533 U.S. 431 (2001)
FEDERAL ELECTION COMMISSION
COLORADO REPUBLICAN FEDERAL CAMPAIGN COMMITTEE
United States Supreme Court.
Argued February 28, 2001.
Decided June 25, 2001.
CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE TENTH CIRCUIT Souter, J., delivered the opinion of the Court, in which Stevens, O'Connor, Ginsburg, and Breyer, JJ., joined. Thomas, J., filed a dissenting opinion, in which Scalia and Kennedy, JJ., joined, and in which Rehnquist, C. J., joined as to Part II, post, p. 465.
Acting Solicitor General Underwood argued the cause for petitioner. With her on the briefs were former Solicitor General Waxman, Malcolm L. Stewart, Lawrence M. Noble, Richard B. Bader, and David Kolker.
Jan Witold Baran argued the cause for respondent. With him on the brief were Thomas W. Kirby and Carol A. Laham.[*]
Justice Souter, delivered the opinion of the Court.
In Colorado Republican Federal Campaign Comm. v. Federal Election Comm'n, 518 U. S. 604 (1996) (Colorado I), we held that spending limits set by the Federal Election Campaign Act were unconstitutional as applied to the Colorado Republican Party's independent expenditures in connection with a senatorial campaign. We remanded for consideration of the party's claim that all limits on expenditures by a political party in connection with congressional campaigns are facially unconstitutional and thus unenforceable even as to spending coordinated with a candidate. Today we reject that facial challenge to the limits on parties' coordinated expenditures.
We first examined the Federal Election Campaign Act of 1971 in Buckley v. Valeo, 424 U. S. 1 (1976) (per curiam), where we held that the Act's limitations on contributions to a candidate's election campaign were generally constitutional, but that limitations on election expenditures were not. Id., at 12-59. Later cases have respected this line between contributing and spending. See, e. g., Nixon v. Shrink Missouri Government PAC, 528 U. S. 377, 386-388 (2000); Colorado I, supra, at 610, 614-615; Federal Election Comm'n v. Massachusetts Citizens for Life, Inc., 479 U. S. 238, 259-260 (1986).
The simplicity of the distinction is qualified, however, by the Act's provision for a functional, not formal, definition of "contribution," which includes "expenditures 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," 2 U. S. C. § 441a(a) (7)(B)(i). Expenditures coordinated with a candidate, that is, are contributions under the Act.
The Federal Election Commission (FEC or Commission) originally took the position that any expenditure by a political party in connection with a particular election for federal office was presumed to be coordinated with the party's candidate. See Federal Election Comm'n v. Democratic Senatorial Campaign Comm., 454 U. S. 27, 28-29, n. 1 (1981); Brief for Petitioner 6-7. The Commission thus operated on the assumption that all expenditure limits imposed on political parties were, in essence, contribution limits and therefore constitutional. Brief for Respondent in Colorado I, O. T. 1995, No. 95-489, pp. 28-30. Such limits include 2 U. S. C. § 441a(d)(3), which provides that in elections for the United States Senate, each national or state party committee is limited to spending the greater of $20,000 (adjusted for inflation, § 441a(c)) or two cents multiplied by the voting age population of the State in which the election is held, § 441a(d)(3)(A).
Colorado I was an as-applied challenge to § 441a(d)(3) (which we spoke of as the Party Expenditure Provision), occasioned by the Commission's enforcement action against the Colorado Republican Federal Campaign Committee (Party) for exceeding the campaign spending limit through its payments for radio advertisements attacking Democratic Congressman and senatorial candidate Timothy Wirth. 518 U. S., at 612-613. The Party defended in part with the claim that the party expenditure limitations violated the First Amendment, and the principal opinion in Colorado I agreed that the limitations were unconstitutional as applied to the advertising expenditures at issue. Unlike the Commission, the Members of the Court who joined the principal opinion thought the payments were "independent expenditures" as that term had been used in our prior cases, owing to the facts that the Party spent the money before selecting its own senatorial candidate and without any arrangement with potential nominees. Id., at 613-614 (opinion of Breyer, J.).
The Party's broader claim remained: that although prior decisions of this Court had upheld the constitutionality of limits on coordinated expenditures by political speakers other than parties, the congressional campaign expenditure limitations on parties themselves are facially unconstitutional, and so are incapable of reaching party spending even when coordinated with a candidate. Id., at 623-626. We remanded that facial challenge, which had not been fully briefed or considered below. Ibid. On remand the District Court held for the Party, 41 F. Supp. 2d 1197 (1999), and a divided panel of the Court of Appeals for the Tenth Circuit affirmed, 213 F. 3d 1221 (2000). We granted certiorari to resolve the question left open by Colorado I, see 531 U. S. 923 (2000), and we now reverse.
Spending for political ends and contributing to political candidates both fall within the First Amendment's protection of speech and political association. Buckley, 424 U. S., at 14-23. But ever since we first reviewed the 1971 Act, we have understood that limits on political expenditures deserve closer scrutiny than restrictions on political contributions. Ibid.; see also, e. g., Shrink Missouri, 528 U. S., at 386-388; Colorado I, supra, at 610, 614-615; Massachusetts Citizens for Life, supra, at 259-260. Restraints on expenditures generally curb more expressive and associational activity than limits on contributions do. Shrink Missouri, supra, at 386-388; Colorado I, supra, at 615; Buckley, supra, at 19-23. A further reason for the distinction is that limits on contributions are more clearly justified by a link to political corruption than limits on other kinds of unlimited political spending are (corruption being understood not only as quid pro quo agreements, but also as undue influence on an officeholder's judgment, and the appearance of such influence, Shrink Missouri, supra, at 388-389). At least this is so where the spending is not coordinated with a candidate or his campaign. Colorado I, supra, at 615; Buckley, 424 U. S., at 47. In Buckley we said that:
"[u]nlike contributions, . . . independent expenditures may well provide little assistance to the candidate's campaign and indeed may prove counterproductive. The absence of prearrangement and coordination of an expenditure with the candidate or his agent not only undermines the value of the expenditure to the candidate, but also alleviates the danger that expenditures will be given as a quid pro quo for improper commitments from the candidate." Ibid.
Given these differences, we have routinely struck down limitations on independent expenditures by candidates, other individuals, and groups, see Federal Election Comm'n v. National Conservative Political Action Comm., 470 U. S. 480, 490-501 (1985) (political action committees); Buckley, supra, at 39-58 (individuals, groups, candidates, and campaigns), while repeatedly upholding contribution limits, see Shrink Missouri, supra (contributions by political action committees); California Medical Assn. v. Federal Election Comm'n, 453 U. S. 182, 193-199 (1981) (contributions by individuals and associations); Buckley, supra, at 23-36 (contributions by individuals, groups, and political committees).
