These consolidated appeals involve the initiative process in Montana. Montana state officials and the League of Women Voters appeal the district court's decision, following a bench trial, that Initiative 125 (I-125), which prohibits direct corporate expenditures in ballot initiative campaigns, violates the First Amendment.2 The Montana Mining Association and various organizations subject to I-125 sought to delay, and then to invalidate, the election in which voters approved Initiative 137 (I-137) (restricting certain types of mining) on the ground that I-125 unconstitutionally constrained their participation in the election process. They appeal the district court's refusal to do either.3

We conclude that the constitutionality of I-125 is controlled by First National Bank of Boston v. Bellotti, 435 U.S. 765 (1978), and that the district court did not clearly err in its findings that corporate wealth had not distorted the ballot initiative process in Montana, and that in light of those findings, the First Amendment does not permit restricting corporate speech on public issues. Accordingly, we affirm the declaration that I-125 is unconstitutional.

Whether the court should have considered (and granted) MMA's request to delay the I-137 election is moot, and we cannot say that the court erred by later declining to invalidate it. We therefore affirm the judgment in the I-137 appeal as well.

* Montana voters approved I-125 in November 1996. I-125 amended Montana's elections law by prohibiting direct corporate spending in connection with ballot issues.4 Specifically, the initiative amended the statutory provision concerning "[p]rohibited contributions from corporations." Mont. Code Ann. S 13-35-227. Under I-125, corporations (other than nonprofit corporations formed for solely political purposes) are prohibited from making a contribution or an expenditure in connection with a ballot issue. Corporations may establish and administer a separate, segregated fund that is allowed to solicit contributions from shareholders, employees, or members of the corporation, but not from the company itself.

The Chamber brought an action in federal district court seeking a declaratory judgment that the initiative was unconstitutional and an injunction restraining its enforcement. The court held on summary judgment that I-125 restricted core political speech, but that a trial was necessary to determine whether a compelling state interest justified the restriction.

Meanwhile, in July 1998, I-137 was certified for the November 1998 ballot. MMA, which opposed I-137, brought suit in September 1998 requesting a preliminary injunction that would either waive I-125 as applied to it, or delay the vote on I-137 until after the I-125 case was resolved. The district court consolidated the two actions.

The I-125 trial focused on the health of the Montana initiative process. I-125 Proponents presented evidence on the effect of corporate money in four unsuccessful Montana initiatives. Witnesses testified that corporate opponents substantially outspent initiative proponents in these races. Advocates for the initiatives testified that they believed the defeats resulted from large corporate expenditures in opposition. Political scientists testified that large scale spending was very effective in initiative campaigns, especially when used in opposition to a ballot issue, and poll results showed that Montanans believe corporations had too much influence in elections. I-125 Opponents, on the other hand, produced evidence that a variety of factors influence election results and that the side spending less money prevailed in 50% of initiative elections. Further, they showed that Montana voter turnout was much higher than the national average, ballot drop-off was low, and the number of ballot issues remained constant over the past 20 years. Finally, I-125 Opponents offered expert testimony that the Montana political system was healthy and free from corruption.

In the I-137 phase, MMA showed that I-125 limited mining companies' ability to oppose I-137 and that I-137 was a significant economic threat to these companies. It also adduced expert testimony that a successful challenge to I-137 was no longer possible, even if the I-125 restriction were lifted, because the I-137 election was only two weeks away. Evi-dence on the other side showed that the mining companies spent approximately the same amount fighting I-137 as they did in their challenge to the Clean Water Bill, where the I-125 restrictions did not apply.

The district court found the I-125 Opponents' evidence credible and persuasive, accepting their expert's opinion that there is no corruption or appearance of corruption in Montana ballot issue elections. It held that I-125, perhaps facially and certainly as applied, infringes upon the First Amendment rights of speech and association of those subject to its prohibitions; that it was not narrowly tailored to address only the campaign contributions and expenditures of large corporations; that requiring corporations to fund ballot issue campaign speech through separate, segregated funds (consisting of voluntary contributions from employees, officers, directors, and shareholders) deprives corporations of their ability to communicate political ideas directly to the electorate, which impermissibly chills their speech and association rights and precludes corporations from directly resisting potential laws that could put them out of business; that it prevents the electorate from being exposed to diverse viewpoints on public policy issues; and that the anecdotal evidence presented by I125 proponents fails to prove that corporations could overwhelm the political speech of individual citizens in Montana to the detriment of the ballot initiative process. Accordingly, the court held, corporations are entitled to defend their economic interests by using the corporate treasury to fund their participation in ballot initiative campaigns. It therefore declared that I-125 is unconstitutional.

