted for numerous felonies. (Canaan, supra, 40 Cal.3d at pp. 708-709.) This predicament was central to the Canaan court’s holding: “If the candidate who has represented an individual’s interests and views is forced to withdraw from the campaign, alters his or her positions or is indicted for alleged felonies, [an] individual may feel compelled to become a candidate in order to fill the void. Rather than ‘doing violence’ to the election process, the availability of a write-in candidacy provides the flexibility to deal with unforeseen political developments and may help ensure that the voters are given meaningful options on election day.” (Id. at pp. 718-719, italics added.) Stated in other terms, “it is possible that the ban on write-in voting in this case may have prevented the election of a candidate who enjoyed the support of the majority of . . . voters.” (Id. at p. 722, italics omitted.)
While the relative importance of the “expressive function” of voting is debatable, the importance of preventing this kind of widespread disenfranchisement is not. Moreover, as the majority recognizes, the very election we are considering, the San Francisco mayoral election of 1999, is a testimony to the power of a write-in candidacy. One of the candidates who qualified for the runoff election, Tom Ammiano, was himself a write-in candidate.
Thus, one of the essential concerns of Canaan, as it was of the dissent in Burdick, is not just that voters will be unable to express dissident views *186through write-in votes, but that potentially large numbers of voters will be effectively disenfranchised, deprived of “the opportunity to cast a meaningful ballot.” (Burdick, supra, 504 U.S. 428, 447 [112 S.Ct. 2059, 2070] (dis. opn. of Kennedy, J.).) In Burdick, the dissent identified this threat of disenfranchisement as coming from Hawaii’s total ban on write-in voting combined with other ballot restrictions and the one-party system then virtually in effect in that state. (Id. at pp. 446-448 [112 S.Ct. at pp. 2070-2071].) In Canaan, the primary/general election scheme, which narrowed the field to two candidates in the general election, combined with the write-in ban, deprived a large number of voters of a meaningful choice when circumstances turned one of the candidates into an indicted felon. Such disenfranchisement “strike[s] at the heart of representative government.” (Reynolds v. Sims (1964) 377 U.S. 533, 555 [84 S.Ct. 1362, 1378, 12 L.Ed.2d 506].)
Against this interest in preventing widespread disenfranchisement, the majority evokes a countervailing state interest in ensuring that the winning candidate obtains a majority of votes cast. As the majority recognizes, this interest, while not illegitimate, is not very substantial. Indeed, all of our nonmunicipal elections—for President, the United States Congress, for California statewide office and for the California Legislature—have no such voter-majority requirement. It has been proven time and again that elected representatives can function effectively in a democracy even if they were elected by less than a majority of the votes cast. And although the majority quotes James Madison, Alexander Hamilton and other of this country’s founders as generally endorsing the importance of majority governance (maj. opn., ante, at p. 183, fn. 8), the requirements of election by a majority of voters is nowhere found in the Constitution they helped draft and ratify.
Moreover, the goal of achieving a voter majority can often, if not always, be achieved without a write-in ban. The restriction on the number of candidates on the ballot to two in a runoff election, as mandated by Proposition D (S.F. Charter, former art. XIII, § 13.102, repealed Mar. 5, 2000), will generally ensure that the winning candidate obtains a majority of votes cast without such a ban. This is because write-in candidates generally do not receive a high proportion of the vote, perhaps because of the inconvenience of casting write-in votes and the marginal political position that write-in candidates usually occupy. For example, the March 5, 2002, San Francisco election results that we judicially notice (see maj. opn., ante, at p. 171, fn. 3) reveal that in contested elections the write-in votes virtually never exceeded 2 percent, and in most cases even 1 percent, of the votes cast. Therefore, unless there is deep and widespread discontent among voters with the two qualifying candidates for municipal elections and/or a very close election, write-in candidates usually will not prevent the winning candidate from *187obtaining a majority. Indeed, the amici curiae brief by the Cities of Los Angeles, Monterey and Redlands in support of San Francisco acknowledges that none of these three cities have banned write-in voting in runoff elections and that only five of 105 charter cities have, to their knowledge, enacted such a ban. The fact that runoff elections can generally perform their “winnowing” function without prohibiting write-in voting makes the City’s interest in such prohibition that much less compelling.
