Full opinion text
Chief Justice Roberts delivered the opinion of the Court. The State of Washington allows its citizens to challenge state laws by referendum. Roughly four percent of Washington voters must sign a petition to place such a referendum on the ballot. That petition, which by law must include the names and addresses of the signers, is then submitted to the government for verification and canvassing, to ensure that only lawful signatures are counted. The Washington Public Records Act (PRA) authorizes private parties to obtain copies of government documents, and the State construes the PRA to cover submitted referendum petitions. This case arises out of a state law extending certain benefits to same-sex couples, and a corresponding referendum petition to put that law to a popular vote. Respondent intervenors invoked the PRA to obtain copies of the petition, with the names and addresses of the signers. Certain petition signers and the petition sponsor objected, arguing that such public disclosure would violate their rights under the First Amendment. The course of this litigation, however, has framed the legal question before us more broadly. The issue at this stage of the case is not whether disclosure of this particular petition would violate the First Amendment, but whether disclosure of referendum petitions in general would do so. We conclude that such disclosure does not as a general matter violate the First Amendment, and we therefore affirm the judgment of the Court of Appeals. We leave it to the lower courts to consider in the first instance the signers’ more focused claim concerning disclosure of the information on this particular petition, which is pending before the District Court. I The Washington Constitution reserves to the people the power to reject any bill, with a few limited exceptions not relevant here, through the referendum process. Wash. Const., Art. II, § 1(b). To initiate a referendum, proponents must file a petition with the secretary of state that contains valid signatures of registered Washington voters equal to or exceeding four percent of the votes cast for the office of Governor at the last gubernatorial election. §§ 1(b), (d). A valid submission requires not only a signature, but also the signer’s address and the county in which he is registered to vote. Wash. Rev. Code §29A.72.130 (2008). In May 2009, Washington Governor Christine Gregoire signed into law Senate Bill 5688, which “expanded] the rights and responsibilities” of state-registered domestic partners, including same-sex domestic partners. 586 F. 3d 671, 675 (CA9 2009). That same month, Protect Marriage Washington, one of the petitioners here, was organized as a “State Political Committee” for the purpose of collecting the petition signatures necessary to place a referendum on the ballot, which would give the voters themselves an opportunity to vote on SB 5688. App. 8-9. If the referendum made it onto the ballot, Protect Marriage Washington planned to encourage voters to reject SB 5688. Id., at 7, 9. On July 25, 2009, Protect Marriage Washington submitted to the secretary of state a petition containing over 137,000 signatures. See 586 F. 3d, at 675; Brief for Respondent Washington Families Standing Together 6. The secretary of state then began the verification and canvassing process, as required by Washington law, to ensure that only legal signatures were counted. Wash. Rev. Code §29A.72.230. Some 120,000 valid signatures were required to place the referendum on the ballot. Sam Reed, Washington Secretary of State, Certification of Referendum 71 (Sept. 2, 2009). The secretary of state determined that the petition contained a sufficient number of valid signatures, and the referendum (R-71) appeared on the November 2009 ballot. The voters approved SB 5688 by a margin of 53 percent to 47 percent. The PRA, Wash. Rev. Code § 42.56.001 et seq. (2008), makes all “public records” available for public inspection and copying. §42.56.070(1). The Act defines “[pjublic record” as “any writing containing information relating to the conduct of government or the performance of any- governmental or proprietary function prepared, owned, used, or retained by any state or local agency.” §42.56.010(2). Washington takes the position that referendum petitions are “public records.” Brief for Respondent Reed 5. By August 20, 2009, the secretary had received requests for copies of the R-71 petition from an individual and four entities, including Washington Coalition for Open Government (WCOG) and Washington Families Standing Together (WFST), two of the respondents here. 586 F. 3d, at 675. Two entities, WhoSigned.org and KnowThyNeighbor.org, issued a joint press release stating their intention.to post the names of the R-71 petition signers online, in a searchable format. See App. 11; 586 F. 3d, at 675. The referendum petition sponsor and certain signers filed a complaint and a motion for a preliminary injunction in the United States District Court for the Western District of Washington, seeking to enjoin the secretary of state from publicly releasing any documents that would reveal the names and contact information of the R-71 petition signers. App. 4. Count I of the complaint alleges that “[t]he Public Records Act is unconstitutional as applied to referendum petitions.” Id., at 16. Count II of the complaint alleges that “the Public Records Act is unconstitutional as applied to the Referendum 71 petition because there is a reasonable probability that the signatories of the Referendum 71 petition will be subjected to threats, harassment, and reprisals.” Id., at 17. Determining that the PRA burdened core political speech, the District Court held that plaintiffs were likely to succeed on the merits of Count I and granted them a preliminary injunction on that count, enjoining release of the information on the petition. 661 F. Supp. 2d 1194, 1205-1206 (WD Wash. 2009). The United States Court of Appeals for the Ninth Circuit reversed. Reviewing only Count I of the complaint, the Court of Appeals held that plaintiffs were unlikely to succeed on their claim that the PRA is unconstitutional as applied to referendum petitions generally. It therefore reversed the District Court’s grant of the preliminary injunction. 586 F. 3d, at 681. We granted certiorari. 558 U. S. 1142 (2010). II It is important at the outset to define the scope of the challenge before us. As noted, Count I of the complaint contends that the PRA “violates the First Amendment as applied to referendum petitions.” App. 16. Count II asserts that the PRA “is unconstitutional as applied to the Referendum 71 petition.” Id., at 17. The District Court decision was based solely on Count I; the Court of Appeals decision reversing the District Court was similarly limited. 