In this respect the court ruled in favor of MMA in the I-137 action, but it refused to enjoin the I-137 election on the ground that the relief sought was premature as of the close of evidence, October 22. On November 3, 1998, Montana voters adopted I-137 by a 53% to a 47% margin. MMA then moved to enjoin Montana from validating I-137 by certifying the election results, but the court indicated that it had already ruled and denied the motion.

Both judgments -that I-125 is unconstitutional and that the election in I-137 stands -were appealed.

By the time we heard oral argument in these consolidated appeals, the Supreme Court had granted certiorari in a campaign contribution case, Shrink Missouri Gov't PAC v. Adams, 161 F.3d 519 (8th Cir. 1998), cert. granted, 525 U.S. 1121 (1999). We thought it prudent to defer our own decision until the Court had rendered its in Shrink. As it turns out, the Court's opinion, Nixon v. Shrink Missouri Gov't PAC, 120 S. Ct. 897 (2000), does not affect the disposition in these appeals.

We determine the constitutionality of I-125 de novo. California Democratic Party v. Jones, 169 F.3d 646, 647 (9th Cir. 1999), rev'd on other grounds, 120 S. Ct. 2402 (2000). The district court's findings of fact are reviewed for clear error. Sana v. Hawaiian Cruises, Ltd., 181 F.3d 1041, 1044 (9th Cir. 1999). While we have not previously addressed the question, we agree with other circuits that have considered similar issues that review of the district court's decision not to delay or void the I-137 election should be for abuse of discretion. Gjersten v. Board of Election Comm'rs, 791 F.2d 472, 479 (7th Cir. 1986) (decision to invalidate election is reviewed for an abuse of discretion); Griffin v. Burns, 570 F.2d 1065, 1079 (1st Cir. 1978) (same).

Montana and League argue that corporations possess unique characteristics and legal advantages that permit them to maximize the accumulation of capital in the economic marketplace. They indicate that I-125 was designed to assure that corporate participation in ballot initiative campaigns reflects actual public support for the corporation's political views, rather than the sheer economic power that a corporation derives from these state-created advantages. To that end, they point out, I-125 prohibits the use of corporate general treasury funds in ballot issue campaigns, while it permits corporations to participate through the use of segregated funds collected from employees, members and shareholders who wish to support the corporation's political activities. Further, they submit, I-125 is carefully tailored to adhere to the Supreme Court's rulings in Austin v. Michigan Chamber of Commerce, 494 U.S. 652 (1990), and FEC v. Massachusetts Citizens for Life, Inc., 479 U.S. 238 (1986) (MCFL). They fault the district court's decision on a number of fronts: for failing to recognize the compelling state interests supporting I-125 in that the court believed the only type of corruption that would justify I-125's limits was quid pro quo corruption, whereas in Austin the Court declined to rely on that limited anti-corruption rationale; for holding that I-125 could be justified only by evidence that money was the most important variable controlling election outcome in Montana, instead of deferring to the legislative judgement that the special benefit conferred by the corporate form presents the potential for distorting the political process; and for finding that voter turnout in Montana is higher than the national average, that I-125 improperly "silences" corporations under Bellotti because they retain the right to speak through segregated funds, and that I-125 is unconstitutional even under Bellotti.

The Chamber counters that I-125 abridges core FirstAmendment rights of political speech and association on issues of public policy. This abridgment of political speech in the context of ballot issue campaigns (by contrast with candidate contributions or expenditures) is contrary to Bellotti and unjustified by any compelling state interest. Additionally, the Chamber maintains, there is no evidence of corruption or that corporate contributions or expenditures have reduced voter turnout, caused voter fall-off on the ballot, or been the principal determinant of the outcome of a ballot issue campaign; to the contrary, it submits, the evidence shows that the Montana electoral process is healthy and that the most important factor which influences outcome is the development of a credible, concise and understandable campaign message. Finally, it contends, the use of the segregated fund permitted by I-125 is neither a constitutionally permissible alternative to direct corporate speech nor an effective means to communicate core political speech on public policy issues.

