Arizona Free Enterprise Club’s Freedom Club PAC v. Bennett

NOTE: Where it is feasible, a syllabus (headnote) will be released, as is being done in connection with this case, at the time the opinion is issued. The syllabus constitutes no part of the opinion of the Court but has been prepared by the Reporter of Decisions for the convenience of the reader. See United States v. Detroit Timber & Lumber Co., 200 U. S. 321, 337 . SUPREME COURT OF THE UNITED STATES Syllabus ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC ET AL. v. BENNETT, SECRETARY OF STATE OF ARIZONA, ET AL. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT No. 10–238. Argued March 28, 2011—Decided June 27, 2011* The Arizona Citizens Clean Elections Act created a public financing system to fund the primary and general election campaigns of candi dates for state office. Candidates who opt to participate, and who ac cept certain campaign restrictions and obligations, are granted an initial outlay of public funds to conduct their campaign. They are also granted additional matching funds if a privately financed candi date’s expenditures, combined with the expenditures of independent groups made in support of the privately financed candidate or in op position to a publicly financed candidate, exceed the publicly financed candidate’s initial state allotment. Once matching funds are trig gered, a publicly financed candidate receives roughly one dollar for every dollar raised or spent by the privately financed candidate— including any money of his own that a privately financed candidate spends on his campaign—and for every dollar spent by independent groups that support the privately financed candidate. When there are multiple publicly financed candidates in a race, each one receives matching funds as a result of the spending of privately financed can didates and independent expenditure groups. Matching funds top out at two times the initial grant to the publicly financed candidate. Petitioners, past and future Arizona candidates and two independ ent expenditure groups that spend money to support and oppose Ari zona candidates, challenged the constitutionality of the matching —————— * Together with No. 10–239, McComish et al. v. Bennett, Secretary of State of Arizona, et al., also on certiorari to the same court. 2 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Syllabus funds provision, arguing that it unconstitutionally penalizes their speech and burdens their ability to fully exercise their First Amend ment rights. The District Court entered a permanent injunction against the enforcement of the matching funds provision. The Ninth Circuit reversed, concluding that the provision imposed only a mini mal burden and that the burden was justified by Arizona’s interest in reducing quid pro quo political corruption. Held: Arizona’s matching funds scheme substantially burdens political speech and is not sufficiently justified by a compelling interest to survive First Amendment scrutiny. Pp. 8–30. (a) The matching funds provision imposes a substantial burden on the speech of privately financed candidates and independent expendi ture groups. Pp. 8–22. (1) Petitioners contend that their political speech is substantially burdened in the same way that speech was burdened by the so-called “Millionaire’s Amendment” of the Bipartisan Campaign Reform Act of 2002, which was invalidated in Davis v. Federal Election Comm’n, 554 U. S. 724 . That law—which permitted the opponent of a candi date who spent over $350,000 of his personal funds to collect triple the normal contribution amount, while the candidate who spent the personal funds remained subject to the original contribution cap— unconstitutionally forced a candidate “to choose between the First Amendment right to engage in unfettered political speech and subjec tion to discriminatory fundraising limitations.” Id., at 739 . This “unprecedented penalty” “impose[d] a substantial burden on the ex ercise of the First Amendment right to use personal funds for cam paign speech” that was not justified by a compelling government in terest. Id., at 739–740. Pp. 8–10. (2) The logic of Davis largely controls here. Once a privately fi nanced candidate has raised or spent more than the State’s initial grant to a publicly financed candidate, each personal dollar the pri vately financed candidate spends results in an award of almost one additional dollar to his opponent. The privately financed candidate must “shoulder a special and potentially significant burden” when choosing to exercise his First Amendment right to spend funds on his own candidacy. 554 U. S., at 739 . If the law at issue in Davis im posed a burden on candidate speech, the Arizona law unquestionably does so as well. The differences between the matching funds provision and the law struck down in Davis make the Arizona law more constitutionally problematic, not less. First, the penalty in Davis consisted of raising the contribution limits for one candidate, who would still have to raise the additional funds. Here, the direct and automatic release of public money to a publicly financed candidate imposes a far heavier burden. Second, in elections where there are multiple publicly fi nanced candidates—a frequent occurrence in Arizona—the matching funds provision can create a multiplier effect. Each dollar spent by the privately funded candidate results in an additional dollar of fund ing to each of that candidate’s publicly financed opponents. Third, unlike the law in Davis, all of this is to some extent out of the pri vately financed candidate’s hands. Spending by independent expen diture groups to promote a privately financed candidate’s election triggers matching funds, regardless whether such support is welcome or helpful. Those funds go directly to the publicly funded candidate to use as he sees fit. That disparity in control—giving money directly to a publicly financed candidate, in response to independent expendi tures that cannot be coordinated with the privately funded candi date—is a substantial advantage for the publicly funded candidate. The burdens that matching funds impose on independent expendi ture groups are akin to those imposed on the privately financed can didates themselves. The more money spent on behalf of a privately financed candidate or in opposition to a publicly funded candidate, the more money the publicly funded candidate receives from the State. The effect of a dollar spent on election speech is a guaranteed financial payout to the publicly funded candidate the group opposes, and spending one dollar can result in the flow of dollars to multiple candidates. In some ways, the burdens imposed on independent groups by matching funds are more severe than the burdens imposed on privately financed candidates. Independent groups, of course, are not eligible for public financing. As a result, those groups can only avoid matching funds by changing their message or choosing not to speak altogether. Presenting independent expenditure groups with such a choice—trigger matching funds, change your message, or do not speak—makes the matching funds provision particularly burden some to those groups and certainly contravenes “the fundamental rule of protection under the First Amendment, that a speaker has the autonomy to choose the content of his own message.” Hurley v. IrishAmerican Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U. S. 557, 573 . Pp. 10–14. (3) The arguments of Arizona, the Clean Elections Institute, and amicus United States attempting to explain away the existence or significance of any burden imposed by matching funds are unpersua sive. Arizona correctly points out that its law is different from the law invalidated in Davis, but there is no doubt that the burden on speech is significantly greater here than in Davis. Arizona argues that the provision actually creates more speech. But even if that were the case, only the speech of publicly financed candidates is increased by 4 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Syllabus the state law. And burdening the speech of some—here privately fi nanced candidates and independent expenditure groups—to increase the speech of others is a concept “wholly foreign to the First Amend ment,” Buckley v. Valeo, 424 U. S. 1 , 48–49; cf. Miami Herald Pub lishing Co. v. Tornillo, 418 U. S. 241, 244, 258 . That no candidate or group is forced to express a particular message does not mean that the matching funds provision does not burden their speech, especially since the direct result of that speech is a state-provided monetary subsidy to a political rival. And precedents upholding government subsidies against First Amendment challenge provide no support for matching funds; none of the subsidies at issue in those cases were granted in response to the speech of another. The burden on privately financed candidates and independent ex penditure groups also cannot be analogized to the burden placed on speakers by the disclosure and disclaimer requirements upheld in Citizens United v. Federal Election Comm’n, 558 U. S. ___ . A political candidate’s disclosure of his funding resources does not result in a cash windfall to his opponent, or affect their respective disclosure ob ligations. The burden imposed by the matching funds provision is evident and inherent in the choice that confronts privately financed candi dates and independent expenditure groups. Indeed every court to have considered the question after Davis has concluded that a candi date or independent group might not spend money if the direct result of that spending is additional funding to political adversaries. Ari zona is correct that the candidates do not complain that providing a lump sum payment equivalent to the maximum state financing that a candidate could obtain through matching funds would be imper missible. But it is not the amount of funding that the State provides that is constitutionally problematic. It is the manner in which that funding is provided—in direct response to the political speech of pri vately financed candidates and independent expenditure groups. Pp. 14–22. (b) Arizona’s matching funds provision is not “ ‘justified by a com pelling state interest,’ ” Davis, supra, at 740 . Pp. 22–28. (1) There is ample support for the argument that the purpose of the matching funds provision is to “level the playing field” in terms of candidate resources. The clearest evidence is that the provision op erates to ensure that campaign funding is equal, up to three times the initial public funding allotment. The text of the Arizona Act con firms this purpose. The provision setting up the matching funds re gime is titled “Equal funding of candidates,” Ariz. Rev. Stat. Ann. §16–952; and the Act and regulations refer to the funds as “equaliz ing funds,” e.g., §16–952(C)(4). This Court has repeatedly rejected the argument that the government has a compelling state interest in “leveling the playing field” that can justify undue burdens on political speech, see, e.g., Citizens United, supra, at ___, and the burdens im posed by matching funds cannot be justified by the pursuit of such an interest. Pp. 22–25. (2) Even if the objective of the matching funds provision is to combat corruption—and not “level the playing field”—the burdens that the matching funds provision imposes on protected political speech are not justified. Burdening a candidate’s expenditure of his own funds on his own campaign does not further the State’s anticor ruption interest. Indeed, “reliance on personal funds reduces the threat of corruption.” Davis, supra, at 740–741; see Buckley, supra, at 53 . The burden on independent expenditures also cannot be sup ported by the anticorruption interest. Such expenditures are “politi cal speech . . . not coordinated with a candidate.” Citizens United, 558 U. S., at ___. That separation negates the possibility that the ex penditures will result in the sort of quid pro quo corruption with which this Court’s case law is concerned. See e.g., id., at ___–___. Moreover, “[t]he interest in alleviating the corrupting influence of large contributions is served by . . . contribution limitations.” Buck ley, supra, at 55. Given Arizona’s contribution limits, some of the most austere in the Nation, its strict disclosure requirements, and the general availability of public funding, it is hard to imagine what marginal corruption deterrence could be generated by the matching funds provision. The State and the Clean Elections Institute contend that even if the matching funds provision does not directly serve the anticorrup tion interest, it indirectly does so by ensuring that enough candidates participate in the State’s public funding system, which in turn helps combat corruption. But the fact that burdening constitutionally pro tected speech might indirectly serve the State’s anticorruption inter est, by encouraging candidates to take public financing, does not es tablish the constitutionality of the matching funds provision. The matching funds provision substantially burdens speech, to an even greater extent than the law invalidated in Davis. Those burdens cannot be justified by a desire to “level the playing field,” and much of the speech burdened by the matching funds provision does not pose a danger of corruption. The fact that the State may feel that the matching funds provision is necessary to allow it to calibrate its pub lic funding system to achieve its desired level of participation— without an undue drain on public resources—is not a sufficient justi fication for the burden. The flaw in the State’s argument is apparent in what its reasoning would allow. By the State’s logic it could award publicly financed 6 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Syllabus candidates five dollars for every dollar spent by a privately financed candidate, or force candidates who wish to run on private funds to pay a $10,000 fine, in order to encourage participation in the public funding regime. Such measures might well promote such participa tion, but would clearly suppress or unacceptably alter political speech. How the State chooses to encourage participation in its pub lic funding system matters, and the Court has never held that a State may burden political speech—to the extent the matching funds provi sion does—to ensure adequate participation in a public funding sys tem. Pp. 25–28. (c) Evaluating the wisdom of public financing as a means of fund ing political candidacy is not the Court’s business. But determining whether laws governing campaign finance violate the First Amend ment is. The government “may engage in public financing of election campaigns,” and doing so can further “significant governmental in terest[s].” Buckley, 424 U. S., at 57, n. 65 , 92–93, 96. But the goal of creating a viable public financing scheme can only be pursued in a manner consistent with the First Amendment. Arizona’s program gives money to a candidate in direct response to the campaign speech of an opposing candidate or an independent group. It does this when the opposing candidate has chosen not to accept public financing, and has engaged in political speech above a level set by the State. This goes too far; Arizona’s matching funds provision substantially bur dens the speech of privately financed candidates and independent expenditure groups without serving a compelling state interest. Pp. 28–30. 611 F. 3d 510 , reversed. ROBERTS, C. J., delivered the opinion of the Court, in which SCALIA, KENNEDY, THOMAS, and ALITO, JJ., joined. KAGAN, J., filed a dissenting opinion, in which GINSBURG, BREYER, and SOTOMAYOR, JJ., joined. NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Wash ington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press. SUPREME COURT OF THE UNITED STATES _________________ Nos. 10–238 and 10–239 _________________ ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC, ET AL., PETITIONERS 10–238 v. KEN BENNETT, IN HIS OFFICIAL CAPACITY AS ARIZONA SECRETARY OF STATE, ET AL. JOHN MCCOMISH, ET AL., PETITIONERS 10–239 v. KEN BENNETT, IN HIS OFFICIAL CAPACITY AS ARIZONA SECRETARY OF STATE, ET AL. ON WRITS OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT [June 27, 2011] CHIEF JUSTICE ROBERTS delivered the opinion of the Court. Under Arizona law, candidates for state office who ac cept public financing can receive additional money from the State in direct response to the campaign activities of privately financed candidates and independent expendi ture groups. Once a set spending limit is exceeded, a publicly financed candidate receives roughly one dollar for every dollar spent by an opposing privately financed can didate. The publicly financed candidate also receives roughly one dollar for every dollar spent by independent expenditure groups to support the privately financed candidate, or to oppose the publicly financed candidate. 