with whom Justice Scalia and Justice Kennedy join, and with whom The Chief Justice joins as to Part II, dissenting.
The Party Expenditure Provision, 2 U. S. C. §441a(d)(3), severely limits the amount of money that a national or state committee of a political party can spend in coordination with its own candidate for the Senate or House of Representatives. See ante, at 438-439, and n. 3. Because this provision sweeps too broadly, interferes with the party-candidate relationship, and has not been proved necessary to combat corruption, I respectfully dissent.
I
As an initial matter, I continue to believe that Buckley v. Valeo, 424 U. S. 1 (1976) (per curiam), should be overruled. See Nixon v. Shrink Missouri Government PAC, 528 U. S. 377, 410 (2000) (Thomas, J., dissenting); Colorado Republican Federal Campaign Comm. v. Federal Election Comm’n, 518 U. S. 604, 631 (1996) (Colorado I) (Thomas, J., concurring in judgment and dissenting in part). “Political speech is the primary object of First Amendment protection,” Shrink Missouri, supra, at 410-411 (Thomas, J., dissenting); see also Eu v. San Francisco County Democratic Central *466Comm., 489 U. S. 214, 228 (1989); Mills v. Alabama, 384 U. S. 214, 218 (1966), and it is the lifeblood of a self-governing people, see Shrink Missouri, supra, at 405 (Kennedy, X, dissenting) (“[Political speech in the course of elections [is] the speech upon which democracy depends”)- I remain baffled that this Court has extended the most generous First Amendment safeguards to filing lawsuits, wearing profane jackets, and exhibiting drive-in movies with nudity,1 but has offered only tepid protection to the core speech and associational rights that our Founders sought to defend.
In this case, the Government does not attempt to argue that the Party Expenditure Provision satisfies strict scrutiny, see Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 45 (1983) (providing that, under strict scrutiny, a restriction on speech is constitutional only if it is narrowly tailored to serve a compelling governmental interest). Nor could it. For the reasons explained in my separate opinions in Colorado I, supra, at 641-644, and Shrink Missouri, supra, at 427-430, the campaign financing law at issue fails strict scrutiny.
II
We need not, however, overrule Buckley and apply strict scrutiny in order to hold the Party Expenditure Provision unconstitutional. Even under Buckley, which described the requisite scrutiny as “exacting” and “rigorous,” 424 U. S., at 16, 29, the regulation cannot pass constitutional muster. In practice, Buckley scrutiny has meant that restrictions on contributions by individuals and political committees do not violate the First Amendment so long as they are “closely drawn” to match a “sufficiently important” government interest, Shrink Missouri, supra, at 387-389; see also Buckley, supra, at 58, but that restrictions on independent expendi*467tures are constitutionally invalid, see Buckley, supra, at 58-59; see also Federal Election Comm’n v. National Conservative Political Action Comm., 470 U. S. 480, 501 (1985). The rationale for this distinction between contributions and independent expenditures has been that, whereas ceilings on contributions by individuals and political committees “en-tai[l] only a marginal restriction” on First Amendment interests, Buckley, 424 U. S., at 20, limitations on independent expenditures “impose significantly more severe restrictions on protected freedoms of political expression and association,” id., at 23.
A
The Court notes this existing rationale and attempts simply to treat coordinated expenditures by political parties as equivalent to contributions by individuals and political committees. Thus, at least implicitly, the Court draws two conclusions: coordinated expenditures are no different from contributions, and political parties are no different from individuals and political committees. Both conclusions are flawed.