We do not wish to appear to give either position short shrift, because all parties have researched, briefed and argued the issues with exceptional ability and professionalism. Yet as we see it, the constitutionality of I-125 comes down to whether restricting corporate expenditures in the ballot issue process is controlled by Bellotti, even though I-125 (unlike the Massachusetts statute at issue in Bellotti ) permits corporations to establish segregated funds through which others may contribute.

Like this case, Bellotti involved a limitation on corporate contributions or expenditures in the ballot issue process (there, in connection with a referendum). A Massachusetts statute prohibited such expenditures "for the purpose of . . . influencing or affecting the vote on any question submitted to the voters, other than one materially affecting any of the property, business or assets of the corporation," and further provided that no issue submitted to the voters which solely concerned individual taxation issues could be deemed to affect a corporation's property, business or assets. When a graduated income tax proposal was put on the ballot, several corporate entities wanted to spend money to publicize their views in opposition. Applying strict scrutiny and requiring the state to show both a compelling interest in prohibiting the "exposition of ideas" and narrowly chosen means to avoid unnecessary abridgment, the Court held that the statute must be invalidated. In doing so, it clearly distinguished elections involving issues from elections involving candidates. As Montana does here, the state argued in Bellotti that preserving the integrity of the electoral process, preventing corruption, and sustaining the active involvement of citizens in good government were important state interests that were threatened by corporate participation in discussion of a referendum issue. The Court noted that if the state's

arguments were supported by record or legislative findings that corporate advocacy threatened immi nently to undermine democratic processes, thereby denigrating rather than serving First Amendment interests, these arguments would merit our consider ation. But there has been no showing that the relative voice of corporations has been overwhelming or even significant in influencing referenda in Massa chusetts, or that there has been any threat to the con fidence of the citizenry in government.

Bellotti, 435 U.S. at 789-90 (citations omitted). It added:

Nor are [the state's] arguments inherently persuasive or supported by the precedents of this Court. Refer enda are held on issues, not candidates for public office. The risk of corruption perceived in cases involving candidate elections simply is not present in a popular vote on a public issue. To be sure, corpo rate advertising may influence the outcome of the vote; this would be its purpose. But the fact that advocacy may persuade the electorate is hardly a reason to suppress it: The Constitution "protects expression which is eloquent no less than that which is unconvincing."

Id. at 790 (citations omitted).5 Concluding that the statute prohibited protected speech in a manner unjustified by a compelling state interest, the Court invalidated it.

In Austin, the Court considered a Michigan statute that prohibited corporations from using corporate treasury funds for independent expenditures in support of, or in opposition to, any candidate in elections for state office, but did not prohibit corporations from making such expenditures from segregated funds used solely for political purposes. The Court acknowledged that requiring corporations to make independent political expenditures only through special segregated funds burdens corporate freedom of expression because "the corporation is not free to use its general funds for campaign advocacy purposes," Austin, 494 U.S. at 658 (quoting MCFL, 479 U.S. at 252), but held that the state's rationale in supporting a regulation aimed at "the corrosive and distorting effects of immense aggregations of wealth that are accumulated with the help of the corporate form and that have little or no correlation to the public's support for the corporation's political ideas" was compelling. Austin, 494 U.S. at 660. The Court concluded that "the State's decision to regulate only corporations [was] precisely tailored to serve the compelling state interest of eliminating . . . the corrosive effect of political `war chests.' " Id. at 666. It therefore upheld Michigan's restriction on independent corporate expenditures in connection with candidate elections for state office.

Justice Brennan, who concurred in the majority's opinion, wrote separately, distinguishing the particular provisions at issue in Austin that prohibit corporations from using treasury funds only for independent expenditures in candidate elections, from the restrictions at issue in Bellotti, noting that "[a] corporation remains free, for example, to use general treasury funds to support an initiative proposal in a state referendum." Id. at 676 (Brennan, J. concurring). So did Justice Stevens, who observed that, "as we recognized in First National Bank of Boston v. Bellotti, 435 U.S. 765, 98 S. Ct. 1407, 55 L. Ed. 2d 707 (1978), there is a vast difference between lobbying and debating public issues on the one hand, and political campaigns for election to public office on the other. " Austin, 494 U.S. at 678 (Stevens, J., concurring).

