concurring in part and dissenting in part.
I concur in the judgment that the Socialist Workers Party (SWP) has sufficiently demonstrated a reasonable probability that disclosure of contributors will subject those persons to threats, harassment, or reprisals, and thus under Buckley v. Valeo, 424 U. S. 1 (1976), the State of Ohio cannot constitutionally compel the disclosure. Further, I agree that the broad concerns of Buckley apply to the required disclosure of recipients of campaign expenditures. But, as I view the record presented here, the SWP has failed to carry its burden of showing that there is a reasonable probability that disclosure of recipients of expenditures will subject the recipients themselves or the SWP to threats, harassment, or reprisals. Moreover, the strong public interest in fair and honest elections outweighs any damage done to the associational rights of the party and its members by application of the State’s expenditure disclosure law.
*108I
Buckley upheld the validity of the Federal Election Campaign Act of 1971, which requires the disclosure of names of both contributors to a campaign and recipients of expenditures from the campaign. Buckley recognized three major governmental interests in disclosure requirements: deterrence of corruption; enhancement of voters’ knowledge about a candidate’s possible allegiances and interests; and provision of the data and means necessary to detect violations of any statutory limitations on contributions or expenditures. The precise challenge that the Buckley Court faced, however, was the overbreadth of the Act’s requirements “insofar as they apply to contributions to minor parties and independent candidates.” Id., at 68-69 (emphasis added).1 Since the appellants in Buckley did not challenge the application to minor parties of requirements of disclosure of expenditures, the Court had no occasion to consider directly the First Amendment interests of a minor political party in preventing disclosure of expenditures, much less to weigh them against the governmental interests in disclosure. The test adopted by Buckley, quoted by the majority, ante, at 93, reflects this limitation, for it contemplates only assessing possible harassment of contributors, without a word about considering the harassment of recipients of expenditures if their names are disclosed or any effects this harassment may have on the party.
This is not to say that Buckley provides no guidance for resolving this claim. I agree with the majority that appellants *109have overstated their argument in declaring that Buckley has no application to the disclosure of recipients of expenditures. Certainly, Buckley enunciates the general governmental interest in regulating minor parties, who, although unlikely to win, can often affect the outcome of an election. 424 U. S., at 70. Buckley also emphasizes the sensitive associational rights of minor parties.
Nevertheless, there are important differences between disclosure of contributors and disclosure of recipients of campaign expenditures — differences that the Buckley Court had no occasion to address, but that compel me to conclude that the balance should not necessarily be calibrated identically. First, unlike the government’s interest in disclosure of contributions, its interest in disclosure of expenditures does not decrease significantly for small parties. The Court in Buckley recognized that knowing the identity of contributors would not significantly increase the voters’ ability to determine the political ideology of the minor-party candidate, for the stance of the minor-party candidate is usually well known. Ibid.2 Nor would identifying a minor party’s contributors further the interest in preventing the “buying” of a candidate, because of the improbability of the minor-party candidate’s winning the election. Ibid. Thus, these two major government interests in disclosure of contributions are significantly reduced for minor parties.3
In sharp contrast, however, the governmental interest in disclosure of expenditures remains significant for minor parties. The purpose of requiring parties to disclose expenditures is to deter improper influencing of voters. Corruption *110of the electoral process can take many forms: the actual buying of votes; the use of “slush funds;” dirty tricks; and bribes of poll watchers and other election officials. Certainly, a “persuasive” campaign worker on election day can corral voters for his minor-party candidate with even a modest “slush fund.”4 Even though such improper practices are unlikely to be so successful as to attract enough votes to elect the minor-party candidate, a minor party, whose short-term goal is merely recognition, may be as tempted to resort to impermissible methods as are major parties, and the resulting deflection of votes can determine the outcome of the election of other candidates.5 The requirement of a full and verifiable report of expenditures is important in deterring such practices, for otherwise the party could hide the improper transactions through an accounting sleight of hand.6
On the other side of the balance, disclosure of recipients of expenditures will have a lesser impact on a minority party’s First Amendment interests than will disclosure of contribu*111tors. As the majority states, ante, at 91, the First Amendment interest here is “[t]he right to privacy in one’s political associations and beliefs.” We have never drawn sharp distinctions between members and contributors, Buckley, 424 U. S., at 66. As we recognized in Buckley, the privacy rights of contributors are especially sensitive, since many seek to express their political views privately through their pocketbook rather than publicly through other means. Disclosure of contributors directly implicates the contributors’ associational rights.
The impact on privacy interests arising from disclosure of expenditures is of a quite different — and generally lesser— dimension. Many expenditures of the minority party will be for quite mundane purposes to persons not intimately connected with the organization. Payments for such things as office supplies, telephone service, bank charges, printing and photography costs would generally fall in this category. The likelihood that such business transactions would dry up if disclosed is remote at best. Unlike silent contributors, whom disclosure would reveal to the public as supporters of the party’s ideological positions, persons providing business services to a minor party are not generally perceived by the public as supporting the party’s ideology, and thus are unlikely to be harassed if their names are disclosed. Consequently, the party’s associational interests are unlikely to be affected by disclosure of recipients of such expenditures.