The First Amendment line between spending and donating is easy to draw when it falls between independent expenditures by individuals or political action committees (PACs) without any candidate's approval (or wink or nod), and contributions in the form of cash gifts to candidates. See, e. g., Shrink Missouri, supra, at 386-388; Buckley, supra, at 19 23. But facts speak less clearly once the independence of the spending cannot be taken for granted, and money spent by an individual or PAC according to an arrangement with a candidate is therefore harder to classify. As already seen, Congress drew a functional, not a formal, line between contributions and expenditures when it provided that coordinated expenditures by individuals and nonparty groups are subject to the Act's contribution limits, 2 U. S. C. § 441a(a)(7)(B)(i); Colorado I, 518 U. S., at 611. In Buckley, the Court acknowledged Congress's functional classification, 424 U. S., at 46-47, and n. 53, and observed that treating coordinated expenditures as contributions "prevent[s] attempts to circumvent the Act through prearranged or coordinated expenditures amounting to disguised contributions," id., at 47. Buckley, in fact, enhanced the significance of this functional treatment by striking down independent expenditure limits on First Amendment grounds while upholding limitations on contributions (by individuals and nonparty groups), as defined to include coordinated expenditures, id., at 23-59.
Colorado I addressed the FEC's effort to stretch the functional treatment of coordinated expenditures further than the plain application of the statutory definition. As we said, the FEC argued that parties and candidates are coupled so closely that all of a party's expenditures on an election campaign are coordinated with its candidate; because Buckley had treated some coordinated expenditures like contributions and upheld their limitation, the argument went, the Party Expenditure Provision should stand as applied to all party election spending. See Brief for Respondent in Colorado I, O. T. 1995, No. 95-489, at 28-30; see also Colorado I, supra, at 619-623. Colorado I held otherwise, however, the principal opinion's view being that some party expenditures could be seen as "independent" for constitutional purposes. 518 U. S., at 614. The principal opinion found no reason to see these expenditures as more likely to serve or be seen as instruments of corruption than independent expenditures by anyone else. So there was no justification for subjecting party election spending across the board to the kinds of limits previously invalidated when applied to individuals and nonparty groups. The principal opinion observed that "[t]he independent expression of a political party's views is `core' First Amendment activity no less than is the independent expression of individuals, candidates, or other political committees." Id., at 616. Since the FEC did not advance any other convincing reason for refusing to draw the independent-coordinated line accepted since Buckley, see National Conservative Political Action Comm., 470 U. S., at 497-498; Buckley, supra, at 46-47, that was the end of the case so far as it concerned independent spending. Colorado I, supra, at 617-623.
But that still left the question whether the First Amendment allows coordinated election expenditures by parties to be treated functionally as contributions, the way coordinated expenditures by other entities are treated. Colorado I found no justification for placing parties at a disadvantage when spending independently; but was there a case for leaving them entirely free to coordinate unlimited spending with candidates when others could not? The principal opinion in Colorado I noted that coordinated expenditures "share some of the constitutionally relevant features of independent expenditures." 518 U. S., at 624. But it also observed that "many [party coordinated expenditures] are . . . virtually indistinguishable from simple contributions." Ibid. Coordinated spending by a party, in other words, covers a spectrum of activity, as does coordinated spending by other political actors. The issue in this case is, accordingly, whether a party is otherwise in a different position from other political speakers, giving it a claim to demand a generally higher standard of scrutiny before its coordinated spending can be limited. The issue is posed by two questions: does limiting coordinated spending impose a unique burden on parties, and is there reason to think that coordinated spending by a party would raise the risk of corruption posed when others spend in coordination with a candidate? The issue is best viewed through the positions developed by the Party and the Government in this case.
The Party's argument that its coordinated spending, like its independent spending, should be left free from restriction under the Buckley line of cases boils down to this: because a party's most important speech is aimed at electing candidates and is itself expressed through those candidates, any limit on party support for a candidate imposes a unique First Amendment burden. See Brief for Respondent 26-31. The point of organizing a party, the argument goes, is to run a successful candidate who shares the party's policy goals. Id., at 26. Therefore, while a campaign contribution is only one of several ways that individuals and nonparty groups speak and associate politically, see Shrink Missouri, 528 U. S., at 386-387; Buckley, supra, at 20-22, financial support of candidates is essential to the nature of political parties as we know them. And coordination with a candidate is a party's natural way of operating, not merely an option that can easily be avoided. Brief for Respondent 26. Limitation of any party expenditure coordinated with a candidate, the Party contends, is therefore a serious, rather than incidental, imposition on the party's speech and associative purpose, and that justifies a stricter level of scrutiny than we have applied to analogous limits on individuals and nonparty groups. But whatever level of scrutiny is applied, the Party goes on to argue, the burden on a party reflects a fatal mismatch between the effects of limiting coordinated party expenditures and the prevention of corruption or the appearance of it. Brief for Respondent 20-22, 25-32; see also 213 F. 3d, at 1227.
The Government's argument for treating coordinated spending like contributions goes back to Buckley. There, the rationale for endorsing Congress's equation of coordinated expenditures and contributions was that the equation "prevent[s] attempts to circumvent the Act through prearranged or coordinated expenditures amounting to disguised contributions." 424 U. S., at 47. The idea was that coordinated expenditures are as useful to the candidate as cash, and that such "disguised contributions" might be given "as a quid pro quo for improper commitments from the candidate" (in contrast to independent expenditures, which are poor sources of leverage for a spender because they might be duplicative or counterproductive from a candidate's point of view). Ibid. In effect, therefore, Buckley subjected limits on coordinated expenditures by individuals and nonparty groups to the same scrutiny it applied to limits on their cash contributions. The standard of scrutiny requires the limit to be "`closely drawn' to match a `sufficiently important interest,' . . . though the dollar amount of the limit need not be `fine tun[ed],' " Shrink Missouri, supra, at 387-388 (quoting Buckley, supra, at 25, 30).