Montana and League are correct that I-125 is similar to the statutory scheme approved in Austin to the extent that both statutes allow corporations to set up segregated funds. However, as we read the Court's opinions, Austin does not turn on this difference from Bellotti so much as it does on the difference between expenditures for candidate elections and ballot issues. As the Court explained, in the context of candidate elections, "[c]orporate wealth can unfairly influence elections when it is deployed in the form of independent expenditures, just as it can when it assumes the guise of political contributions." Id. at 660. It is also true that the Court indicated in Bellotti that through a record or findings, Congress or a state might be able to demonstrate the danger of real or apparent corruption posed by corporate expenditures, 435 U.S. at 787 n. 26 and 789-90, whereas in Austin the Court appears to have accepted the state's articulated interest in avoiding corruption without a record or findings, 494 U.S. 659-60, 666. But again, this appears to reflect the difference between restrictions on the corporate voice in public issue elections, and in partisan candidate elections.

Even if Austin may plausibly be read as undermining Bellotti, this is for the Supreme Court, not us, to say. As the Court has instructed, " `[i]f a precedent of this Court has direct application in a case, yet appears to rest on reasons rejected in some other line of decisions, the Court of Appeals should follow the case which directly controls, leaving to this Court the prerogative of overruling its own decisions.' " Agostini v. Felton, 521 U.S. 203, 237 (1997) (quoting Rodriquez de Quijas v. Shearson/American Express, Inc., 490 U.S. 477, 484 (1989)). Austin cited Bellotti and did not overrule it. Therefore, we believe that Bellotti is still good law, and controls.

There is no question that a law requiring corporations to make independent expenditures (even for candidates) through a segregated fund burdens corporate expression. Austin and MCFL so recognize. The burden is even greater when the limitation has to do with independent expenditures in the ballotissue process, as speech on such issues is "at the heart of the First Amendment's protection." Bellotti, 435 U.S. at 776.

Because the state's interest in preventing corporate wealth from distorting the political process is not obvious or compelling in connection with ballot issue elections, Bellotti, 435 U.S. at 789-92, the district court in this case heard considerable evidence, both empirical and expert. It found that "the state has failed to produce evidence that would support a judicial finding that corporate wealth has dominated citizen voices to the detriment of the ballot initiative process." As there was evidence pointing in each direction, we cannot say the court clearly erred in finding that there was no imminent threat to the democratic process.

It follows that I-125 unconstitutionally restricts public discussion in the ballot issue (initiative) process. As the Court concluded in Bellotti with respect to the Massachusetts statute, we conclude here that, " `[a] restriction so destructive of the right of public discussion [as I-125], without greater or more imminent danger to the public interest than existed in this case, is incompatible with the freedoms secured by the First Amendment.' " Id. at 792 (quoting Thomas v. Collins, 323 U.S. 516, 537 (1945)).6 Its enforcement was, therefore, properly enjoined.

Having determined that I-125 is unconstitutional, we must decide whether the district court should have delayed or invalidated the I-137 election. Although we have previously held that challenges to election procedures should be made before the election occurs, see Soules v. Kauaians for Nukolii Campaign Committee, 849 F.2d 1176, 1180-81 (9th Cir. 1988) (post-election challenge came too late); Chinese for Affirmative Action v. Leguennec, 580 F.2d 1006, 1008 (9th Cir. 1978) (parties must bring complaints forward for preelection adjudication), we have not expressly considered whether it is error for a court to refuse to resolve such challenges on the footing that a pre-election challenge is premature. Nor do we need to do so now, for this issue, together with the question whether MMA's pre-election request for injunctive relief should have been granted, are moot.

This leaves only the question whether the district court erred in refusing to invalidate the election. A state election violates due process "if it is conducted in a manner that is fundamentally unfair." Bennett v. Yoshina, 140 F.3d 1218, 1226 (9th Cir. 1998). Of course, "the voiding of a state election isa `drastic if not staggering' remedy." Soules, 849 F.2d at 1180 (quoting Bell v. Southwell, 376 F.2d 659, 662 (5th Cir. 1967)). In determining whether to void an election, we balance the constitutional violation against the "countervailing equitable factors" such as "the extremely disruptive effect of election invalidation and the havoc it wreaks upon local political continuity." Soules, 849 F.2d at 1180.

