announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I and II, and an opinion with respect to Parts III and IV, in which Justice Auto joins.
Section 203 of the Bipartisan Campaign Reform Act of 2002 (BCRA), 116 Stat. 91, 2 U. S. C. §441b(b)(2) (2000 ed., Supp. IV), makes it a federal crime for any corporation to *456broadcast, shortly before an election, any communication that names a federal candidate for elected office and is targeted to the electorate. In McConnell v. Federal Election Comm’n, 540 U. S. 93 (2003), this Court considered whether §203 was facially overbroad under the First Amendment because it captured within its reach not only campaign speech, or “express advocacy,” but also speech about public issues more generally, or “issue advocacy,” that mentions a candidate for federal office. The Court concluded that there was no overbreadth concern to the extent the speech in question was the “functional equivalent” of express campaign speech. Id., at 204-205, 206. On the other hand, the Court “assume[d]” that the interests it had found to “justify the regulation of campaign speech might not apply to the regulation of genuine issue ads.” Id., at 206, n. 88. The Court nonetheless determined that §203 was not facially overbroad. Even assuming §203 “inhibited] some constitutionally protected corporate and union speech,” the Court concluded that those challenging the law on its face had failed to carry their “heavy burden” of establishing that all enforcement of the law should therefore be prohibited. Id., at 207.
Last Term, we reversed a lower court ruling, arising in the same litigation before us now, that our decision in McConnell left “no room” for as-applied challenges to §203. App. to Juris. Statement 52a. We held on the contrary that “[i]n upholding §203 against a facial challenge, we did not purport to resolve future as-applied challenges.” Wisconsin Right to Life, Inc. v. Federal Election Comm’n, 546 U. S. 410, 412 (2006) (per curiam) (WRTL I).
We now confront such an as-applied challenge. Resolving it requires us first to determine whether the speech at issue is the “functional equivalent” of speech expressly advocating the election or defeat of a candidate for federal office, or instead a “genuine issue a[d].” McConnell, supra, at 206, and n. 88. We have long recognized that the distinction between campaign advocacy and issue advocacy “may often dissolve in practical application. Candidates, especially incumbents, *457are intimately tied to public issues involving legislative proposals and governmental actions.” Buckley v. Valeo, 424 U. S. 1, 42 (1976) (per curiam). Our development of the law in this area requires us, however, to draw such a line, because we have recognized that the interests held to justify the regulation of campaign speech and its “functional equivalent” “might not apply” to the regulation of issue advocacy. McConnell, supra, at 206, and n. 88.
In drawing that line, the First Amendment requires us to err on the side of protecting political speech rather than suppressing it. We conclude that the speech at issue in this as-applied challenge is not the “functional equivalent” of express campaign speech. We further conclude that the interests held to justify restricting corporate campaign speech or its functional equivalent do not justify restricting issue advocacy, and accordingly we hold that BCRA §203 is unconstitutional as applied to the advertisements at issue in these cases.
I
Prior to BCRA, corporations were free under federal law to use independent expenditures to engage in political speech so long as that speech did not expressly advocate the election or defeat of a clearly identified federal candidate. See Federal Election Comm’n v. Massachusetts Citizens for Life, Inc., 479 U. S. 238, 249 (1986) (MCFL); Buckley, supra, at 44-45; 2 U. S. C. §§441b(a), (b)(2) (2000 ed. and Supp. IV).
BCRA significantly cut back on corporations’ ability to engage in political speech. BCRA §203, at issue in these cases, makes it a crime for any labor union or incorporated entity — whether the United Steelworkers, the American Civil Liberties Union, or General Motors — to use its general treasury funds to pay for any “electioneering communication.” §441b(b)(2) (2000 ed., Supp. IV). BCRA’s definition of “electioneering communication” is clear and expansive. It encompasses any broadcast, cable, or satellite communication that refers to a candidate for federal office and that is *458aired within 30 days of a federal primary election or 60 days of a federal general election in the jurisdiction in which that candidate is running for office. § 434(f)(3)(A).1
Appellee Wisconsin Right to Life, Inc. (WRTL), is a nonprofit, nonstock, ideological advocacy corporation recognized by the Internal Revenue Service as tax exempt under § 501(c)(4) of the Internal Revenue Code. On July 26, 2004, as part of what it calls a “grassroots lobbying campaign,” Brief for Appellee 8, WRTL began broadcasting a radio advertisement entitled ‘Wedding.” The transcript of ‘Wedding” reads as follows:
“ ‘PASTOR: And who gives this woman to be married to this man?
“‘BRIDE’S FATHER: Well, as father of the bride, I certainly could. But instead, I’d like to share a few tips on how to properly install, drywall. Now you put the drywall up ...
“ ‘VOICE-OVER: Sometimes it’s just not fair to delay an important decision.
“ ‘But in Washington it’s happening. A group of Senators is using the filibuster delay tactic to block federal *459judicial nominees from a simple “yes” or “no” vote. So qualified candidates don’t get a chance to serve.
“ ‘It’s politics at work, causing gridlock and backing up some of our courts to a state of emergency.
“‘Contact Senators Feingold and Kohl and tell them to oppose the filibuster.
“ ‘Visit: BeFair.org
“‘Paid for by Wisconsin Right to Life (befair.org), which is responsible for the content of this advertising and not authorized by any candidate or candidate’s committee.’” 466 F. Supp. 2d 195,198, n. 3 (DC 2006).
On the same day, WRTL aired a similar radio ad entitled “Loan.”2 It had also invested treasury funds in producing a television ad entitled “Waiting,”3 ****8which is similar in substance and format to “Wedding” and “Loan.”
*460WRTL planned on running “Wedding,” “Waiting,” and “Loan” throughout August 2004 and financing the ads with funds from its general treasury. It recognized, however, that as of August 15,30 days prior to the Wisconsin primary, the ads would be illegal “electioneering communieation[s]” under BCRA §203.
