Justice O’Connor, with whom Justice Breyer joins except as to Part III, concurring in part and concurring in the judgment.
I join the Court’s opinion except for Part II–A. Although I agree with most of the Court’s reasoning, I write separately to emphasize two points. First, I think respondents’ claim implicates important associational interests, and I see no reason to minimize those interests to dispose of this case. Second, I agree with the Court that only Oklahoma’s semiclosed primary law is properly before us, that standing alone it imposes only a modest, nondiscriminatory burden on respondents’ associational rights, and that this burden is justified by the State’s legitimate regulatory interests. I note, however, that there are some grounds for concern that other state laws may unreasonably restrict voters’ ability to change party registration so as to participate in the Libertarian Party of Oklahoma’s (LPO) primary. A realistic assessment of regulatory burdens on associational rights would, in an appropriate case, require examination of the cumulative effects of the State’s overall scheme governing primary elections; and any finding of a more severe burden would trigger more probing review of the justifications offered by the State.
I
Nearly every State in the Nation now mandates that political parties select their candidates for national or statewide office by means of primary elections. See Galderisi & Ezra, Congressional Primaries in Historical and Theoretical Context, in Congressional Primaries and the Politics of Representation 17, and n. 34 (P. Galderisi, M. Ezra, & M. Lyons eds. 2001). Primaries constitute both a “ ‘crucial juncture’ ” in the electoral process, California Democratic Party v. Jones, 530 U. S. 567, 575 (2000) (quoting Tashjian v. Republican Party of Conn., 479 U. S. 208, 216 (1986)), and a vital forum for expressive association among voters and political parties, see Kusper v. Pontikes, 414 U. S. 51, 58 (1973) (“[A] basic function of a political party is to select the candidates for public office to be offered to the voters at general elections[, and a] prime objective of most voters in associating themselves with a particular party must surely be to gain a voice in that selection process”). It is here that the parties invite voters to join in selecting their standard bearers. The outcome is pivotal, of course, for it dictates the range of choices available at—and often the presumptive winner of—the general election.
“No right is more precious in a free country than that of having a voice in the election of those who make the laws under which, as good citizens, we must live,” Wesberry v. Sanders, 376 U. S. 1, 17 (1964), and “[t]he right to associate with the political party of one’s choice is an integral part of this basic constitutional freedom,” Kusper, supra, at 57. The Court has repeatedly reaffirmed that the First and Fourteenth Amendments protect the rights of voters and parties to associate through primary elections. See, e.g., California Democratic Party, supra, at 574–575; Tashjian, supra, at 214; Kusper, supra, at 56–57. Indeed, constitutional protection of associational rights is especially important in this context because the aggregation of votes is, in some sense, the essence of the electoral process. To have a meaningful voice in this process, the individual voter must join together with likeminded others at the polls. And the choice of who will participate in selecting a party’s candidate obviously plays a critical role in determining both the party’s message and its prospects of success in the electoral contest. See California Democratic Party, supra, at 575; see also Democratic Party of United States v. Wisconsin ex rel. La Follette, 450 U. S. 107, 122 (1981) (“[T]he freedom to associate for the ‘common advancement of political beliefs’ necessarily presupposes the freedom to identify the people who constitute the association” (quoting Kusper, supra, at 56)).
The majority questions whether the LPO and voters registered with another party have any constitutionally cognizable interest in associating with one another through the LPO’s primary. See ante, at 5–6. Its doubts on this point appear to stem from two implicit premises: first, that a voter forms a cognizable association with a political party only by registering with that party; and second, that a voter can only form a cognizable association with one party at a time. Neither of these premises is sound, in my view. As to the first, registration with a political party surely may signify an important personal commitment, which may be accompanied by faithful voting and even activism beyond the polls. But for many voters, registration serves principally as a mandatory (and perhaps even ministerial) prerequisite to participation in the party’s primaries. The act of casting a ballot in a given primary may, for both the voter and the party, constitute a form of association that is at least as important as the act of registering. See La Follette, supra, at 130, n. 2 (Powell, J., dissenting) (“[T]he act of voting in the Democratic primary fairly can be described as an act of affiliation with the Democratic Party”). The fact that voting is episodic does not, in my judgment, undermine its associational significance; it simply reflects the special character of the electoral process, which allows citizens to join together at regular intervals to shape government through the choice of public officials.