1
The Court considers a coordinated expenditure to be an “ ‘expenditure] made by any person in cooperation, consultation, or concert, with, or at the request or suggestion of, a candidate, his authorized political committees, or their agents.’ ” Ante, at 438 (quoting 2 U. S. C. § 441a(a)(7)(B)(i)). This definition covers a broad array of conduct, some of which is akin to an independent expenditure. At one extreme, to be sure, are outlays that are “virtually indistinguishable from simple contributions.” Colorado I, 518 U. S., at 624 (opinion of Breyer, J.). An example would be “a donation of money with direct payment of a candidate’s media bills.” Ibid. But toward the other end of the spectrum are expenditures that largely resemble, and should be entitled to the same protection as, indepéndent expenditures. *468Take, for example, a situation in which the party develops a television advertising campaign touting a candidate’s record on education, and the party simply “consults],” 2 U. S. C. § 441a(a)(7)(B)(i), with the candidate on which time slot the advertisement should run for maximum effectiveness. I see no constitutional difference between this expenditure and a purely independent one. In the language of Buckley, the advertising campaign is not a mere “general expression of support for the candidate and his views,” but a communication of “the underlying basis for the support.” 424 U. S., at 21. It is not just “symbolic expression,” ibid., but a clear manifestation of the party’s most fundamental political views. By restricting such speech, the Party Expenditure Provision undermines parties’ “freedom to discuss candidates and issues,” ibid., and cannot be reconciled with our campaign finance jurisprudence.
2
Even if I were to ignore the breadth of the statutory text, and to assume that all coordinated expenditures are functionally equivalent to contributions,2 I still would strike down the Party Expenditure Provision. The source of the “contribution” at issue is a political party, not an individual or a political committee, as in Buckley and Shrink Missouri. *469Restricting contributions by individuals and political committees may, under Buckley, entail only a “marginal restriction,” Buckley, supra, at 20, but the same cannot be said about limitations on political parties.
Political parties and their candidates are “inextricably intertwined” in the conduct of an election. Colorado I, supra, at 630 (Kennedy, J., concurring in judgment and dissenting in part). A party nominates its candidate; a candidate often is identified by party affiliation throughout the election and on the ballot; and a party’s public image is largely defined by what its candidates say and do. See, e. g., California Democratic Party v. Jones, 530 U. S. 567, 575 (2000) (“Some political parties — such as President Theodore Roosevelt’s Bull Moose Party, the La Follette Progressives of 1924, the Henry Wallace Progressives of 1948, and the George Wallace American Independent Party of 1968 — are virtually inseparable from their nominees (and tend not to outlast them”); see also M. Zak, Back to Basics for the Republican Party 1 (2000) (noting that the Republican Party has been identified as the “Party of Lincoln”). Most importantly, a party’s success or failure depends in large part on whether its candidates get elected. Because of this unity of interest, it is natural for a party and its candidate to work together and consult with one another during the course of the election. See, e.g., App. 137 (declaration of Herbert E. Alexander, Director of the Citizens’ Research Foundation at the University of Southern California). Indeed, “it would be impractical and imprudent... for a party to support its own candidates without some form of 'cooperation’ or ‘consultation.’ ” See Colorado I, 518 U. S., at 630 (Kennedy, J., concurring in judgment and dissenting in part). “[Candidates are necessary to make the party’s message known and effective, and vice versa.” Id., at 629. Thus, the ordinary means for a party to provide support is to make coordinated expenditures, see, e. g., App. 137-138 (declaration of Herbert E. Alexander), as the Government itself maintained just five years ago, see *470Brief for Respondent in Colorado I, O. T. 1995, No. 95-489, p. 27 (contending that Congress had made an “empirical judgment that party officials will as a matter of course consult with the party’s candidates before funding communications intended to influence the outeome,of a federal election”); see also FEC Advisory Opinion 1985-14, CCH Fed. Election Camp. Fin. Guide ¶ 5819, p. 11,186, n. 4 (1985) (“Party political committees are incapable of making independent expenditures”).