Believing that it nonetheless possessed a First Amendment right to broadcast these ads, WRTL filed suit against the Federal Election Commission (FEC) on July 28, 2004, seeking declaratory and injunctive relief before a three-judge District Court. See note following 2 U. S. C. §437h (2000 ed., Supp. IV); 28 U. S. C. §2284. WRTL alleged that BCRA’s prohibition on the use of corporate treasury funds for “electioneering communication^]” as defined in the Act is unconstitutional as applied to “Wedding,” “Loan,” and “Waiting,” as well as any materially similar ads it might seek to run in the future.
Just before the BCRA blackout period was to begin, the District Court denied a preliminary injunction, concluding that “the reasoning of the McConnell Court leaves no room for the kind of ‘as applied’ challenge WRTL propounds before us.” App. to Juris. Statement 52a. In response to this ruling, WRTL did not run its ads during the blackout period. The District Court subsequently dismissed WRTL’s complaint. See id., at 47a-48a (“WRTL’s ‘as-applied’ challenge to BCRA [§203] is foreclosed by the Supreme Court’s decision in McConnell”). On appeal, we vacated the District Court’s judgment, holding that McConnell “did not purport to resolve future as-applied challenges” to BCRA §203, and remanded “for the District Court to consider the merits of WRTL’s as-applied challenge in the first instance.” WRTL I, 546 U. S., at 412.
On remand, after allowing four Members of Congress to intervene as defendants, the three-judge District Court granted summary judgment for WRTL, holding BCRA §203 unconstitutional as applied to the three advertisements *461WRTL planned to run during the 2004 blackout period. The District Court first found adjudication of the dispute not barred by mootness because the controversy was “ ‘capable of repetition, yet evading review.’ ” 466 F. Supp. 2d, at 202. Turning to the merits, the court began by noting that under McConnell, BCRA could constitutionally proscribe “express advocacy” — defined as ads that expressly advocate the election or defeat of a candidate for federal office — and the “functional equivalent” of such advocacy. 466 F. Supp. 2d, at 204. Stating that it was limiting its inquiry to “language within the four corners” of the ads, id., at 207, the District Court concluded that the ads were not express advocacy or its functional equivalent, but instead “genuine issue ads,” id., at 205-208. Then, reaching a question “left open in McConnell,” the court held that no compelling interest justified BCRA’s regulation of genuine issue ads such as those WRTL sought to run. Id., at 208-210.
One judge dissented, contending that the majority’s “plain facial analysis of the text in WRTL’s 2004 advertisements” ignored “the context in which the text was developed.” Id., at 210 (opinion of Roberts, J.). In that judge’s view, a contextual analysis of the ads revealed “deep factual rifts between the parties concerning the purpose and intended effects of the ads” such that neither side was entitled to summary judgment. Id., at 210, 211.
The FEC and intervenors filed separate notices of appeal and jurisdictional statements. We consolidated the two appeals and set the matter for briefing and argument, postponing further consideration of jurisdiction to the hearing on the merits. 549 U. S. 1177 (2007).
II
Article Ill’s “case-or-controversy requirement subsists through all stages of federal judicial proceedings .... [I]t is not enough that a dispute was very much alive when suit was filed.” Lewis v. Continental Bank Corp., 494 U. S. 472, *462477 (1990). Based on these principles, the FEC argues (though the intervenors do not) that these cases are moot because the 2004 election has passed and WRTL “does not assert any continuing interest in running [its three] advertisements, nor does it identify any reason to believe that a significant dispute over Senate filibusters of judicial nominees will occur in the foreseeable future.” Brief for Appellant FEC 21.
As the District Court concluded, however, these cases fit comfortably within the established exception to mootness for disputes capable of repetition, yet evading review. See Los Angeles v. Lyons, 461 U. S. 95, 109 (1983); Southern Pacific Terminal Co. v. ICC, 219 U. S. 498, 515 (1911). The exception applies where “(1) the challenged action is in its duration too short to be fully litigated prior to cessation or expiration, and (2) there is a reasonable expectation that the same complaining party will be subject to the same action again.” Spencer v. Kemna, 523 U. S. 1, 17 (1998) (internal quotation marks and brackets omitted). Both circumstances are present here.
As the District Court found, it would be “entirely unreasonable ... to expect that [WRTL] could have obtained complete judicial review of its claims in time for it to air its ads” during the BCRA blackout periods. 466 F. Supp. 2d, at 202. The FEC contends that the 2-year window between elections provides ample time for parties to litigate their rights before each BCRA blackout period. But groups like WRTL cannot predict what issues will be matters of public concern during a future blackout period. In these cases, WRTL had no way of knowing well in advance that it would want to run ads on judicial filibusters during the BCRA blackout period. In any event, despite BCRA’s command that the cases be expedited “to the greatest possible extent,” § 403(a)(4), 116 Stat. 113, note following 2 U. S. C. §437h (2000 ed., Supp. IV), two BCRA blackout periods have come and gone during the pendency of this action. “[A] decision allowing the desired *463expenditures would be an empty gesture unless it afforded appellants sufficient opportunity prior to the election date to communicate their views effectively.” First Nat. Bank of Boston v. Bellotti, 435 U. S. 765, 774 (1978).
The second prong of the “capable of repetition” exception requires a “‘reasonable expectation’” or a “‘demonstrated probability ”• that “the same controversy will recur involving the same complaining party.” Murphy v. Hunt, 455 U. S. 478, 482 (1982) (per curiam). Our cases find the same controversy sufficiently likely to recur when a party has a reasonable expectation that it “will again be subjected to the alleged illegality,” Lyons, supra, at 109, or “will be subject to the threat of prosecution” under the challenged law, Bellotti, supra, at 774-775 (citing Weinstein v. Bradford, 423 U. S. 147, 149 (1975) (per curiam)). The FEC argues that in order to prove likely recurrence of the same controversy, WRTL must establish that it will run ads in the future sharing all “the characteristics that the district court deemed legally relevant.” Brief for Appellant FEC 23.