Here, even though there was evidence that I-125 had affected MMA's ability to campaign, there was also evidence that it had no substantial impact. Eleven days remained before the election after the district court lifted the I-125 restrictions. And the state has a significant interest in avoiding the costs of a special election. In these circumstances, we cannot say that the district court abused its discretion in failing to void the results of the election.

AFFIRMED.

I join in the opinion but write separately to underscore that First Amendment protection of political contributions is not absolute. Here, in light of the specific evidence presented and in light of the requirements of First National Bank of Boston v. Bellotti, 435 U.S. 765 (1978), the state failed to sustain its burden of articulating a compelling interest for the restrictions. This is not to say, however, that such compelling interests do not exist, even in the context of initiative elections. The Supreme Court's recent treatment of the evidentiary burden is instructive:

The quantum of empirical evidence needed to satisfy heightened judicial scrutiny of legislative judgments will vary up or down with the novelty and plausibil ity of the justification raised. Buckley demonstrates that the dangers of large, corrupt contributions and the suspicion that large contributions are corrupt are neither novel nor implausible.

Nixon v. Shrink Missouri Government PAC, 120 S. Ct. 897, 907 (2000). Although Shrink concerned candidate elections, its reasoning on judicial scrutiny of state-asserted compelling interests applies with similar force to campaigns for ballot measures.

Legitimate policy concerns have prompted legislators toaddress the power of the corporate contributor. I believe that a statute may be carefully crafted and tailored to meet a compelling interest in preserving the integrity of the electoral process, such as where contributions so distort the political process that they undermine our democratic system. But that is a case for another day.

We occasionally hear a case falling so squarely between two conflicting Supreme Court opinions that no matter which way we veer, we are likely to run aground. This is one of those cases. With Bellotti and Austin , the Supreme Court has laid out a difficult course for us to navigate in determining whether restrictions on corporate campaign spending are constitutional. Judge Rymer works her way admirably through this challenge, but in the end I am not convinced. At the risk of running off course myself, I choose a different line because I think it is more consistent with the development of Supreme Court precedent in this area and reflects a more sensitive understanding of the ways in which corporate spending candistort the electoral process. Although certainly not dispositive, this approach has the added benefit of upholding the expressed discernment of the people of Montana about these matters.

In Buckley v. Valeo, 424 U.S. 1 (1976), the Court's first major campaign spending case, the plaintiffs challenged the Federal Election Campaign Act, which placed dollar limits on political contributions to candidates and on individual expenditures "relative to a clearly identified candidate."1 The Court held that the contribution limits were constitutional because the government has a compelling interest in eliminating corruption (or its appearance) in the electoral process, and the restrictions did not completely preclude individuals or groups from expressing their support of a particular candidate. See id. at 26-29. But the limits on individual expenditures were struck down. These restrictions, the Court held, could not be justified by the government's interest in battling corruption because independent spending in candidate elections presents little risk of quid pro quo bargaining. See id. at 45-47.

Shortly afterward, the Court decided First Nat'l Bank of Boston v. Bellotti, 435 U.S. 765 (1978), which involved a complete ban on corporate spending in referenda campaigns. Here, as in Buckley, the government argued that the restriction was justified by its interest in battling corruption and in preserving the integrity of the electoral process. See id. at 788-89. The Court disagreed. "Referenda are held on issues, not candidates for public office," the Court stated. See id. at 790. And while a candidate may offer political favors in return for campaign contributions, a referendum cannot. As a result, "[t]he risk of corruption perceived in cases involving candidate elections simply is not present in a popular vote on a public issue." Id.

Buckley and Bellotti focused much attention on the government's asserted interest in preventing quid pro quo corruption. Indeed, in a subsequent case, the Court stated that "preventing corruption or the appearance of corruption are the only legitimate and compelling government interests thus far identified for restricting campaign finances." Federal Election Comm'n v. National Conservative Political Action Comm., 470 U.S. 480, 496-97 (1985).

But twelve years after Bellotti, the Court was presented with a new asserted interest. In Austin v. Michigan Chamber of Commerce, 494 U.S. 652 (1990), a state statute prohibited corporations from using general treasury funds to make independent expenditures in candidate elections, but allowed them to set up segregated campaign funds to which employees, shareholders, and members could contribute. See id. at 655-56. Because the statute targeted campaign expenditures, not contributions, the state could not justify it as a measure tofight quid pro quo corruption. See Buckley, 424 U.S. at 45-47. So instead, the state argued that "the unique legal and economic characteristics of corporations" created the risk of a different kind of corruption not based on political dealmaking.