As the District Court explained, to break this link between the party and its candidates would impose “additional costs and burdens to promote the party message.” 41 F. Supp. 2d 1197, 1210 (Colo. 1999). This observation finds full support in the record. See, e. g., App. 218 (statement of Anthony Corrado, Associate Professor of Government, Colby College) (explaining that, to ensure that expenditures were independent, party organizations had to establish legally separate entities, which in turn had to “rent and furnish an office, hire staff, and pay other administrative costs,” as well as “engage additional consulting services” and “duplicate many of the functions already being undertaken by other party offices”); id., at 52 (statement by Federal Election Commission admitting that national party established separate entities that made independent expenditures); id., at 217 (statement of Anthony Corrado) (explaining that reliance on independent expenditures would increase fundraising demands on party organizations because independent expenditures are less effective means of communication); id., at 219 (“[Independent expenditures do not qualify for the lowest unit rates on the purchase of broadcasting time”); App. in No. 99-1211 (CA10), p. 512 (report of Frank J. Sorauf, professor at University of Minnesota, and Jonathan S. Krasno, professor at Princeton University) (noting inefficiency of independent expenditures). Establishing and maintaining independence also tends to create voter confusion and to undermine the candidate that the party sought to support. App. 220 (statement of Anthony Corrado); App. in No. 99-1211 (CA10), at 623-624 *471(deposition of John Heubusch); App. 159 (affidavit of Donald K. Bain) (“[0]ur communications can be more focused, understandable, and effective if the Party and its candidates can work together”). Finally, because of the ambiguity in the term “coordinated expenditure,” the Party Expenditure Provision chills permissible speech as well. See, e.g., id., at 159-160 (affidavit of Donald K. Bain). Thus, far from being a mere “marginal” restraint on speech, Buckley, 424 U. S., at 20, the Party Expenditure Provision has restricted the party’s most natural, form of communication; has precluded parties “from effectively amplifying the voice of their adherents,” id., at 22; and has had a “stifling effect on the ability of the party to do what it exists to do.”3 Colorado I, supra, at 630 (Kennedy, J., concurring in judgment and dissenting in part).
The Court nevertheless concludes that these concerns of inhibiting party speech are rendered “implausible” by the nearly 30 years of history in which coordinated spending has been statutorily limited. Ante, at 449. Without a single citation to the record, the Court rejects the assertion “that for almost three decades political parties have not been func*472tional or have been functioning in systematic violation of the law.” Ibid. I am unpersuaded by the Court’s attempts to downplay the extent of the burden on political parties’ First Amendment rights. First, the Court does not examine the record or the findings of the District Court, but instead relies wholly on the “observations]” of the “political scientists” who happen to have written an amicus brief in support of the petitioner. Ibid. I find more convincing, and more relevant, the record evidence that the parties have developed, which, as noted above, indicates that parties have suffered as a result of the Party Expenditure Provision.4 See supra, at 470-471. Second, we have never before upheld a limitation on speech simply because speakers have coped with the limitation for 30 years. See, e. g., Bartnicki v. Vopper, 532 U. S. 514, 517 (2001) (holding unconstitutional under the First Amendment restrictions on the disclosure of the contents of an illegally intercepted communication, even though federal law had prohibited such disclosure for 67 years). And finally, if the passage of time were relevant to the constitutional inquiry, I would wonder why the Court adopted *473a “30-year” rule rather than the possible countervailing “200-year” rule. For nearly 200 years, this country had congressional elections without limitations on coordinated expenditures by political parties. Nowhere does the Court suggest that these elections were not “functional,” ante, at 449, or that they were marred by corruption.