The FEC asks for too much. We have recognized that the “ ‘capable of repetition, yet evading review’ doctrine, in the context of election cases, is appropriate when there are ‘as applied’ challenges as well as in the more typical case involving only facial attacks.” Storer v. Brown, 415 U. S. 724, 737, n. 8 (1974). Requiring repetition of every “legally relevant” characteristic of an as-applied challenge — down to the last detail — would effectively overrule this statement by making this exception unavailable for virtually all as-applied challenges. History repeats itself, but not at the level of specificity demanded by the FEC. Here, WRTL credibly claimed that it planned on running “ ‘materially similar’ ” future targeted broadcast ads mentioning a candidate within the blackout period, 466 F. Supp. 2d, at 197, and there is no reason to believe that the FEC will “refrain from prosecuting violations” of BCRA, Bellotti, supra, at 775. Under the circumstances, particularly where WRTL sought another prelimi*464nary injunction based on an ad it planned to run during the 2006 blackout period, see 466 F. Supp. 2d, at 203, n. 15, we hold that there exists a reasonable expectation that the same controversy involving the same party will recur. We have jurisdiction to decide these cases.
Ill
WRTL rightly concedes that its ads are prohibited by BCRA §203. Each ad clearly identifies Senator Feingold, who was running (unopposed) in the Wisconsin Democratic primary on September 14, 2004, and each ad would have been “targeted to the relevant electorate,” see 2 U. S. C. § 434(f)(3)(C) (2000 ed., Supp. IV), during the BCRA blackout period. WRTL further concedes that its ads do not fit under any of BCRA’s exceptions to the term “electioneering communication.” See § 434(f)(3)(B). The only question, then, is whether it is consistent with the First Amendment for BCRA § 203 to prohibit WRTL from running these three ads.
A
Appellants contend that WRTL should be required to demonstrate that BCRA is unconstitutional as applied to the ads. Reply Brief for Appellant Sen. John McCain et al. in No. 06-970, p. 5, n. 4; Brief for Appellant FEC 34. After all, appellants reason, McConnell already held that BCRA § 203 was facially valid. These cases, however, present the separate question whether §203 may constitutionally be applied to these specific ads. Because BCRA §203 burdens political speech, it is subject to strict scrutiny. See McConnell, 540 U. S., at 205; Austin v. Michigan Chamber of Commerce, 494 U. S. 652, 658 (1990); MCFL, 479 U. S., at 252 (plurality opinion); Bellotti, supra, at 786; Buckley, 424 U. S., at 44-45. Under strict scrutiny, the Government must prove that applying BCRA to WRTL’s ads furthers a compelling interest and is narrowly tailored to achieve that interest. See Bellotti, supra, at 786 (“Especially where, as *465here, a prohibition is directed at speech itself, and the speech is intimately related to the process of governing, . . . ‘the burden is on the government to show the existence of [a compelling] interest’ ” (footnote omitted)).
The strict scrutiny analysis is, of course, informed by our precedents. This Court has already ruled that BCRA survives strict scrutiny to the extent it regulates express advocacy or its functional equivalent. McConnell, supra, at 206. So to the extent the ads in these cases fit this description, the FEC’s burden is not onerous; all it need do is point to McConnell and explain why it applies here. If, on the other hand, WRTL’s ads are not express advocacy or its equivalent, the Government’s task is more formidable. It must then demonstrate that banning such ads during the blackout periods is narrowly tailored to serve a compelling interest. No precedent of this Court has yet reached that conclusion.
B
The FEC, intervenors, and the dissent below contend that McConnell already established the constitutional test for determining if an ad is the functional equivalent of express advocacy: whether the ad is intended to influence elections and has that effect. See, e. g., 466 F. Supp. 2d, at 214 (opinion of Roberts, J.). Here is the relevant portion of our opinion in McConnell-.
“[P]laintiffs argue that the justifications that adequately support the regulation of express advocacy do not apply to significant quantities of speech encompassed by the definition of electioneering communications.
“This argument fails to the extent that the issue ads broadcast during the 30- and 60-day periods preceding federal primary and general elections are the functional equivalent of express advocacy. The justifications for the regulation of express advocacy apply equally to ads aired during those periods if the ads are intended to *466influence the voters’ decisions and have that effect.” 540 U. S., at 205-206.
WRTL and the District Court majority, on the other hand, claim that McConnell did not adopt any test as the standard for future as-applied challenges. We agree. McConnell’s analysis was grounded in the evidentiary record before the Court. Two key studies in the McConnell record constituted “the central piece of evidence marshaled by defenders of BCRA’s electioneering communication provisions in support of their constitutional validity.” McConnell v. FEC, 251 F. Supp. 2d 176, 307, 308 (DC 2003) (opinion of Henderson, J.) (internal quotation marks and brackets omitted). Those studies asked “student coders” to separate ads based on whether the students thought the “purpose” of the ad was “to provide information about or urge action on a bill or issue,” or “to generate support or opposition for a particular candidate.” Id., at 308-309 (internal quotation marks omitted; emphasis deleted); see Brief for Appellee 38. The studies concluded “‘that BCRA’s definition of Electioneering Communications accurately captures those ads that have the purpose or effect of supporting candidates for election to office.” Ibid, (emphasis in original).
When the McConnell Court considered the possible facial overbreadth of §203, it looked to the studies in the record analyzing ads broadcast during the blackout periods, and those studies had classified the ads in terms of intent and effect. The Court’s assessment was accordingly phrased in the same terms, which the Court regarded as sufficient to conclude, on the record before it, that the plaintiffs had not “carried their heavy burden of proving” that §203 was facially overbroad and could not be enforced in any circumstances. 540 U. S., at 207. The Court did not explain that it was adopting a particular test for determining what constituted the “functional equivalent” of express advocacy. The fact that the student coders who helped develop the evidentiary record before the Court in McConnell looked to intent and effect in doing so, and that the Court dealt with the *467record on that basis in deciding the facial overbreadth claim, neither compels nor warrants accepting that same standard as the constitutional test for separating, in an as-applied challenge, political speech protected under the First Amendment from that which may be banned.4
More importantly, this Court in Buckley had already rejected an intent-and-effeet test for distinguishing between discussions of issues and candidates. See 424 U. S., at 43-44. After noting the difficulty of distinguishing between discussion of issues on the one hand and advocacy of election or defeat of candidates on the other, the Buckley Court explained that analyzing the question in terms “‘of intent and of effect’” would afford “‘no security for free discussion.’” Id., at 43 (quoting Thomas v. Collins, 323 U. S. 516, 535 (1945)). It therefore rejected such an approach, and McConnell did not purport to overrule Buckley on this point— or even address what Buckley had to say on the subject.