As to the question of dual associations, I fail to see why registration with one party should negate a voter’s First Amendment interest in associating with a second party. We surely would not say, for instance, that a registered Republican or Democrat has no protected interest in associating with the Libertarian Party by attending meetings or making political contributions. The validity of voters’ and parties’ interests in dual associations seems particularly clear where minor parties are concerned. For example, a voter may have a longstanding affiliation with a major party that she wishes to maintain, but she may nevertheless have a substantial interest in associating with a minor party during particular election cycles or in elections for particular offices. The voter’s refusal to disaffiliate from the major party may reflect her abiding commitment to that party (which is not necessarily inconsistent with her desire to associate with a second party), the objective costs of disaffiliation, see, e.g., infra, at 9–10, or both. The minor party, for its part, may have a significant interest in augmenting its voice in the political process by associating with sympathetic members of the major parties.
None of this is to suggest that the State does not have a superseding interest in restricting certain forms of association. We have never questioned, for example, the States’ authority to restrict voters’ public registration to a single party or to limit each voter to participating in a single party’s primary. But the fact that a State’s regulatory authority may ultimately trump voters’ or parties’ associational interests in a particular context is no reason to dismiss the validity of those interests. As a more general matter, I question whether judicial inquiry into the genuineness, intensity, or duration of a given voter’s association with a given party is a fruitful way to approach constitutional challenges to regulations like the one at issue here. Primary voting is an episodic and sometimes isolated act of association, but it is a vitally important one and should be entitled to some level of constitutional protection. Accordingly, where a party invites a voter to participate in its primary and the voter seeks to do so, we should begin with the premise that there are significant associational interests at stake. From this starting point, we can then ask to what extent and in what manner the State may justifiably restrict those interests.
II
As to the remainder of the constitutional analysis, I am substantially in accord with the Court’s reasoning. Our constitutional system assigns the States broad authority to regulate the electoral process, and we have recognized that, “as a practical matter, there must be substantial regulation of elections if they are to be fair and honest and if some sort of order, rather than chaos, is to accompany the democratic processes,” Storer v. Brown, 415 U. S. 724, 730 (1974). We have sought to balance the associational interests of parties and voters against the States’ regulatory interests through the flexible standard of review reaffirmed by the Court today. See ante, at 3. Under that standard, “the rigorousness of our inquiry into the propriety of a state election law depends upon the extent to which a challenged regulation burdens First and Fourteenth Amendment rights.” Burdick v. Takushi, 504 U. S. 428, 434 (1992). Regulations imposing severe burdens on associational rights must be narrowly tailored to advance a compelling government interest. Timmons v. Twin Cities Area New Party, 520 U. S. 351, 358 (1997) . Regulations imposing lesser burdens are subject to less intensive scrutiny, and reasonable, nondiscriminatory restrictions ordinarily will be sustained if they serve important regulatory interests. Ibid.
This regime reflects the limited but important role of courts in reviewing electoral regulation. Although the State has a legitimate—and indeed critical—role to play in regulating elections, it must be recognized that it is not a wholly independent or neutral arbiter. Rather, the State is itself controlled by the political party or parties in power, which presumably have an incentive to shape the rules of the electoral game to their own benefit. Recognition of that basic reality need not render suspect most electoral regulations. Where the State imposes only reasonable and genuinely neutral restrictions on associational rights, there is no threat to the integrity of the electoral process and no apparent reason for judicial intervention. As such restrictions become more severe, however, and particularly where they have discriminatory effects, there is increasing cause for concern that those in power may be using electoral rules to erect barriers to electoral competition. In such cases, applying heightened scrutiny helps to ensure that such limitations are truly justified and that the State’s asserted interests are not merely a pretext for exclusionary or anticompetitive restrictions.