The Court’s only other response to the argument that parties are linked to candidates and that breaking this link would impose significant costs on speech is no response at all. The Court contends that parties are not organized simply to “elec[t] particular candidates” as evidenced by the fact that many political action committees donate money to both parties and sometimes even opposing candidates. Ante, at 451. According to the Court, “[pjarties are thus necessarily the instruments of some contributors whose object is not to support the party’s message or to elect party candidates across the board.” Ante, at 451-452. There are two flaws in the Court’s analysis. First, no one argues that a party’s role is merely to get particular candidates elected. Surely, among other reasons, parties also exist to develop and promote a platform. See, e. g., Brief for Respondent 23. The point is simply that parties and candidates have shared interests, that it is natural for them to work together, and that breaking the connection between parties and their candidates inhibits the promotion of the party’s message. Second, the mere fact that some donors contribute to both parties and their candidates does not necessarily imply that the donors control the parties or their candidates. It certainly does not mean that the parties are mere “instruments” or “agents,” ante, at 452, of the donors. Indeed, if a party receives money from donors on both sides of an issue, how can it be a tool of both donors? If the Green Party were to receive a donation from an industry that pollutes, would the Green Party necessarily become, through no choice of its own, an instrument of the polluters? The Court proffers no evidence that parties have become pawns of wealthy contrib*474utors. Parties might be the target of the speech of donors, but that does not suggest that parties are influenced (let alone improperly influenced) by the speech. Thus, the Court offers no explanation for why political parties should be treated the same as individuals and political committees.
B
But even if I were to view parties’ coordinated expenditures as akin to contributions by individuals and political committees, I still would hold the Party Expenditure Provision constitutionally invalid. Under Shrink Missouri, a contribution limit is constitutional only if the Government demonstrates that the regulation is “closely drawn” to match a “sufficiently important interest.” 528 U. S., at 387-388 (quoting Buckley, 424 U. S., at 25) (internal quotation marks omitted). In this case, there is no question that the Government has asserted a sufficient interest, that of preventing corruption. See Shrink Missouri, supra, at 388 (“ ‘[T]he prevention of corruption and the appearance of corruption’ was found to be a 'constitutionally sufficient justification’ ”) (quoting Buckley, supra, at 25-26). The question is whether the Government has demonstrated both that coordinated expenditures by parties give rise to corruption and that the restriction is “closely drawn” to curb this corruption. I believe it has not.
1
As this Court made clear just last Term, “[w]e have never accepted mere conjecture as adequate to carry a First Amendment burden.” Shrink Missouri, 528 U. S., at 392. Some “quantum of empirical evidence [is] needed to satisfy heightened judicial scrutiny of legislative judgments.” Id., at 391. Precisely how much evidence is required will “vary up or down with the novelty and plausibility of the justification raised.” Ibid. Today, the Court has jettisoned this evidentiary requirement.
*475Considering that we have never upheld an expenditure limitation against political parties, I would posit that substantial evidence is necessary to justify the infringement of parties’ First Amendment interests. But we need not accept this high evidentiary standard to strike down the Party Expenditure Provision for want of evidence. Under the least demanding evidentiary requirement, the Government has failed to carry its burden, for it has presented no evidence at all of corruption or the perception of corruption. The Government does not, and indeed cannot, point to any congressional findings suggesting that the Party Expenditure Provision is necessary, or even helpful, in reducing corruption or the perception of corruption. In fact, this Court has recognized that “Congress wrote the Party Expenditure Provision not so much because of a special concern about the potentially ‘corrupting’ effect of party expenditures, but rather for the constitutionally insufficient purpose of reducing what it saw as wasteful and excessive campaign spending.”5 Colorado I, 518 U. S., at 618. See also ibid. (“[R]ather than indicating a special fear of the corruptive influence of political parties, the legislative history demonstrates Congress’ general desire to enhance what was seen as an important and legitimate role for political parties in American elections”).