The Government develops this rationale a step further in applying it here. Coordinated spending by a party should be limited not only because it is like a party contribution, but for a further reason. A party's right to make unlimited expenditures coordinated with a candidate would induce individual and other nonparty contributors to give to the party in order to finance coordinated spending for a favored candidate beyond the contribution limits binding on them. The Government points out that a degree of circumvention is occurring under present law (which allows unlimited independent spending and some coordinated spending). Individuals and nonparty groups who have reached the limit of direct contributions to a candidate give to a party with the understanding that the contribution to the party will produce increased party spending for the candidate's benefit. The Government argues that if coordinated spending were unlimited, circumvention would increase: because coordinated spending is as effective as direct contributions in supporting a candidate, an increased opportunity for coordinated spending would aggravate the use of a party to funnel money to a candidate from individuals and nonparty groups, who would thus bypass the contribution limits that Buckley upheld.
Each of the competing positions is plausible at first blush. Our evaluation of the arguments, however, leads us to reject the Party's claim to suffer a burden unique in any way that should make a categorical difference under the First Amendment. On the other side, the Government's contentions are ultimately borne out by evidence, entitling it to prevail in its characterization of party coordinated spending as the functional equivalent of contributions.
In assessing the Party's argument, we start with a word about what the Party is not saying. First, we do not understand the Party to be arguing that the line between independent and coordinated expenditures is conceptually unsound when applied to a political party instead of an individual or other association. See, e. g., Brief for Respondent 29 (describing "independent party speech"). Indeed, the good sense of recognizing the distinction between independence and coordination was implicit in the principal opinion in Colorado I, which did not accept the notion of a "metaphysical identity" between party and candidate, 518 U. S., at 622 623, but rather decided that some of a party's expenditures could be understood as being independent and therefore immune to limitation just as an individual's independent expenditure would be, id., at 619-623.
Second, we do not understand the Party to be arguing that associations in general or political parties in particular may claim a variety of First Amendment protection that is different in kind from the speech and associational rights of their members. The Party's point, rather, is best understood as a factual one: coordinated spending is essential to parties because "a party and its candidate are joined at the hip," Brief for Respondent 31, owing to the very conception of the party as an organization formed to elect candidates. Parties, thus formed, have an especially strong working relationship with their candidates, id., at 26, and the speech this special relationship facilitates is much more effective than independent speech, id., at 29.
There are two basic arguments here. The first turns on the relationship of a party to a candidate: a coordinated relationship between them so defines a party that it cannot function as such without coordinated spending, the object of which is a candidate's election. We think political history and political reality belie this argument. The second argument turns on the nature of a party as uniquely able to spend in ways that promote candidate success. We think that this argument is a double-edged sword, and one hardly limited to political parties.
The assertion that the party is so joined at the hip to candidates that most of its spending must necessarily be coordinated spending is a statement at odds with the history of nearly 30 years under the Act. It is well to remember that ever since the Act was amended in 1974, coordinated spending by a party committee in a given race has been limited by the provision challenged here (or its predecessor). See 18 U. S. C. § 608(f) (1970 ed., Supp. IV); see also Buckley, 424 U. S., at 194 (reprinting then-effective Party Expenditure Provision). It was not until 1996 and the decision in Colorado I that any spending was allowed above that amount, and since then only independent spending has been unlimited. As a consequence, the Party's claim that coordinated spending beyond the limit imposed by the Act is essential to its very function as a party amounts implicitly to saying that for almost three decades political parties have not been functional or have been functioning in systematic violation of the law. The Party, of course, does not in terms make either statement, and we cannot accept either implication. There is no question about the closeness of candidates to parties and no doubt that the Act affected parties' roles and their exercise of power. But the political scientists who have weighed in on this litigation observe that "there is little evidence to suggest that coordinated party spending limits adopted by Congress have frustrated the ability of political parties to exercise their First Amendment rights to support their candidates," and that "[i]n reality, political parties are dominant players, second only to the candidates themselves, in federal elections." Brief for Paul Allen Beck et al. as Amici Curiae 5-6. For the Party to claim after all these years of strictly limited coordinated spending that unlimited coordinated spending is essential to the nature and functioning of parties is in reality to assert just that "metaphysical identity," 518 U. S., at 623, between free-spending party and candidate that we could not accept in Colorado I.
There is a different weakness in the seemingly unexceptionable premise that parties are organized for the purpose of electing candidates, Brief for Respondent 26 ("Parties exist precisely to elect candidates that share the goals of their party"), so that imposing on the way parties serve that function is uniquely burdensome. The fault here is not so much metaphysics as myopia, a refusal to see how the power of money actually works in the political structure.
When we look directly at a party's function in getting and spending money, it would ignore reality to think that the party role is adequately described by speaking generally of electing particular candidates. The money parties spend comes from contributors with their own personal interests. PACs, for example, are frequent party contributors who (according to one of the Party's own experts) "do not pursue the same objectives in electoral politics" that parties do. App. 180 (statement of Professor Anthony Corrado). PACs "are most concerned with advancing their narrow interest[s]" and therefore "provide support to candidates who share their views, regardless of party affiliation." Ibid. In fact, many PACs naturally express their narrow interests by contributing to both parties during the same electoral cycle, and sometimes even directly to two competing candidates in the same election, L. Sabato, PAC Power, Inside the World of Political Action Committees 88 (1984). Parties 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, but rather to support a specific candidate for the sake of a position on one narrow issue, or even to support any candidate who will be obliged to the contributors.
Parties thus perform functions more complex than simply electing candidates; whether they like it or not, they act as agents for spending on behalf of those who seek to produce obligated officeholders. It is this party role, which functionally unites parties with other self-interested political actors, that the Party Expenditure Provision targets. This party role, accordingly, provides good reason to view limits on coordinated spending by parties through the same lens applied to such spending by donors, like PACs, that can use parties as conduits for contributions meant to place candidates under obligation.
Insofar as the Party suggests that its strong working relationship with candidates and its unique ability to speak in coordination with them should be taken into account in the First Amendment analysis, we agree. It is the accepted understanding that a party combines its members' power to speak by aggregating contributions and broadcasting messages more widely than individual contributors generally could afford to do, and the party marshals this power with greater sophistication than individuals generally could, using such mechanisms as speech coordinated with a candidate. In other words, the party is efficient in generating large sums to spend and in pinpointing effective ways to spend them. Cf. Colorado I, 518 U. S., at 637 (Thomas, J., concurring in judgment and dissenting in part) ("Political associations allow citizens to pool their resources and make their advocacy more effective").