Throughout the proceedings in the lower courts, respondents framed their suit as a facial challenge to Oklahoma’s semiclosed primary law. The sum of their argument was that, by requiring voters to register either as Libertarians or Independents in order to participate in the LPO’s primary, state law imposes a severe and unjustified burden on the LPO’s and Oklahoma voters’ associational rights. For the reasons explained by the Court, ante, at 14–15, that is the only claim properly before us. Assuming (as I believe we must under the circumstances) that Oklahoma provides reasonable avenues for voters to reregister as Independents or Libertarians, I agree with the Court that the semiclosed primary law imposes only a modest and politically neutral burden on associational rights. The burden is not altogether trivial: A voter with a significant commitment to a major party (for example) must forfeit registration with that party in order to participate in the LPO primary in any given election cycle, and the LPO cannot define the bounds of the association as broadly as it would like. Seepost, at 3, and n. 1 (Stevens, J., dissenting); see alsosupra, at 4 (discussing the interest in dual associations). But neither is it severe or discriminatory.
Oklahoma’s semiclosed primary law simply requires that voters wishing to participate in the LPO’s primary do what they would have to do in order to participate in any other party’s primary. By providing a reasonably fixed party-related electoral base from the close of registration until the date of the vote, this requirement facilitates campaign planning. And assuming the availability of reasonable reregistration procedures, a party’s inability to persuade a voter to disaffiliate from a rival party would suggest not the presence of anticompetitive regulatory restrictions, but rather the party’s failure to win the voter’s allegiance. The semiclosed primary law, standing alone, does not impose a significant obstacle to participation in the LPO’s primary, nor does it indicate partisan self dealing or a lockup of the political process that would warrant heightened judicial scrutiny.
For essentially the reasons explained by the Court, see ante, at 10–14, I agree that Oklahoma has a legitimate interest in requiring voters to disaffiliate from one party before participating in another party’s primary. On the record before us, I also agree that the State’s regulatory interests are adequate to justify the limited burden the semiclosed primary law imposes on respondents’ freedom of association. And finally, I agree that this case is distinguishable from Tashjian. See ante, at 8–10. I joined the dissent in that case, and I think the Court’s application of strict scrutiny there is difficult to square with the flexible standard of review articulated in our more recent cases, see supra, at 5–6. But Tashjian is entitled to respect under principles of stare decisis, and it can be fairly distinguished on the grounds that the closed primary law in that case imposed a greater burden on associational interests than does Oklahoma’s semiclosed primary law, see ante, at 9, while the State’s regulatory interests in Tashjian were weaker than they are here, compare ante, at 10–14, with Tashjian, 479 U. S., at 217–225.
III
In briefing and oral argument before this Court, respondents raise for the first time the claim that Oklahoma’s semiclosed primary law severely burdens their associational rights not through the law’s own operation, but rather because other state laws make it quite difficult for voters to reregister as Independents or Libertarians so as to participate in the LPO primary. See Brief for Respondents 12–24. Respondents characterize Oklahoma’s regulatory scheme as follows.