Without explanation, the Court departs from this earlier, well-considered understanding of the Party Expenditure Provision. Were there any evidence of corruption in the *476record that the parties have since developed, such a departure might be justified. But as the District Court found, “[t]he facts which [the] FEC contends support its position ... do not establish that the limit on party coordinated expenditures is necessary to prevent corruption or the appearance thereof.” 41 F. Supp. 2d, at 1211. Indeed, “[n]one of the FEC’s examples [of alleged corruption] involve[s] coordinated expenditures.” Ibid. See also App. in No. 99-1211 (CA10), at 346 (declaration of Herbert E. Alexander) (“In the decades since 1974, when coordinated expenditures were allowed for both presidential and congressional campaigns, there has not been any dispute relating to them, no charges of corruption or the appearance thereof . . .”); id., at 430 (statement of Anthony Corrado) (“There is no academic analysis or scholarly study conducted to date that demonstrates that parties are corrupted by the federally regulated contributions, the so-called ‘hard-money funds,’ they receive from donors. None of the studies of party finance or party coordinated spending contend[s] that these funds are corruptive or generate the appearance of corruption in the political process”); id., at 624 (deposition of John Heubusch) (testifying that, in his experience, political party spending was not a source of corruption of Members of the United States Senate).6
The dearth of evidence is unsurprising in light of the unique relationship between a political party and its candidates: “The very aim of a political party is to influence its candidate’s stance on issues and, if the candidate takes office or is reelected, his votes.” Colorado I, 518 U. S., at *477646 (Thomas, J., concurring in judgment and dissenting in part). If coordinated expenditures help achieve this aim, the achievement “does not. . . constitute ‘a subversion of the political process.’” Ibid, (quoting Federal Election Cornm’n, 470 U. S., at 497). It is simply the essence of our Nation’s party system of government. One can speak of an individual citizen or a political action committee corrupting or coercing a candidate, but “[w]hat could it mean for a party to ‘corrupt’ its candidate or to exercise ‘coercive’ influence over him?” 518 U. S., at 646.
Apparently unable to provide an answer to this question, the Court relies upon an alternative theory of corruption. According to the Court, the Party Expenditure Provision helps combat circumvention of the limits on individual donors’ contributions, which limits are necessary to reduce corruption by those donors.7 See ante, at 452-455. The primary problem with this contention, however, is that it too is plainly contradicted by the findings of the District Court, see 41 F. Supp. 2d, at 1211, and the overwhelming evidence in the record, see supra, at 475.8 And this contention is particularly surprising in light of Colorado I, in which we discussed the same opportunity for corruption through circumvention, and, far from finding it dispositive, concluded *478that any opportunity for corruption was, “at best, attenuated.” 518 U. S., at 616.
Without addressing the District Court’s determination or reflecting on this Court’s understanding in Colorado I, the Court today asserts that its newfound position is supported by “substantial evidence.” The best evidence the Court can come up with, however, is the Democratic Senatorial Campaign Committee’s (DSCC) use of the “tally system,” which “connects] donors to candidates through the accommodation of a party.” Ante, at 459. The tally system is not evidence of corruption-by-circumvention. In actuality, the DSCC is not acting as a mere conduit, allowing donors to contribute money in excess of the legal limits. The DSCC instead has allocated money based on a number of factors, including “the financial strength of the campaign,” “what [the candidate’s] poll numbers looked like,” and “who had the best chance of winning or who needed the money most.” App. 250-251 (declaration of Robert Hickmott, former Democratic fundraiser and National Finance Director for Timothy Wirth’s Senate campaign); see also App. in No. 99-1211 (CA10), at 430 (statement of Anthony Corrado) (“When parties are deciding whether to spend fluids on behalf of a candidate, they chiefly examine the competitiveness of the district or race, the political situation of the incumbent, and the strength of the party contender’s candidacy”); id., at 563 (deposition of Donald Bain) (stating that the party generally did not support someone who has a safe seat or is clearly not going to win). As the District Court found, “the primary consideration in allocating funds is which races are marginal — that is, which races are ones where party money could be the difference between winning and losing.” 41 F. Supp. 2d, at 1203. “Maintaining party control over seats is paramount to the parties’ pursuits.” Ibid.; see also App. in No. 99-1211 (CA10), at 483 (stating that primary goal of legislative campaign committees is “to win or maintain control of the chamber and the powers of the majority legislative party”). The *479“bottom line” of the tally system is that “some candidates get back more money than they raise, and others get back less.” App. 250 (declaration of Robert Hickmott).