It does not, however, follow from a party's efficiency in getting large sums and spending intelligently that limits on a party's coordinated spending should be scrutinized under an unusually high standard, and in fact any argument from sophistication and power would cut both ways. On the one hand, one can seek the benefit of stricter scrutiny of a law capping party coordinated spending by emphasizing the heavy burden imposed by limiting the most effective mechanism of sophisticated spending. And yet it is exactly this efficiency culminating in coordinated spending that (on the Government's view) places a party in a position to be used to circumvent contribution limits that apply to individuals and PACs, and thereby to exacerbate the threat of corruption and apparent corruption that those contribution limits are aimed at reducing. As a consequence, what the Party calls an unusual burden imposed by regulating its spending is not a simple premise for arguing for tighter scrutiny of limits on a party; it is the premise for a question pointing in the opposite direction. If the coordinated spending of other, less efficient and perhaps less practiced political actors can be limited consistently with the Constitution, why would the Constitution forbid regulation aimed at a party whose very efficiency in channeling benefits to candidates threatens to undermine the contribution (and hence coordinated spending) limits to which those others are unquestionably subject?
The preceding question assumes that parties enjoy a power and experience that sets them apart from other political spenders. But in fact the assumption is too crude. While parties command bigger spending budgets than most individuals, some individuals could easily rival party committees in spending. Rich political activists crop up, and the United States has known its Citizens Kane. Their money speaks loudly, too, and they are therefore burdened by restrictions on its use just as parties are. And yet they are validly subject to coordinated spending limits, Buckley, 424 U. S., at 46-47, and so are PACs, id., at 35-36, 46-47, which may amass bigger treasuries than most party members can spare for politics.
Just as rich donors, media executives, and PACs have the means to speak as loudly as parties do, they would also have the capacity to work effectively in tandem with a candidate, just as a party can do. While a candidate has no way of coordinating spending with every contributor, there is nothing hard about coordinating with someone with a fortune to donate, any more than a candidate would have difficulty in coordinating spending with an inner circle of personal political associates or with his own family. Yet all of them are subject to coordinated spending limits upheld in Buckley, supra, at 53, n. 59. A party, indeed, is now like some of these political actors in yet another way: in its right under Colorado I to spend money in support of a candidate without legal limit so long as it spends independently. A party may spend independently every cent it can raise wherever it thinks its candidate will shine, on every subject and any viewpoint.
A party is not, therefore, in a unique position. It is in the same position as some individuals and PACs, as to whom coordinated spending limits have already been held valid, Buckley, supra, at 46-47; and, indeed, a party is better off, for a party has the special privilege the others do not enjoy, of making coordinated expenditures up to the limit of the Party Expenditure Provision.
The Party's arguments for being treated differently from other political actors subject to limitation on political spending under the Act do not pan out. Despite decades of limitation on coordinated spending, parties have not been rendered useless. In reality, parties continue to organize to elect candidates, and also function for the benefit of donors whose object is to place candidates under obligation, a fact that parties cannot escape. Indeed, parties' capacity to concentrate power to elect is the very capacity that apparently opens them to exploitation as channels for circumventing contribution and coordinated spending limits binding on other political players. And some of these players could marshal the same power and sophistication for the same electoral objectives as political parties themselves.
We accordingly apply to a party's coordinated spending limitation the same scrutiny we have applied to the other political actors, that is, scrutiny appropriate for a contribution limit, enquiring whether the restriction is "closely drawn" to match what we have recognized as the "sufficiently important" government interest in combating political corruption. Shrink Missouri, 528 U. S., at 387-388 (quoting Buckley, supra, at 25, 30). With the standard thus settled, the issue remains whether adequate evidentiary grounds exist to sustain the limit under that standard, on the theory that unlimited coordinated spending by a party raises the risk of corruption (and its appearance) through circumvention of valid contribution limits. Indeed, all Members of the Court agree that circumvention is a valid theory of corruption; the remaining bone of contention is evidentiary.
Since there is no recent experience with unlimited coordinated spending, the question is whether experience under the present law confirms a serious threat of abuse from the unlimited coordinated party spending as the Government contends. Cf. Burson v. Freeman, 504 U. S. 191, 208 (1992) (opinion of Blackmun, J.) (noting difficulty of mustering evidence to support long-enforced statutes). It clearly does. Despite years of enforcement of the challenged limits, substantial evidence demonstrates how candidates, donors, and parties test the limits of the current law, and it shows beyond serious doubt how contribution limits would be eroded if inducement to circumvent them were enhanced by declaring parties' coordinated spending wide open.
Under the Act, a donor is limited to $2,000 in contributions to one candidate in a given election cycle. The same donor may give as much as another $20,000 each year to a national party committee supporting the candidate. What a realist would expect to occur has occurred. Donors give to the party with the tacit understanding that the favored candidate will benefit. See App. 247 (declaration of Robert Hickmott, former Democratic fundraiser and National Finance Director for Timothy Wirth's Senate campaign) ("We . . . told contributors who had made the maximum allowable contribution to the Wirth campaign but who wanted to do more that they could raise money for the DSCC so that we could get our maximum [Party Expenditure Provision] allocation from the DSCC"); id., at 274 (declaration of Timothy Wirth) ("I 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"); id., at 166 (declaration of Leon G. Billings, former Executive Director of the Democratic Senatorial Campaign Committee (DSCC)) ("People often contribute to party committees because they have given the maximum amount to a candidate, and want to help the candidate indirectly by contributing to the party"); id., at 99-100 (fund raising letter from Congressman Wayne Allard, dated Aug. 27, 1996, explaining to contributor that "you are at the limit of what you can directly contribute to my campaign," but "you can further help my campaign by assisting the Colorado Republican Party").
Although the understanding between donor and party may involve no definite commitment and may be tacit on the donor's part, the frequency of the practice and the volume of money involved has required some manner of informal bookkeeping by the recipient. In the Democratic Party, at least, the method is known as "tallying," a system that helps to connect donors to candidates through the accommodation of a party. See App. 246-247 (Hickmott declaration) ("[The tally system] is an informal agreement between the DSCC and the candidates' campaigns that if you help the DSCC raise contributions, we will turn around and help your campaign"); id., at 268 (declaration of former Senator Paul Simon) ("Donors would be told the money they contributed could be credited to any Senate candidate. The callers would make clear that this was not a direct contribution, but it was fairly close to direct"); id., at 165-166 (Billings declaration) ("There appeared to be an understanding between the DSCC and the Senators that the amount of money they received from the DSCC was related to how much they raised for the Committee").