2 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court We hold that Arizona’s matching funds scheme substan tially burdens protected political speech without serving a compelling state interest and therefore violates the First Amendment. I A The Arizona Citizens Clean Elections Act, passed by initiative in 1998, created a voluntary public financing system to fund the primary and general election cam paigns of candidates for state office. See Ariz. Rev. Stat. Ann. §16–940 et seq. (West 2006 and Supp. 2010). All eligible candidates for Governor, secretary of state, attor ney general, treasurer, superintendent of public instruc tion, the corporation commission, mine inspector, and the state legislature (both the House and Senate) may opt to receive public funding. §16–950(D) (West Supp. 2010). Eligibility is contingent on the collection of a specified number of five-dollar contributions from Arizona voters, §§16–946(B) (West 2006), 16–950 (West Supp. 2010),1 and the acceptance of certain campaign restrictions and obliga tions. Publicly funded candidates must agree, among other things, to limit their expenditure of personal funds to $500, §16–941(A)(2) (West Supp. 2010); participate in at least one public debate, §16–956(A)(2); adhere to an over all expenditure cap, §16–941(A); and return all unspent public moneys to the State, §16–953. In exchange for accepting these conditions, participating candidates are granted public funds to conduct their cam paigns.2 In many cases, this initial allotment may be the —————— 1 The number of qualifying contributions ranges from 200 for a candi date for the state legislature to 4,000 for a candidate for Governor. Ariz. Rev. Stat. Ann. §16–950(D) (West Supp. 2010). 2 Publicly financed candidates who run unopposed, or who run as the representative of a party that does not have a primary, may receive less funding than candidates running in contested elections. See §§16– whole of the State’s financial backing of a publicly funded candidate. But when certain conditions are met, publicly funded candidates are granted additional “equalizing” or matching funds. §§16–952(A), (B), and (C)(4)–(5) (provid ing for “[e]qual funding of candidates”). Matching funds are available in both primary and gen eral elections. In a primary, matching funds are triggered when a privately financed candidate’s expenditures, com bined with the expenditures of independent groups made in support of the privately financed candidate or in opposi tion to a publicly financed candidate, exceed the primary election allotment of state funds to the publicly financed candidate. §§16–952(A), (C). During the general election, matching funds are triggered when the amount of money a privately financed candidate receives in contributions, combined with the expenditures of independent groups made in support of the privately financed candidate or in opposition to a publicly financed candidate, exceed the general election allotment of state funds to the publicly fi nanced candidate. §16–952(B). A privately financed can didate’s expenditures of his personal funds are counted as contributions for purposes of calculating matching funds during a general election. See ibid.; Citizens Clean Elections Commission, Ariz. Admin. Rule R2–20– 113(B)(1)(f) (Sept. 2009). Once matching funds are triggered, each additional dol lar that a privately financed candidate spends during the primary results in one dollar in additional state fund ing to his publicly financed opponent (less a 6% reduction meant to account for fundraising expenses). §16–952(A). During a general election, every dollar that a candidate receives in contributions—which includes any money of his own that a candidate spends on his campaign—results in roughly one dollar in additional state funding to his —————— 951(A)(2)–(3) and (D) (West 2006). 4 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court publicly financed opponent. In an election where a pri vately funded candidate faces multiple publicly financed candidates, one dollar raised or spent by the privately fi nanced candidate results in an almost one dollar increase in public funding to each of the publicly financed candi dates. Once the public financing cap is exceeded, additional expenditures by independent groups can result in dollar for-dollar matching funds as well. Spending by independ ent groups on behalf of a privately funded candidate, or in opposition to a publicly funded candidate, results in matching funds. §16–952(C). Independent expenditures made in support of a publicly financed candidate can result in matching funds for other publicly financed candi dates in a race. Ibid. The matching funds provision is not activated, however, when independent expenditures are made in opposition to a privately financed candidate. Matching funds top out at two times the initial authorized grant of public funding to the publicly financed candidate. §16–952(E). Under Arizona law, a privately financed candidate may raise and spend unlimited funds, subject to state-imposed contribution limits and disclosure requirements. Contri butions to candidates for statewide office are limited to $840 per contributor per election cycle and contributions to legislative candidates are limited to $410 per contribu tor per election cycle. See §§16–905(A)(1), 16–941(B)(1); Ariz. Dept. of State, Office of the Secretary of State, 2009– 2010 Contribution Limits (rev. Aug. 14, 2009), http:// www.azsos.gov/election/2010/Info/Campaign_Contribution _Limits_2010.htm (all Internet materials as visited June 24, 2011, and available in Clerk of Court’s case file). An example may help clarify how the Arizona matching funds provision operates. Arizona is divided into 30 dis tricts for purposes of electing members to the State’s House of Representatives. Each district elects two repre sentatives to the House biannually. In the last general election, the number of candidates competing for the two available seats in each district ranged from two to seven. See State of Arizona Official Canvass, 2010 General Elec tion Report (compiled and issued by the Arizona secretary of state). Arizona’s Fourth District had three candidates for its two available House seats. Two of those candidates opted to accept public funding; one candidate chose to operate his campaign with private funds. In that election, if the total funds contributed to the privately funded candidate, added to that candidate’s expenditure of personal funds and the expenditures of supportive independent groups, exceeded $21,479—the allocation of public funds for the general election in a contested State House race—the matching funds provision would be triggered. See Citizens Clean Elections Commis sion, Participating Candidate Guide 2010 Election Cycle 30 (Aug. 10, 2010). At that point, a number of differ ent political activities could result in the distribution of matching funds. For example: • If the privately funded candidate spent $1,000 of his own money to conduct a direct mailing, each of his publicly funded opponents would receive $940 ($1,000 less the 6% offset). • If the privately funded candidate held a fundraiser that generated $1,000 in contributions, each of his publicly funded opponents would receive $940. • If an independent expenditure group spent $1,000 on a brochure expressing its support for the pri vately financed candidate, each of the publicly fi nanced candidates would receive $940 directly. • If an independent expenditure group spent $1,000 on a brochure opposing one of the publicly financed candidates, but saying nothing about the privately 6 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court financed candidate, the publicly financed candidates would receive $940 directly. • If an independent expenditure group spent $1,000 on a brochure supporting one of the publicly fi nanced candidates, the other publicly financed can didate would receive $940 directly, but the privately financed candidate would receive nothing. • If an independent expenditure group spent $1,000 on a brochure opposing the privately financed can didate, no matching funds would be issued. A publicly financed candidate would continue to receive additional state money in response to fundraising and spending by the privately financed candidate and inde pendent expenditure groups until that publicly financed candidate received a total of $64,437 in state funds (three times the initial allocation for a State House race).3 B Petitioners in this case, plaintiffs below, are five past and future candidates for Arizona state office—four mem bers of the House of Representatives and the Arizona state treasurer—and two independent groups that spend money to support and oppose Arizona candidates. They filed suit challenging the constitutionality of the matching funds provision. The candidates and independent expenditure groups argued that the matching funds provision uncon —————— 3 Maine and North Carolina have both passed matching funds stat utes that resemble Arizona’s law. See Me. Rev. Stat. Ann., Tit. 21–A, §§1125(8), (9) (2008); N. C. Gen. Stat. Ann. §163–278.67 (Lexis 2009). Minnesota, Connecticut, and Florida have also adopted matching funds provisions, but courts have enjoined the enforcement of those schemes after concluding that their operation violates the First Amendment. See Day v. Holahan, 34 F. 3d 1356, 1362 (CA8 1994); Green Party of Conn. v. Garfield, 616 F. 3d 213, 242 (CA2 2010); Scott v. Roberts, 612 F. 3d 1279 , 1297–1298 (CA11 2010). stitutionally penalized their speech and burdened their ability to fully exercise their First Amendment rights. The District Court agreed that this provision “consti tute[d] a substantial burden” on the speech of privately financed candidates because it “award[s] funds to a [pri vately financed] candidate’s opponent” based on the pri vately financed candidate’s speech. App. to Pet. for Cert. in No. 10–239, p. 69 (internal quotation marks omitted). That court further held that “no compelling interest [was] served by the” provision that might justify the burden imposed. Id., at 69, 71. The District Court entered a permanent injunction against the enforcement of the matching funds provision, but stayed implementation of that injunction to allow the State to file an appeal. Id., at 76–81. The Court of Appeals for the Ninth Circuit stayed the District Court’s injunction pending appeal. Id., at 84–85.4 After hearing the case on the merits, the Court of Appeals reversed the District Court. The Court of Appeals con cluded that the matching funds provision “imposes only a minimal burden on First Amendment rights” because it “does not actually prevent anyone from speaking in the first place or cap campaign expenditures.” 611 F. 3d 510, 513, 525 (2010). In that court’s view, any burden imposed by the matching funds provision was justified because the provision “bears a substantial relation to the State’s im portant interest in reducing quid pro quo political corrup tion.” Id., at 513 .5 —————— 4 Judge Bea dissented from the stay of the District Court’s injunction, stating that the Arizona public financing system unconstitutionally prefers publicly financed candidates and that under the matching funds scheme “it makes no more sense for [a privately financed candidate or independent expenditure group] to spend money now than for a poker player to make a bet if he knows the house is going to match his bet for his opponent.” App. to Pet. for Cert. in No. 10–239, p. 87; see id., at 89. 5 One judge concurred, relying primarily on his view that “the Arizona 8 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court We stayed the Court of Appeals’ decision, vacated the stay of the District Court’s injunction, see 560 U. S. ___ (2010), and later granted certiorari, 562 U. S. ___ (2010). II “Discussion of public issues and debate on the qualifica tions of candidates are integral to the operation” of our system of government. Buckley v. Valeo, 424 U. S. 1, 14 (1976) (per curiam). As a result, the First Amendment “ ‘has its fullest and most urgent application’ to speech ut tered during a campaign for political office.” Eu v. San Francisco County Democratic Central Comm., 489 U. S. 214, 223 (1989) (quoting Monitor Patriot Co. v. Roy, 401 U. S. 265, 272 (1971)). “Laws that burden political speech are” accordingly “subject to strict scrutiny, which requires the Government to prove that the restriction furthers a compelling interest and is narrowly tailored to achieve that interest.” Citizens United v. Federal Election Comm’n, 558 U. S. ___ , ___ (2010) (slip op., at 23) (internal quotation marks omitted); see Federal Election Comm’n v. Massachusetts Citizens for Life, Inc., 479 U. S. 238, 256 (1986). Applying these principles, we have invalidated govern ment-imposed restrictions on campaign expenditures, Buckley, supra, at 52–54, restraints on independent ex penditures applied to express advocacy groups, Massachu setts Citizens for Life, supra, at 256–265, limits on uncoor dinated political party expenditures, Colorado Republican Federal Campaign Comm. v. Federal Election Comm’n, 518 U. S. 604, 608 (1996) (opinion of BREYER, J.) (Colorado I), and regulations barring unions, nonprofit and other associations, and corporations from making independent expenditures for electioneering communication, Citizens —————— public financing scheme imposes no limitations whatsoever on a candi date’s speech.” 611 F. 3d, at 527 (Kleinfeld, J.). United, supra, at ___ (slip op., at 57). At the same time, we have subjected strictures on campaign-related speech that we have found less onerous to a lower level of scrutiny and upheld those restrictions. For example, after finding that the restriction at issue was “closely drawn” to serve a “sufficiently important interest,” see, e.g., McConnell v. Federal Election Comm’n, 540 U. S. 93, 136 (2003) (internal quotation marks omitted); Nixon v. Shrink Missouri Government PAC, 528 U. S. 377 , 387– 388 (2000) (internal quotation marks omitted), we have upheld government-imposed limits on contributions to can didates, Buckley, supra, at 23–35, caps on coordinated party expenditures, Federal Election Comm’n v. Colorado Republican Federal Campaign Comm., 533 U. S. 431, 437 (2001) (Colorado II), and requirements that political fund ing sources disclose their identities, Citizens United, su pra, at ___–___ (slip op., at 55–56). Although the speech of the candidates and independent expenditure groups that brought this suit is not directly capped by Arizona’s matching funds provision, those par ties contend that their political speech is substantially burdened by the state law in the same way that speech was burdened by the law we recently found invalid in Davis v. Federal Election Comm’n, 554 U. S. 724 (2008). In Davis, we considered a First Amendment challenge to the so-called “Millionaire’s Amendment” of the Bipartisan Campaign Reform Act of 2002, 2 U. S. C. §441a–1(a). Under that Amendment, if a candidate for the United States House of Representatives spent more than $350,000 of his personal funds, “a new, asymmetrical regulatory scheme [came] into play.” 554 U. S., at 729 . The opponent of the candidate who exceeded that limit was permitted to collect individual contributions up to $6,900 per contributor—three times the normal contribu tion limit of $2,300. See ibid. The candidate who spent more than the personal funds limit remained subject to 10 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court the original contribution cap. Davis argued that this scheme “burden[ed] his exercise of his First Amendment right to make unlimited expenditures of his personal funds because” doing so had “the effect of enabling his opponent to raise more money and to use that money to finance speech that counteract[ed] and thus diminishe[d] the effectiveness of Davis’ own speech.” Id., at 736 . In addressing the constitutionality of the Millionaire’s Amendment, we acknowledged that the provision did not impose an outright cap on a candidate’s personal expendi tures. Id., at 738–739. We nonetheless concluded that the Amendment was unconstitutional because it forced a candidate “to choose between the First Amendment right to engage in unfettered political speech and subjection to discriminatory fundraising limitations.” Id., at 739 . Any candidate who chose to spend more than $350,000 of his own money was forced to “shoulder a special and poten tially significant burden” because that choice gave fund raising advantages to the candidate’s adversary. Ibid. We determined that this constituted an “unprecedented pen alty” and “impose[d] a substantial burden on the exercise of the First Amendment right to use personal funds for campaign speech,” and concluded that the Government had failed to advance any compelling interest that would justify such a burden. Id., at 739–740. A [1] The logic of Davis largely controls our approach to this case. Much like the burden placed on speech in Davis, the matching funds provision “imposes an unprecedented penalty on any candidate who robustly exercises [his] First Amendment right[s].” Id., at 739 . Under that provi sion, “the vigorous exercise of the right to use personal funds to finance campaign speech” leads to “advantages for opponents in the competitive context of electoral poli tics.” Ibid. Once a privately financed candidate has raised or spent more than the State’s initial grant to a publicly financed candidate, each personal dollar spent by the privately financed candidate results in an award of almost one additional dollar to his opponent. That plainly forces the privately financed candidate to “shoulder a special and potentially significant burden” when choosing to exercise his First Amendment right to spend funds on behalf of his candidacy. Ibid. If the law at issue in Davis imposed a burden on candidate speech, the Arizona law unques tionably does so as well. The penalty imposed by Arizona’s matching funds provi sion is different in some respects from the penalty imposed by the law we struck down in Davis. But those differences make the Arizona law more constitutionally problematic, not less. See Green Party of Conn. v. Garfield, 616 F. 3d 213 , 244–245 (CA2 2010). First, the penalty in Davis consisted of raising the contribution limits for one of the candidates. The candidate who benefited from the in creased limits still had to go out and raise the funds. He may or may not have been able to do so. The other candi date, therefore, faced merely the possibility that his oppo nent would be able to raise additional funds, through contribution limits that remained subject to a cap. And still the Court held that this was an “unprecedented pen alty,” a “special and potentially significant burden” that had to be justified by a compelling state interest—a rigor ous First Amendment hurdle. 