Moreover, the Court does not explain how the tally system could constitute evidence of corruption. Both the initial contribution to the party and the subsequent expenditure by the party on the candidate are currently legal. In essence, the Court is asserting that it is corrupt for parties to do what is legal to enhance their participation in the political process. Each step in the process is permitted, but the combination of those steps, the Court apparently believes, amounts to corruption sufficient to, silence those who wish to support a candidate. In my view, the First Amendment demands a more coherent explication of the evidence of corruption.9
Finally, even if the tally system were evidence of corruption-through-circumvention, it is only evidence of what is occurring under the current system, not of additional “corruption” that would arise in the absence of the Party Expenditure Provision. The Court speculates that, if we invalidated the Party Expenditure Provision, “the inducement to circumvent would almost certainly intensify.” Ante, at 460. But that is nothing more than supposition, which is insufficient under our precedents to sustain a restriction on First Amendment interests. See Shrink Missouri, 528 U. S., at 892 (“We have never accepted mere conjecture as adequate to carry a First Amendment burden”). See also United States v. Playboy Entertainment Group, Inc., 529 U. S. 803, 822 (2000) (concluding that the government “must present more than anecdote and supposition”). And it is weak supposition at that. The Court does not contend that *480the DSCC’s alleged efforts to channel money through the tally system were restricted in any way by the Party Expenditure Provision. On the contrary, the Court suggests that a donation to the DSCC was increased by the party; in other words, the candidate got more than the initial donation. See ante, at 458 (quoting declaration of Timothy Wirth) (“ T understood that when I raised funds for the DSCC, the donors expected that I would receive the amount of their donations multiplied by a certain number that the DSCC had determined in advance, assuming the DSCC has raised other funds’” (emphasis added)). Because I am un- . persuaded by weak speculation ungrounded in any evidence, I disagree with the Court’s conclusion that the Party Expenditure Provision furthers the Government interest of reducing corruption.10
*4812
Even if the Government had presented evidence that the Party Expenditure Provision affects corruption, the statute still would be unconstitutional, because there are better tailored alternatives for addressing the corruption. In addition to bribery laws and disclosure laws, see Shrink Missouri, supra, at 428 (Thomas, J., dissenting), the Government has two options that would not entail the restriction of political parties’ First Amendment rights.
First, the Government could enforce the earmarking rule of 2 U. S. C. §441a(a)(8), under which contributions that “are in any way earmarked or otherwise directed through an intermediary or conduit to [a] candidate” are treated as contributions to the candidate. Vigilant enforcement of this provision is a precise response to the Court’s circumvention concerns. If a donor contributes $2,000 to a candidate (the maximum donation in an election cycle), he cannot direct the political party to funnel another dime to the candidate without confronting the Federal Election Campaign Act’s civil and criminal penalties, see 2 U. S. C. § 437g(a)(6)(C) (civil); §437g(d) (criminal).
According to the Court, reliance on this earmarking provision “ignores the practical difficulty of identifying and directly combating circumvention” and “would reach only the most clumsy attempts to pass contributions through to candidates.” Ante, at 462. The Court, however, does not cite any evidence to support this assertion. Nor does it articulate what failed steps the Government already has taken. Nor does it explain why the burden that the Government allegedly would have to bear in uncovering circumvention justifies the infringement of political parties’ First Amendment rights. In previous cases, we have not been so willing to overlook such failures. See, e. g., Bartnicki, 532 U. S., at 530-531 (“[T]here is no empirical evidence to support the assumption that the prohibition against disclosures reduces the number of illegal interceptions”).