Other recipients of expenditures may have closer ideological ties to the party. The majority suggests that campaign workers receiving per diem, travel, or room expenses may fit in this category. Ante, at 97, n. 12. It is certainly conceivable that such persons may be harassed or threatened for their conduct. Laws requiring disclosure of recipients of expenditures, however, are not likely to contribute to this harassment. Once an individual has openly shown his close ties to the organization by campaigning for it, disclosure of receipt of expenditures is unlikely to increase the degree of *112harassment so significantly as to deter the individual from campaigning for the party. Further, in striking the balance, the governmental concerns are greatest precisely for the actions of campaign workers that might improperly influence voters. Thus, whatever marginal deterrence that may arise from disclosure of expenditures is outweighed by the heightened governmental interest.
In sum, the heightened governmental interest in disclosure of expenditures and the reduced marginal deterrent effect on associational interests demand a separately focused inquiry into whether there exists a reasonable probability that disclosure will subject recipients or the party itself to threats, harassment, or reprisals.7
*113► — I HH
Turning to the evidence in this case, it is important to remember that, even though proof requirements must be flexible, Buckley, supra, at 74, the minor party carries the burden of production and persuasion to show that its First Amendment interests outweigh the governmental interests. Additionally, the application of the Buckley standard to the historical evidence is most properly characterized as a mixed question of law and fact, for which we normally assess the record independently to determine if it supports the conclusion of unconstitutionality as applied.8
Here, there is no direct evidence of harassment of either contributors or recipients of expenditures. Rather, as the majority accurately represents it, the evidence concerns harassment and reprisals of visible party members, including violence at party headquarters and loss of jobs. I concur in the majority’s conclusion that this evidence, viewed in its entirety, supports the conclusion that there will be a reasonable probability of harassment of contributors if their names are disclosed. This evidence is sufficiently linked to disclosure of contributors in large part because any person publicly known to support the SWP’s unpopular ideological position may suffer the reprisals that this record shows active party members suffer, and the disclosure of contributors may lead the public to presume these people support the party’s ideology.
*114In contrast, the record, read in its entirety, does not suggest that disclosure of recipients of expenditures would lead to harassment of recipients or reprisals to the party or its members. Appellees gave no breakdown of the types of expenditures they thought would lead to harassment if disclosed. The record does contain the expenditure statements of the SWP, which itemize each expenditure with its purpose while usually omitting the name and address of the recipient. The majority of expenditures, both in number and dollar amount, are for business transactions such as office supplies, food, printing, photographs, telephone service, and books. There is virtually no evidence that disclosure of the recipients of these expenditures will impair the SWP’s ability to obtain needed services.9 Even if we assume that a portion *115of expenditures went to temporary campaign workers or others whom the public might identify as supporting the party’s ideology,10 these persons have already publicly demonstrated their support by their campaign work. There is simply no basis for inferring that such persons would thereafter be harassed or threatened or otherwise deterred from working for the party by virtue of inclusion of their names in later expenditure reports, or that if any such remote danger existed, it would outweigh the concededly important governmental interests in disclosure of recipients of expenditures.
It is plain that appellees did not carry their burden of production and persuasion insofar as they challenge the expenditure disclosure provisions. I would therefore uphold the constitutionality of those portions of the Ohio statute that require the SWP to disclose the recipients of expenditures.11
Of course, the plaintiffs in Buckley challenged many aspects of the federal Act, including expenditure limitations and the disclosure requirements for independent contributions and expenditures. The Court upheld all disclosure requirements, including disclosure of independent expenditures “for communications that expressly advocate the election or defeat of a clearly identified candidate.” 424 U. S., at 80. The plaintiffs in Buckley did not challenge, however, the federal requirement that all political parties, including minor political parties, disclose the recipients of their expenditures.
Certainly, that is true in this instance. The general political stance of the SWP and its candidates is readily discernible from the most cursory glance at its constitution or literature.
The majority is obviously correct in noting that the third governmental interest articulated in Buckley — using disclosures to police limitations on contributions and expenditures — has no application to either contributions or expenditures in Ohio, since the Ohio statute sets no limitations on them.
As Justice White noted in partial dissent in Buckley, 424 U. S., at 264-265, citing Burroughs v. United States, 290 U. S. 534 (1934):
“[T]he corrupt use of money by candidates is as much to be feared as the corrosive influence of large contributions. There are many illegal ways of spending money to influence elections. One would be blind to history to deny that unlimited money tempts people to spend it on whatever money can buy to influence an election.” (Emphasis in original.)
Certainly the SWP could have this effect. For example, appellants noted at oral argument that the SWP candidate in the 1974 Ohio gubernatorial election received some 95,000 votes. The Republican candidate’s margin of victory over the Democratic candidate was only some 13,500 votes. Tr. of Oral Arg. 18. The impact of minor parties on elections in the United States is well documented. See generally W. Hesseltine, Third-Party Movements in the United States (1962).