554 U. S., at 739–740. Here the benefit to the publicly financed candidate is the direct and automatic release of public money. That is a far heavier burden than in Davis. Second, depending on the specifics of the election at issue, the matching funds provision can create a multiplier effect. In the Arizona Fourth District House election pre viously discussed, see supra, at 4–6, if the spending cap 12 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court were exceeded, each dollar spent by the privately funded candidate would result in an additional dollar of campaign funding to each of that candidate’s publicly financed oppo nents. In such a situation, the matching funds provi sion forces privately funded candidates to fight a political hydra of sorts. Each dollar they spend generates two adversarial dollars in response. Again, a markedly more significant burden than in Davis. Third, unlike the law at issue in Davis, all of this is to some extent out of the privately financed candidate’s hands. Even if that candidate opted to spend less than the initial public financing cap, any spending by independent expenditure groups to promote the privately financed candidate’s election—regardless whether such support was welcome or helpful—could trigger matching funds. What is more, that state money would go directly to the publicly funded candidate to use as he saw fit. That disparity in control—giving money directly to a publicly financed candidate, in response to independent expendi tures that cannot be coordinated with the privately funded candidate—is a substantial advantage for the publicly funded candidate. That candidate can allocate the money according to his own campaign strategy, which the pri vately financed candidate could not do with the independ ent group expenditures that triggered the matching funds. Cf. Citizens United, 558 U. S., at ___ (slip op., at 41) (“ ‘The absence of prearrangement and coordination of an expen diture with the candidate or his agent . . . undermines the value of the expenditure to the candidate’ ” (quoting Buck ley, 424 U. S., at 47 )). The burdens that this regime places on independent expenditure groups are akin to those imposed on the privately financed candidates themselves. Just as with the candidate the independent group supports, the more money spent on that candidate’s behalf or in opposition to a publicly funded candidate, the more money the publicly funded candidate receives from the State. And just as with the privately financed candidate, the effect of a dollar spent on election speech is a guaranteed financial payout to the publicly funded candidate the group opposes. More over, spending one dollar can result in the flow of dollars to multiple candidates the group disapproves of, dollars directly controlled by the publicly funded candidate or candidates. In some ways, the burden the Arizona law imposes on independent expenditure groups is worse than the burden it imposes on privately financed candidates, and thus substantially worse than the burden we found constitu tionally impermissible in Davis. If a candidate contem plating an electoral run in Arizona surveys the campaign landscape and decides that the burdens imposed by the matching funds regime make a privately funded campaign unattractive, he at least has the option of taking public financing. Independent expenditure groups, of course, do not. Once the spending cap is reached, an independent ex penditure group that wants to support a particular candi date—because of that candidate’s stand on an issue of con cern to the group—can only avoid triggering matching funds in one of two ways. The group can either opt to change its message from one addressing the merits of the candidates to one addressing the merits of an issue, or refrain from speaking altogether. Presenting independent expenditure groups with such a choice makes the match ing funds provision particularly burdensome to those groups. And forcing that choice—trigger matching funds, change your message, or do not speak—certainly contra venes “the fundamental rule of protection under the First Amendment, that a speaker has the autonomy to choose the content of his own message.” Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U. S. 557, 573 (1995); cf. Citizens United, supra, at ___ (slip op., 14 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court at 24) (“the First Amendment stands against attempts to disfavor certain subjects or viewpoints”); Federal Election Comm’n v. Wisconsin Right to Life, Inc., 551 U. S. 449, 477, n. 9 (2007) (opinion of ROBERTS, C. J.) (the argument that speakers can avoid the burdens of a law “by changing what they say” does not mean the law complies with the First Amendment).6 [2] Arizona, the Clean Elections Institute, and the United States offer several arguments attempting to explain away the existence or significance of any burden imposed by matching funds. None is persuasive. Arizona contends that the matching funds provision is distinguishable from the law we invalidated in Davis. The State correctly points out that our decision in Davis fo cused on the asymmetrical contribution limits imposed by the Millionaire’s Amendment. See 554 U. S., at 729 . But that is not because—as the State asserts—the reach of that opinion is limited to asymmetrical contribution limits. Brief for State Respondents 26–32. It is because that was the particular burden on candidate speech we faced in Davis. And whatever the significance of the distinction in general, there can be no doubt that the burden on speech is significantly greater in this case than in Davis: That means that the law here—like the one in Davis—must be justified by a compelling state interest. —————— 6 The dissent sees “chutzpah” in candidates exercising their right not to participate in the public financing scheme, while objecting that the system violates their First Amendment rights. See post, at 12 (opinion of KAGAN, J.). The charge is unjustified, but, in any event, it certainly cannot be leveled against the independent expenditure groups. The dissent barely mentions such groups in its analysis, and fails to address not only the distinctive burdens imposed on these groups—as set forth above—but also the way in which privately financed candidates are particularly burdened when matching funds are triggered by independ ent group speech. The State argues that the matching funds provision actually results in more speech by “increas[ing] debate about issues of public concern” in Arizona elections and “promot[ing] the free and open debate that the First Amendment was intended to foster.” Brief for State Re spondents 41; see Brief for Respondent Clean Elections Institute 55. In the State’s view, this promotion of First Amendment ideals offsets any burden the law might impose on some speakers. Not so. Any increase in speech resulting from the Ari zona law is of one kind and one kind only—that of publicly financed candidates. The burden imposed on privately financed candidates and independent expenditure groups reduces their speech; “restriction[s] on the amount of money a person or group can spend on political communi cation during a campaign necessarily reduces the quantity of expression.” Buckley, 424 U. S., at 19 . Thus, even if the matching funds provision did result in more speech by publicly financed candidates and more speech in general, it would do so at the expense of impermissibly burdening (and thus reducing) the speech of privately financed candidates and independent expenditure groups. This sort of “beggar thy neighbor” approach to free speech— “restrict[ing] the speech of some elements of our society in order to enhance the relative voice of others”—is “wholly foreign to the First Amendment.” Id., at 48–49.7 —————— 7 The dissent also repeatedly argues that the Arizona matching funds regime results in “more political speech,” post, at 9 (emphasis in origi nal); see post, at 2, 10, 13, 16, 32, but—given the logic of the dissent’s position—that is only as a step to less speech. If the matching funds provision achieves its professed goal and causes candidates to switch to public financing, post, at 25, 30, there will be less speech: no spending above the initial state-set amount by formerly privately financed candidates, and no associated matching funds for anyone. Not only that, the level of speech will depend on the State’s judgment of the desirable amount, an amount tethered to available (and often scarce) state resources. 16 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court We have rejected government efforts to increase the speech of some at the expense of others outside the cam paign finance context. In Miami Herald Publishing Co. v. Tornillo, 418 U. S. 241, 244, 258 (1974), we held unconsti tutional a Florida law that required any newspaper assail ing a political candidate’s character to allow that candi date to print a reply. We have explained that while the statute in that case “purported to advance free discussion, . . . its effect was to deter newspapers from speaking out in the first instance” because it “penalized the newspaper’s own expression.” Pacific Gas & Elec. Co. v. Public Util. Comm’n of Cal., 475 U. S. 1, 10 (1986) (plurality opinion). Such a penalty, we concluded, could not survive First Amendment scrutiny. The Arizona law imposes a similar penalty: The State grants funds to publicly financed can didates as a direct result of the speech of privately fi nanced candidates and independent expenditure groups. The argument that this sort of burden promotes free and robust discussion is no more persuasive here than it was in Tornillo.8 Arizona asserts that no “candidate or independent ex penditure group is ‘obliged personally to express a mes sage he disagrees with’ ” or “ ‘required by the government to subsidize a message he disagrees with.’ ” Brief for State —————— 8 Along the same lines, we have invalidated government mandates that a speaker “help disseminate hostile views” opposing that speaker’s message. Pacific Gas & Elec. Co. v. Public Util. Comm’n of Cal., 475 U. S. 1, 14 (1986) (plurality opinion). In Pacific Gas, we found a public utility commission order forcing a utility company to disseminate in its billing envelopes views that the company opposed ran afoul of the First Amendment. That case is of course distinguishable from the instant case on its facts, but the central concern—that an individual should not be compelled to “help disseminate hostile views”—is implicated here as well. Ibid. If a candidate uses his own money to engage in speech above the initial public funding threshold, he is forced to “help dissemi nate hostile views” in a most direct way—his own speech triggers the release of state money to his opponent. Respondents 32 (quoting Johanns v. Livestock Marketing Assn., 544 U. S. 550, 557 (2005)). True enough. But that does not mean that the matching funds provision does not burden speech. The direct result of the speech of pri vately financed candidates and independent expenditure groups is a state-provided monetary subsidy to a political rival. That cash subsidy, conferred in response to political speech, penalizes speech to a greater extent and more directly than the Millionaire’s Amendment in Davis. The fact that this may result in more speech by the other candidates is no more adequate a justification here than it was in Davis. See 554 U. S., at 741–742. In disagreeing with our conclusion, the dissent relies on cases in which we have upheld government subsidies against First Amendment challenge, and asserts that “[w]e have never, not once, understood a viewpoint-neutral subsidy given to one speaker to constitute a First Amend ment burden on another.” Post, at 16. But none of those cases—not one—involved a subsidy given in direct re sponse to the political speech of another, to allow the recipient to counter that speech. And nothing in the analysis we employed in those cases suggests that the challenged subsidies would have survived First Amend ment scrutiny if they were triggered by someone else’s political speech.9 The State also argues, and the Court of Appeals con cluded, that any burden on privately financed candidates and independent expenditure groups is more analogous to —————— 9 The dissent cites Buckley in response, see post, at 12, n. 3, but the funding in Buckley was of course not triggered by the speech of a publicly funded candidate’s political opponent, or the speech of anyone else for that matter. See 424 U. S., at 91–95. Whether Arizona’s matching funds provision comports with the First Amendment is not simply a question of whether the State can give a subsidy to a candi date to fund that candidate’s election, but whether that subsidy can be triggered by the speech of another candidate or independent group. 18 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court the burden placed on speakers by the disclosure and disclaimer requirements we recently upheld in Citizens United than to direct restrictions on candidate and inde pendent expenditures. See 611 F. 3d, at 525 ; Brief for State Respondents 21, 35; Brief for Respondent Clean Elections Institute 16–17. This analogy is not even close. A political candidate’s disclosure of his funding resources does not result in a cash windfall to his opponent, or affect their respective disclosure obligations. The State and the Clean Elections Institute assert that the candidates and independent expenditure groups have failed to “cite specific instances in which they decided not to raise or spend funds,” Brief for State Respondents 11; see id., at 11–12, and have “failed to present any reliable evidence that Arizona’s triggered matching funds deter their speech,” Brief for Respondent Clean Elections Insti tute 6; see id., at 6–8. The record in this case, which we must review in its entirety, does not support those asser tions. See Bose Corp. v. Consumers Union of United States, Inc., 466 U. S. 485, 499 (1984). That record contains examples of specific candidates curtailing fundraising efforts, and actively discouraging supportive independent expenditures, to avoid triggering matching funds. See, e.g., App. 567 (Rick Murphy), 578 (Dean Martin); App. to Pet. for Cert. in No. 10–239, at 329 (John McComish), 300 (Tony Bouie). The record also includes examples of independent expenditure groups deciding not to speak in opposition to a candidate, App. 569 (Arizona Taxpayers Action Committee), or in support of a candidate, id., at 290 (Club for Growth), to avoid triggering matching funds. In addition, Dr. David Primo, an expert involved in the case, “found that privately fi nanced candidates facing the prospect of triggering match ing funds changed the timing of their fundraising activi ties, the timing of their expenditures, and, thus, their overall campaign strategy.” Reply Brief for Petitioner Arizona Free Enterprise Club’s (AFEC) Freedom Club PAC et al. 12; see also id., at 11–17 (listing additional sources of evidence detailing the burdens imposed by the matching funds provision); Brief for Petitioner AFEC’s Freedom Club PAC et al. 14–21 (AFEC Brief) (same); Brief for Petitioner McComish et al. 30–37 (same). The State contends that if the matching funds provision truly burdened the speech of privately financed candidates and independent expenditure groups, spending on behalf of privately financed candidates would cluster just below the triggering level, but no such phenomenon has been observed. Brief for State Respondents 39; Brief for Re spondent Clean Elections Institute 18–19. That should come as no surprise. The hypothesis presupposes a pri vately funded candidate who would spend his own money just up to the matching funds threshold, when he could have simply taken matching funds in the first place. Furthermore, the Arizona law takes into account all manner of uncoordinated political activity in awarding matching funds. If a privately funded candidate wanted to hover just below the triggering level, he would have to make guesses about how much he will receive in the form of contributions and supportive independent expenditures. He might well guess wrong. In addition, some candidates may be willing to bear the burden of spending above the cap. That a candidate is willing to do so does not make the law any less burden some. See Davis, 554 U. S., at 739 (that candidates may choose to make “personal expenditures to support their campaigns” despite the burdens imposed by the Million aire’s Amendment does not change the fact that “they must shoulder a special and potentially significant burden if they make that choice”). If the State made privately funded candidates pay a $500 fine to run as such, the fact that candidates might choose to pay it does not make the fine any less burdensome. 