586 F. 3d, at 676, n. 6. Neither court addressed Count II. The parties disagree about whether Count I is properly viewed as a facial or as-applied challenge. Compare Reply Brief for Petitioners 8 (“Count I expressly made an as-applied challenge”) with Brief for Respondent Reed 1 (“This is a facial challenge to Washington’s Public Records Act”). It obviously has characteristics of both: The claim is “as applied” in the sense that it does not seek to strike the PRA in all its applications, but only to the extent it covers referendum petitions. The claim is “facial” in that it is not limited to plaintiffs’ particular case, but challenges application of the law more broadly to all referendum petitions. The label is not what matters. The important point is that plaintiffs’ claim and the relief that would follow — an injunction barring the secretary of state “from making referendum petitions available to the public,” App. 16 (Complaint Count I) — reach beyond the particular circumstances of these plaintiffs. They must therefore satisfy our standards for a facial challenge to the extent of that reach. See United States v. Stevens, 559 U. S. 460, 472-473 (2010). III A The compelled disclosure of signatory information on referendum petitions is subject to review under the First Amendment. An individual expresses a view on a political matter when he signs a petition under Washington’s referendum procedure. In most cases, the individual’s signature will express the view that the law subject to the petition should be overturned. Even if the signer is agnostic as to the merits of the underlying law, his signature still expresses the political view that the question should be considered “by the whole electorate.” Meyer v. Grant, 486 U. S. 414, 421 (1988). In either case, the expression of a political view implicates a First Amendment right. The State, having “cho[sen] to tap the energy and the legitimizing power of the democratic process,. . . must accord the participants in that process the First Amendment rights that attach to their roles.” Republican Party of Minn. v. White, 536 U. S. 765, 788 (2002) (internal quotation marks and ellipsis omitted). Respondents counter that signing a petition is a legally operative legislative act and therefore “does not involve any significant expressive element.” Brief for Respondent Reed 31. It is true that signing a referendum petition may ultimately have the legal consequence of requiring the secretary of state to place the referendum on the ballot. But we do not see how adding such legal effect to an expressive activity somehow deprives that activity of its expressive component, taking it outside the scope of the First Amendment. Respondents themselves implicitly recognize that the signature expresses a particular viewpoint, arguing that one purpose served by disclosure is to allow the public to engage signers in a debate on the merits of the underlying law. See, e. g., id., at 45; Brief for Respondent WCOG 49; Brief for Respondent WFST 58. Petition signing remains expressive even when it has legal effect in the electoral process. But that is not to say that the electoral context is irrelevant to the nature of our First Amendment review. We allow States significant flexibility in implementing their own voting systems. See Burdick v. Takushi, 504 U. S. 428, 433-434 (1992). To the extent a regulation concerns the legal effect of a particular activity in that process, the government will be afforded substantial latitude to enforce that regulation. Also pertinent to our analysis is the fact that the PRA is not a prohibition on speech, but instead a disclosure requirement. “[Disclosure requirements may burden the ability to speak, but they . . . do not prevent anyone from speaking.” Citizens United v. Federal Election Comm’n, 558 U. S. 310, 366 (2010) (intern We have a series of precedents considering First Amendment challenges to disclosure requirements in the electoral context. These precedents have reviewed such challenges under what has been termed “exacting scrutiny.” See, e. g., Buckley v. Valeo, 424 U. S. 1, 64 (1976) (per curiam) (“Since NAACP v. Alabama [ex rel. Patterson, 357 U. S. 449 (1958),] we have required that the subordinating interests of the State [offered to justify compelled disclosure] survive exacting scrutiny”); Citizens United, supra, at 366 (“The Court has subjected [disclosure] requirements to ‘exacting scrutiny’” (quoting Buckley, supra, at 64)); Davis v. Federal Election Comm’n, 554 U. S. 724, 744 (2008) (governmental interest in disclosure “‘must survive exacting scrutiny’” (quoting Buckley, supra, at 64)); Buckley v. American Constitutional Law Foundation, Inc., 525 U. S. 182, 204 (1999) (ACLF) (finding that disclosure rules “fail[ed] exacting scrutiny” (internal quotation marks omitted)). That standard “requires a ‘substantial relation’ between the disclosure requirement and a ‘sufficiently important’ governmental interest.” Citizens United, supra, at 366-367 (quoting Buckley, supra, at 64, 66). To withstand this scrutiny, “the strength of the governmental interest must reflect the seriousness of the actual burden on First Amendment rights.” Davis, supra, at 744 (citing Buckley, supra, at 68, 71). B Respondents assert two interests to justify the burdens of compelled disclosure under the PRA on First Amendment rights: (1) preserving the integrity of the electoral process by combating fraud, detecting invalid signatures, and fostering government transparency and accountability; and (2) providing information to the electorate about who supports the petition. See, e. g., Brief for Respondent Reed 39-42, 44-45. Because we determine that the State’s interest in preserving the integrity of the electoral process suffices to defeat the argument that the PRA is unconstitutional with respect to referendum petitions in general, we need not, and do not, address the State’s “informational” interest. The State’s interest in preserving the integrity of the electoral process is undoubtedly important. “States allowing ballot initiatives have considerable leeway to protect the integrity and reliability of the initiative process, as they have with respect to election processes generally.” ACLF, supra, at 191. The State’s interest is particularly strong with respect to efforts to root out fraud, which not only may produce fraudulent outcomes, but has a systemic effect as well: It “drives honest citizens out of the democratic process and breeds distrust of our government.” Purcell v. Gonzalez, 549 U. S. 1, 4 (2006) (per curiam)-, see also Crawford v. Marion County Election Bd., 553 U. S. 181, 196 (2008) (opinion of Stevens, J.). The threat of fraud in this context is not merely hypothetical; respondents and their amici cite a number of cases of petition-related fraud across the country to support the point. See Brief for Respondent Reed 43; Brief for State of Ohio et al. as Amici Curiae 22-24. But the State’s interest in preserving electoral integrity is not limited to combating fraud. That interest extends to efforts to ferret out invalid signatures caused not by fraud but by simple mistake, such as duplicate signatures or signatures of individuals who are not registered to vote in the State. See Brief for Respondent Reed 42. That interest also extends more generally to promoting transparency and accountability in the electoral process, which the State argues is “essential to the proper functioning of a democracy.” Id., at 39. Plaintiffs contend that the disclosure requirements of the PRA are not “sufficiently related” to the interest of protecting the integrity of the electoral process. Brief for Petitioners 51. They argue that disclosure is not necessary because the secretary of state is already charged with verifying and canvassing the names on a petition, advocates and opponents of a measure can observe that process, and any citizen can challenge the secretary’s actions in court. See Wash. Rev. Code §§29A.72.230, 29A.72.240. They also stress that existing criminal penalties reduce the danger of fraud in the petition process. See Brief for Petitioners 50; §§29A.84.210, 29A.84.230, 29A.84.250. But