1 therefore disagree with the majority’s suggestion, ante, at 98-99, n. 16, that the government interest in deterring corruption is not furthered by disclosure of all expenditures, including those for commercial services. Even if improprieties are unlikely to occur in expenditures for commercial services, full and verifiable disclosure is needed to ensure that other, improper expenditures are not hidden in commercial accounts.
According to the majority, “the question whether the Buckley test applies to the compelled disclosure of recipients of expenditures is properly before us.” Ante, at 94, n. 9. The majority declares that, in answering this question, “the District Court necessarily held (1) that the Buckley standard, which permits flexible proof of the reasonable probability of threats, harassment, or reprisals, applies to both contributions and expenditures, and (2) that the evidence was sufficient to show a reasonable probability that disclosure would subject both contributors and recipients to public hostility and harassment.” Ibid, (emphasis added).
Justice Blackmun, ante, at 102, however, more accurately characterizes the District Court’s action as assuming that the Buckley standard applies to disclosure of expenditures and holding the evidence sufficient to meet this standard. The District Court’s assumption is understandable, since appellants did not question it below. Thus, this is not the appropriate case to determine whether a different test or standard of proof should be employed in determining the constitutional validity of required disclosure of expenditures.
Even assuming the general applicability of the Buckley standard, though, the question presented here requires us to inquire whether the evidence of harassment establishes a “reasonable probability” that the Ohio law would trigger “threats, harassment, or reprisals” against recipients of expenditures that in turn may harm the party’s associational interests. This inquiry is necessarily distinct from the inquiry whether the evidence establishes a reasonable probability that disclosure would trigger threats, harassment, or reprisals against contributors. Although the proof requirements guiding this separate inquiry remain flexible, and direct proof *113of harm from disclosure is not required, ultimately the party must prove that the harm to it from disclosure of recipients outweighs the governmental interest in disclosure. This separately focused inquiry does not necessarily alter Buckley’s “reasonable probability” test or “flexible proof” standard. It does, however, plainly require a different result.
See Pullman-Standard v. Swint, 456 U. S. 273, 289, n. 19 (1982). The majority does not clearly articulate the standard of review it is applying. By determining that the District Court “properly concluded” that the evidence established a reasonable probability of harassment, ante, at 100, the majority seems to apply an independent-review standard.
The District Court admitted Exhibit 129 into the record, which is a certified copy of findings of fact made by the Federal Election Commission pursuant to a 1977 court order in Socialist Workers 197h National Campaign Committee v. Jennings, No. 74-1338 (DC, stipulated judgment entered Jan. 3, 1979). The FEC in that case analyzed affidavits submitted by SWP members and other documentary evidence of public and private harassment of SWP members. In finding No. 126, the FEC accepted the SWP’s proposed finding that in 1971 a landlady in San Francisco rejected the application of two SWP members for an apartment, because the FBI had visited the landlady and warned her of the dangers of the SWP. In finding No. 127, the FEC accepted the SWP’s proposed finding that in 1974 a landlady in Chicago evicted a SWP member from her apartment. The landlady explained, “they told me all about you,” refusing to identify who “they” were.
These two incidents are, of course, remote in time and place, and do not suggest that the party itself has had difficulty in finding office space. Nor do they suggest that the general public is likely to engage in similar activity. Moreover, the FBI’s actions against the SWP have long been ended, see Final Report of the Select Committee to Study Governmental Operations with Respect to Intelligence Activities, S. Rep. No. 94-755, Vol. 4-5, pp. 3-4 (1976), and Congress has since instituted more rigorous oversight of FBI and other intelligence activities, see 50 U. S. C. §413 (1976 ed., Supp. IV). An inference from these two incidents that disclosure of recipients of expenditures would increase any difficulty the party might have in obtaining office space would be tenuous, and is plainly outweighed by tfhe “substantial public interest in disclosure,” Buckley, 424 U. S., at 72.
As the majority notes, ante, at 97, n. 12, some entries in the expenditure forms are designated as per diem, travel expenses, and room rental. At least until 1978, the expenditure statements gave the names of persons receiving per diem funds from the SWP. Apparently, party treasurers and party candidates received per diem payments. There is no evidence that filing these statements with the Ohio Secretary of State caused any harassment of the named persons, and indeed it is highly unlikely that this disclosure would increase the exposure of persons already so publicly identified with the party.
In holding a state statute unconstitutional as applied, a court must sever and apply constitutional portions unless the legislature would not have intended to have applied ‘“those provisions which are within its power, independently of that which is not. . . Buckley, supra, at 108 (severing constitutional portions of Federal Election Campaign Act after holding other portions unconstitutional on their face), quoting Champlin Refining Co. v. Corporation Comm’n of Okla., 286 U. S. 210, 234 (1932). Clearly, the expenditure disclosure requirements of the Ohio statute should be severed and applied even though the contribution disclosure requirements cannot be applied in this instance, for the two requirements are analytically and practically distinct.