20 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court While there is evidence to support the contention of the candidates and independent expenditure groups that the matching funds provision burdens their speech, “it is never easy to prove a negative”—here, that candidates and groups did not speak or limited their speech because of the Arizona law. Elkins v. United States, 364 U. S. 206, 218 (1960). In any event, the burden imposed by the matching funds provision is evident and inherent in the choice that confronts privately financed candidates and independent expenditure groups. Cf. Davis, 554 U. S., at 738–740. Indeed even candidates who sign up for public funding recognize the burden matching funds impose on private speech, stating that they participate in the program be cause “matching funds . . . discourage[ ] opponents, special interest groups, and lobbyists from campaigning against” them. GAO, Campaign Finance Reform: Experiences of Two States that Offered Full Public Funding for Political Candidates 27 (GAO–10–390, 2010). As in Davis, we do not need empirical evidence to determine that the law at issue is burdensome. See 554 U. S., at 738–740 (requiring no evidence of a burden whatsoever). It is clear not only to us but to every other court to have considered the question after Davis that a candidate or independent group might not spend money if the direct result of that spending is additional funding to political adversaries. See, e.g., Green Party of Conn., 616 F. 3d, at 242 (matching funds impose “a substantial burden on the exercise of First Amendment rights” (internal quotation marks omitted)); McComish v. Bennett, 611 F. 3d, at 524 (matching funds create “potential chilling effects” and “impose some First Amendment burden”); Scott v. Roberts, 612 F. 3d 1279, 1290 (CA11 2010) (“we think it is obvious that the [matching funds] subsidy imposes a burden on [privately financed] candidates”); id., at 1291 (“we know of no court that doubts that a [matching funds] subsidy like the one at issue here burdens” the speech of privately financed candidates); see also Day v. Holahan, 34 F. 3d 1356, 1360 (CA8 1994) (it is “clear” that matching funds provisions infringe on “protected speech because of the chilling effect” they have “on the political speech of the person or group making the [triggering] expenditure” (cited in Davis, supra, at 739 )). The dissent’s disagree ment is little more than disagreement with Davis. The State correctly asserts that the candidates and independent expenditure groups “do not . . . claim that a single lump sum payment to publicly funded candidates,” equivalent to the maximum amount of state financing that a candidate can obtain through matching funds, would impermissibly burden their speech. Brief for State Re spondents 56; see Tr. of Oral Arg. 5. The State reasons that if providing all the money up front would not burden speech, providing it piecemeal does not do so either. And the State further argues that such incremental admini stration is necessary to ensure that public funding is not under- or over-distributed. See Brief for State Respon dents 56–57. These arguments miss the point. It is not the amount of funding that the State provides to publicly financed candi dates that is constitutionally problematic in this case. It is the manner in which that funding is provided—in direct response to the political speech of privately financed can didates and independent expenditure groups. And the fact that the State’s matching mechanism may be more effi cient than other alternatives—that it may help the State in “finding the sweet-spot” or “fine-tuning” its financing system to avoid a drain on public resources, post, at 26 (KAGAN, J., dissenting)—is of no moment; “the First Amendment does not permit the State to sacrifice speech for efficiency.” Riley v. National Federation of Blind of N. C., Inc., 487 U. S. 781, 795 (1988). The United States as amicus contends that “[p]roviding additional funds to petitioners’ opponents does not make 22 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court petitioners’ own speech any less effective” and thus does not substantially burden speech. Brief for United States 27. Of course it does. One does not have to subscribe to the view that electoral debate is zero sum, see AFEC Brief 30, to see the flaws in the United States’ perspective. All else being equal, an advertisement supporting the election of a candidate that goes without a response is often more effective than an advertisement that is directly contro verted. And even if the publicly funded candidate decides to use his new money to address a different issue alto gether, the end goal of that spending is to claim electoral victory over the opponent that triggered the additional state funding. See Davis, 554 U. S., at 736 . B Because the Arizona matching funds provision imposes a substantial burden on the speech of privately financed candidates and independent expenditure groups, “that pro vision cannot stand unless it is ‘justified by a compelling state interest,’ ” id., at 740 (quoting Massachusetts Citizens for Life, 479 U. S., at 256 ). There is a debate between the parties in this case as to what state interest is served by the matching funds provi sion. The privately financed candidates and independent expenditure groups contend that the provision works to “level[ ] electoral opportunities” by equalizing candidate “resources and influence.” Brief for Petitioner McComish et al. 64; see AFEC Brief 23. The State and the Clean Elections Institute counter that the provision “furthers Arizona’s interest in preventing corruption and the ap pearance of corruption.” Brief for State Respondents 42; Brief for Respondent Clean Elections Institute 47. [1] There is ample support for the argument that the matching funds provision seeks to “level the playing field” in terms of candidate resources. The clearest evidence is of course the very operation of the provision: It ensures that campaign funding is equal, up to three times the initial public funding allotment. The text of the Citizens Clean Elections Act itself confirms this purpose. The sta tutory provision setting up the matching funds regime is titled “Equal funding of candidates.” Ariz. Rev. Stat. Ann. §16–952 (West Supp. 2010). The Act refers to the funds doled out after the Act’s matching mechanism is triggered as “equalizing funds.” See §§16–952(C)(4), (5). And the regulations implementing the matching funds provision refer to those funds as “equalizing funds” as well. See Citizens Clean Elections Commission, Ariz. Admin. Rule R2–20–113. Other features of the Arizona law reinforce this under standing of the matching funds provision. If the Citizens Clean Election Commission cannot provide publicly fi nanced candidates with the moneys that the matching funds provision envisions because of a shortage of funds, the statute allows a publicly financed candidate to “accept private contributions to bring the total monies received by the candidate” up to the matching funds amount. Ariz. Rev. Stat. Ann. §16–954(F) (West 2006). Limiting contri butions, of course, is the primary means we have upheld to combat corruption. Buckley, 424 U. S., at 23–35, 46–47. Indeed the State argues that one of the principal ways that the matching funds provision combats corruption is by eliminating the possibility of any quid pro quo between private interests and publicly funded candidates by elimi nating contributions to those candidates altogether. See Brief for State Respondents 45–46. But when confronted with a choice between fighting corruption and equalizing speech, the drafters of the matching funds provision chose the latter. That significantly undermines any notion that the “Equal funding of candidates” provision is meant to serve some interest other than an interest in equalizing 24 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court funds.10 We have repeatedly rejected the argument that the government has a compelling state interest in “leveling the playing field” that can justify undue burdens on politi cal speech. See, e.g., Citizens United, 558 U. S., at ___ (slip op., at 34). In Davis, we stated that discriminatory contribution limits meant to “level electoral opportunities for candidates of different personal wealth” did not serve “a legitimate government objective,” let alone a compelling one. 554 U. S., at 741 (internal quotation marks omitted). And in Buckley, we held that limits on overall campaign expenditures could not be justified by a purported govern ment “interest in equalizing the financial resources of candidates.” 424 U. S., at 56 ; see id., at 56–57. After all, equalizing campaign resources “might serve not to equal ize the opportunities of all candidates, but to handicap a candidate who lacked substantial name recognition or exposure of his views before the start of the campaign.” Id., at 57 . “Leveling electoral opportunities means making and implementing judgments about which strengths should be permitted to contribute to the outcome of an election,” Davis, supra, at 742—a dangerous enterprise and one that cannot justify burdening protected speech. The dissent essentially dismisses this concern, see post, at 27–29, but it needs to be taken seriously; we have, as noted, held that it is not legitimate for the government to attempt to equal ize electoral opportunities in this manner. And such basic —————— 10 Priorto oral argument in this case, the Citizens Clean Elections Commission’s Web site stated that “ ‘The Citizens Clean Elections Act was passed by the people of Arizona in 1998 to level the playing field when it comes to running for office.’ ” AFEC Brief 10, n. 3 (quoting http://www.azcleanelections.gov/about-us/get-involved.aspx); Tr. of Oral Arg. 48. The Web site now says that “The Citizens Clean Elections Act was passed by the people of Arizona in 1998 to restore citizen participa tion and confidence in our political system.” intrusion by the government into the debate over who should govern goes to the heart of First Amendment values. “Leveling the playing field” can sound like a good thing. But in a democracy, campaigning for office is not a game. It is a critically important form of speech. The First Amendment embodies our choice as a Nation that, when it comes to such speech, the guiding principle is freedom— the “unfettered interchange of ideas”—not whatever the State may view as fair. Buckley, supra, at 14 (internal quotation marks omitted). [2] As already noted, the State and the Clean Elections Institute disavow any interest in “leveling the playing field.” They instead assert that the “Equal funding of candidates” provision, Ariz. Rev. Stat. Ann. §16–952 (West Supp. 2010), serves the State’s compelling interest in combating corruption and the appearance of corruption. See, e.g., Davis, supra, at 740 ; Wisconsin Right to Life, 551 U. S., at 478–479 (opinion of ROBERTS, C. J.). But even if the ultimate objective of the matching funds provision is to combat corruption—and not “level the playing field”—the burdens that the matching funds provision imposes on protected political speech are not justified. Burdening a candidate’s expenditure of his own funds on his own campaign does not further the State’s anticor ruption interest. Indeed, we have said that “reliance on personal funds reduces the threat of corruption” and that “discouraging [the] use of personal funds[ ] disserves the anticorruption interest.” Davis, supra, at 740–741. That is because “the use of personal funds reduces the candi date’s dependence on outside contributions and thereby counteracts the coercive pressures and attendant risks of abuse” of money in politics. Buckley, supra, at 53 . The matching funds provision counts a candidate’s expendi 26 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court tures of his own money on his own campaign as contribu tions, and to that extent cannot be supported by any anti corruption interest. We have also held that “independent expenditures . . . do not give rise to corruption or the appearance of corrup tion.” Citizens United, 558 U. S., at ___ (slip op., at 42). “By definition, an independent expenditure is political speech presented to the electorate that is not coordinated with a candidate.” Id., at ___ (slip op., at 44). The candidate-funding circuit is broken. The separation between candidates and independent expenditure groups negates the possibility that independent expenditures will result in the sort of quid pro quo corruption with which our case law is concerned. See id., at ___–___ (slip op., at 42–45); cf. Buckley, 424 U. S., at 46 . Including independent ex penditures in the matching funds provision cannot be supported by any anticorruption interest. We have observed in the past that “[t]he interest in alleviating the corrupting influence of large contributions is served by . . . contribution limitations.” Id., at 55 . Arizona already has some of the most austere contribution limits in the United States. See Randall v. Sorrell, 548 U. S. 230 , 250–251 (2006) (plurality opinion). Contribu tions to statewide candidates are limited to $840 per contributor per election cycle and contributions to legisla tive candidates are limited to $410 per contributor per election cycle. See Ariz. Rev. Stat. Ann. §§16–905(A)(1), 941(B)(1); Ariz. Dept. of State, Office of the Secretary of State, 2009–2010 Contribution Limits, see supra, at 4. Arizona also has stringent fundraising disclosure re quirements. In the face of such ascetic contribution limits, strict disclosure requirements, and the general availability of public funding, it is hard to imagine what marginal corruption deterrence could be generated by the matching funds provision. Perhaps recognizing that the burdens the matching funds provision places on speech cannot be justified in and of themselves, either as a means of leveling the play ing field or directly fighting corruption, the State and the Clean Elections Institute offer another argument: They contend that the provision indirectly serves the anticor ruption interest, by ensuring that enough candidates participate in the State’s public funding system, which in turn helps combat corruption.11 See Brief for State Re spondents 46–47; Brief for Respondent Clean Elections Institute 47–49. We have said that a voluntary system of “public financing as a means of eliminating the improper influence of large private contributions furthers a signifi cant governmental interest.” Buckley, supra, at 96 . But the fact that burdening constitutionally protected speech might indirectly serve the State’s anticorruption interest, by encouraging candidates to take public financing, does not establish the constitutionality of the matching funds provision. We have explained that the matching funds provision substantially burdens the speech of privately financed candidates and independent groups. It does so to an even greater extent than the law we invalidated in Davis. We have explained that those burdens cannot be justified by a desire to “level the playing field.” We have also explained that much of the speech burdened by the matching funds provision does not, under our precedents, pose a danger of corruption. In light of the foregoing analysis, the fact that the State may feel that the matching funds provision is —————— 11 The State claims that the Citizens Clean Elections Act was passed in response to rampant corruption in Arizona politics—elected officials “literally taking duffle bags full of cash in exchange for sponsoring legislation.” Brief for State Respondents 45. That may be. But, as the candidates and independent expenditure groups point out, the corrup tion that plagued Arizona politics is largely unaddressed by the match ing funds regime. AFEC Brief 11, n. 4. Public financing does nothing to prevent politicians from accepting bribes in exchange for their votes. 