For the reasons regarded as sufficient in Buckley, we decline to adopt a test for as-applied challenges turning on the speaker’s intent to affect an election. The test to distinguish constitutionally protected political speech from speech that BCRA may proscribe should provide a safe harbor for those who wish to exercise First Amendmént rights. The test should also “reflec[t] our ‘profound national commitment to the principle that debate on public issues should be uninhibited, robust, and wide-open.’” Buckley, supra, at 14 *468(quoting New York Times Co. v. Sullivan, 376 U. S. 254, 270 (1964)). A test turning on the intent of the speaker does not remotely fit the bill.
Far from serving the values the First Amendment is meant to protect, an intent-based test would chill core political speech by opening the door to a trial on every ad within the terms of §203, on the theory that the speaker actually intended to affect an election, no matter how compelling the indications that the ad concerned a pending legislative or policy issue. No reasonable speaker would choose to run an ad covered by BCRA if its only defense to a criminal prosecution would be that its motives were pure. An intent-based standard “blankets with uncertainty whatever may be said,” and “offers no security for free discussion.” Buckley, supra, at 43 (internal quotation marks omitted). The FEC does not disagree. In its brief filed in the first appeal in this litigation, it argued that a “constitutional standard that turned on the subjective sincerity of a speaker’s message would likely be incapable of workable application; at a minimum, it would invite costly, fact-dependent litigation.” Brief for Appellee in WRTL I, O. T. 2005, No. 04-1581, p. 39.5
A test focused on the speaker’s intent could lead to the bizarre result that identical ads aired at the same time could be protected speech for one speaker, while leading to criminal penalties for another. See M. Redish, Money Talks: Speech, Economic Power, and the Values of Democracy 91 (2001) (“[U]nder well-accepted First Amendment doctrine, a speaker’s motivation is entirely irrelevant to the question of constitutional protection”). “First Amendment freedoms *469need breathing space to survive.” NAACP v. Button, 371 U. S. 415, 433 (1963). An intent test provides none.
Buckley also explains the flaws of a test based on the actual effect speech will have on an election or on a particular segment of the target audience. Such a test “‘puts the speaker ... wholly at the mercy of the varied understanding of his hearers.’ ” 424 U. S., at 43. It would also typically lead to a burdensome, expert-driven inquiry, with an indeterminate result. Litigation on such a standard may or may not accurately predict electoral effects, but it will unquestionably chill a substantial amount of political speech.
C
“The freedom of speech... guaranteed by the Constitution embraces at the least the liberty to discuss publicly and truthfully all matters of public concern without previous restraint or fear of subsequent punishment.” Bellotti, 435 U. S., at 776 (internal quotation marks omitted). See Consolidated Edison Co. of N. Y. v. Public Serv. Comm’n of N. Y, 447 U. S. 530, 534 (1980). To safeguard this liberty, the proper standard for an as-applied challenge to BCRA § 203 must be objective, focusing on the substance of the communication rather than amorphous considerations of intent and effect. See Buckley, supra, at 43-44. It must entail minimal if any discovery, to allow parties to resolve disputes quickly without chilling speech through the threat of burdensome litigation. See Virginia v. Hicks, 539 U. S. 113, 119 (2003). And it must eschew “the open-ended rough-and-tumble of factors,” which “invit[es] complex argument in a trial court and a virtually inevitable appeal.” Jerome B. Grubart, Inc. v. Great Lakes Dredge & Dock Co., 513 U. S. 527, 547 (1995). In short, it must give the benefit of any doubt to protecting rather than stifling speech. See New York Times Co. v. Sullivan, supra, at 269-270.
In light of these considerations, a court should find that an ad is the functional equivalent of express advocacy only if *470the ad is susceptible of no reasonable interpretation other than as an appeal to vote for or against a specific candidate. Under this test, WRTL’s three ads are plainly not the functional equivalent of express advocacy. First, their content is consistent with that of a genuine issue ad: The ads focus on a legislative issue, take a position on the issue, exhort the public to adopt that position, and urge the public to contact public officials with respect to the matter. Second, their content lacks indicia of express advocacy: The ads do not mention an election, candidacy, political party, or challenger; and they do not take a position on a candidate’s character, qualifications, or fitness for office.
Despite these characteristics, appellants assert that the content of WRTL’s ads alone betrays their electioneering nature. Indeed, the FEC suggests that any ad covered by §203 that includes “an appeal to citizens to contact their elected representative” is the “functional equivalent” of an ad saying defeat or elect that candidate. Brief for Appellant FEC 31; see Brief for Appellant Sen. John McCain et al. in No. 06-970, pp. 21-23 (hereinafter McCain Brief). We do not agree. To take just one example, during a blackout period the House considered the proposed Universal National Service Act. See App. to Brief for American Center for Law and Justice et al. as Amici Curiae B-3. There would be no reason to regard an ad supporting or opposing that Act, and urging citizens to contact their Representative about it, as the equivalent of an ad saying vote for or against the Representative. Issue advocacy conveys information and educates. An issue ad’s impact on an election, if it exists at all, will come only after the voters hear the information and choose — uninvited by the ad — to factor it into their voting decisions.6
*471The FEC and intervenors try to turn this difference to their advantage, citing McConnell’s statements “that the most effective campaign ads, like the most effective commercials for products ..., avoid the [Buckley] magic words [expressly advocating the election or defeat of a candidate],” 540 U. S., at 127, and that advertisers “would seldom choose to use such words even if permitted,” id., at 193. See McCain Brief 19. An expert for the FEC in these cases relied on those observations to argue that WRTL’s ads are especially effective electioneering ads because they are “subtl[e],” focusing on issues rather than simply exhorting the electorate to vote against Senator Feingold. App. 56-57. Rephrased a bit, the argument perversely maintains that the less an issue ad resembles express advocacy, the more likely it is to be the functional equivalent of express advocacy. This “heads I win, tails you lose” approach cannot be correct. It would effectively eliminate First Amendment protection for genuine issue ads, contrary to our conclusion in WRTL I that as-applied challenges to §203 are available, and our assumption in McConnell that “the interests that justify the regulation of campaign speech might not apply to the regulation of genuine issue ads,” 540 U. S., at 206, n. 88. Under appel*472lants’ view, there can be no such thing as a genuine issue ad during the blackout period — it is simply a very effective electioneering ad.