Four years earlier, the Court had itself identified this "new" form of corruption. In FEC v. Massachusetts Citizens for Life, Inc., 479 U.S. 238, 257 (1986) ("MCFL"), the Court explained that state laws grant corporations special advantages that not only allow them to play a dominant role in the economy, but permit them to use "resources amassed in the economic marketplace" to obtain "an unfair advantage in the political marketplace." The Court elaborated:

The resources in the treasury of a business corpora tion, however, are not an indication of popular sup port for the corporation's political ideas. They reflect instead the economically motivated decisions of investors and customers. The availability of these resources may make a corporation a formidable political presence, even though the power of the cor poration may be no reflection of the power of its ideas.

Id. at 258.

The Court nonetheless struck down the corporate spending limits in MCFL as applied because the plaintiff was an ideological organization "formed to disseminate political ideas, not to amass capital." Id. at 259. But the plaintiff in Austin, the Michigan Chamber of Commerce, possessed several characteristics that made it more like a corporation and less like a nonprofit ideological group. See Austin, 494 U.S. at 662665. So the Court applied the analysis it had articulated in MCFL to uphold the statute.

Michigan's statute, the Court made clear, did not aim at traditional quid pro quo corruption. See id. at 659. It targeted "a different type of corruption in the political arena: the corrosive and distorting effects of immense aggregations of wealth that are accumulated with the help of the corporate form and that have little or no correlation to the public's support for the corporation's political ideas." Id. at 660. "The unique stateconferred corporate structure that facilitates the amassing of large treasuries warrants the limit on independent expenditures," the Court stated. Id. "Corporate wealth can unfairly influence elections when it is deployed in the form of independent expenditures, just as it can when it assumes the guise of political contributions. We therefore hold that the state has articulated a sufficiently compelling rationale to support itsrestriction on independent expenditures by corporations." Id.

Austin did not expressly overrule Bellotti; in fact, Justice Marshall's opinion remained conspicuously silent about the relevance of its holding for state referenda.2 But most observers have concluded that the Court's analysis undermines the distinction between candidate elections and referenda, which was the key to Bellotti. Although the traditional quid pro quo corruption may not be a problem in referenda, it is difficult to see why the distorting effects of corporate wealth -the "new corruption," as it has been called -would not be equally problematic in both contexts. Once Austin was decided, therefore, the natural conclusion was that a similar restriction on corporate spending in state referenda would be upheld in spite of Bellotti. See Thomas C. Goldstein, Corporate Influence in Referenda: A Comment About the Prescription, 1996 Ann. Surv. Am. L. 469, 473 (1996) (stating that "the Court in Austin substantially reversed course from Bellotti" and that "the expenditures in Austin, which were not coordinated with the candidate, were the rough equivalent of spending in referenda."); James B. Raskin, Direct Democracy, Corporate Power and Judicial Review of Popularly-Enacted Campaign Finance Reform, 1996 Ann. Surv. Am. L. Rev. 393, 408 (stating that a ban on corporate spending in ballot initiatives is "perfectly congruent with the Court's new and far more nuanced conception of `corruption' "); Gerald G. Ashdown, Controlling Campaign Spending and the New Corruption: Waiting for the Court, 44 Vanderbilt Law Review 767, 779 (1991) ("The conclusion is inescapable that legislatures are now free to restrict corporations to spend only from separate political funds in ballot measures as well as candidate elections.") (emphasis added); David Cole, First Amendment Antitrust: The End of Laissez-Faire in Campaign Finance, 9 Yale L. & Pol'y Rev. 236, 252 (1991) ("[I]f corruption includes the distorting effects of large amounts of wealth on the political debate, corruption is no less troubling in referenda than in candidate elections. Thus, this distinction also seems destined to fall in light of the Austin Court's redefinition of corruption.").