the secretary’s verification and canvassing will not catch all invalid signatures: The job is large and difficult (the secretary ordinarily checks “only 3 to 5% of signatures,” Brief for Respondent WFST 54), and the secretary can make mistakes, too, see Brief for Respondent Reed 42. Public disclosure can help cure the inadequacies of the verification and canvassing process. Disclosure also helps prevent certain types of petition fraud otherwise difficult to detect, such as outright forgery and “bait and switch” fraud, in which an individual signs the petition based on a misrepresentation of the underlying issue. See Brief for Respondent WFST 9-11, 53-54; cf. Brief for Massachusetts Gay and Lesbian Political Caucus et al. as Amici Curiae 18-22 (detailing “bait and switch” fraud in a petition drive in Massachusetts). The signer is in the best position to detect these types of fraud, and public disclosure can bring the issue to the signer’s attention. Public disclosure thus helps ensure that the only signatures counted are those that should be, and that the only referenda placed on the ballot are those that garner enough valid signatures. Public disclosure also promotes transparency and accountability in the electoral process to an extent other measures cannot. In light of the foregoing, we reject plaintiffs’ argument and conclude that public disclosure of referendum petitions in general is substantially related to the important interest of preserving the integrity of the electoral process. C Plaintiffs’ more significant objection is that “the strength of the governmental interest” does not “reflect the seriousness of the actual burden on First Amendment rights.” Davis, 554 U. S., at 744 (citing Buckley, 424 U. S., at 68, 71); see, e. g., Brief for Petitioners 12-13, 30. According to plaintiffs, the objective of those seeking disclosure of the R-71 petition is not to prevent fraud, but to publicly identify those who had validly signed and to broadcast the signers’ political views on the subject of the petition. Plaintiffs allege, for example, that several groups plan to post the petitions in searchable form on the Internet, and then encourage other citizens to seek out the R-71 signers. See App. 11; Brief for Petitioners 8, 46-47. Plaintiffs explain that once on the Internet, the petition signers’ names.and addresses “can be combined with publicly available phone numbers and maps,” in what will effectively become a blueprint for harassment and intimidation. Id., at 46. To support their claim that they will be subject to reprisals, plaintiffs cite examples from the history of a similar proposition in California, see, e. g., id., at 2-6, 31-32, and from the experience of one of the petition sponsors in this case, see App. 9. In related contexts, we have explained that those resisting disclosure can prevail under the First Amendment if they can show “a reasonable probability that the compelled disclosure [of personal information] will subject them to threats, harassment, or reprisals from either Government officials or private parties.” Buckley, supra, at 74; see also Citizens United, 558 U. S., at 367. The question before us, however, is not whether PRA disclosure violates the First Amendment with respect to those who signed the R-71 petition, or other particularly controversial petitions. The question instead is whether such disclosure in general violates the First Amendment rights of those who sign referendum petitions. The problem for plaintiffs is that their argument rests almost entirely on the specific harm they say would attend disclosure of the information on the R-71 petition, or on similarly controversial ones. See, e. g., Brief for Petitioners 10, 26-29, 46, 56. But typical referendum petitions “concern tax policy, revenue, budget, or other state law issues.” Brief for Respondent WFST 36 (listing referenda); see also App. 26 (stating that in recent years the State has received PRA requests for petitions supporting initiatives concerning limiting motor vehicle charges; government regulation of private property; energy resource use by certain electric utilities; long-term care services for the elderly and persons with disabilities; and state, county, and city revenue); id., at 26-27 (stating that in the past 20 years, referendum measures that have qualified for the ballot in the State concerned land-use regulation; unemployment insurance; charter public schools; and insurance coverage and benefits). Voters care about such issues, some quite deeply — but there is no reason to assume that any burdens imposed by disclosure of typical referendum petitions would be remotely like the burdens plaintiffs fear in this case. Plaintiffs have offered little in response. They have provided us scant evidence or argument beyond the burdens they assert disclosure would impose on R-71 petition signers or the signers of other similarly controversial petitions. Indeed, what little plaintiffs do offer with respect to typical petitions in Washington hurts, not helps: Several other petitions in the State “have been subject to release in recent years,” plaintiffs tell us, Brief for Petitioners 50, but apparently that release has come without incident. Cf. Citizens United, supra, at 370 (“Citizens United has been disclosing its donors for years and has identified no instance of harassment or retaliation”). Faced with the State’s unrebutted arguments that only modest burdens attend the disclosure of a typical petition, we must reject plaintiffs’ broad challenge to the PR A. In doing so, we note — as we have in other election law disclosure cases — that upholding the law against a broad-based challenge does not foreclose a litigant’s success in a narrower one. See Buckley, supra, at 74 (“minor parties” may be exempt from disclosure requirements if they can show “a reasonable probability that the compelled disclosure of a party’s contributors’ names will subject them to threats, harassment, or reprisals from either Government officials or private parties”); Citizens United, supra, at 370 (disclosure “would be unconstitutional as applied to an organization if there were a reasonable probability that the group’s members would face threats, harassment, or reprisals if their names were disclosed” (citing McConnell v. Federal Election Comm’n, 540 U. S. 93, 198 (2003))). The secretary of state acknowledges that plaintiffs may press the narrower challenge in Count II of their complaint in proceedings pending before the District Court. Brief for Respondent Reed 17. * * * We conclude that disclosure under the PRA would not violate the First Amendment with respect to referendum petitions in general and therefore affirm the judgment of the Court of Appeals. It is so ordered. Justice Scalia doubts whether petition signing is entitled to any First Amendment protection at all. Post, at 219 (opinion concurring in judgment). His skepticism is based on the view that petition signing has “legal effects” in the legislative process, while other aspects of political participation — with respect to which we have held there is a First Amendment interest, see supra, at 194-196 — do not. See post, at 221-222, and n. 3. That line is not as sharp as Justice Scalia would have it; he himself recognizes “the existence of a First Amendment interest in voting,” post, at 224, which of course also can have legal effect. The distinction becomes even fuzzier given that only some petition signing has legal effect, and any such legal effect attaches only well after the expressive act of signing, if the secretary determines that the petition satisfies the requirements for inclusion on the ballot. See post, at 221. Petitions that do not qualify for the ballot of course carry no legal effect. Justice Thomas’s contrary assessment of the relationship between the disclosure of referendum petitions generally and the State’s interests in this ease is based on his determination that strict scrutiny applies, post, at 232 (dissenting opinion), rather than the standard of review that we have concluded is appropriate, see supra, at 196.