28 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court necessary to allow it to “find[ ] the sweet-spot” and “fine tun[e]” its public funding system, post, at 26 (KAGAN, J., dissenting), to achieve its desired level of participation without an undue drain on public resources, is not a suffi cient justification for the burden. The flaw in the State’s argument is apparent in what its reasoning would allow. By the State’s logic it could grant a publicly funded candidate five dollars in matching funds for every dollar his privately financed opponent spent, or force candidates who wish to run on private funds to pay a $10,000 fine in order to encourage participation in the public funding regime. Such measures might well pro mote participation in public financing, but would clearly suppress or unacceptably alter political speech. How the State chooses to encourage participation in its public funding system matters, and we have never held that a State may burden political speech—to the extent the matching funds provision does—to ensure adequate par ticipation in a public funding system. Here the State’s chosen method is unduly burdensome and not sufficiently justified to survive First Amendment scrutiny. III We do not today call into question the wisdom of public financing as a means of funding political candidacy. That is not our business. But determining whether laws gov erning campaign finance violate the First Amendment is very much our business. In carrying out that responsibil ity over the past 35 years, we have upheld some restric tions on speech and struck down others. See, e.g., Buckley, supra, at 35–38, 51–54 (upholding contribution limits and striking down expenditure limits); Colorado I, 518 U. S., at 608 (opinion of BREYER, J.) (invalidating ban on inde pendent expenditures for electioneering communication); Colorado II, 533 U. S., at 437 (upholding caps on coordi nated party expenditures); Davis, 554 U. S., at 736 (in validating asymmetrical contribution limits triggered by candidate spending). We have said that governments “may engage in public financing of election campaigns” and that doing so can further “significant governmental interest[s],” such as the state interest in preventing corruption. Buckley, 424 U. S., at 57, n. 65 , 92–93, 96. But the goal of creating a viable public financing scheme can only be pursued in a manner consistent with the First Amendment. The dis sent criticizes the Court for standing in the way of what the people of Arizona want. Post, at 2–3, 31–32. But the whole point of the First Amendment is to protect speakers against unjustified government restrictions on speech, even when those restrictions reflect the will of the major ity. When it comes to protected speech, the speaker is sovereign. Arizona’s program gives money to a candidate in direct response to the campaign speech of an opposing candidate or an independent group. It does this when the opposing candidate has chosen not to accept public financing, and has engaged in political speech above a level set by the State. The professed purpose of the state law is to cause a sufficient number of candidates to sign up for public fi nancing, see post, at 5, which subjects them to the various restrictions on speech that go along with that program. This goes too far; Arizona’s matching funds provision substantially burdens the speech of privately financed candidates and independent expenditure groups without serving a compelling state interest. “[T]here is practically universal agreement that a major purpose of” the First Amendment “was to protect the free discussion of governmental affairs,” “includ[ing] discus sions of candidates.” Buckley, 424 U. S., at 14 (internal quotation marks omitted; second alteration in original). That agreement “reflects our ‘profound national commit ment to the principle that debate on public issues should 30 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT Opinion of the Court be uninhibited, robust, and wide-open.’ ” Ibid. (quoting New York Times Co. v. Sullivan, 376 U. S. 254, 270 (1964)). True when we said it and true today. Laws like Arizona’s matching funds provision that inhibit robust and wide-open political debate without sufficient justification cannot stand. The judgment of the Court of Appeals for the Ninth Circuit is reversed. It is so ordered. SUPREME COURT OF THE UNITED STATES _________________ Nos. 10–238 and 10–239 _________________ ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC, ET AL., PETITIONERS 10–238 v. KEN BENNETT, IN HIS OFFICIAL CAPACITY AS ARIZONA SECRETARY OF STATE, ET AL. JOHN MCCOMISH, ET AL., PETITIONERS 10–239 v. KEN BENNETT, IN HIS OFFICIAL CAPACITY AS ARIZONA SECRETARY OF STATE, ET AL. ON WRITS OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT [June 27, 2011] JUSTICE KAGAN, with whom JUSTICE GINSBURG, JUSTICE BREYER, and JUSTICE SOTOMAYOR join, dissenting. Imagine two States, each plagued by a corrupt political system. In both States, candidates for public office accept large campaign contributions in exchange for the promise that, after assuming office, they will rank the donors’ interests ahead of all others. As a result of these bargains, politicians ignore the public interest, sound public policy languishes, and the citizens lose confidence in their government. Recognizing the cancerous effect of this corruption, voters of the first State, acting through referendum, enact several campaign finance measures previously approved by this Court. They cap campaign contributions; require disclosure of substantial donations; and create an optional 2 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting public financing program that gives candidates a fixed public subsidy if they refrain from private fundraising. But these measures do not work. Individuals who “bun dle” campaign contributions become indispensable to candidates in need of money. Simple disclosure fails to prevent shady dealing. And candidates choose not to participate in the public financing system because the sums provided do not make them competitive with their privately financed opponents. So the State remains af flicted with corruption. Voters of the second State, having witnessed this fail ure, take an ever-so-slightly different tack to cleaning up their political system. They too enact contribution limits and disclosure requirements. But they believe that the greatest hope of eliminating corruption lies in creating an effective public financing program, which will break can didates’ dependence on large donors and bundlers. These voters realize, based on the first State’s experience, that such a program will not work unless candidates agree to participate in it. And candidates will participate only if they know that they will receive sufficient funding to run competitive races. So the voters enact a program that carefully adjusts the money given to would-be officehold ers, through the use of a matching funds mechanism, in order to provide this assurance. The program does not discriminate against any candidate or point of view, and it does not restrict any person’s ability to speak. In fact, by providing resources to many candidates, the program creates more speech and thereby broadens public debate. And just as the voters had hoped, the program accom plishes its mission of restoring integrity to the political system. The second State rids itself of corruption. A person familiar with our country’s core values—our devotion to democratic self-governance, as well as to “un inhibited, robust, and wide-open” debate, New York Times Co. v. Sullivan, 376 U. S. 254, 270 (1964)—might expect this Court to celebrate, or at least not to interfere with, the second State’s success. But today, the majority holds that the second State’s system—the system that produces honest government, working on behalf of all the people— clashes with our Constitution. The First Amendment, the majority insists, requires us all to rely on the measures employed in the first State, even when they have failed to break the stranglehold of special interests on elected officials. I disagree. The First Amendment’s core purpose is to foster a healthy, vibrant political system full of robust discussion and debate. Nothing in Arizona’s anti corruption statute, the Arizona Citizens Clean Elections Act, violates this constitutional protection. To the con trary, the Act promotes the values underlying both the First Amendment and our entire Constitution by enhanc ing the “opportunity for free political discussion to the end that government may be responsive to the will of the people.” Id., at 269 (internal quotation marks omitted). I therefore respectfully dissent. I A Campaign finance reform over the last century has focused on one key question: how to prevent massive pools of private money from corrupting our political system. If an officeholder owes his election to wealthy contributors, he may act for their benefit alone, rather than on behalf of all the people. As we recognized in Buckley v. Valeo, 424 U. S. 1, 26 (1976) (per curiam), our seminal campaign finance case, large private contributions may result in “political quid pro quo[s],” which undermine the integrity of our democracy. And even if these contributions are not converted into corrupt bargains, they still may weaken confidence in our political system because the public per ceives “the opportunities for abuse[s].” Id., at 27 . To 4 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting prevent both corruption and the appearance of corrup tion—and so to protect our democratic system of govern ance—citizens have implemented reforms designed to curb the power of special interests. Among these measures, public financing of elections has emerged as a potentially potent mechanism to preserve elected officials’ independence. President Theodore Roo sevelt proposed the reform as early as 1907 in his State of the Union address. “The need for collecting large cam paign funds would vanish,” he said, if the government “pro vided an appropriation for the proper and legitimate ex penses” of running a campaign, on the condition that a “party receiving campaign funds from the Treasury” would forgo private fundraising. 42 Cong. Rec. 78 (1907). The idea was—and remains—straightforward. Candidates who rely on public, rather than private, moneys are “be holden [to] no person and, if elected, should feel no post election obligation toward any contributor.” Republican Nat. Comm. v. FEC, 487 F. Supp. 280, 284 (SDNY), aff’d 445 U. S. 955 (1980). By supplanting private cash in elections, public financing eliminates the source of politi cal corruption. For this reason, public financing systems today dot the national landscape. Almost one-third of the States have adopted some form of public financing, and so too has the Federal Government for presidential elections. See R. Garrett, Congressional Research Service Report for Congress, Public Financing of Congressional Campaigns: Overview and Analysis 2, 32 (2009). The federal pro gram—which offers presidential candidates a fixed public subsidy if they abstain from private fundraising— originated in the campaign finance law that Congress enacted in 1974 on the heels of the Watergate scandal. Congress explained at the time that the “potentia[l] for abuse” inherent in privately funded elections was “all too clear.” S. Rep. No. 93–689, p. 4 (1974). In Congress’s view, public financing represented the “only way . . . [to] eliminate reliance on large private contributions” and its attendant danger of corruption, while still ensuring that a wide range of candidates had access to the ballot. Id., at 5 (emphasis deleted). We declared the presidential public financing system con stitutional in Buckley v. Valeo. Congress, we stated, had created the program “for the ‘general welfare’—to reduce the deleterious influence of large contributions on our political process,” as well as to “facilitate communica tion by candidates with the electorate, and to free candi dates from the rigors of fundraising.” 424 U. S., at 91 . We reiterated “that public financing as a means of eliminat ing the improper influence of large private contributions furthers a significant governmental interest.” Id., at 96 . And finally, in rejecting a challenge based on the First Amendment, we held that the program did not “restrict[] or censor speech, but rather . . . use[d] public money to facilitate and enlarge public discussion and participation in the electoral process.” Id., at 92–93. We declared this result “vital to a self-governing people,” and so concluded that the program “further[ed], not abridge[d], pertinent First Amendment values.” Id., at 93 . We thus gave state and municipal governments the green light to adopt public financing systems along the presidential model. But this model, which distributes a lump-sum grant at the beginning of an election cycle, has a significant weak ness: It lacks a mechanism for setting the subsidy at a level that will give candidates sufficient incentive to par ticipate, while also conserving public resources. Public financing can achieve its goals only if a meaningful num ber of candidates receive the state subsidy, rather than raise private funds. See 611 F. 3d 510, 527 (CA9 2010) (“A public financing system with no participants does nothing to reduce the existence or appearance of quid pro quo corruption”). But a public funding program must be vol 6 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting untary to pass constitutional muster, because of its re strictions on contributions and expenditures. See Buckley, 424 U. S., at 57, n. 65, 95 . And candidates will choose to sign up only if the subsidy provided enables them to run competitive races. If the grant is pegged too low, it puts the participating candidate at a disadvantage: Because he has agreed to spend no more than the amount of the sub sidy, he will lack the means to respond if his privately funded opponent spends over that threshold. So when lump-sum grants do not keep up with campaign expendi tures, more and more candidates will choose not to par ticipate.1 But if the subsidy is set too high, it may impose an unsustainable burden on the public fisc. See 611 F. 3d, at 527 (noting that large subsidies would make public funding “prohibitively expensive and spell its doom”). At the least, hefty grants will waste public resources in the many state races where lack of competition makes such funding unnecessary. The difficulty, then, is in finding the Goldilocks solu tion—not too large, not too small, but just right. And this in a world of countless variables—where the amount of money needed to run a viable campaign against a pri —————— 1 The problem is apparent in the federal system. In recent years, the number of presidential candidates opting to receive public financing has declined because the subsidy has not kept pace with spending by privately financed candidates. See Corrado, Public Funding of Presidential Campaigns, in The New Campaign Finance Sourcebook 180, 200 (A. Corrado, T. Mann, D. Ortiz, & T. Potter eds. 2005). The last election cycle offers a stark example: Then-candidate Barack Obama raised $745.7 million in private funds in 2008, Federal Election Commission, 2008 Presidential Campaign Financial Activity Summa rized, June 8, 2009, online at http://www.fec.gov/press/press2009/ 20090608PresStat.shtml, in contrast with the $105.4 million he could have received in public funds, see Federal Election Commission, Presi dential Election Campaign Fund, online at http://www.fec.gov/press/ bkgnd/fund.shtml (all Internet materials as visited June 24, 2011, and available in Clerk of Court’s case file). vately funded candidate depends on, among other things, the district, the office, and the election cycle. A state may set lump-sum grants district-by-district, based on spend ing in past elections; but even that approach leaves out many factors—including the resources of the privately funded candidate—that alter the competitiveness of a seat from one election to the next. See App. 714–716 (record evidence chronicling the history of variation in campaign spending levels in Arizona’s legislative districts). In short, the dynamic nature of our electoral system makes ex ante predictions about campaign expenditures almost impossi ble. And that creates a chronic problem for lump-sum public financing programs, because inaccurate estimates produce subsidies that either dissuade candidates from participating or waste taxpayer money. And so States have made adjustments to the lump-sum scheme that we approved in Buckley, in attempts to more effectively re duce corruption. B The people of Arizona had every reason to try to develop effective anti-corruption measures. Before turning to pub lic financing, Arizonans voted by referendum to estab lish campaign contribution limits. See Ariz. Rev. Stat. Ann. §16–905 (West Supp. 2010). But that effort to abate corruption, standing alone, proved unsuccessful. Five years after the enactment of these limits, the State suf fered “the worst public corruption scandal in its history.” Brief for State Respondents 1. In that scandal, known as “AzScam,” nearly 10% of the State’s legislators were caught accepting campaign contributions or bribes in ex change for supporting a piece of legislation. Following that incident, the voters of Arizona decided that further reform was necessary. Acting once again by referendum, they adopted the public funding system at issue here. The hallmark of Arizona’s program is its inventive 8 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting approach to the challenge that bedevils all public financ ing schemes: fixing the amount of the subsidy. For each electoral contest, the system calibrates the size of the grant automatically to provide sufficient—but no more than sufficient—funds to induce voluntary participation. In effect, the program’s designers found the Goldilocks solution, which produces the “just right” grant to ensure that a participant in the system has the funds needed to run a competitive race. As the Court explains, Arizona’s matching funds ar rangement responds to the shortcoming of the lump-sum model by adjusting the public subsidy in each race to re flect the expenditures of a privately financed candidate and the independent groups that support him. See Ariz. Rev. Stat. Ann. §16–940 et seq. (West 2006 and West Supp. 2010). A publicly financed candidate in Arizona receives an initial lump-sum to get his campaign off the ground. See §16–951 (West 2006). But for every dollar his privately funded opponent (or the opponent’s supporters) spends over the initial subsidy, the publicly funded candi date will—to a point—get an additional 94 cents. See §16–952 (West Supp. 2010). Once the publicly financed candidate has received three times the amount of the initial disbursement, he gets no further public funding, see ibid., and remains barred from receiving private contribu tions, no matter how much more his privately funded opponent spends, see §16–941(A). This arrangement, like the lump-sum model, makes use of a pre-set amount to provide financial support to partici pants. For example, all publicly funded legislative candi dates collect an initial grant of $21,479 for a general elec tion race. And they can in no circumstances receive more than three times that amount ($64,437); after that, their privately funded competitors hold a marked advantage. But the Arizona system improves on the lump-sum model in a crucial respect. By tying public funding to private spending, the State can afford to set a more generous upper limit—because it knows that in each campaign it will only have to disburse what is necessary to keep a par ticipating candidate reasonably competitive. Arizona can therefore assure candidates that, if they accept public funds, they will have the resources to run a viable race against those who rely on private money. And at the same time, Arizona avoids wasting taxpayers’ dollars. In this way, the Clean Elections Act creates an effective and sustainable public financing system. The question here is whether this modest adjustment to the public financing program that we approved in Buckley makes the Arizona law unconstitutional. The majority contends that the matching funds provision “substantially burdens protected political speech” and does not “serv[e] a compelling state interest.” Ante, at 2. But the Court is wrong on both counts. II Arizona’s statute does not impose a “restriction,” ante, at 15, or “substantia[l] burde[n],” ante, at 2, on expression. The law has quite the opposite effect: It subsidizes and so produces more political speech. We recognized in Buckley that, for this reason, public financing of elections “facili tate[s] and enlarge[s] public discussion,” in support of First Amendment values. 424 U. S., at 92–93. And what we said then is just as true today. Except in a world gone topsy-turvy, additional campaign speech and electoral competition is not a First Amendment injury. A At every turn, the majority tries to convey the im pression that Arizona’s matching fund statute is of a piece with laws prohibiting electoral speech. The majority invokes the language of “limits,” “bar[s],” and “restraints.” Ante, at 8–9. It equates the law to a “restrictio[n] on the 10 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting amount of money a person or group can spend on political communication during a campaign.” Ante, at 15 (internal quotation marks omitted). It insists that the statute “re strict[s] the speech of some elements of our society” to enhance the speech of others. Ibid. (internal quotation marks omitted). And it concludes by reminding us that the point of the First Amendment is to protect “against unjustified government restrictions on speech.” Ante, at 29. There is just one problem. Arizona’s matching funds provision does not restrict, but instead subsidizes, speech. The law “impose[s] no ceiling on [speech] and do[es] not prevent anyone from speaking.” Citizens United v. Fed eral Election Comm’n, 558 U. S. ___ , ___ (2010) (slip op., at 51) (citation and internal quotation marks omitted); see Buckley, 424 U. S., at 92 (holding that a public financing law does not “abridge, restrict, or censor” expression). The statute does not tell candidates or their supporters how much money they can spend to convey their message, when they can spend it, or what they can spend it on. Rather, the Arizona law, like the public financing statute in Buckley, provides funding for political speech, thus “facilitat[ing] communication by candidates with the elec torate.” Id., at 91 . By enabling participating candidates to respond to their opponents’ expression, the statute expands public debate, in adherence to “our tradition that more speech, not less, is the governing rule.” Citizens United, 558 U. S., at ___ (slip op., at 45). What the law does—all the law does—is fund more speech.2 And under the First Amendment, that makes all the —————— 2 And the law appears to do that job well. Between 1998 (when the statute was enacted) and 2006, overall candidate expenditures in creased between 29% and 67%; overall independent expenditures rose by a whopping 253%; and average candidate expenditures grew by 12% to 40%. App. to Pet. for Cert. in No. 10–239, pp. 284–285; App. 916– 917. difference. In case after case, year upon year, we have distinguished between speech restrictions and speech subsidies. “ ‘There is a basic difference,’ ” we have held, “ ‘between direct state interference with [First Amend ment] protected activity and state encouragement’ ” of other expression. Rust v. Sullivan, 500 U. S. 173, 193 (1991) (quoting Maher v. Roe, 432 U. S. 464, 475 (1977)); see also, e.g., Federal Election Comm’n v. Massachusetts Citizens for Life, Inc., 479 U. S. 238, 256, n. 9 (1986); Regan v. Taxation With Representation of Wash., 461 U. S. 540, 550 (1983); National Endowment for Arts v. Finley, 524 U. S. 569 , 587–588 (1998); id., at 599 (SCALIA, J., concurring in judgment) (noting the “fundamental divide” between “ ‘abridging’ speech and funding it”). Government subsidies of speech, designed “to stimulate . . . expres sion[,] . . . [are] consistent with the First Amendment,” so long as they do not discriminate on the basis of viewpoint. Board of Regents of Univ. of Wis. System v. Southworth, 529 U. S. 217, 234 (2000); see, e.g., Rosenberger v. Rector and Visitors of Univ. of Va., 515 U. S. 819, 834 (1995); Finley, 524 U. S., at 587–588. That is because subsidies, by definition and contra the majority, do not restrict any speech. No one can claim that Arizona’s law discriminates against particular ideas, and so violates the First Amend ment’s sole limitation on speech subsidies. The State throws open the doors of its public financing program to all candidates who meet minimal eligibility requirements and agree not to raise private funds. Republicans and Democrats, conservatives and liberals may participate; so too, the law applies equally to independent expenditure groups across the political spectrum. Arizona disburses funds based not on a candidate’s (or supporter’s) ideas, but on the candidate’s decision to sign up for public funding. So under our precedent, Arizona’s subsidy statute should 12 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting easily survive First Amendment scrutiny.3 This suit, in fact, may merit less attention than any challenge to a speech subsidy ever seen in this Court. In the usual First Amendment subsidy case, a person com plains that the government declined to finance his speech, while bankrolling someone else’s; we must then decide whether the government differentiated between these speakers on a prohibited basis—because it preferred one speaker’s ideas to another’s. See, e.g., id., at 577–578; Regan, 461 U. S., at 543–545. But the candidates bringing this challenge do not make that claim—because they were never denied a subsidy. Arizona, remember, offers to support any person running for state office. Petitioners here refused that assistance. So they are making a novel argument: that Arizona violated their First Amendment rights by disbursing funds to other speakers even though they could have received (but chose to spurn) the same financial assistance. Some people might call that chutzpah. Indeed, what petitioners demand is essentially a right to —————— 3 The majority claims that none of our subsidy cases involved the funding of “respons[ive]” expression. See ante, at 17. But the majority does not explain why this distinction, created to fit the facts of this case, should matter so long as the government is not discriminating on the basis of viewpoint. Indeed, the difference the majority highlights should cut in the opposite direction, because facilitating responsive speech fosters “uninhibited, robust, and wide-open” public debate. New York Times Co. v. Sullivan, 376 U. S. 254, 270 (1964). In any event, the majority is wrong to say that we have never approved funding to “allow the recipient to counter” someone else’s political speech. Ante, at 17. That is exactly what we approved in Buckley v. Valeo, 424 U. S. 1 (1976) (per curiam). See supra, at 5. The majority notes that the public financing scheme in Buckley lacked the trigger mechanism used in the Arizona law. See ante, at 17, n. 9. But again, that is just to describe a difference, not to say why it matters. As I will show, the trigger is constitutionally irrelevant—as we made clear in the very case (Davis v. Federal Election Comm’n, 554 U. S. 724 (2008)) on which the majority principally relies. See infra, at 17–19, 21–22. quash others’ speech through the prohibition of a (univer sally available) subsidy program. Petitioners are able to convey their ideas without public financing—and they would prefer the field to themselves, so that they can speak free from response. To attain that goal, they ask this Court to prevent Arizona from funding electoral speech—even though that assistance is offered to every state candidate, on the same (entirely unobjectionable) basis. And this Court gladly obliges. If an ordinary citizen, without the hindrance of a law degree, thought this result an upending of First Amend ment values, he would be correct. That Amendment protects no person’s, nor any candidate’s, “right to be free from vigorous debate.” Pacific Gas & Elec. Co. v. Public Util. Comm’n of Cal., 475 U. S. 1, 14 (1986) (plurality opinion). Indeed, the Amendment exists so that this de bate can occur—robust, forceful, and contested. It is the theory of the Free Speech Clause that “falsehood and fallacies” are exposed through “discussion,” “education,” and “more speech.” Whitney v. California, 274 U. S. 357, 377 (1927) (Brandeis, J., concurring). Or once again from Citizens United: “[M]ore speech, not less, is the governing rule.” 558 U. S., at ___ (slip op., at 45). And this is no place more true than in elections, where voters’ ability to choose the best representatives depends on debate—on charge and countercharge, call and response. So to invali date a statute that restricts no one’s speech and dis criminates against no idea—that only provides more voices, wider discussion, and greater competition in elec tions—is to undermine, rather than to enforce, the First Amendment.4 —————— 4 The majority argues that more speech will quickly become “less speech,” as candidates switch to public funding. Ante, at 15, n. 7. But that claim misunderstands how a voluntary public financing system works. Candidates with significant financial resources will likely decline public funds, so that they can spend in excess of the system’s 14 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting We said all this in Buckley, when we upheld the presi dential public financing system—a ruling this Court has never since questioned. The principal challenge to that system came from minor-party candidates not eligible for benefits—surely more compelling plaintiffs than petition ers, who could have received funding but refused it. Yet we rejected that attack in part because we understood the federal program as supporting, rather than interfering with, expression. See 424 U. S., at 90–108; see also Regan, 461 U. S., at 549 (relying on Buckley to hold that selective subsidies of expression comport with the First Amendment if they are viewpoint neutral). Buckley re jected any idea, along the lines the majority proposes, that a subsidy of electoral speech was in truth a restraint. And more: Buckley recognized that public financing of elections fosters First Amendment principles. “[T]he central pur pose of the Speech and Press Clauses,” we explained, “was to assure a society in which ‘uninhibited, robust, and wide open’ public debate concerning matters of public interest would thrive, for only in such a society can a healthy representative democracy flourish.” 424 U. S., at 93 , n. 127 (quoting New York Times, 376 U. S., at 270 ). And we continued: “[L]aws providing financial assistance to the exercise of free speech”—including the campaign finance statute at issue—“enhance these First Amendment val ues.” 424 U. S., at 93, n. 127 . We should be saying the same today. B The majority has one, and only one, way of separating this case from Buckley and our other, many precedents —————— expenditure caps. Other candidates accept public financing because they believe it will enhance their communication with voters. So the system continually pushes toward more speech. That is exactly what has happened in Arizona, see n. 2, supra, and the majority offers no counter-examples. involving speech subsidies. According to the Court, the special problem here lies in Arizona’s matching funds mechanism, which the majority claims imposes a “sub stantia[l] burde[n]” on a privately funded candidate’s speech. Ante, at 2. Sometimes, the majority suggests that this “burden” lies in the way the mechanism “ ‘diminish[es] the effectiveness’ ” of the privately funded candidate’s expression by enabling his opponent to respond. Ante, at 10 (quoting Davis v. Federal Election Comm’n, 554 U. S. 724, 736 (2008)); see ante, at 21–22. At other times, the majority indicates that the “burden” resides in the deter rent effect of the mechanism: The privately funded candi date “might not spend money” because doing so will trig ger matching funds. Ante, at 20. Either way, the majority is wrong to see a substantial burden on expression.5 Most important, and as just suggested, the very notion that additional speech constitutes a “burden” is odd and unsettling. Here is a simple fact: Arizona imposes nothing remotely resembling a coercive penalty on privately funded candidates. The State does not jail them, fine them, or subject them to any kind of lesser disability. (So the majority’s analogies to a fine on speech, ante, at 19, 28, are inapposite.) The only “burden” in this case comes from the grant of a subsidy to another person, and the opportu nity that subsidy allows for responsive speech. But that —————— 5 The majority’s error on this score extends both to candidates and to independent expenditure groups. Contrary to the majority’s sugges tion, see ante, at 14, n. 6, nearly all of my arguments showing that the Clean Elections Act does not impose a substantial burden apply to both sets of speakers (and apply regardless of whether independent or candidate expenditures trigger the matching funds). That is also true of every one of my arguments demonstrating the State’s compelling interest in this legislation. See infra, at 22–26. But perhaps the best response to the majority’s view that the Act inhibits independent expenditure groups lies in an empirical fact already noted: Expendi tures by these groups have risen by 253% since Arizona’s law was enacted. See n. 2, supra. 