Judge Rymer resists this conclusion. Although the Montana initiative is identical to the Austin statute in that it permits corporations to spend from a segregated fund -something the Bellotti statute did not allow -she asserts that the relevant distinction is between laws that regulate spending in candidate elections and laws that target spending in ballot initiatives or referenda. (Majority Op. at 12341). She does not explain how this traditional distinction survives Austin's identification of a new form of corruption unrelated to quid pro quo bargaining. She only quotes Austin's statement that "[c]orporate wealth can unfairly influence elections when it is deployed in the form of independent expenditures, just as it can when it assumes the guise of political contributions." (Majority Op. at 12341). This statement, however, does not suggest that corporate wealth plays a greater role in candidate elections than in referenda. It merely describes the effect ofcorporate wealth in elections generally.

Judge Rymer appears to recognize this difficulty. As a result, she attempts to find support for her position in the concurring opinions of Justices Brennan and Stevens in Austin. She suggests that Justice Brennan relied on the same distinction between candidate elections and referenda campaigns that she relies on here. Justice Brennan, she points out, noted that under the Austin statute, "a corporation remains free, for example to use general treasury funds to support an initiative proposal in a state referendum." (Majority Op. at 12340). The implication is that Justice Brennan's vote turned on the limited scope of the Austin statute.

A closer reading of Justice Brennan's statement undermines this implication. When Justice Brennan made this statement, he was responding to the dissent's claim that the statute was under-inclusive. He first acknowledged that the statute "is concededly under-inclusive insofar as it does not ban other types of political expenditures to which a dissenting . . . corporate shareholder might object." Austin, 494 U.S. at 675-76. He then explained: "The particular provision at issue prohibits corporations from using treasury funds only for making independent expenditures in support of, or in opposition to, any candidate in state elections. A corporation remains free, for example, to use general treasury funds to support an initiative proposal in a state referendum." Id. at 676.

The context of Justice Brennan's statement makes clear that he did not think the limited scope of the statute (i.e. that it did not restrict corporate spending in state referenda) was what saved it. In fact, he recognized that the statute's underinclusiveness was problematic. He then argued that this under-inclusiveness was not fatal because candidate elections are at the heart of political debate and the state can choose to single out corporate spending in that context. See id. at 676-77. He never suggested, however, that the state could not go further and limit corporate spending in referenda or ballot initiatives. Nor did he suggest that what distinguished Austin from Bellotti was that the latter involved referenda. To the contrary, he explained that Bellotti was different because it involved a complete ban on corporate spending, while Austin allowed for a segregated fund. See id. at 670 n.1 ("In [MCFL], we observed that the requirement that expenditures be made through PACs `is of course distinguishable from the complete foreclosure of any opportunity for political speech that we invalidated in First National Bank of Boston v. Bellotti.").

Justice Stevens did distinguish Bellotti from Austin on the ground that one involved referenda and the other involved candidate elections. See id. at 678. But that is because he rejected Buckley's conclusion that restrictions on campaign expenditures were not justified by the government's interest in preventing quid pro quo corruption. See id. Therefore, he did not need to rely on the government's new asserted interest -that corporate wealth distorts the electoral process. And because he relied on the traditional corruption rationale, he could make the same distinction the Court made in Bellotti. Justice Stevens' views are interesting, but because he was the sixth vote to uphold the statute, they shed little light on the majority's holding.

Finally, Judge Rymer argues that when a Supreme Court opinion directly applies to a case, we must follow it even if the opinion's reasoning was undermined by a later Court decision. (Majority Op. at 12341). This is a serious argument no doubt. Bellotti has not been overruled, and the facts of this case are similar to Bellotti: both involve restrictions on corporate spending in state referenda. The problem is that while the facts of the two cases are similar, there is also a significant difference. The Montana initiative -like the statute in Austin -allows for corporate spending through a segregated fund, while the Bellotti statute banned corporate spending entirely. Thus, to say we must follow the case that directly applies begs the question: which case applies when both share a common and important feature with the case before us? In such a situation, we should not simply assert that one case directly applies and then ignore the other. We must do our best to reconcile the two. And in my opinion, the best way to reconcile these cases is to acknowledge that Austin identified a new rationale for limiting corporate campaign spending that does not turn on whether candidate elections or ballot initiatives are at issue. In addition, the statute in Austin was less objectionable than the statute in Bellotti because it allowed for corporate spending through a segregated fund. Because the initiative also allows for a segregated fund, I think it is justified by Montana's asserted interest in eliminating what its people have determined to be distorting effects of corporate wealth on the electoral process.