Looking beyond the content of WRTL’s ads, the FEC and intervenors argue that several “contextual” factors prove that the ads are the equivalent of express advocacy. First, appellants cite evidence that during the same election cycle, WRTL and its Political Action Committee (PAC) actively opposed Senator Feingold’s reelection and identified filibusters as a campaign issue. This evidence goes to WRTL’s subjective intent in running the ads, and we have already explained that WRTL’s intent is irrelevant in an as-applied challenge. Evidence of this sort is therefore beside the point, as it should be — WRTL does not forfeit its right to speak on issues simply because in other aspects of its work it also opposes candidates who are involved with those issues.
Next, the FEC and intervenors seize on the timing of WRTL’s ads. They observe that the ads were to be aired near elections but not near actual Senate votes on judicial nominees, and that WRTL did not run the ads after the elections. To the extent this evidence goes to WRTL’s subjective intent, it is again irrelevant. To the extent it nonetheless suggests that the ads should be interpreted as express advocacy, it falls short. That the ads were run close to an election is unremarkable in a challenge like this. Every ad covered by BCRA §203 will by definition air just before a primary or general election. If this were enough to prove that an ad is the functional equivalent of express advocacy, then BCRA would be constitutional in all of its applications. This Court unanimously rejected this contention in WRTL /.
That the ads were run shortly after the Senate had recessed is likewise unpersuasive. Members of Congress often return to their districts during recess, precisely to determine the views of their constituents; an ad run at that time may succeed in getting more constituents to contact the Representative while he or she is back home. In any event, *473a group can certainly choose to run an issue ad to coincide with public interest rather than a floor vote. Finally, WRTL did not resume running its ads after the BCRA blackout period because, as it explains, the debate had changed. Brief for Appellee 16. The focus of the Senate was on whether a majority would vote to change the Senate rules to eliminate the filibuster — not whether individual Senators would continue filibustering. Given this change, WRTL’s decision not to continue running its ads after the blackout period does not support an inference that the ads were the functional equivalent of electioneering.
The last piece of contextual evidence the FEC and intervenors highlight is the ads’ “specific and repeated cross-reference” to a Web site. Reply Brief for Appellant FEC 15. In the middle of the Web site’s homepage, in large type, were the addresses, phone numbers, fax numbers, and e-mail addresses of Senators Feingold and Kohl. Wisconsinites who viewed “Wedding,” “Loan,” or ‘Waiting” and wished to contact their Senators — as the ads requested — would be able to obtain the pertinent contact information immediately upon visiting the Web site. This is fully consistent with viewing WRTL's ads as genuine issue ads. The Web site also stated both Wisconsin Senators’ positions on judicial filibusters, and allowed visitors to sign up for “e-alerts,” some of which contained exhortations to vote against Senator Feingold. These details lend the electioneering interpretation of the ads more credence, but again, WRTL’s participation in express advocacy in other aspects of its work is not a justification for censoring its issue-related speech. Any express advocacy on the Web site, already one step removed from the text of the ads themselves, certainly does not render an interpretation of the ads as genuine issue ads unreasonable.
Given the standard we have adopted for determining whether an ad is the “functional equivalent” of express advocacy, contextual factors of the sort invoked by appellants *474should seldom play a significant role in the inquiry. Courts need not ignore basic background information that may be necessary to put an ad in context — such as whether an ad “describes a legislative issue that is either currently the subject of legislative scrutiny or likely to be the subject of such scrutiny in the near future,” 466 F. Supp. 2d, at 207 — but the need to consider such background should not become an excuse for discovery or a broader inquiry of the sort we have just noted raises First Amendment concerns.
At best, appellants have shown what we have acknowledged at least since Buckley: that “the distinction between discussion of issues and candidates and advocacy of election or defeat of candidates may often dissolve in practical application.” 424 U. S., at 42. Under the test set forth above, that is not enough to establish that the ads can only reasonably be viewed as advocating or opposing a candidate in a federal election. “Freedom of discussion, if it would fulfill its historic function in this nation, must embrace all issues about which information is needed or appropriate to enable the members of society to cope with the exigencies of their period.” Thornhill v. Alabama, 310 U. S. 88, 102 (1940). Discussion of issues cannot be suppressed simply because the issues may also be pertinent in an election. Where the First Amendment is implicated, the tie goes to the speaker, not the censor.7
*475We confronted a similar issue in Ashcroft v. Free Speech Coalition, 535 U. S. 234 (2002), in which the Government argued that virtual images of child pornography were difficult to distinguish from real images. The Government’s solution was “to prohibit both kinds of images.” Id., at 254-255. We rejected the argument that “protected speech may be banned as a means to ban unprotected speech,” concluding that it “turns the First Amendment upside down.” Id., at 255. As we explained: “The Government may not suppress lawful speech as the means to suppress unlawful speech. Protected speech does not become unprotected merely because it resembles the latter. The Constitution requires the reverse.” Ibid.