Such is the state of affairs under the current law, which requires most party spending on a candidate's behalf to be done independently, and thus less desirably from the point of view of a donor and his favored candidate. If suddenly every dollar of spending could be coordinated with the candidate, the inducement to circumvent would almost certainly intensify. Indeed, if a candidate could be assured that donations through a party could result in funds passed through to him for spending on virtually identical items as his own campaign funds, a candidate enjoying the patronage of affluent contributors would have a strong incentive not merely to direct donors to his party, but to promote circumvention as a step toward reducing the number of donors requiring time-consuming cultivation. If a candidate could arrange for a party committee to foot his bills, to be paid with $20,000 contributions to the party by his supporters, the number of donors necessary to raise $1,000,000 could be reduced from 500 (at $2,000 per cycle) to 46 (at $2,000 to the candidate and $20,000 to the party, without regard to donations outside the election year).
While this evidence rules out denying the potential for corruption by circumvention, the Party does try to minimize the threat. It says that most contributions to parties are small, with negligible corrupting momentum to be carried through the party conduit. Brief for Respondent 14. But some contributions are not small; they can go up to $20,000, 2 U. S. C. § 441a(a)(1)(B), and the record shows that even under present law substantial donations turn the parties into matchmakers whose special meetings and receptions give the donors the chance to get their points across to the candidates. The Party again discounts the threat of outflanking contribution limits on individuals and nonparty groups by stressing that incumbent candidates give more excess campaign funds to parties than parties spend on coordinated expenditures. Brief for Respondent 34. But the fact that parties may do well for themselves off incumbents does not defuse concern over circumvention; if contributions to a party were not used as a funnel from donors to candidates, there would be no reason for using the tallying system the way the witnesses have described it.
Finally, the Party falls back to claiming that, even if there is a threat of circumvention, the First Amendment demands a response better tailored to that threat than a limitation on spending, even coordinated spending. Id., at 46-48. The Party has two suggestions.
First, it says that better crafted safeguards are in place already, in particular the earmarking rule of § 441a(a)(8), which provides that 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. The Party says that this provision either suffices to address any risk of circumvention or would suffice if clarified to cover practices like tallying. Id., at 42, 47; see also 213 F. 3d, at 1232. This position, however, ignores the practical difficulty of identifying and directly combating circumvention under actual political conditions. Donations are made to a party by contributors who favor the party's candidates in races that affect them; donors are (of course) permitted to express their views and preferences to party officials; and the party is permitted (as we have held it must be) to spend money in its own right. When this is the environment for contributions going into a general party treasury, and candidate-fundraisers are rewarded with something less obvious than dollar-for-dollar pass-throughs (distributed through contributions and party spending), circumvention is obviously very hard to trace. The earmarking provision, even if it dealt directly with tallying, would reach only the most clumsy attempts to pass contributions through to candidates. To treat the earmarking provision as the outer limit of acceptable tailoring would disarm any serious effort to limit the corrosive effects of what Chief Judge Seymour called "`understandings' regarding what donors give what amounts to the party, which candidates are to receive what funds from the party, and what interests particular donors are seeking to promote," id., at 1241 (dissenting opinion); see also Briffault, Political Parties and Campaign Finance Reform, 100 Colum. L. Rev. 620, 652 (2000) (describing "web of relations linking major donors, party committees, and elected officials").
The Party's second preferred prescription for the threat of an end run calls for replacing limits on coordinated expenditures by parties with limits on contributions to parties, the latter supposedly imposing a lesser First Amendment burden. Brief for Respondent 46-48. The Party thus invokes the general rule that contribution limits take a lesser First Amendment toll, expenditure limits a greater one. That was one strand of the reasoning in Buckley itself, which rejected the argument that limitations on independent expenditures by individuals, groups, and candidates were justifiable in order to avoid circumvention of contribution limitations. 424 U. S., at 44. It was also one strand of the logic of the Colorado I principal opinion in rejecting the Party Expenditure Provision's application to independent party expenditures. 518 U. S., at 617.
In each of those cases, however, the Court's reasoning contained another strand. The analysis ultimately turned on the understanding that the expenditures at issue were not potential alter egos for contributions, but were independent and therefore functionally true expenditures, qualifying for the most demanding First Amendment scrutiny employed in Buckley. Colorado I, supra, at 617; Buckley, supra, at 44-47. Thus, in Colorado I we could not assume, "absent convincing evidence to the contrary," that the Party's independent expenditures formed a link in a chain of corruptionby-conduit. 518 U. S., at 617. "The absence of prearrangement and coordination of an expenditure with the candidate or his agent not only undermines the value of the expenditure to the candidate, but also alleviates the danger that expenditures will be given as a quid pro quo for improper commitments from the candidate," Buckley, supra, at 47; therefore, "the constitutionally significant fact" in Colorado I was "the lack of coordination between the candidate and the source of the expenditure," 518 U. S., at 617.
Here, however, just the opposite is true. There is no significant functional difference between a party's coordinated expenditure and a direct party contribution to the candidate, and there is good reason to expect that a party's right of unlimited coordinated spending would attract increased contributions to parties to finance exactly that kind of spending. Coordinated expenditures of money donated to a party are tailor-made to undermine contribution limits. Therefore the choice here is not, as in Buckley and Colorado I, between a limit on pure contributions and pure expenditures. The choice is between limiting contributions and limiting expenditures whose special value as expenditures is also the source of their power to corrupt. Congress is entitled to its choice.
* * *
We hold that a party's coordinated expenditures, unlike expenditures truly independent, may be restricted to minimize circumvention of contribution limits. We therefore reject the Party's facial challenge and, accordingly, reverse the judgment of the United States Court of Appeals for the Tenth Circuit.
It is so ordered. Justice Thomas, 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.
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 Comm., 489 U. S. 214, 223 (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, J., dissenting) ("[P]olitical 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, 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.
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 expenditures 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 "entai[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.
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.
The Court considers a coordinated expenditure to be an "`expenditur[e] 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, independent expenditures. Take, for example, a situation in which the party develops a television advertising campaign touting a candidate's record on education, and the party simply "consult[s]," 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.
Even if I were to ignore the breadth of the statutory text, and to assume that all coordinated expenditures are functionally equivalent to contributions, 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. Restricting 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 partiessuch 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 1968are 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). "[C]andidates 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 Brief 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 outcome 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 fund raising demands on party organizations because independent expenditures are less effective means of communication); id., at 219 ("[I]ndependent 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 (deposition of John Heubusch); App. 159 (affidavit of Donald K. Bain) ("[O]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."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 functional 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 "observ[ations]" 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. 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 a "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, "[p]arties 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 contributors. 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.
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.
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.
Considering 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."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 record 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).
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 646 (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 Comm'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. 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. 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 that 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 "connect[s] 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 funds 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 marginalthat 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 "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.
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 392 ("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 the 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) ("`I 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 unpersuaded 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.