*482In any event, there is a second, well-tailored option for combating corruption that does not entail the reduction of parties’ First Amendment freedoms. The heart of the Court’s circumvention argument is that, whereas individuals can donate only $2,000 to a candidate in a given election cycle, they can donate $20,000 to the national committees of a political party, an amount that is allegedly large enough to corrupt the candidate. See ante, at 453. If indeed $20,000 is enough to corrupt a candidate (an assumption that seems implausible on its face and is, in any event, unsupported by any evidence), the proper response is to lower the cap. That way, the speech restriction is directed at the source of the alleged corruption — the individual donor — and not the party. “The normal method of deterring unlawful conduct is to impose an appropriate punishment on the person who engages in it.” Bartnicki, 532 U. S., at 529. “[I]t would be quite remarkable to hold that speech by a law-abiding [entity] can be suppressed in order to deter conduct by a non-law-abiding third party.” Id., at 529-530. The Court takes that unorthodox path today, a decision that is all the more remarkable considering that the controlling opinion in Colorado I expressly rejected it just five years ago. 518 U. S., at 617 (“We could understand how Congress, were it to conclude that the potential for evasion of the individual contribution limits was a serious matter, might decide to change the statute’s limitations on contributions to political parties. But we do not believe that the risk of corruption present here could justify the ‘markedly greater burden on basic freedoms caused by’ the statute’s limitations on expenditures” (citations omitted)).
In my view, it makes no sense to contravene a political party’s core First Amendment rights because of what a third party might unlawfully try to do. Instead of broadly restricting political parties’ speech, the Government should have pursued better-tailored alternatives for combating the alleged corruption.
NAACP v. Button, 371 U. S. 415, 444 (1968); Cohen v. California, 403 U. S. 15, 26 (1971); Erznoznik v. Jacksonville, 422 U. S. 205, 208-215 (1975).
The Court makes this very assumption. See ante, at 464 (“There is no significant functional difference between a party’s coordinated expenditure and a direct party contribution to the candidate”). To the extent the Court has not defined the universe of coordinated expenditures and leaves open the possibility that there are such expenditures that would not be functionally identical to direct contributions, the constitutionality of the Party Expenditure Provision as applied to such expenditures remains unresolved. See, e. g., ante, at 456, n. 17. At oral argument, the Government appeared to suggest that the Party Expenditure Provision might not reach expenditures that are not functionally identical to contributions. See Tr. of Oral Arg. 15 (stating that the purpose of the Party Expenditure Provision is simply to prevent someone “from making contributions in the form of paying the candidate’s bills”).
The Court contends that, notwithstanding this burden, “it is nonetheless possible for parties, like individuals and nonparty groups, to speak independently.” Ante, at 450, n. 11 (emphasis added). That is correct, but it does not render the restriction constitutional. If Congress were to pass a law imposing a $1,000 tax on every political newspaper editorial, the law would surely constitute an unconstitutional restraint on speech, even though it would still be possible for newspapers to print such editorials.
The Court’s holding presents an additional First Amendment problem. Because of the close relationship between parties and candidates, lower courts will face a difficult, if not insurmountable, task in trying to determine whether particular party expenditures are in fact coordinated or independent. As the American Civil Liberties Union points out, “[e]ven if such an inquiry is feasible, it inevitably would involve an intrusive and constitutionally troubling investigation of the inner workings of political parties.” Brief for American Civil Liberties Union et al. as Amici Curiae 18.
Moreover, were I to depart from the record, as does the Court, I could consider sources suggesting that parties in fact have lost power in recent years. See, e. g., M. Wattenberg, The Decline of American Political Parties, 1952-1996, p. 174 (1998) (indicating that percentage of voters who identify with a party has declined while percentage of split tickets has increased); Maisel, American Political Parties: Still Central to a Functioning Democracy?, in American Political Parties: Decline or Resurgence?, 103, 107-111 (J. Cohen, R. Fleisher, & P. Kantor eds. 2001) (describing weaknesses of modern political parties). I also could explore how political parties have coped with the restrictions on coordinated expenditures. As Justice Kennedy has explained, “[t]he Court has forced a substantial amount, of political speech underground, as contributors and candidates devise ever more elaborate methods of avoiding contribution limits.” Nixon v. Shrink Missouri Government PAC, 528 U. S. 377, 406 (2000) (dissenting opinion). Perhaps political parties have survived, not because the regulation at issue imposes less than a substantial burden on speech, but simply because the parties have found “underground” alternatives for communication.