*476Because WRTL’s ads may reasonably be interpreted as something other than as an appeal to vote for or against a specific candidate, we hold they are not the functional equivalent of express advocacy, and therefore fall outside the scope of McConnell's holding.8
IV
BCRA § 203 can be constitutionally applied to WRTL’s ads only if it is narrowly tailored to further a compelling interest. McConnell, 540 U. S., at 205; Bellotti, 435 U. S., at 786; Buckley, supra, at 44-45. This Court has never recognized a compelling interest in regulating ads, like WRTL’s, that are neither express advocacy nor its functional equivalent. The District Court below considered interests that might justify regulating WRTL’s ads here, and found none sufficiently *477compelling. 466 F. Supp. 2d, at 208-210. We reach the same conclusion.9
At the outset, we reject the contention that issue advocacy may be regulated because express election advocacy may be, and “the speech involved in so-called issue advocacy is [not] any more core political speech than are words of express advocacy.” McConnell, supra, at 205. This greater-includes-the-lesser approach is not how strict scrutiny works. A corporate ad expressing support for the local football team could not be regulated on the ground that such *478speech is less “core” than corporate speech about an election, which we have held may be restricted. A court applying strict scrutiny must ensure that a compelling interest supports each application of a statute restricting speech. That a compelling interest justifies restrictions on express advocacy tells us little about whether a compelling interest justifies restrictions on issue advocacy; the McConnell Court itself made just that point. See 540 U. S., at 206, n. 88. Such a greater-includes-the-lesser argument would dictate that virtually all corporate speech can be suppressed, since few kinds of speech can lay claim to being as central to the First Amendment as campaign speech. That conclusion is clearly foreclosed by our precedent. See, e. g., Bellotti, supra, at 776-777.
This Court has long recognized “the governmental interest in preventing corruption and the appearance of corruption” in election campaigns. Buckley, 424 U. S., at 45. This interest has been invoked as a reason for upholding contribution limits. As Buckley explained, “[t]o the extent that large contributions are given to secure a political quid pro quo from current and potential office holders, the integrity of our system of representative democracy is undermined.” Id., at 26-27. We have suggested that this interest might also justify limits on electioneering expenditures because it may be that, in some circumstances, “large independent expenditures pose the same dangers of actual or apparent quid pro quo arrangements as do large contributions.” Id., at 45.
McConnell arguably applied this interest — which this Court had only assumed could justify regulation of express advocacy — to ads that were the “functional equivalent” of express advocacy. See 540 U. S., at 204-206. But to justify regulation of WRTL’s ads, this interest must be stretched yet another step to ads that are not the functional equivalent of express advocacy. Enough is enough. Issue ads like WRTL’s are by no means equivalent to contributions, and *479the quid-pro-quo corruption interest cannot justify regulating them. To equate WRTL’s ads with contributions is to ignore their value as political speech.
Appellants argue that an expansive definition of “functional equivalent” is needed to ensure that issue advocacy does not circumvent the rule against express advocacy, which in turn helps protect against circumvention of the rule against contributions. Cf. McConnell, supra, at 205 (“[R]ecent cases have recognized that certain restrictions on corporate electoral involvement permissibly hedge against circumvention of [valid] contribution limits” (internal quotation marks omitted; brackets in original)). But such a prophylaxis-upon-prophylaxis approach to regulating expression is not consistent with strict scrutiny. “[T]he desire for a bright-line rule ... hardly constitutes the compelling state interest necessary to justify any infringement on First Amendment freedom.” MCFL, 479 U. S., at 263. See Free Speech Coalition, 535 U. S., at 255 (“The Government may. not suppress lawful speech as the means to suppress unlawful speech”); Buckley, supra, at 44 (expenditure limitations “cannot be sustained simply by invoking the interest in maximizing the effectiveness of the less intrusive contribution limitations”).
A second possible compelling interest recognized by this Court lies in addressing a “different type of corruption in the political arena: the corrosive and distorting effects of immense aggregations of wealth that are accumulated with the help of the corporate form and that have little or no correlation to the public’s support for the corporation’s political ideas.” Austin, 494 U. S., at 660. Austin invoked this interest to uphold a state statute making it a felony for corporations to use treasury funds for independent expenditures on express election advocacy. Id., at 654-655. McConnell also relied on this interest in upholding regulation not just of express advocacy, but also its “functional equivalent.” 540 U. S., at 205-206.
*480These cases did not suggest, however, that the interest in combating “a different type of corruption” extended beyond campaign speech. Quite the contrary. Two of the Justices who joined the 6-to-3 majority in Austin relied, in upholding the constitutionality of the ban on campaign speech, on the fact that corporations retained freedom to speak on issues as distinct from election campaigns. See 494 U. S., at 675-678 (Brennan, J., concurring) (describing fact that campaign speech ban “does not regulate corporate expenditures in referenda or other corporate expression” as “reflecting] the requirements of our decisions”); id., at 678 (Stevens, J., concurring) (“[T]here is a vast difference between lobbying and debating public issues on the one hand, and political campaigns for election to public office on the other”). The McConnell Court similarly was willing to “assume that the interests that justify the regulation of campaign speech might not apply to the regulation of genuine issue ads.” 540 U. S., at 206, n. 88. And our decision in WRTL I reinforced the validity of that assumption by holding that BCRA §203 is susceptible to as-applied challenges. 546 U. S., at 412.
Accepting the notion that a ban on campaign speech could also embrace issue advocacy would call into question our holding in Bellotti that the corporate identity of a speaker does not strip corporations of all free speech rights. 435 U. S., at 778. It would be a constitutional “bait and switch” to conclude that corporate campaign speech may be banned in part because corporate issue advocacy is not, and then assert that corporate issue advocacy may be banned as well, pursuant to the same asserted compelling interest, through á broad conception of what constitutes the functional equivalent of campaign speech, or by relying on the inability to distinguish campaign speech from issue advocacy.
The FEC and intervenors do not argue that the Austin interest justifies regulating genuine issue ads. Instead, they focus on establishing that WRTL’s ads are the functional equivalent of express advocacy — a contention we have *481already rejected. We hold that the interest recognized in Austin as justifying regulation of corporate campaign speech and extended in McConnell to the functional equivalent of such speech has no application to issue advocacy of the sort engaged in by WRTL.10
Because WRTL’s ads are not express advocacy or its functional equivalent, and because appellants identify no interest sufficiently compelling to justify burdening WRTL’s speech, we hold that BCRA §203 is unconstitutional as applied to WRTL’s “Wedding,” “Loan,” and “Waiting” ads.