16 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting means the majority cannot get out from under our subsidy precedents. Once again: We have never, not once, under stood a viewpoint-neutral subsidy given to one speaker to constitute a First Amendment burden on another. (And that is so even when the subsidy is not open to all, as it is here.) Yet in this case, the majority says that the prospect of more speech—responsive speech, competitive speech, the kind of speech that drives public debate—counts as a constitutional injury. That concept, for all the reasons previously given, is “wholly foreign to the First Amend ment.” Buckley, 424 U. S., at 49 . But put to one side this most fundamental objection to the majority’s argument; even then, has the majority shown that the burden resulting from the Arizona statute is “substantial”? See Clingman v. Beaver, 544 U. S. 581, 592 (2005) (holding that stringent judicial review is “ap propriate only if the burden is severe”). I will not quarrel with the majority’s assertion that responsive speech by one candidate may make another candidate’s speech less effective, see ante, at 21–22; that, after all, is the whole idea of the First Amendment, and a benefit of having more responsive speech. See Abrams v. United States, 250 U. S. 616, 630 (1919) (Holmes., J., dissenting) (“[T]he best test of truth is the power of the thought to get itself accepted in the competition of the market”). And I will assume that the operation of this statute may on occasion deter a pri vately funded candidate from spending money, and con veying ideas by that means.6 My guess is that this does —————— 6I will note, however, that the record evidence of this effect is spotty at best. The majority finds anecdotal evidence supporting its argument on just 6 pages of a 4500-page summary judgment record. See ante, at 18–19. (The majority also cites sections of petitioners’ briefs, which cite the same 6 pages in the record. See ante, at 19.) That is consistent with the assessment of the District Court Judge who presided over the proceedings in this case: He stated that petitioners had presented only “vague” and “scattered” evidence of the law’s deterrent impact. App. to not happen often: Most political candidates, I suspect, have enough faith in the power of their ideas to prefer speech on both sides of an issue to speech on neither. But I will take on faith that the matching funds provision may lead one or another privately funded candidate to stop spending at one or another moment in an election. Still, does that effect count as a severe burden on expression? By the measure of our prior decisions—which have upheld campaign reforms with an equal or greater impact on speech—the answer is no. Number one: Any system of public financing, including the lump-sum model upheld in Buckley, imposes a similar burden on privately funded candidates. Suppose Arizona were to do what all parties agree it could under Buckley— provide a single upfront payment (say, $150,000) to a participating candidate, rather than an initial payment (of $50,000) plus 94% of whatever his privately funded opponent spent, up to a ceiling (the same $150,000). That system would “diminis[h] the effectiveness” of a privately funded candidate’s speech at least as much, and in the same way: It would give his opponent, who presumably would not be able to raise that sum on his own, more money to spend. And so too, a lump-sum system may deter speech. A person relying on private resources might well choose not to enter a race at all, because he knows he will face an adequately funded opponent. And even if he —————— Pet. for Cert. in No. 10–239, p. 54. The appellate court discerned even less evidence of any deterrent effect. Id., at 30 (“No Plaintiff . . . has pointed to any specific instance in which she or he has declined a contribution or failed to make an expenditure for fear of triggering matching funds”); see also id., at 28, 31, 34. I understand the majority to essentially concede this point (“ ‘it is never easy to prove a negative,’ ” ante, at 20) and to say it does not matter (“we do not need empirical evidence,” ibid.). So I will not belabor the issue by detailing the sub stantial testimony (much more than 6 pages worth) that the matching funds provision has not put a dent in privately funded candidates’ spending. 18 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting decides to run, he likely will choose to speak in different ways—for example, by eschewing dubious, easy-to-answer charges—because his opponent has the ability to respond. Indeed, privately funded candidates may well find the lump-sum system more burdensome than Arizona’s (as suming the lump is big enough). Pretend you are financ ing your campaign through private donations. Would you prefer that your opponent receive a guaranteed, upfront payment of $150,000, or that he receive only $50,000, with the possibility—a possibility that you mostly get to con trol—of collecting another $100,000 somewhere down the road? Me too. That’s the first reason the burden on speech cannot command a different result in this case than in Buckley. Number two: Our decisions about disclosure and dis claimer requirements show the Court is wrong. Starting in Buckley and continuing through last Term, the Court has repeatedly declined to view these requirements as a sub stantial First Amendment burden, even though they dis courage some campaign speech. “It is undoubtedly true,” we stated in Buckley, that public disclosure obligations “will deter some individuals” from engaging in ex pressive activity. 424 U. S., at 68 ; see Davis, 554 U. S., at 744 . Yet we had no difficulty upholding these require ments there. And much more recently, in Citizens United and Doe v. Reed, 561 U. S. ___ (2010), we followed that precedent. “ ‘Disclosure requirements may burden the ability to speak,” we reasoned, but they “do not prevent anyone from speaking.’ ” Id., at ___ (slip op., at 7) (quoting Citizens United, 558 U. S., at ___ (slip op., at 51)). So too here. Like a disclosure rule, the matching funds provision may occasionally deter, but “impose[s] no ceiling” on elec toral expression. Id., at ___ (slip op., at 51). The majority breezily dismisses this comparison, label ing the analogy “not even close” because disclosure re quirements result in no payment of money to a speaker’s opponent. Ante, at 18. That is indeed the factual distinc tion: A matching fund provision, we can all agree, is not a disclosure rule. But the majority does not tell us why this difference matters. Nor could it. The majority strikes down the matching funds provision because of its ostensi ble effect—most notably, that it may deter a person from spending money in an election. But this Court has ac knowledged time and again that disclosure obligations have the selfsame effect. If that consequence does not trigger the most stringent judicial review in the one case, it should not do so in the other. Number three: Any burden that the Arizona law im poses does not exceed the burden associated with contri bution limits, which we have also repeatedly upheld. Con tribution limits, we have stated, “impose direct quantity restrictions on political communication and association,” Buckley, 424 U. S., at 18 (emphasis added), thus “ ‘signifi cant[ly] interfer[ing]’ ” with First Amendment interests, Nixon v. Shrink Missouri Government PAC, 528 U. S. 377, 387 (2000) (quoting Buckley, 424 U. S., at 25 ). Rather than potentially deterring or “ ‘diminish[ing] the effective ness’ ” of expressive activity, ante, at 10 (quoting Davis, 554 U. S., at 736 ), these limits stop it cold. Yet we have never subjected these restrictions to the most stringent review. See Buckley, 424 U. S., at 29–38. I doubt I have to reiterate that the Arizona statute imposes no restraints on any expressive activity. So the majority once again has no reason here to reach a different result. In this way, our campaign finance cases join our speech subsidy cases in supporting the constitutionality of Ari zona’s law. Both sets of precedents are in accord that a statute funding electoral speech in the way Arizona’s does imposes no First Amendment injury. C The majority thinks it has one case on its side—Davis v. 20 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting Federal Election Comm’n, 554 U. S. 724—and it pegs everything on that decision. See ante, at 9–12. But Davis relies on principles that fit securely within our First Amendment law and tradition—most unlike today’s opinion. As the majority recounts, Davis addressed the constitu tionality of federal legislation known as the Millionaire’s Amendment. Under that provision (which applied in elec tions not involving public financing), a candidate’s expen diture of more than $350,000 of his own money activated a change in applicable contribution limits. Before, each candidate in the race could accept $2,300 from any donor; but now, the opponent of the self-financing candidate could accept three times that much, or up to $6,900 per contributor. So one candidate’s expenditure of personal funds on campaign speech triggered discriminatory con tribution restrictions favoring that candidate’s opponent. Under the First Amendment, the similarity between Davis and this case matters far less than the differences. Here is the similarity: In both cases, one candidate’s cam paign expenditure triggered . . . something. Now here are the differences: In Davis, the candidate’s expenditure triggered a discriminatory speech restriction, which Con gress could not otherwise have imposed consistent with the First Amendment; by contrast, in this case, the candi date’s expenditure triggers a non-discriminatory speech subsidy, which all parties agree Arizona could have pro vided in the first instance. In First Amendment law, that difference makes a difference—indeed, it makes all the difference. As I have indicated before, two great fault lines run through our First Amendment doctrine: one, between speech restrictions and speech subsidies, and the other, between discriminatory and neutral government action. See supra, at 10–11. The Millionaire’s Amend ment fell on the disfavored side of both divides: To reiter ate, it imposed a discriminatory speech restriction. The Arizona Clean Elections Act lands on the opposite side of both: It grants a non-discriminatory speech subsidy.7 So to say that Davis “largely controls” this case, ante, at 10, is to decline to take our First Amendment doctrine seriously. And let me be clear: This is not my own idiosyncratic or post hoc view of Davis; it is the Davis Court’s self expressed, contemporaneous view. That decision began, continued, and ended by focusing on the Millionaire Amendment’s “discriminatory contribution limits.” 554 U. S., at 740 . We made that clear in the very first sen tence of the opinion, where we summarized the question presented. Id., at 728 (“In this appeal, we consider the constitutionality of federal election law provisions that . . . impose different campaign contribution limits on candi dates”). And our focus on the law’s discriminatory restric tions was evident again when we examined how the Court’s prior holdings informed the case. Id., at 738 (“We have never upheld the constitutionality of a law that imposes different contribution limits for candidates”). And then again, when we concluded that the Millionaire’s Amendment could not stand. Id., at 740 (explaining that the “the activation of a scheme of discriminatory contribu tion limits” burdens speech). Our decision left no doubt (because we repeated the point many times over, see also id., at 729, 730, 739, 740, n. 7, 741, 744 ): The constitu tional problem with the Millionaire’s Amendment lay in its use of discriminatory speech restrictions. —————— 7 Of course, only publicly funded candidates receive the subsidy. But that is because only those candidates have agreed to abide by stringent spending caps (which privately funded candidates can exceed by any amount). And Buckley specifically approved that exchange as consis tent with the First Amendment. See 424 U. S., at 57, n. 65, 95 . By contrast, Davis involved a scheme in which one candidate in a race received concrete fundraising advantages, in the form of asymmetrical contribution limits, just because his opponent had spent a certain amount of his own money. 22 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting But what of the trigger mechanism—in Davis, as here, a candidate’s campaign expenditures? That, after all, is the only thing that this case and Davis share. If Davis had held that the trigger mechanism itself violated the First Amendment, then the case would support today’s holding. But Davis said nothing of the kind. It made clear that the trigger mechanism could not rescue the discriminatory contribution limits from constitutional invalidity; that the limits went into effect only after a candidate spent sub stantial personal resources rendered them no more per missible under the First Amendment. See id., at 739. But Davis did not call into question the trigger mechanism itself. Indeed, Davis explained that Congress could have used that mechanism to activate a non-discriminatory (i.e., across-the-board) increase in contribution limits; in that case, the Court stated, “Davis’ argument would plainly fail.” Id., at 737.8 The constitutional infirmity in Davis was not the trigger mechanism, but rather what lay on the other side of it—a discriminatory speech restriction. The Court’s response to these points is difficult to fathom. The majority concedes that “our decision in Davis focused on the asymmetrical contribution limits imposed by the Millionaire’s Amendment.” Ante, at 14. That was because, the majority explains, Davis presented only that issue. See ante, at 14. And yet, the majority insists (with out explaining how this can be true), the reach of Davis is not so limited. And in any event, the majority claims, the burden on speech is “greater in this case than in Davis.” —————— 8 Notably, the Court found this conclusion obvious even though an across-the-board increase in contribution limits works to the compara tive advantage of the non-self-financing candidate—that is, the candi date who actually depends on contributions. Such a system puts the self-financing candidate to a choice: Do I stop spending, or do I allow the higher contribution limits (which will help my opponent) to kick in? That strategic choice parallels the one that the Arizona statute forces. See supra, at 15. Ante, at 14. But for reasons already stated, that is not so. The burden on speech in Davis—the penalty that cam paign spending triggered—was the discriminatory contri bution restriction, which Congress could not otherwise have imposed. By contrast, the thing triggered here is a non-discriminatory subsidy, of a kind this Court has ap proved for almost four decades. Maybe the majority is saying today that it had something like this case in mind all the time. But nothing in the logic of Davis controls this decision.9 III For all these reasons, the Court errs in holding that the government action in this case substantially burdens speech and so requires the State to offer a compelling in terest. But in any event, Arizona has come forward with just such an interest, explaining that the Clean Elections Act attacks corruption and the appearance of corruption in the State’s political system. The majority’s denigration of this interest—the suggestion that it either is not real or does not matter—wrongly prevents Arizona from protect ing the strength and integrity of its democracy. A Our campaign finance precedents leave no doubt: Pre venting corruption or the appearance of corruption is a —————— 9 The majority also briefly relies on Miami Herald Publishing Co. v. Tornillo, 418 U. S. 241 (1974), but that case is still wider of the mark. There, we invalidated a law compelling newspapers (by threat of criminal sanction) to print a candidate’s rejoinder to critical commen tary. That law, we explained, overrode the newspaper’s own editorial judgment and forced the paper both to pay for and to convey a message with which it disagreed. See id., at 256–258. An analogy might be if Arizona forced privately funded candidates to purchase their opponents’ posters, and then to display those posters in their own campaign offices. But that is very far from this case. The Arizona statute does not require petitioners to disseminate or fund any opposing speech; nor does it in any way associate petitioners with that speech. 24 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting compelling government interest. See, e.g., Davis, 554 U. S., at 741 ; Federal Election Comm’n v. National Con servative Political Action Comm., 470 U. S. 480 , 496–497 (1985) (NCPAC). And so too, these precedents are clear: Public financing of elections serves this interest. See supra, at 4–5. As Buckley recognized, and as I earlier described, public financing “reduce[s] the deleterious influence of large contributions on our political process.” 