The Court contends that I “ignorfe] [a] distinction,” ante, at 457, n. 19: Whereas Congress may not have been concerned with corruption insofar as independent expenditures were implicated, Congress was concerned with corruption insofar as coordinated expenditures were implicated. This “distinction” must have been lost on Congress as well, which made no finding that the Party Expenditure Provision serves different purposes for different expenditures. It also was lost on the Court in Colorado I, which stated in no uncertain terms that Congress was not motivated by “the potentially ‘corrupting’ effect of party expenditures.” 518 U. S., at 618.
In Missouri Republican Party v. Lamb, 227 F. 3d 1070 .(2000), the Eighth Circuit held that the State of Missouri’s restrictions on contributions by political parties violated the First Amendment. In accord with the Tenth Circuit in this case, the Eighth Circuit concluded that “the record is wholly devoid of any evidence that limiting parties’ campaign contributions will either reduce corruption or measurably decrease the number of occasions on which limitations on individuals’ campaign contributions are circumvented.” Id., at 1073.
The Court does not argue that the Party Expenditure Provision is necessary to reduce the perception of corruption. Nor could the record sustain such an argument. See 41 F. Supp. 2d 1197, 1211 (Colo. 1999).
Contrary to the Court’s suggestion, ante, at 459, n. 21, the District Court did not simply conclude that “Colorado I had rejected the anti-circumvention rationale as a matter of law.” Instead, the District Court first concluded there was no evidence of corruption, 41 F. Supp. 2d, at 1211. Only after the District Court made this factual finding did it, in a footnote; cite Colorado I to support the legal conclusion. See 41 F. Supp. 2d, at 1211, n. 9 (“Moreover, if the skirting of contribution limits is the issue with which the FEC is concerned ... there are more tailored means of addressing such a concern than limiting the coordinated expenditure limits” (citing Colorado 1)).
Ironically, earlier this Term, this Court was less willing to uphold a speech restriction based on inference of circumvention. See, e. g., Bart-nicki v. Vopper, 532 U. S. 514, 529-535 (2001) (holding unconstitutional the prohibition on disclosure of illegally intercepted conversation even though the initial step in the disclosure process, the interception, was illegal and harmful to those whose privacy was invaded).
The other “evidence” on which the Court relies is less compelling than the tally system. The Court presents four quotations, two of which do not even support the proposition that donations are funneled through parties to candidates. See ante, at 458 (quoting declaration of Leon G. Billings, former Executive Director of the DSCC); ante, at 458. These comments simply reflect the obvious fact that a candidate benefits when his party receives money. Neither comment suggests that the candidate is aided through the surreptitious laundering of money, as opposed to issue advertisements, get-out-the-vote campaigns, and independent expenditures.
The other two quotations are somewhat suspect in that they are made by Timothy Wirth, who was the object of the negative advertisements giving rise to this lawsuit, and by his national finance director. See ibid. (quoting App. 274 (declaration of Timothy Wirth)); App. 247 (declaration of Robert Hickmott, former Democratic fundraiser and National Finance Director for Timothy Wirth’s Senate campaign). Moreover, neither Wirth nor his finance director described how donations were actually treated by the DSCC, either in general or in Wirth’s particular case; instead Wirth and his finance director simply reflected on their understandings of how the money would be used in Wirth’s election. As noted above, the District Court found that “the primary consideration in allocating funds is which races are marginal.” 41 F. Supp. 2d, at 1203. And the evidence in the record supports this finding. See supra, at 477.