Justice Breyer, concurring. In circumstances where, as here, “a law significantly implicates competing constitutionally protected interests in complex ways,” the Court balances interests. Nixon v. Shrink Missouri Government PAC, 528 U. S. 377, 402 (2000) (Breyer, J., concurring). “And in practice that has meant asking whether the statute burdens any one such interest in a manner out of proportion to the statute’s salutary effects upon the others.” Ibid. As I read their opinions, this is what both the Court and Justice Stevens do. See ante, at 196 (opinion of the Court); post, at 217-218 (Stevens, J., concurring in part and concurring in judgment). And for the reasons stated in those opinions (as well as many of the reasons discussed by Justice Sotomayor), I would uphold the statute challenged in this case. With this understanding, I join the opinion of the Court and Justice Stevens’ opinion.
Justice Alito, concurring. The Court holds that the disclosure under the Washington Public Records Act (PRA), Wash. Rev. Code §42.56.001 et seq. (2008), of the names and addresses of persons who sign referendum petitions does not as a general matter violate the First Amendment, ante this page, and I agree with that conclusion. Many referendum petitions concern relatively uncontroversial matters, see ante, at 200-201, and plaintiffs have provided no reason to think that disclosure of signatory information in those contexts would significantly chill the willingness of voters to sign. Plaintiffs’ facial challenge therefore must fail. See ante, at 191, 194. Nonetheless, facially valid disclosure requirements can impose heavy burdens on First Amendment rights in individual cases. Acknowledging that reality, we have long held that speakers can obtain as-applied exemptions from disclosure requirements if they can show “a reasonable probability that the compelled disclosure of [personal information] will subject them to threats, harassment, or reprisals from either Government officials or private parties.” Buckley v. Valeo, 424 U. S. 1, 74 (1976) (per curiam); see also Citizens United v. Federal Election Comm’n, 558 U. S. 310, 367 (2010); McConnell v. Federal Election Comm’n, 540 U. S. 93, 197-198 (2003); Brown v. Socialist Workers ’74 Campaign Comm. (Ohio), 459 U. S. 87, 93 (1982). Because compelled disclosure can “burden the ability to speak,” Citizens United, supra, at 366, and “seriously infringe on privacy of association and belief guaranteed by the First Amendment,” Buckley, supra, at 64, the as-applied exemption plays a critical role in safeguarding First Amendment rights. I The possibility of prevailing in an as-applied challenge provides adequate protection for First Amendment rights only if (1) speakers can obtain the exemption sufficiently far in advance to avoid chilling protected speech and (2) the showing necessary to obtain the exemption is not overly burdensome. With respect to the first requirement, the as-applied exemption becomes practically worthless if speakers cannot obtain the exemption quickly and well in advance of speaking. To avoid the possibility that a disclosure requirement might chill the willingness of voters to sign a referendum petition (and thus burden a circulator’s ability to collect the necessary number of signatures, cf. Meyer v. Grant, 486 U. S. 414, 423 (1988)), voters must have some assurance at the time when they are presented with the petition that their names and identifying information will not be released to the public. The only way a circulator can provide such assurance, however, is if the circulator has sought and obtained an as-applied exemption from the disclosure requirement well before circulating the petition. Otherwise, the best the circulator could do would be to tell voters that an exemption might be obtained at some point in the future. Such speculation would often be insufficient to alleviate voters’ concerns about the possibility of being subjected to threats, harassment, or reprisals. Cf. Citizens United, supra, at 484-485 (Thomas, J., concurring in part and dissenting in part). Additionally, speakers must be able to obtain an as-applied exemption without clearing a high evidentiary hurdle. We acknowledged as much in Buckley, where we noted that “unduly strict requirements of proof could impose a heavy burden” on speech. 424 U. S., at 74. Recognizing that speakers “must be allowed sufficient flexibility in the proof of injury to assure a fair consideration of their claim,” we emphasized that speakers “need show only a reasonable probability” that disclosure will lead to threats, harassment, or reprisals. Ibid, (emphasis added). We stated that speakers could rely on a wide array of evidence to meet that standard, including “specific evidence of past or present harassment of [group] members,” “harassment directed against the organization itself,” or a “pattern of threats or specific manifestations of public hostility.” Ibid. Significantly, we also made clear that “[n]ew [groups] that have no history upon which to draw may be able to offer evidence of reprisals and threats directed against individuals or organizations holding similar views.” Ibid. From its inception, therefore, the as-applied exemption has not imposed onerous burdens of proof on speakers who fear that disclosure might lead to harassment or intimidation. II. In light of those principles, the plaintiffs in this case have a strong argument that the PRA violates the First Amendment as applied to the Referendum 71 petition. A Consider first the burdens on plaintiffs’ First Amendment rights. The widespread harassment and intimidation suffered by supporters of California’s Proposition 8 provides strong support for an as-applied exemption in the present case. See Buckley, supra, at 74 (explaining that speakers seeking as-applied relief from a disclosure requirement can rely on “evidence of reprisals and threats directed against individuals or organizations holding similar views”). Proposition 8 amended the California Constitution to provide that “[o]nly marriage between a man and a woman is valid or recognized in California,” Cal. Const., Art. I, § 7.5, and plaintiffs submitted to the District Court substantial evidence of the harassment suffered by Proposition 8 supporters, see Declaration of Scott F. Bieniek in No. C:09-5456 (WD Wash.), Exhs. 12, 13. Members of this Court have also noted that harassment. See Hollingsworth v. Perry, 558 U. S. 183, 185-186 (2010) (per curiam); Citizens United, supra, at 481-482 (opinion of Thomas, J.). Indeed, if the evidence relating to Proposition 8 is not sufficient to obtain an as-applied exemption in this case, one may wonder whether that vehicle provides any meaningful protection for the First Amendment rights of persons who circulate and sign referendum and initiative petitions. What is more, when plaintiffs return