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").
In 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 corruptionthe individual donorand 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.
[*] Briefs of amici curiae urging reversal were filed for the State of Missouri et al. by Jeremiah W. (Jay) Nixon, Attorney General of Missouri, and James R. Layton, State Solicitor, joined by the Attorneys General for their respective States as follows: Ken Salazar of Colorado, Earl I. Anzai of Hawaii, Joseph P. Mazurek of Montana, Eliot Spitzer of New York, W. A. Drew Edmondson of Oklahoma, and William H. Sorrell of Vermont; for Common Cause et al. by Roger M. Witten, Daniel H. Squire, Donald J. Simon, and Fred Wertheimer; for the National Voting Rights Institute by David A. Wilson, John C. Bonifaz, Brenda Wright, and Gregory G. Luke; for Senator John F. Reed et al. by Donald B. Verrilli, Jr., and Deanne E. Maynard; and for Paul Allen Beck et al. by Burt Neuborne.
Briefs of amici curiae urging affirmance were filed for the American Civil Liberties Union et al. by Mark J. Lopez, Steven R. Shapiro, and Joel M. Gora; for the California Republican Party by Charles H. Bell, Jr.; for the Missouri Republican Party by D. Bruce La Pierre and W. Bevis Schock; and for the National Republican Congressional Committee by Benjamin L. Ginsberg.
 "Contribution" is otherwise defined as "any gift, subscription, loan, advance, or deposit of money or anything of value made by any person for the purpose of influencing any election for Federal office"; or "the payment by any person of compensation for the personal services of another person which are rendered to a political committee without charge for any purpose." 2 U. S. C. § 431(8).
The Act defines "expenditure" as "any purchase, payment, distribution, loan, advance, deposit, or gift of money or anything of value, made by any person for the purpose of influencing any election for Federal office." § 431(9)(A)(i). A "written contract, promise, or agreement to make an expenditure" also counts as an expenditure. § 431(9)(A)(ii).
 A political party's "national committee" is the "organization which, by virtue of the bylaws of a political party, is responsible for the day-to-day operation of such political party at the national level, as determined by the [Federal Election] Commission." § 431(14). A "state committee" fills the same role at the state level. § 431(15).
 The same limits apply to campaigns for House of Representatives from States entitled to only one Representative. § 441a(d)(3)(A). For other States, the limit on party expenditures in connection with House campaigns is $10,000 preadjustment. § 441a(d)(3)(B). As adjusted for inflation, the 2000 Senate limits ranged from $67,560 to $1,636,438; House limits ranged from $33,780 to $67,560. 26 FEC Record 14-15 (Mar. 2000).
The FEC reads the Act to permit parties to make campaign contributions within the otherwise-applicable contribution limits, in addition to the expenditures permitted by § 441a(d). See n. 16, infra.
 The limits applicable to Presidential campaigns were not at issue in Colorado I, 518 U. S. 604, 610-611 (1996), and are not at issue here, Brief for Respondent 49, n. 30.
 Along with its constitutional claim, the Party argued to the District Court that the Party Expenditure Provision's application to independent expenditures was not severable from the other possible applications of the provision, a nonconstitutional basis for resolving the case that the Colorado I principal opinion suggested should be explored on remand. Colorado I, supra, at 625-626. The District Court rejected the nonseverability argument, 41 F. Supp. 2d, at 1207, and the Party did not renew it on appeal, 213 F. 3d, at 1225, n. 3.
 The expenditure limits invalidated in Buckley applied to candidates and their campaigns, and to "persons." See 424 U. S., at 39-40, 51, 54, 58. "Person" was defined as "an individual, partnership, committee, association, corporation, or any other organization or group of persons." 18 U. S. C. § 591(g) (1970 ed., Supp. IV); see also Buckley, 424 U. S., at 144 235 (appendix reprinting then-current Act). Although this language is broad enough to cover political parties, id., at 19, and n. 19, 39, parties with a candidate on the ballot were covered instead by the special Party Expenditure Provision, which was not challenged on First Amendment grounds, id., at 58, n. 66.
 The contribution limits at issue in Buckley applied to "persons" ("person" again defined as "an individual, partnership, committee, association, corporation or any other organization or group of persons," id., at 23). Certain groups (referred to under current law as "multicandidate political committees") that registered with the FEC and met other qualifications, including making contributions to five or more candidates for federal office, were subject to a higher limit. Id., at 35.
The current contribution limits appear in 2 U. S. C. § 441a(a). They provide that "persons" (still broadly defined, see § 431(11)) may contribute no more than $1,000 to a candidate "with respect to any election for Federal office," $5,000 to any political committee in any year, and $20,000 to the national committees of a political party in any year. § 441a(a)(1). Individuals are limited to a yearly contribution total of $25,000. § 441a(a)(3). "[M]ulticandidate political committees" are limited to a $5,000 contribution to a candidate "with respect to any election," $5,000 to any political committee in any year, and $15,000 to the national committees of a political party in any year. § 441a(a)(2). Unlike the party expenditure limits, these contribution limits are not adjusted for inflation.
 The Party does not challenge the constitutionality of limits on cash contributions from parties to candidates, Brief for Respondent 49, n. 31, which, on the FEC's reading of the Act, are imposed on parties by the generally applicable contribution limits of 2 U. S. C. § 441a(a), see n. 16, infra. And the Party, unlike Justice Thomas, post, at 465 (dissenting opinion), does not call for the overruling of Buckley. Nor does the FEC ask us to revisit Buckley `s general approach to expenditure limits, although some have argued that such limits could be justified in light of post-Buckley developments in campaign finance, see, e. g., Blasi, Free Speech and the Widening Gyre of Fundraising, 94 Colum. L. Rev. 1281 (1994); cf. Nixon v. Shrink Missouri Government PAC, 528 U. S. 377, 409 (2000) (Kennedy, J., dissenting) ("I would leave open the possibility that Congress, or a state legislature, might devise a system in which there are some limits on both expenditures and contributions, thus permitting officeholders to concentrate their time and efforts on official duties rather than on fund raising"); id., at 405 (Breyer, J., concurring) ("Suppose Buckley denies the political branches sufficient leeway to enact comprehensive solutions to the problems posed by campaign finance. If so, like Justice Kennedy, I believe the Constitution would require us to reconsider Buckley ").
 As noted, n. 6, supra, the Party Expenditure Provision itself was not challenged on First Amendment grounds in Buckley, supra, at 58, n. 66.