Nonetheless, a write-in ban in runoff elections is not without justification. Given that municipal elections in California are nonpartisan (Cal. Const., art. II, § 6), the major political parties are not able to perform their traditional function of choosing a single candidate to represent them in the general election and of organizing voting constituencies, thereby reducing voter fragmentation. (See Pomper, The Contributions of Political Parties to American Democracy in Party Renewal in America in Theory and Practice (1980) pp. 5-6 [political parties serve to aggregate diverse interests].) The advocates of Proposition D expressed concern that San Francisco mayors were being or could be elected by “a very small minority of the voters—even a scant twenty percent or less.” (S.F. Ballot Pamp., Gen. Mun. Elec. (Nov. 6, 1973) pp. 60-61.) In other words, a municipality with a history of political fractiousness may legitimately conclude, as San Francisco has, that without a write-in ban, severe voter fragmentation may persist into the runoff stage of a nonpartisan election, leading to the selection of mayoral candidates who are less able to govern because they have received a small proportion of the vote. Stated in these terms, San Francisco’s interest in preventing this type of fragmentation appears somewhat weightier than the mere abstract interest in having a winning candidate receive 50 percent or more of the vote.
Therefore, we have on one hand, the potential that a write-in ban will severely impact the electorate’s ability to choose viable candidates, particularly when circumstances change drastically after a qualifying election, and on the other hand, San Francisco’s legitimate if not compelling interest in having its candidates elected by a majority of votes cast. Balancing these interests, I conclude that in this case, unlike in Canaan, the balance tips in favor of upholding the write-in prohibition, because the potential for disenfranchisement due to changed circumstances between the first and second elections is diminished by the brevity of the interval between them. Simply stated, the possibility of changed circumstances that would effectively eliminate one of the two chosen candidates is significantly smaller within the six-week period at issue in this case than within the five-month period at issue in Canaan.
Furthermore, as stated above, a court “ ‘must not only determine the legitimacy and strength of each of [the state’s] interests; it also must *188consider the extent to which those interests make it necessary to burden the plaintiffs rights.’ ” (Canaan, supra, 40 Cal.3d at p. 712, quoting Anderson, supra, 460 U.S. at pp. 789-790 [103 S.Ct. at p. 1570].) In Canaan, we indicated that a write-in ban might be justified by the “ ‘important and legitimate interest in voter education’ ” (Canaan, supra, 40 Cal.3d at p. 720) but that prohibiting write-in candidates who did not enter the contest more than five months prior to the general election was excessive; “ ‘ “Given modem communications, and given the clear indication that campaign spending and voter education occur largely during the month before an election” ’ ([Anderson, supra, 460 U.S.] at p. 797 [75 L.Ed.2d at p. 563] . . .) it is unnecessary to impose a cutoff date some five months before the general election.” (Canaan, supra, 40 Cal.3d at p. 720; see also Anderson, supra, 460 U.S. 780, 797-798 [103 S.Ct. 1564, 1574-1575] [excessive to require filing for Ohio’s presidential ballot seven months before the election].) In light of the recognition in Canaan and Anderson that most campaign activity occurs in the month prior to an election, and in light of the practical realities of mounting an election campaign, San Francisco’s six-week period between the general and runoff elections seems close to the minimum necessary for a meaningful campaign in which the two qualifying candidates can be closely scmtinized. It would be an exercise in judicial overreaching for a court to insist on a shorter interval.
I therefore conclude that the election scheme at issue in this case, by minimizing the interval between the general and runoff elections, burdens the electorate’s interest in choosing candidates of choice no more than is reasonably necessary to accomplish the legitimate goal of reducing voter fragmentation. In this respect, it differs significantly from the election scheme at issue in Canaan. On that basis, and without overruling Canaan, I would hold that San Francisco’s write-in prohibition in runoff elections is constitutional.
Kennard, J., and Werdegar, J., concurred.
Appellants’ petition for a rehearing was denied January 15, 2003.