to the District Court, they will have the opportunity to develop evidence of intimidation and harassment of Referendum 71 supporters — an opportunity that was pretermitted because of the District Court’s decision to grant a preliminary injunction on count I of plaintiffs’ complaint. See 661 F. Supp. 2d 1194, 1205-1206 (WD Wash. 2009); Tr. of Oral Arg. 40-41. For example, plaintiffs allege that the campaign manager for one of the plaintiff groups received threatening e-mails and phone calls, and that the threats were so severe that the manager filed a-complaint with the local sheriff and had his children sleep in an interior room of his home. App. 9-10. B The inadequacy of the State’s interests in compelling public disclosure of referendum signatory information further confirms that courts should be generous in granting as-applied relief in this context. See Buckley, supra, at 71 (recognizing that the weakness of the State’s interests in an individual case can require exempting speakers from compelled disclosure); Brown, 459 U. S., at 92-93 (same). As the Court notes, respondents rely on two interests to justify compelled disclosure in this context: (1) providing information to voters about who supports a referendum petition; and (2) preserving the integrity of the referendum process by detecting fraudulent and mistaken signatures. Ante, at 197. 1 In my view, respondents’ asserted informational interest will not in any case be sufficient to trump the First Amendment rights of signers and circulators who face a threat of harassment. Respondents maintain that publicly disclosing the names and addresses of referendum signatories provides the voting public with “insight into whether support for holding a vote comes predominantly from particular interest groups, political or religious organizations, or other group[s] of citizens,” and thus allows voters to draw inferences about whether they should support or oppose the referendum. Brief for Respondent Washington Families Standing Together 58; see also Brief for Respondent Reed 46-48. Additionally, respondents argue that disclosure “allows Washington voters to engage in discussion of referred measures with persons whose acts secured the election and suspension of state law.” Id., at 45; see also Brief for Respondent Washington Families Standing Together 58. The implications of accepting such an argument are breathtaking. Were we to accept respondents’ asserted informational interest, the State would be free to require petition signers to disclose all kinds of demographic information, including -the signer’s race, religion, political affiliation, sexual orientation, ethnic background, and interest-group memberships. Requiring such disclosures, however, runs headfirst into a half century of our case law, which firmly establishes that individuals have a right to privacy of belief and association. See Rumsfeld v. Forum for Academic and Institutional Rights, Inc., 547 U. S. 47, 69 (2006); Brown, supra, at 91; Buckley, 424 U. S., at 64; DeGregory v. Attorney General of N. H., 383 U. S. 825, 829 (1966); Gibson v. Florida Legislative Investigation Comm., 372 U. S. 539, 544 (1963); NAACP v. Alabama ex rel. Patterson, 357 U. S. 449, 462 (1958). Indeed, the State’s informational interest paints such a chilling picture of the role of government in our lives that at oral argument the Washington attorney general balked when confronted with the logical implications of accepting such an argument, conceding that the State could not require petition signers to disclose their religion or ethnicity. Tr. of Oral Arg. 37, 56. Respondents’ informational interest is no more legitimate when viewed as a means of providing the public with information needed to locate and contact supporters of a referendum. In the name of pursuing such an interest, the State would be free to require petition signers to disclose any information that would more easily enable members of the voting public to contact them and engage them in discussion, including telephone numbers, e-mail addresses, and Internet aliases. Once again, permitting the government to require speakers to disclose such information runs against the current of our associational privacy cases. But more important, when speakers are faced with a reasonable probability of harassment or intimidation, the State no longer has any interest in enabling the public to locate and contact supporters of a particular measure — for in that instance, disclosure becomes a means of facilitating harassment that impermissibly chills the exercise of First Amendment rights. In this case, two groups proposed to place on the Internet the names and addresses of all those who signed Referendum 71, and it is alleged that their express aim was to encourage “uncomfortable conversation^].” 661 F. Supp. 2d, at 1199 (internal quotation marks omitted). If this information is posted on the Internet, then anyone with access to a computer could compile a wealth of information about all of those persons, including in many cases all of the following: the names of their spouses and neighbors, their telephone numbers, directions to their homes, pictures of their homes, information about their homes (such as size, type of construction, purchase price, and mortgage amount), information about any motor vehicles that they own, any court case in which they were parties, any information posted on a social networking site, and newspaper articles in which their names appeared (including such things as wedding announcements, obituaries, and articles in local papers about their children’s school and athletic activities). The potential that such information could be used for harassment is vast. 2 Respondents also maintain that the State has an interest in preserving the integrity of the referendum process and that public disclosure furthers that interest by helping the State detect fraudulent and mistaken signatures. I agree with the Court that preserving the integrity of the referendum process constitutes a sufficiently important state interest. Ante, at 197. But I harbor serious doubts as to whether public disclosure of signatory information serves that interest in a way that always “reflect[s] the seriousness of the actual burden on First Amendment rights.” Davis v. Federal Election Comm’n, 554 U. S. 724, 744 (2008). First, the realities of Washington law undermine the State’s argument that public disclosure is necessary to ensure the integrity of the referendum process. The State of Washington first authorized voter initiatives via constitutional amendment in 1912, and the following year the Washington Legislature passed a statute specifying the particulars of the referendum process. See State ex rel. Case v. Superior Ct.for Thurston Cty., 81 Wash. 623, 628, 143 P. 461, 462 (1914). Significantly, Washington’s laws pertaining to initiatives and referenda did not then and do not now authorize the public disclosure of signatory information. See Wash. Rev. Code §29A.72.010 et seq.; 1913 Wash. Laws pp. 418-437. Instead, the public disclosure requirement stems from the PRA, which was enacted in 1972 and which requires the public disclosure of state documents generally, not referendum documents specifically. See Wash. Rev. Code §42.56.001 et seq. Indeed, if anything, Washington’s referenda and initiative laws suggest that signatory information should remain confidential: Outside observers are permitted to observe the secretary of state’s