424 U. S., at 91 ; see id., at 96 . When private contributions fuel the political system, candidates may make corrupt bargains to gain the money needed to win election. See NCPAC, 470 U. S., at 497 . And voters, seeing the depend ence of candidates on large contributors (or on bundlers of smaller contributions), may lose faith that their represen tatives will serve the public’s interest. See Shrink Mis souri, 528 U. S., at 390 (the “assumption that large donors call the tune [may] jeopardize the willingness of voters to take part in democratic governance”). Public financing addresses these dangers by minimizing the importance of private donors in elections. Even the majority appears to agree with this premise. See ante, at 27 (“We have said that . . . ‘public financing as a means of eliminating the improper influence of large private contributions furthers a significant governmental interest’ ”). This compelling interest appears on the very face of Arizona’s public financing statute. Start with the title: The Citizens Clean Elections Act. Then proceed to the statute’s formal findings. The public financing program, the findings state, was “inten[ded] to create a clean elec tions system that will improve the integrity of Arizona state government by diminishing the influence of special interest money.” §16–940(A) (West 2006). That measure was needed because the prior system of private fundrais ing had “[u]ndermine[d] public confidence in the integrity of public officials;” allowed those officials “to accept large campaign contributions from private interests over which they [had] governmental jurisdiction;” favored “a small number of wealthy special interests” over “the vast major ity of Arizona citizens;” and “[c]os[t] average taxpayers millions of dollars in the form of subsidies and special privileges for campaign contributors.” §16–940(B).10 The State, appearing before us, has reiterated its important anti-corruption interest. The Clean Elections Act, the State avers, “deters quid pro quo corruption and the ap pearance of corruption by providing Arizona candidates with an option to run for office without depending on outside contributions.” Brief for State Respondents 19. And so Arizona, like many state and local governments, has implemented public financing on the theory (which this Court has previously approved, see supra, at 5), that the way to reduce political corruption is to diminish the role of private donors in campaigns.11 And that interest justifies the matching funds provision —————— 10 The legislative findings also echo what the Buckley Court found true of public financing—that it “encourage[s] citizen participation in the political process” and “promote[s] freedom of speech” by enhancing the ability of candidates to “communicat[e] to voters.” §§16–940(A), (B). 11 The majority briefly suggests that the State’s “austere contribution limits” lessen the need for public financing, see ante, at 26, but provides no support for that dubious claim. As Arizona and other jurisdictions have discovered, contribution limits may not eliminate the risk of corrupt dealing between candidates and donors, especially given the widespread practice of bundling small contributions into large pack ages. See Brief for United States as Amicus Curiae 31. For much this reason, Buckley upheld both limits on contributions to federal candi dates and public financing of presidential campaigns. See 424 U. S., at 23–38, 90–108. Arizona, like Congress, was “surely entitled to con clude” that contribution limits were only a “partial measure,” id., at 28, and that a functional public financing system was also necessary to eliminate political corruption. In stating otherwise, the Court substi tutes its judgment for that of Arizona’s voters, contrary to our practice of declining to “second-guess a . . . determination as to the need for prophylactic measures where corruption is the evil feared.” Federal Election Comm’n v. National Right to Work Comm., 459 U. S. 197, 210 (1982). 26 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting at issue because it is a critical facet of Arizona’s public financing program. The provision is no more than a dis bursement mechanism; but it is also the thing that makes the whole Clean Elections Act work. As described earlier, see supra, at 5–6, public financing has an Achilles heel— the difficulty of setting the subsidy at the right amount. Too small, and the grant will not attract candidates to the program; and with no participating candidates, the pro gram can hardly decrease corruption. Too large, and the system becomes unsustainable, or at the least an unneces sary drain on public resources. But finding the sweet-spot is near impossible because of variation, across districts and over time, in the political system. Enter the matching funds provision, which takes an ordinary lump-sum amount, divides it into thirds, and disburses the last two of these (to the extent necessary) via a self-calibrating mechanism. That provision is just a fine-tuning of the lump-sum program approved in Buckley—a fine-tuning, it bears repeating, that prevents no one from speaking and discriminates against no message. But that fine-tuning can make the difference between a wholly ineffectual program and one that removes corruption from the politi cal system.12 If public financing furthers a compelling interest—and according to this Court, it does—then so too does the disbursement formula that Arizona uses to make public financing effective. The one conclusion follows directly from the other. —————— 12 For this reason, the majority is quite wrong to say that the State’s interest in combating corruption does not support the matching fund provision’s application to a candidate’s expenditure of his own money or to an independent expenditure. Ante, at 25–26. The point is not that these expenditures themselves corrupt the political process. Rather, Arizona includes these, as well as all other, expenditures in the pro gram to ensure that participating candidates receive the funds neces sary to run competitive races—and so to attract those candidates in the first instance. That is in direct service of the State’s anti-corruption interest. Except in this Court, where the inescapable logic of the State’s position is . . . virtually ignored. The Court, to be sure, repeatedly asserts that the State’s interest in pre venting corruption does not “sufficiently justif[y]” the mechanism it has chosen to disburse public moneys. Ante, at 28; see ante, at 27. Only one thing is missing from the Court’s response: any reasoning to support this conclusion. Nowhere does the majority dispute the State’s view that the success of its public financing system depends on the matching funds mechanism; and nowhere does the major ity contest that, if this mechanism indeed spells the differ ence between success and failure, the State’s interest in preventing corruption justifies its use. And so the major ity dismisses, but does not actually answer the State’s contention—even though that contention is the linchpin of the entire case. Assuming (against reason and precedent) that the matching funds provision substantially burdens speech, the question becomes whether the State has of fered a sufficient justification for imposing that burden. Arizona has made a forceful argument on this score, based on the need to establish an effective public fi nancing system. The majority does not even engage that reasoning. B The majority instead devotes most of its energy to trying to show that “level[ing] the playing field,” not fighting corruption, was the State’s real goal. Ante, at 22–23 (in ternal quotation marks omitted); see ante, at 22–24. But the majority’s distaste for “leveling” provides no excuse for striking down Arizona’s law. [1] For starters, the Court has no basis to question the sincerity of the State’s interest in rooting out political corruption. As I have just explained, that is the interest 28 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting the State has asserted in this Court; it is the interest predominantly expressed in the “findings and declara tions” section of the statute; and it is the interest univer sally understood (stretching back to Teddy Roosevelt’s time) to support public financing of elections. See supra, at 4, 23–24. As against all this, the majority claims to have found three smoking guns that reveal the State’s true (and nefarious) intention to level the playing field. But the only smoke here is the majority’s, and it is the kind that goes with mirrors. The majority first observes that the matching funds provision is titled “ ‘Equal funding of candidates’ ” and that it refers to matching grants as “ ‘equalizing funds.’ ” Ante, at 23 (quoting §16–952). Well, yes. The statute provides for matching funds (above and below certain thresholds); a synonym for “match” is “equal”; and so the statute uses that term. In sum, the statute describes what the statute does. But the relevant question here (according to the majority’s own analysis) is why the statute does that thing—otherwise said, what interest the statute serves. The State explains that its goal is to prevent corruption, and nothing in the Act’s descriptive terms suggests any other objective. Next, the majority notes that the Act allows participat ing candidates to accept private contributions if (but only if) the State cannot provide the funds it has promised (for example, because of a budget crisis). Ante, at 23 (citing §16–954(F)). That provision, the majority argues, shows that when push comes to shove, the State cares more about “leveling” than about fighting corruption. Ante, at 23. But this is a plain misreading of the law. All the statute does is assure participating candidates that they will not be left in the lurch if public funds suddenly be come unavailable. That guarantee helps persuade candi dates to enter the program by removing the risk of a state default. And so the provision directly advances the Act’s goal of combating corruption. Finally, the Court remarks in a footnote that the Clean Elections Commission’s website once stated that the “ ‘Act was passed by the people of Arizona . . . to level the play ing field.’ ” Ante, at 24, n. 10. I can understand why the majority does not place much emphasis on this point. Some members of the majority have ridiculed the practice of relying on subsequent statements by legislators to demonstrate an earlier Congress’s intent in enacting a statute. See, e.g., Sullivan v. Finkelstein, 496 U. S. 617 , 631–632 (1990) (SCALIA, J., concurring in part); United States v. Hayes, 555 U. S. 415 , 434–435 (2009) (ROBERTS, C. J., dissenting). Yet here the majority makes a much stranger claim: that a statement appearing on a govern ment website in 2011 (written by who-knows-whom?) reveals what hundreds of thousands of Arizona’s voters sought to do in 1998 when they enacted the Clean Elec tions Act by referendum. Just to state that proposition is to know it is wrong. So the majority has no evidence—zero, none—that the objective of the Act is anything other than the interest that the State asserts, the Act proclaims, and the history of public financing supports: fighting corruption. [2] But suppose the majority had come up with some evi dence showing that Arizona had sought to “equalize elec toral opportunities.” Ante, at 24. Would that discovery matter? Our precedent says no, so long as Arizona had a compelling interest in eliminating political corruption (which it clearly did). In these circumstances, any interest of the State in “leveling” should be irrelevant. That inter est could not support Arizona’s law (assuming the law burdened speech), but neither would the interest invali date the legislation. To see the point, consider how the matter might arise. 30 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting Assume a State has two reasons to pass a statute affecting speech. It wants to reduce corruption. But in addition, it wishes to “level the playing field.” Under our First Amendment law, the interest in preventing corruption is compelling and may justify restraints on speech. But the interest in “leveling the playing field,” according to well established precedent, cannot support such legislation.13 So would this statute (assuming it met all other constitu tional standards) violate the First Amendment? The answer must be no. This Court, after all, has never said that a law restricting speech (or any other constitu tional right) demands two compelling interests. One is enough. And this statute has one: preventing corruption. So it does not matter that equalizing campaign speech is an insufficient interest. The statute could violate the First Amendment only if “equalizing” qualified as a forbidden motive—a motive that itself could annul an otherwise constitutional law. But we have never held that to be so. And that should not be surprising: It is a “fundamental principle of constitutional adjudication,” from which we have deviated only in exceptional cases, “that this Court will not strike down an otherwise constitutional statute on the basis of an alleged illicit legislative motive.” United States v. O’Brien, 391 U. S. 367, 383 (1968); see id., at 384 (declining to invalidate a statute when “Congress had the undoubted power to enact” it without the suspect motive); accord, Turner Broadcasting System, Inc. v. FCC, 512 —————— 13 I note that this principle relates only to actions restricting speech. See Buckley, 424 U. S., at 48–49 (rejecting the notion “that government may restrict the speech of some . . . to enhance the relative voice of others”). As previously explained, speech subsidies stand on a different constitutional footing, see supra, at 10–11; so long as the government remains neutral among viewpoints, it may choose to assist the speech of persons who might not otherwise be heard. But here I am assuming for the sake of argument that the Clean Elections Act imposes the kind of restraint on expression requiring that the State show a compelling interest. U. S. 622, 652 (1994); Renton v. Playtime Theatres, Inc., 475 U. S. 41 , 47–48 (1986). When a law is otherwise constitutional—when it either does not restrict speech or rests on an interest sufficient to justify any such restric tion—that is the end of the story. That proposition disposes of this case, even if Arizona had an adjunct interest here in equalizing electoral oppor tunities. No special rule of automatic invalidation applies to statutes having some connection to equality; like any other laws, they pass muster when supported by an im portant enough government interest. Here, Arizona has demonstrated in detail how the matching funds provision is necessary to serve a compelling interest in combating corruption. So the hunt for evidence of “leveling” is a waste of time; Arizona’s law survives constitutional scru tiny no matter what that search would uncover. IV This case arose because Arizonans wanted their gov ernment to work on behalf of all the State’s people. On the heels of a political scandal involving the near-routine purchase of legislators’ votes, Arizonans passed a law de signed to sever political candidates’ dependence on large contributors. They wished, as many of their fellow Ameri cans wish, to stop corrupt dealing—to ensure that their representatives serve the public, and not just the wealthy donors who helped put them in office. The legislation that Arizona’s voters enacted was the product of deep thought and care. It put into effect a public financing system that attracted large numbers of candidates at a sustain able cost to the State’s taxpayers. The system discrimi nated against no ideas and prevented no speech. Indeed, by increasing electoral competition and enabling a wide range of candidates to express their views, the system “further[ed] . . . First Amendment values.” Buckley, 424 U. S., at 93 (citing New York Times, 376 U. S., at 270 ). 32 ARIZONA FREE ENTERPRISE CLUB’S FREEDOM CLUB PAC v. BENNETT KAGAN, J., dissenting Less corruption, more speech. Robust campaigns leading to the election of representatives not beholden to the few, but accountable to the many. The people of Arizona might have expected a decent respect for those objectives. Today, they do not get it. The Court invalidates Arizo nans’ efforts to ensure that in their State, “ ‘[t]he people . . . possess the absolute sovereignty.’ ” Id., at 274 (quoting James Madison in 4 Elliot’s Debates on the Federal Con stitution 569–570 (1876)). No precedent compels the Court to take this step; to the contrary, today’s decision is in tension with broad swaths of our First Amendment doctrine. No fundamental principle of our Constitution backs the Court’s ruling; to the contrary, it is the law struck down today that fostered both the vigorous compe tition of ideas and its ultimate object—a government responsive to the will of the people. Arizonans deserve better. Like citizens across this country, Arizonans de serve a government that represents and serves them all. And no less, Arizonans deserve the chance to reform their electoral system so as to attain that most American of goals. Truly, democracy is not a game. See ante, at 25. I respectfully dissent.