* * *
These cases are about political speech. The importance of the cases to speech and debate on public policy issues is reflected in the number of diverse organizations that have joined in supporting WRTL before this Court: the American Civil Liberties Union, the National Rifle Association, the American Federation of Labor and Congress of Industrial Organizations, the Chamber of Commerce of the United States of America, Focus on the Family, the Coalition of Public Charities, the Cato Institute, and many others.
Yet, as is often the case in this Court’s First Amendment opinions, we have gotten this far in the analysis without *482quoting the Amendment itself: “Congress shall make no law ... abridging the freedom of speech.” The Framers’ actual words put these cases in proper perspective. Our jurisprudence over the past 216 years has rejected an absolutist interpretation of those words, but when it comes to drawing difficult lines in the area of pure political speech — between what is protected and what the Government may ban — it is worth recalling the language we are applying. McConnell held that express advocacy of a candidate or his opponent by a corporation shortly before an election may be prohibited, along with the functional equivalent of such express advocacy. We have no occasion to revisit that determination today. But when it comes to defining what speech qualifies as the functional equivalent of express advocacy subject to such a ban — the issue we do have to decide — we give the benefit of the doubt to speech, not censorship. The First Amendment’s command that “Congress shall make no law... abridging the freedom of speech” demands at least that.
The judgment of the United States District Court for the District of Columbia is affirmed.
It is so ordered.
Subparagraph (A) provides:
“(i) The term ‘electioneering communication’ means any broadcast, cable, or satellite communication which—
“(I) refers to a clearly identified candidate for Federal office;
“(II) is made within—
“(aa) 60 days before a general, special, or runoff election for the office sought by the candidate; or
“(bb) 30 days before a primary or preference election, or a convention or caucus of a political party that has authority to nominate a candidate, for the office sought by the candidate; and
“(III) in the case of a communication which refers to a candidate for an office other than President or Vice President, is targeted to the relevant electorate.” 2 U. S. C. § 434(f)(3)(A) (2000 ed., Supp. IV).
Subparagraph (B) defines exceptions to “electioneering communication” not relevant to this litigation. Subparagraph (C) defines the term “targeted to the relevant electorate.”
The radio script for “Loan” differs from “Wedding” only in its lead-in. “Loan” begins:
“ ‘LOAN OFFICER: Welcome Mr. and Mrs. Shulman. We’ve reviewed your loan application, along with your credit report, the appraisal on the house, the inspections, and well. ..
“ ‘COUPLE: Yes, yes ... we’re listening.
“ ‘OFFICER: Well, it all reminds me of a time I went fishing with my father. We were on the Wolf River Waupaca ...
“‘VOICE-OVER: Sometimes it’s just not fair to delay an important decision.
“ ‘But in Washington it’s happening____’ ” 466 F. Supp. 2d, at 198, n. 4.
The remainder of the script is identical to “Wedding.”
In “Waiting,” the images on the television ad depict a “‘middle-aged man being as productive as possible while his professional life is in limbo.’ ” Id., at 198, n. 5. The man reads the morning paper, polishes his shoes, scans through his Rolodex, and does other similar activities. The television script for this ad reads:
“ ‘VOICE-OVER: There are a lot of judicial nominees out there who can’t go to work. Their careers are put on hold because a group of Senators is filibustering — blocking qualified nominees from a simple “yes” or “no” vote.
“ ‘It’s politics at work and it’s causing gridlock.... ’ ” Ibid.
The remainder of the script is virtually identical to “Wedding.”
This is particularly true given that the methodology, data, and conclusions of the two studies were the subject of serious dispute among the District Court judges. Compare McConnell v. FEC, 251 F. Supp. 2d 176, 307-312 (DC 2003) (opinion of Henderson, J.) (stating that the studies were flawed and of limited evidentiary value), with id., at 585,583-588 (opinion of Kollar-Kotelly, J.) (finding the studies generally credible, but stating that “I am troubled by the fact that coders in both studies were asked questions regarding their own perceptions of the advertisements’ purposes, and that [some of] these perceptions were later recoded” by study supervisors). Nothing in this Court’s opinion in McConnell suggests it was resolving the sharp disagreements about the evidentiary record in this respect.
Consider what happened in these cases. The District Court permitted extensive discovery on the assumption that WRTL’s intent was relevant. As a result, the defendants deposed WRTL’s executive director, its legislative director, its political action committee director, its lead communications consultant, and one of its fundraisers. WRTL also had to turn over many documents related to its operations, plans, and finances. Such litigation constitutes a severe burden on political speech.
For these reasons, we cannot agree with Justice Souter’s assertion that “anyone who heard the Feingold ads . . . would know that WRTL’s message was to vote against Feingold.” Post, at 525. The dissent supports this assertion by likening WRTL’s ads to the “Jane Doe” example *471identified in McConnell v. Federal Election Comm’n, 540 U. S. 93 (2003). But that ad “condemned Jane Doe’s record on a particular issue.” Post, at 525 (internal quotation marks omitted). WRTL’s ads do not do so; they instead take a position on the filibuster issue and exhort constituents to contact Senators Feingold and Kohl to advance that position. Indeed, one would not even know from the ads whether Senator Feingold supported or opposed filibusters. Justice Souter is confident Wisconsinites independently knew Senator Feingold’s position on filibusters, but we think that confidence misplaced. A prominent study found, for example, that during the 2000 election cycle, 85 percent of respondents to a survey were not even able to name at least one candidate for the House of Representatives in their own district. See Inter-university Consortium for Political and Social Research, American National Election Study, 2000: Pre- and Post-Election Survey 243 (N. Burns et al. eds. 2002), online at http:// www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03131.xml (as visited June 22, 2007, and available in Clerk of Court’s case file).
Justice Scalia thinks our test impermissibly vague. See post, at 492-494 (opinion concurring in part and concurring in judgment). As should be evident, we agree with Justice Scalia on the imperative for clarity in this area; that is why our test affords protection unless an ad is susceptible of no reasonable interpretation other than as an appeal to vote for or against a specific candidate. It is why we emphasize that (1) there can be no free-ranging intent-and-effect test; (2) there generally should be no discovery or inquiry into the sort of “contextual” factors highlighted by the FEC and intervenors; (3) discussion of issues cannot be banned merely because the issues might be relevant to an election; and (4) in a debatable case, the tie is resolved in favor of protecting speech. And keep in mind this test is only triggered if the speech meets the bright-line requirements of BCRA §203 in the first place. Justice *475Scalia’s criticism of our test is all the more confusing because he accepts WRTL’s proposed three-prong test as “clear.” Post, at 498. We do not think our test any vaguer than WRTL’s, and it is more protective of political speech.