verification and canvassing process only “so long as they make no record of the names, addresses, or other information on the petitions or related records during the verification process,” §29A.72.230, and the State is required to destroy all those petitions that fail to qualify for the ballot, § 29A.72.200. Second, the State fails to come to grips with the fact that public disclosure of referendum signatory information is a relatively recent practice in Washington. Prior to the adoption of the PRA in 1972, the Washington attorney general took the view that referendum petitions were not subject to public disclosure. See Op. Wash. Atty. Gen. 55-57 No. 274, pp. 1-2 (May 28,1956), online at http://www.atg.wa.gov/AGO Opinions/opinion.aspx?section=topic&id=10488 (all Internet materials as visited June 17, 2010, and available in Clerk of Court’s case file) (declaring that public disclosure of initiative petitions would be “contrary to public policy” and would run contrary to “a tendency on the part of the legislature to regard the signing of an initiative petition as a matter concerning only the individual signers except in so far as necessary to safeguard against abuses of the privilege”). Indeed, the secretary of state represents on his Web site that even after the PRA was enacted, “various Secretary of State administrations took the position, from 1973 to 1998, that the personal information on petition sheets were NOT subject to disclosure.” B. Zylstra, The Disclosure History of Petition Sheets (Sept. 17, 2009), online at http://blogs.sos. wa.gov/FromOurCorner/index.php/2009/09/the-disclosurehistory-of-petition-sheets. Although the secretary of state apparently changed this policy in the late 1990’s, it appears that the secretary did not release any initiative petitions until 2006. Ibid. And to date, the secretary has released only a handful of petitions. Ibid.; App. 26. That history substantially undermines the State’s assertion that public disclosure is necessary to ensure the integrity of the referendum process. For nearly a century, Washington’s referendum process operated — and apparently operated successfully — without the public disclosure of signatory information. The State has failed to explain how circumstances have changed so dramatically in recent years that public disclosure is now required. Third, the experiences of other States demonstrate that publicly disclosing the names and identifying information of referendum signatories is not necessary to protect against fraud and mistake. To give but one example, California has had more initiatives on the ballot than any other State save Oregon. See Initiative and Referendum Institute, Initiative Use, p. 1 (Feb. 2009), online at http://www.iandrinstitute.org/ IRI%20Initiative%20Use%20%281904=2008%29.pdf. Nonetheless, California law explicitly protects the privacy of initiative and referendum signatories. See Cal. Elec. Code Ann. §18650 (West 2003); Cal. Govt. Code Ann. §6253.5 (West 2008). It is thus entirely possible for a State to keep signatory information private and maintain a referendum and initiative process free from fraud. Finally, Washington could easily and cheaply employ alternative mechanisms for protecting against fraud and mistake that would be far more protective of circulators’ and signers’ First Amendment rights. For example, the Washington attorney general represented to us at oral argument that “the Secretary of State’s first step after receiving submitted petitions is to take them to his archiving section and to have them digitized.” Tr. of Oral Arg. 30. With a digitized list, it should be relatively easy for the secretary to check for duplicate signatures on a referendum petition. And given that the secretary maintains a “centralized, uniform, interactive computerized statewide voter registration list that contains the name and registration information of every registered voter in the state,” Wash. Rev. Code Ann. §29A.08.125(1) (West Supp. 2010), the secretary could use a computer program to cross-check the names and addresses on the petition with the names and addresses on the voter registration rolls, thus ensuring the accuracy and legitimacy of each signature. Additionally, using the digitized version of the referendum petition, the State could set up a simple system for Washington citizens to check whether their names have been fraudulently signed to a petition. For example, on his Web site, the secretary maintains an interface that allows voters to confirm their voter registration information simply by inputting their name and date of birth. • See http://wei.secstate. wa.gov/osos/VoterVault/Pages/MyVote.aspx. Presumably the secretary could set up a similar interface for referendum petitions. Indeed, the process would seem to be all the more simple given that Washington requires a “unique identifier [to] be assigned to each registered voter in the state.” §29A.08.125(4). * * * As-applied challenges to disclosure requirements play a critical role in protecting First Amendment freedoms. To give speech the breathing room it needs to flourish, prompt judicial remedies must be available well before the relevant speech occurs and the burden of proof must be low. In this case — both through analogy and through their own experiences — plaintiffs have a strong case that they are entitled to as-applied relief, and they will be able to pursue such relief before the District Court.
Justice Sotomayor, with 'whom Justice Stevens and Justice Ginsburg join, concurring. I write separately to emphasize a point implicit in the opinion of the Court and the concurring opinions of Justice Stevens, Justice Scalia, and Justice Breyer: In assessing the countervailing interests at stake in this case, we must be mindful of the character of initiatives and referenda. These mechanisms of direct democracy are not compelled by the Federal Constitution. It is instead up to the people of each State, acting in their sovereign capacity, to decide whether and how to permit legislation by popular action. States enjoy “considerable leeway” to choose the subjects that are eligible for placement on the ballot and to specify the requirements for obtaining ballot access (e. g., the number of signatures required, the time for submission, and the method of verification). Buckley v. American Constitutional Law Foundation, Inc., 525 U. S. 182, 191 (1999). As the Court properly recognizes, each of these structural decisions “inevitably affects — at least to some degree — the individual’s right” to speak about political issues and “to associate with others for political ends.” Anderson v. Celebrezze, 460 U. S. 780, 788 (1983). For instance, requiring petition signers to be registered voters or to use their real names no doubt limits the ability or willingness of some individuals to undertake the expressive act of signing a petition. Regulations of this nature, however, stand “a step removed from the communicative aspect of petitioning,” and the ability of States to impose them can scarcely be doubted. Buckley, 525 U. S., at 215 (O’Connor, J., concurring in judgment in part and dissenting in part); see also McIntyre v. Ohio Elections Comm’n, 514 U. S. 334, 345 (1995) (contrasting measures to “control the mechanics of the electoral process” with the “regulation of pure speech”). It is by no means necessary for a State to prove that such “reasonable, nondiscriminatory restrictions” are narrowly tailored to its interests. Anderson, 460 