Partisan primaries in Oklahoma are held on the last Tuesday in July of each even-numbered year. Okla. Stat. Ann., Tit. 26, §1–102 (West Supp. 2005). To field a party candidate in an election, the LPO must obtain “recognized” party status. See ibid.; see also §§1–107, 5–104 (West 1997 and Supp. 2005). This requires it to submit, no later than May 1 of any even-numbered year ( i.e., any election year), a petition with the signatures of registered voters equal to at least five percent of the total votes cast in the most recent gubernatorial or Presidential election. §1–108 (West Supp. 2005). The State Election Board then has 30 days to determine whether the petition is sufficient. §1–108(3) . The LPO has attained recognized party status in this fashion in every Presidential election year since 1980. However, unless the party’s candidate receives at least 10 percent of the total votes cast for Governor or President in the general election (which no minor party has been able to do in any State in recent history), it loses recognized party status. §1–109. To regain party status, the group must go through the petition process again. Ibid.When a party loses its recognized status, as the LPO has after every general election in which it has participated, the affiliation of any voter registered with the party is changed to Independent. §1–110. As the District Court noted, “it is highly likely that the ranks of independents, and, indeed, of registered Republicans and Democrats, contain numerous voters who sympathize with the LPO but who simply do not wish to go through the motions of re-registering every time they are purged from the rolls.” Memorandum Opinion, Case No. CIV–00–1071–F (WD Okla., Jan. 24, 2003), App. to Pet. for Cert. A–48. And the Republican and Democratic parties in Oklahoma, as it turns out, do not permit voters registered as Independents to participate in their primaries.
Most importantly, according to respondents, the deadline for changing party affiliation makes it quite difficult for the LPO to invite voters to reregister in order to participate in its primary. Assuming the LPO submits its petition for recognized party status on the May 1 deadline, the State has until May 31 to determine whether party status will be conferred. See Okla. Stat. Ann., Tit. 26, §1–108 (West Supp. 2005). But in order to participate in the LPO primary, a voter registered with another party must change her party affiliation to Independent or Libertarian no later than June 1. See §4–119. Moreover, no candidate for office is permitted officially to declare her candidacy with the State Election Board until the period between the first Monday in June and the next succeeding Wednesday. §5–110.
If this characterization of state law is accurate, a registered Democrat or Republican sympathetic to the LPO or to an LPO candidate in a given election year would seem to face a genuine dilemma. On the one hand, she may stick with her major party registration and forfeit the opportunity to participate in the LPO primary. Alternatively, she may reregister as a Libertarian or Independent, thus forfeiting her opportunity to participate in the major party primary, though no candidate will have officially declared yet and the voter may not yet know whether the LPO will even be permitted to conduct a primary. Moreover, she must make this choice roughly eight weeks before the primaries, at a time when most voters have not yet even tuned in to the election, much less decided upon a candidate. See California Democratic Party, 530 U. S., at 586 (Kennedy, J., concurring). That might pose a special difficulty for voters attracted to minor party candidates, for whom support may not coalesce until comparatively late in the election cycle. See Anderson v. Celebrezze, 460 U. S. 780, 791–792 (1983) (discussing emergence of independent candidacies late in the election cycle).
Throughout the proceedings in the lower courts, which included a full bench trial before the District Court, respondents made no attempt to challenge these other electoral requirements or to argue that they were relevant to respondents’ challenge to the semiclosed primary law. The lower courts, accordingly, gave little or no consideration to how these various regulations interrelate or operate in practice, nor did the State seek to justify them. Given this posture, I agree with the Court that it would be neither proper nor prudent for us to rule on the reformulated claim that respondents now urge. See ante, at 14–15.
Nevertheless, respondents’ allegations are troubling, and, if they had been properly raised, the Court would want to examine the cumulative burdens imposed by the overall scheme of electoral regulations upon the rights of voters and parties to associate through primary elections. A panoply of regulations, each apparently defensible when considered alone, may nevertheless have the combined effect of severely restricting participation and competition. Even if each part of a regulatory regime might be upheld if challenged separately, one or another of these parts might have to fall if the overall scheme unreasonably curtails associational freedoms. Oklahoma’s requirement that a voter register as an Independent or a Libertarian in order to participate in the LPO’s primary is not itself unduly onerous; but that is true only to the extent that the State provides reasonable avenues through which a voter can change her registration status. The State’s regulations governing changes in party affiliation are not properly before us now. But if it were shown, in an appropriate case, that such regulations imposed a weighty or discriminatory restriction on voters’ ability to participate in the LPO’s or some other party’s primary, then more probing scrutiny of the State’s justifications would be required.