Justice Scalia also asserts that our test conflicts with Buckley v. Valeo, 424 U. S. 1 (1976) (per curiam). Post, at 495-497. The Buckley Court confronted a statute restricting “any expenditure ... relative to a clearly identified candidate.” 424 U. S., at 42 (internal quotation marks omitted). To avoid vagueness concerns, this Court first narrowed the statute to cover only expenditures expressly “advocating the election or defeat of a candidate” — using the so-called “magic words” of express advocacy. Ibid, (internal quotation marks omitted). The Court then proceeded to strike down the newly narrowed statute under strict scrutiny on the ground that its reach was not broad enough. Id., at 44. From this, Justice Scaua concludes that “[i]f a permissible test short of the magic-words test existed, Buckley would surely have adopted it.” Post, at 495. We are not so sure. The question in Buckley was how a particular statutory provision could be construed to avoid vagueness concerns, not what the constitutional standard for clarity was in the abstract, divorced, from specific statutory language. Buckley’s intermediate step of statutory construction on the way to its constitutional holding does not dictate a constitutional test. The Buckley Court’s “express advocacy restriction was an endpoint of statutory interpretation, not a first principle of constitutional law.” McConnell, 540 U. S., at 190. And despite Justice Scalia’s claim to the contrary, our citation of Buckley along with other decisions in rejecting an intent-and-effect test does not force us to adopt (or reject) Buckley's statutory construction as a constitutional test.
Nothing in McConnell’s statement that the “vast majority” of issue ads broadcast in the periods preceding federal elections had an “electioneering purpose” forecloses this conclusion. 540 U. S., at 206. Courts do not resolve unspecified as-applied challenges in the course of resolving a facial attack, so McConnell could not have settled the issue we address today. See Members of City Council of Los Angeles v. Taxpayers for Vincent, 466 U. S. 789, 803, n. 22 (1984) (“The fact that [a law] is capable of valid applications does not necessarily mean that it is valid as applied to these’ litigants”). Indeed, WRTL I, 546 U. S. 410, 412, confirmed as much. By the same token, in deciding this as-applied challenge, we have no occasion to revisit McConnell’s conclusion that the statute is not facially overbroad.
The “vast majority” language, moreover, is beside the point. The McConnell Court did not find that a “vast majority” of the issue ads considered were the functional equivalent of direct advocacy. Rather, it found that such ads had an “electioneering purpose.” For the reasons we have explained, “pin-pose” is not the appropriate test for distinguishing between genuine issue ads and the functional equivalent of express campaign advocacy. See supra, at 468-469. In addition, the “vast majority” statement was not necessary to the Court’s facial holding in McConnell. The standard required for a statute to survive an overbreadth challenge is not that the “vast majority” of a statute’s applications be legitimate. “[B]road language . . . unnecessary to the Court’s decision . .. cannot be considered binding authority.” Kastigar v. United States, 406 U. S. 441, 454-455 (1972).
The dissent stresses a number of points that, while not central to our decision, nevertheless merit a response. First, the dissent overstates its case when it asserts that the “PAC alternative” gives corporations a constitutionally sufficient outlet to speak. See post, at 532. PACs impose well-documented and onerous burdens, particularly on small nonprofits. See MCFL, 479 U. S. 238, 253-255 (1986) (plurality opinion). McConnell did conclude that segregated funds “pr<md[e] corporations and unions with a constitutionally sufficient opportunity to engage in express advocacy” and its functional equivalent, 540 U. S., at 203, but that holding did not extend beyond functional equivalents — and if it did, the PAC option would justify regulation of all corporate speech, a proposition we have rejected, see Bellotti, 435 U. S., at 777-778. Second, the response that a speaker should just take out a newspaper ad, or use a Web site, rather than complain that it cannot speak through a broadcast communication, see post, at 521, 534, is too glib. Even assuming for the sake of argument that the possibility of using a different medium of communication has relevance in determining the permissibility of a limitation on speech, newspaper ads and Web sites are not reasonable alternatives to broadcast speech in terms of impact and effectiveness. See McConnell v. FEC, 251 F. Supp. 2d, at 569-573, 646 (Kollar-Kotelly, J.). Third, we disagree with the dissent’s view that corporations can still speak by changing what they say to avoid mentioning candidates, post, at 532-533. That argument is akin to telling Cohen that he cannot wear his jacket because he is free to wear one that says “I disagree with the draft,” cf. Cohen v. California, 403 U. S. 15 (1971), or telling 44 Liquormart that it can advertise so long as it avoids mentioning prices, cf. 44 Liquormart, Inc. v. Rhode Island, 517 U. S. 484 (1996). Such notions run afoul of “the fundamental rule of protection under the First Amendment, that a speaker has the autonomy to choose the content of his own message.” Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U. S. 557, 573 (1995).
The interest recognized in Austin v. Michigan Chamber of Commerce, 494 U. S. 652 (1990), stems from a concern that “ ‘[t]he resources in the treasury of a business corporation ... are not an indication of popular support for the corporation’s political ideas.’” Id., at 659 (alteration in original). Some of WRTL’s amici contend that this interest is not implicated here because of WRTL’s status as a nonprofit advocacy organization. They assert that “[s]peech by nonprofit advocacy groups on behalf of their members does not ‘corrupt’ candidates or ‘distort’ the political marketplace,” and that “Monprofit advocacy groups funded by individuals are readily distinguished from for-profit corporations funded by general treasuries.” Brief for Family Research Council et al. as Amici Curiae 3, 4. Cf. MCFL, 479 U. S., at 264. We do not pass on this argument in this as-applied challenge because WRTL’s funds for its ads were not derived solely from individual contributions. See Brief for Appellant FEC 11.