U. S., at 788. The Court today confirms that the State of Washington’s decision to make referendum petition signatures available for public inspection falls squarely within the realm of permissible election-related regulations. Cf. Buckley, 525 U. S., at 200 (describing a state law requiring petition circulators to submit affidavits containing their names and addresses as “exemplif[ying] the type of regulation” that States may adopt). Public disclosure of the identity of petition signers, which is the rule in the overwhelming majority of States that use initiatives and referenda, advances States’ vital interests in “[preserving the integrity of the electoral process, preventing corruption, and sustaining the active, alert responsibility of the individual citizen in a democracy for the wise conduct of government.” First Nat. Bank of Boston v. Bellotti, 435 U. S. 765, 788-789 (1978) (internal quotation marks and alteration omitted); see also Citizens United v. Federal Election Comm’n, 558 U. S. 310, 371 (2010) (“[Transparency enables the electorate to make informed decisions and give proper weight to different speakers and messages”); Brief for Respondent Washington Families Standing Together 34 (reporting that only one State exempts initiative and referendum petitions from public disclosure). In a society “in which the citizenry is the final judge of the proper conduct of public business,” openness in the democratic process is of “critical importance.” Cox Broadcasting Corp. v. Cohn, 420 U. S. 469, 495 (1975); see also post, at 222 (Scalia, J., concurring in judgment) (noting that “[t]he public nature of federal lawmaking is constitutionally required”). On the other side of the ledger, I view the burden of public disclosure on speech and associational rights as minimal in this context. As this Court has observed with respect to campaign-finance regulations, “disclosure requirements . . . ‘do not prevent anyone from speaking.’” Citizens United, 558 U. S., at 366. When it comes to initiatives and referenda, the impact of public disclosure on expressive interests is even more attenuatéd. While campaign-finance disclosure injects the government into what would otherwise have been private political activity, the process of legislating by referendum is inherently public. To qualify a referendum for the ballot, citizens are required to sign a petition and supply identifying information to the State. The act of signing typically occurs in public, and the circulators who collect and submit signatures ordinarily owe signers no guarantee of confidentiality. For persons with the “civic courage” to participate in this process, post, at 228 (opinion of Scalia, J.), the State’s decision to make accessible what they voluntarily place in the public sphere should not deter them from engaging in the expressive act of petition signing. Disclosure of the identity of petition signers, moreover, in no way directly impairs the ability of anyone to speak and associate for political ends either publicly or privately. Given the relative weight of the interests at stake and the traditionally public nature of initiative and referendum processes, the Court rightly rejects petitioners’ constitutional challenge to the State of Washington’s petition disclosure regulations. These same considerations also mean that any party attempting to challenge particular applications of the State’s regulations will bear a heavy burden. Even when a referendum involves a particularly controversial subject and some petition signers fear harassment from nonstate actors, a State’s important interests in “protect[ing] the integrity and reliability of the initiative process” remain undiminished, and the State retains significant discretion in advancing those interests. Buckley, 525 U. S., at 191. Likewise, because the expressive interests implicated by the act of petition signing are always modest, I find it difficult to see how any incremental disincentive to sign a petition would tip the constitutional balance. Case-specific relief may be available when a State selectively applies a facially neutral petition disclosure rule in a manner that discriminates based on the content of referenda or the viewpoint of petition signers, or in the rare circumstance in which disclosure poses a reasonable probability of serious and widespread harassment that the State is unwilling or unable to control. Cf. NAACP v. Alabama ex rel. Patterson, 357 U. S. 449 (1958). Allowing case-specific invalidation under a more forgiving standard would unduly diminish the substantial breathing room States are afforded to adopt and implement reasonable, nondiscriminatory measures like the disclosure requirement now at issue. Accordingly, courts presented with an as-applied challenge to a regulation authorizing the disclosure of referendum petitions should be deeply skeptical of any assertion that the Constitution, which embraces political transparency, compels States to conceal the identity of persons who seek to participate in lawmaking through a state-created referendum process. With this understanding, I join the opinion of the Court.
Justice Stevens, with whom Justice Breyer joins, concurring in part and concurring in the judgment. This is not a hard case. It is not about a restriction on voting or on speech and does not involve a classic disclosure requirement. Rather, the case concerns a neutral, nondiscriminatory policy of disclosing information already in the State’s possession that, it has been alleged, might one day indirectly burden petition signatories. The burden imposed by Washington’s application of the Public Records Act (PRA) to referendum petitions in the vast majority, if not all, its applications is not substantial. And the State has given a more than adequate justification for its choice. For a number of reasons, the application of the PRA to referendum petitions does not substantially burden any individual’s expression. First, it is not “a regulation of pure speech.” McIntyre v. Ohio Elections Comm’n, 514 U. S. 334, 345 (1995); cf. United States v. O’Brien, 391 U. S. 367, 377 (1968). It does not prohibit expression, nor does it require that any person signing a petition disclose or say anything at all. See McIntyre, 514 U. S. 334. Nor does the State’s disclosure alter the content of a speaker’s message. See id., at 342-343. Second, any effect on speech that disclosure might have is minimal. The PRA does not necessarily make it more difficult to circulate or obtain signatures on a petition, see Buckley v. American Constitutional Law Foundation, Inc., 525 U. S. 182, 193-196 (1999); Meyer v. Grant, 486 U. S. 414, 422-423 (1988), or to communicate one’s views generally. Regardless of whether someone signs a referendum petition, that person remains free to say anything to anyone at any time. If disclosure indirectly burdens a speaker, “the amount of speech covered” is small — only a single, narrow message conveying one fact in one place, Watchtower Bible & Tract Soc. of N. Y. Inc. v. Village of Stratton, 536 U. S. 150, 165 (2002); cf. Cox v. New Hampshire, 312 U. S. 569 (1941). And while the democratic act of casting a ballot or signing a petition does serve an expressive purpose, the act does not involve any “interactive communication,” Meyer, 486 U. S., at 422, and is “not principally” a method of “individual expression of political sentiment,” Timmons v. Twin Cities Area New Party, 520 U. S. 351, 373 (1997) (Stevens, J., dissenting); cf. O’Brien, 