 We have repeatedly held that political parties and other associations derive rights from their members. E. g., Norman v. Reed, 502 U. S. 279, 288 (1992); Tashjian v. Republican Party of Conn., 479 U. S. 208, 214-215 (1986); Roberts v. United States Jaycees, 468 U. S. 609, 622-623 (1984); NAACP v. Alabama ex rel. Patterson, 357 U. S. 449, 459-460 (1958); Sweezy v. New Hampshire, 354 U. S. 234, 250 (1957). While some commentators have assumed that associations' rights are also limited to the rights of the individuals who belong to them, e. g., Supreme Court, 1996 Term, Leading Cases, Associational Rights of Political Parties, 111 Harv. L. Rev. 197, 315, n. 50 (1997), that view has been subject to debate, see, e. g., Gottlieb, Fleshing Out the Right of Association, 49 Albany L. Rev. 825, 826, 836-837 (1985); see generally Issacharoff, Private Parties with Public Purposes, 101 Colum. L. Rev. 274 (2001). There is some language in our cases supporting the position that parties' rights are more than the sum of their members' rights, e. g., California Democratic Party v. Jones, 530 U. S. 567, 575 (2000) (referring to the "special place" the First Amendment reserves for the process by which a political party selects a standard bearer); Timmons v. Twin Cities Area New Party, 520 U. S. 351, 373 (1997) (Stevens, J., dissenting), but we have never settled upon the nature of any such difference and have no reason to do so here.
 To say that history and common sense make us skeptical that parties are uniquely incapacitated by the challenged limitations is not to deny that limiting parties' coordinated expenditures while permitting unlimited independent expenditures prompts parties to structure their spending in a way that they would not otherwise choose. See post, at 470. And we acknowledge below, infra, at 453-455, that limiting coordinated expenditures imposes some burden on parties' associational efficiency. But the very evidence cited by the dissent suggests that it is nonetheless possible for parties, like individuals and nonparty groups, to speak independently. E. g., App. 218 (statement of Professor Anthony Corrado) ("[I]t is likely that parties will allocate an increasing amount of money to independent expenditure efforts in the future"); id., at 159 (affidavit of Donald K. Bain, Chairman of the Colorado Republican Federal Campaign Committee) (describing ability to make independent expenditures as "welcome").
 As former Senator Paul Simon explained, "I believe people contribute to party committees on both sides of the aisle for the same reason that Federal Express does, because they want favors. There is an expectation that giving to party committees helps you legislatively." Id., at 270. See also id., at 269-270 (recounting debate over a bill favored by Federal Express during which a colleague exclaimed "we've got to pay attention to who is buttering our bread").
The FEC's public records confirm that Federal Express's PAC (along with many others) contributed to both major parties in recent elections. See, e. g., FEC Disclosure Report, Search Results for Federal Express Political Action Committee (June 20, 2001), http://herndon1.sdrdc.com/ cgi-bin/com supopp/C00068692; FEC Disclosure Report, Search Results for Association of Trial Lawyers of America Political Action Committee (June 20, 2001), http://herndon1.sdrdc.com/cgi-bin/com supopp/C00024521; FEC Disclosure Report, Search Results for Philip Morris Companies, Inc., Political Action Committee (June 20, 2001), http://herndon1.sdrdc.com/ cgi-bin/com supopp/C00089136; FEC Disclosure Report, Search Results for American Medical Association Political Action Committee (June 20, 2001), http://herndon1.sdrdc.com/cgi-bin/com supopp/C00000422; FEC Disclosure Report, Search Results for Letter Carriers Political Action Fund (June 20, 2001), http://herndon1.sdrdc.com/cgi-bin/com supopp/ C00023580.
 For example, the PACs associated with AOL Time Warner Inc. and Philip Morris Companies, Inc., both made contributions to the competing 2000 Senate campaigns of George Allen and Charles Robb. See FEC Disclosure Report, Search Results for AOL Time Warner Inc. Political Action Committee (June 20, 2001), http://herndon1.sdrdc.com/cgi-bin/ com supopp/C00339291; FEC Disclosure Report, Search Results for Philip Morris Companies, Inc., Political Action Committee, supra.
 We have long recognized Congress's concern with this reality of political life. For example, in United States v. Automobile Workers, 352 U. S. 567 (1957), Justice Frankfurter recounted Senator Robinson's explanation for the Federal Corrupt Practices Act's restriction of corporate campaign contributions:
"`We all know . . . that one of the great political evils of the time is the apparent hold on political parties which business interests and certain organizations seek and sometimes obtain by reason of liberal campaign contributions. Many believe that when an individual or association of individuals makes large contributions for the purpose of aiding candidates of political parties in winning the elections, they expect, and sometimes demand, and occasionally, at least, receive, consideration by the beneficiaries of their contributions which not infrequently is harmful to the general public interest.' " Id., at 576 (quoting 65 Cong. Rec. 9507-9508 (1924)).
 By noting that other political actors are validly burdened by limitations on their coordinated spending, we do not mean to take a position as to the wisdom of policies that promote one source of campaign funding or another. Cf. Brief for Respondent 27, n. 17 (citing academic support for expanding the role of parties in campaign finance).
 This is the position of the FEC in the aftermath of Colorado I: that a party committee may make coordinated expenditures up to the amount of its expenditure limit, in addition to the amount of direct contributions permitted by the generally applicable contribution limit. Brief for Petitioner 5-6, and n. 3.
 Whether a different characterization, and hence a different type of scrutiny, could be appropriate in the context of an as-applied challenge focused on application of the limit to specific expenditures is a question that, as Justice Thomas notes, post, at 468, n. 2, we need not reach in this facial challenge. Cf. Brief for Petitioner 9, n. 5 (noting that the FEC has solicited comments regarding possible criteria for identifying coordinated expenditures).
The Party appears to argue that even if the Party Expenditure Provision is justified with regard to coordinated expenditures that amount to no more than payment of the candidate's bills, the limitation is facially invalid because of its potential application to expenditures that involve more of the party's own speech. Brief for Respondent 48-49. But the Party does not tell us what proportion of the spending falls in one category or the other, or otherwise lay the groundwork for its facial over breadth claim. Cf. Broadrick v. Oklahoma, 413 U. S. 601 (1973) (over breadth must be substantial to trigger facial invalidation).
 Apart from circumvention, the FEC also argues that the Party Expenditure Provision is justified by a concern with quid pro quo arrangements and similar corrupting relationships between candidates and parties themselves, see Brief for Petitioner 33-38. We find no need to reach that argument because the evidence supports the long-recognized rationale of combating circumvention of contribution limits designed to combat the corrupting influence of large contributions to candidates from individuals and nonparty groups. The dissent does not take issue with this justification as a theoretical matter. See also 213 F. 3d 1221, 1232 (CA10 2000) (Court of Appeals acknowledging circumvention as a possible "avenue of abuse").