391 U. S., at 377. Weighed against the possible burden on constitutional rights are the State’s justifications for its rule. In this case, the State has posited a perfectly adequate justification: an interest in deterring and detecting petition fraud. Given the pedigree of this interest and of similar regulations, the State need not produce concrete evidence that the PRA is the best way to prevent fraud. See Crawford v. Marion County Election Bd., 553 U. S. 181, 191-200 (2008) (opinion of Stevens, J.) (discussing voting fraud); Nixon v. Shrink Missouri Government PAC, 528 U. S. 377, 391 (2000) (“The quantum of empirical evidence needed to satisfy heightened judicial scrutiny of legislative judgments will vary up or down with the novelty and plausibility of the justification raised”); see also Timmons, 520 U. S., at 375 (Stevens, J., dissenting) (rejecting “imaginative [and] theoretical” justification supported only by “bare assertion”). And there is more than enough evidence to support the State’s election-integrity justification. See ante, at 197-199 (opinion of the Court). There remains the issue of petitioners’ as-applied challenge. As a matter of law, the Court is correct to keep open the possibility that in particular instances in which a policy such as the PRA burdens expression “by the public enmity attending publicity,” Brown v. Socialist Workers ’74 Campaign Comm. (Ohio), 459 U. S. 87, 98 (1982), speakers may have a winning constitutional claim. “ '[F]rom time to time throughout history,’ ” persecuted groups have been able “ 'to criticize oppressive practices and laws either anonymously or not at all.’ ” McIntyre, 514 U. S., at 342. In my view, this is unlikely to occur in cases involving the PRA. Any burden on speech that petitioners posit is speculative as well as indirect. For an as-applied challenge to a law such as the PRA to succeed, there would have to be a significant threat of harassment directed at those who sign the petition that cannot be mitigated by law enforcement measures. Moreover, the character of the law challenged in a referendum does not, in itself, affect the analysis. Debates about tax policy and regulation of private property can become just as heated as debates about domestic partnerships. And as a general matter, it is very difficult to show that by later disclosing the names of petition signatories, individuals will be less willing to sign petitions. Just as we have in the past, I would demand strong evidence before concluding that an indirect and speculative chain of events imposes a substantial burden on speech. A statute “is not to be upset upon hypothetical and unreal possibilities, if it would be good upon the facts as they are.” Pullman Co. v. Knott, 235 U. S. 23, 26 (1914). * * Accordingly, I concur with the opinion of the Court to the extent that it is not inconsistent with my own, and I concur in the judgment. Although a “petition” is a classic means of political expression, the type of petition at issue in this case is not merely a document on which people are expressing their views but rather is a state-created forum with a particular function: sorting those issues that have enough public support to warrant limited space on a referendum ballot. Cf. Widmar v. Vincent, 454 U. S. 263, 278 (1981) (Stevens, J., concurring in judgment). Washington also points out that its disclosure policy informs voters about who supports the particular referendum. In certain election-law contexts, this informational rationale (among others) may provide a basis for regulation; in this ease, there is no need to look beyond the State’s quite obvious antifraud interest. There is no reason to think that our ordinary presumption that the political branches are better suited than courts to weigh a policy’s benefits and burdens is inapplicable in this case. The degree to which we defer to a judgment by the political branches must vary up and down with the degree to which that judgment reflects considered, public-minded decision-making. Thus, when a law appears to have been adopted without reasoned consideration, see, e. g., Salazar v. Buono, 559 U. S. 700, 756-757 (2010) (Stevens, J., dissenting), for discriminatory purposes, see, e. g., Bates v. Little Bock, 361 U. S. 516, 517-518, 524-525 (1960), or to entrench political majorities, see, e. g., Vieth v. Jubelirer, 541 U. S. 267, 317-319, 324-326, 332-333 (2004) (Stevens, J., dissenting), we are less willing to defer to the institutional strengths of the legislature. That one may call into question the process used to create a law is not a reason to “disregard]” “sufficiently strong,” “valid[,] neutral justifications” for an otherwise “nondiscriminatory” policy. Crawford, 553 U. S., at 204. But it is a reason to examine more carefully the justifications for that measure. Justice Scalia conceives of the issue as a right to anonymous speech. See, e. g., post, at 220 (opinion concurring in judgment). But our decision in McIntyre posited no such freewheeling right. The Constitution protects “freedom of speech.” Arndt. 1; see also McIntyre, 514 U. S., at 336 (“The question presented is whether [a] . . . statute that prohibits the distribution of anonymous campaign literature is a ‘law ... abridging the freedom of speech’ within the meaning of the First Amendment”). That freedom can be burdened by a law that exposes the speaker to fines, as much as it can be burdened by a law that exposes a speaker to harassment, changes the content of his speech, or prejudices others against his message. See id., at 342. The right, however, is the right to speak, not the right to speak without being fined or the right to speak anonymously. A rare case may also arise in which the level of threat to any individual is not quite so high but a State’s disclosure would substantially limit a group’s ability to “garner the number of signatures necessary to place [a] matter on the ballot,” thereby “limiting [its] ability to make the matter the focus of statewide discussion.” Meyer v. Grant, 486 U. S. 414, 423 (1988). See, e. g., Bates v. Little Rock, 361 U. S., at 521-522, 523-524; Buckley v. Valeo, 424 U. S. 1, 69-72 (1976) (per curiam); Brown v. Socialist Workers '74 Campaign Comm. (Ohio), 459 U. S. 87, 98-101 (1982); Buckley v. American Constitutional Law Foundation, Inc., 525 U. S. 182, 197-198 (1999).
Justice Scalia, concurring in the judgment. Plaintiffs claim the First Amendment, as applied to the States through the Fourteenth Amendment, forbids the State of Washington to release to the public signed referendum petitions, which they submitted to the State in order to suspend operation of a law and put it to a popular vote. I doubt whether signing a petition that has the effect of suspending a law fits within “the freedom of speech” at all. But even if, as the Court concludes, ante, at 194-195, it does, a long history of practice shows that the First Amendment does not prohibit public disclosure. We should not repeat and extend the mistake of McIntyre v. Ohio Elections Comm’n, 514 U. S. 334 (1995). There, with neither textual support nor precedents requiring the result, the Court invalidated a form of election regulation that had been widely used by the States since the end of the 19th century. Id., at 371 (Scalia, J., dissenting). The Court held that an Ohio statute prohibiting the distribution of anonymous campaign literature violated the First and Fourteenth Amendments. Mrs. McIntyre sought