 In Colorado I, the principal opinion suggested that the Party Expenditure Provision was not enacted out of "a special concern about the potentially `corrupting' effect of party expenditures, but rather for the constitutionally insufficient purpose of reducing what [Congress] saw as wasteful and excessive campaign spending." 518 U. S., at 618. That observation was relevant to our examination of the Party Expenditure Provision as applied to independent expenditures, see id., at 617-618, limits on which were invalidated with regard to other political actors in Buckley in part because they were justified by concern with wasteful campaign spending, Buckley, 424 U. S., at 57. Our point in Colorado I was that there was no evidence that Congress had a special motivation regarding parties that would justify limiting their independent expenditures after similar limits imposed on other spenders had been invalidated. As for the Party Expenditure Provision's application to coordinated expenditures, on the other hand, the evidence discussed in the text suggests that the anticircumvention rationale that justifies other coordinated expenditure limits, see Buckley, supra, at 46-47, is at work here as well. The dissent ignores this distinction, post, at 475, but neither the dissent nor the Party seriously argues that Congress was not concerned with circumvention of contribution limits using parties as conduits. All acknowledge that Congress enacted other measures prompted by just that concern. See post, at 481; Brief for Respondent 41-42 ("FECA provides interlocking multilayered provisions designed to prevent circumvention").
 See n. 7, supra; see generally Federal Election Commission, Campaign Guide for Congressional Candidates and Committees 10 (1999).
 Contrary to the dissent's suggestion, post, at 477-478, we are not closing our eyes to District Court findings rejecting this record evidence. After alluding to the evidence cited above, 41 F. Supp. 2d 1197, 1203-1204 (Colo. 1999), and concluding that it did not support theories of corruption that we do not address here, see id., at 1211; n. 18, supra, the District Court mistakenly concluded that Colorado I had rejected the anticircumvention rationale as a matter of law, 41 F. Supp. 2d, at 1211, n. 9. We explain below, infra, at 463-465, why Colorado I `srejection of the anticircumvention rationale in the context of limits applied to independent party expenditures does not control the outcome of this case.
 The dissent dismisses this evidence as describing "legal" practices. Post, at 479. The dissent may be correct that the FEC considers tallying legal, see Reply Brief for Petitioner 9, n. 3, but one thing is clear: tallying is a sign that contribution limits are being diluted and could be diluted further if the floodgates were open. Why, after all, does a party bother to tally? The obvious answer is that it wants to know who gets the benefit of the contributions to the party, as the record quotations attest. See also n. 23, infra. And the fact that the parties may not fund sure losers, stressed by the dissent (post, at 478-479), is irrelevant. The issue is what would become of contribution limits if parties could use unlimited coordinated spending to funnel contributions to those serious contenders who are favored by the donors.
 Any such dollar-for-dollar pass-through would presumably be too obvious to escape the special provision on earmarking, 2 U. S. C. § 441a(a)(8), see infra, at 462. But the example illustrates the undeniable inducement to more subtle circumvention.
The same enhanced value of coordinated spending that could be expected to promote greater circumvention of contribution limits for the benefit of the candidate-fundraiser would probably enhance the power of the fundraiser to use circumvention as a tactic to increase personal power and a claim to party leadership. The affluent nominee can already do this to a limited extent, by directing donations to the party and making sure that the party knows who raised the money, and that the needier candidates who receive the benefit of party spending know whom to thank. The candidate can thus become a player beyond his own race, and the donor's influence is multiplied. See generally App. 249 (Hickmott declaration) ("Incumbents who were not raising money for themselves because they were not up for reelection would sometimes raise money for other Senators, or for challengers. They would send $20,000 to the DSCC and ask that this be entered on another candidate's tally. They might do this, for example, if they were planning to run for a leadership position and wanted to obtain the support of the Senators they assisted"). If the effectiveness of party spending could be enhanced by limitless coordination, the ties of straitened candidates to prosperous ones and, vicariously, to large donors would be reinforced as well. Party officials who control distribution of coordinated expenditures would obviously form an additional link in this chain. See id., at 164, 168 (Billings declaration) ("[The DSCC's three-member Executive Committee] basically made the decisions as to how to distribute the money. . . . Taking away the limits on coordinated expenditures would result in a fundamental transferal of power to certain individual Senators").
 In 1996, 46 percent of itemized (over $200) individual contributions to the Democratic national party committees and 15 percent of such contributions to the Republican national party committees were $10,000 or more. Biersack & Haskell, Spitting on the Umpire: Political Parties, the Federal Election Campaign Act, and the 1996 Campaigns, in Financing the 1996 Election 155, 160 (J. Green ed. 1999).
 For example, the DSCC has established exclusive clubs for the most generous donors, who are invited to special meetings and social events with Senators and candidates. App. 254-255 (Hickmott declaration).
 The Party's argument for relying on better earmarking enforcement, accepted by the dissent, post, at 481, would invite a corresponding attack on all contribution limits. As we said in Buckley, 424 U. S., at 27-28, and Shrink Missouri, 528 U. S., at 390, the policy supporting contribution limits is the same as for laws against bribery. But we do not throw out the contribution limits for unskillful tailoring; prohibitions on bribery, like the earmarking provision here, address only the "most blatant and specific" attempts at corruption, 424 U. S., at 28.
 The dissent therefore suggests, post, at 482, and the District Court mistakenly concluded, see discussion n. 21, supra, that Colorado I disposed of the tailoring question for purposes of this case.
 The dissent notes a superficial tension between this analysis and our recent statement in Bartnicki v. Vopper, 532 U. S. 514 (2001), that "it 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. Unlike Bartnicki, there is no clear dichotomy here between law abider and lawbreaker. The problem of circumvention is a systemic one, accomplished only through complicity between donor and party.
 Also, again, contrast Bartnicki, where the gulf between the First Amendment implications of two enforcement options was clear. We rejected the decision to penalize disclosure of lawfully obtained information of public interest instead of vigorously enforcing prohibitions on intercepting private conversations. Ibid.
 NAACP v. Button, 371 U. S. 415, 444 (1963); 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 stillbe 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 "ignor[e] [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 anticircumvention 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 I )).
 Ironically, earlier this Term, this Court was less willing to uphold a speech restriction based on inference of circumvention. See, e. g., Bartnicki 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 ofLeon 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.