OPINION

BYBEE, Circuit Judge:

Since 1994, Montana has regulated the amount that individuals, political committees, and political parties can contribute to candidates for state office. Mont.Code Ann. § 13-37-216, as adjusted by Admin. R. Mont. § 44.10.338.1 In 2003, we upheld this provision against a constitutional challenge based on Buckley v. Valeo, 424 U.S. 1, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976), and Nixon v. Shrink Missouri Government PAC, 528 U.S. 377, 120 S.Ct. 897, 145 L.Ed.2d 886 (2000). Mont. Right to Life Ass'n v. Eddleman, 343 F.3d 1085 (9th Cir.2003), cert. denied, 543 U.S. 812, 125 S.Ct. 47, 160 L.Ed.2d 16 (2004). Applying the "analytical framework set forth in Buckley and [Shrink Missouri]," we held that "Montana's interest in purging corruption and the appearance of corruption from its electoral system is sufficiently important to withstand constitutional scrutiny" and that § 13-37-216 was "closely tailored to achieving those ends." Id. at 1098. We concluded that § 13-37-216 was "constitutional and [did] not violate the First Amendment." Id.

On October 3, 2012, with less than five weeks before the general election and after absentee voting in Montana began, the district court concluded that "Montana's contribution limits in Montana Code Annotated § 13-37-216 are unconstitutional under the First Amendment." Order, Lair v. Murry, No. CV 12-12-H-CCL, 2012 WL 4815411, at *1 (D.Mont. Oct. 3, 2012) [hereinafter Order]. The district court permanently enjoined Montana from enforcing its campaign contribution limits. Id. at *2. In an opinion and order issued on October 10, 2012, the district court explained that our decision in Eddleman was "not binding on this Court because the U.S. Supreme Court's intervening decision in Randall [v. Sorrell, 548 U.S. 230, 126 S.Ct. 2479, 165 L.Ed.2d 482 (2006),] compels a different outcome." Opinion and Order, Lair v. Murry, No. CV 12-12-H-CCL, 2012 WL 4815411, at *10 (D.Mont. Oct. 10, 2012).

The State of Montana has sought a stay of the district court's order pending appeal. For the reasons we explain below, we believe that the state is likely to succeed on appeal. We conclude that the State of Montana has made a strong showing that a merits panel of this Court will likely conclude that, absent en banc proceedings or an intervening decision of the Supreme Court, we remain bound by our decision in Eddleman. See Miller v. Gammie, 335 F.3d 889, 892-93 (9th Cir.2003) (en banc). We also conclude that a merits panel is likely to hold that the analytical framework of the Supreme Court's decision in Randall does not alter the analysis of Buckley or Shrink Missouri in a way that affects our decision in Eddleman, for three reasons. First, there is no opinion of the Court in Randall. Thalheimer v. City of San Diego, 645 F.3d 1109, 1127 n. 5 (9th Cir.2011) ("[T]he plurality opinion [in Randall] [i]s persuasive authority, though not a binding precedent." (internal quotation marks omitted)). Second, even if we thought that Justice Breyer's plurality opinion represented the narrowest view of a majority of the Court, it did not depart from the principles of Buckley and Shrink Missouri that we applied in Eddleman. Randall, 548 U.S. at 242, 126 S.Ct. 2479 (opinion of Breyer, J.) ("[T]his Court has repeatedly adhered to Buckley's constraints...."). Third, even if we applied Randall to § 13-37-216, we cannot find, on the basis of the district court's findings, reason to disagree with, much less overturn, Eddleman. In light of Montana's interest in regulating campaign contributions, the lack of evidence that other parties will be substantially injured, and the public's substantial interest in the stability of its electoral system in the final weeks leading to an election, we will stay the order pending the state's appeal. See Nken v. Holder, 556 U.S. 418, 434, 129 S.Ct. 1749, 173 L.Ed.2d 550 (2009).

I. PROCEEDINGS BELOW

The plaintiffs-appellees, various individuals, political action committees, and other political organizations, brought suit in September 2011 to challenge several provisions of Montana's finance and election laws. The defendants-appellants are various officials of the State of Montana. Only one provision, § 13-37-216 of the Montana Code Annotated, which limits contributions that individuals and political committees can make to candidates, is at issue in this case. The district court held a bench trial on September 12-14, 2012. On October 3, 2012, the district court issued a brief order recounting the procedural history of the suit and the fact of the bench trial. The court stated that "[h]aving reviewed and considered the entire record and the parties' arguments and evidence, the Court concludes that Montana's contribution limits in Montana Code Annotated § 13-37-216 are unconstitutional under the First Amendment." Order at 4. The court permanently enjoined the enforcement of § 13-37-216. The district court did not issue an opinion, but stated that "complete and extensive findings of fact and conclusions of law that support this order" would be filed separately. Order at 5. The order was filed before it issued the findings of fact and conclusions of law "so that th[e] order c[ould] be issued before voting begins in the upcoming election." Id.

The following day, October 4, 2012, the state defendants-appellants filed for a stay pending appeal. We ordered an expedited response from the plaintiffs-appellees, which they filed on October 9, 2012. That same day, noting that the district court had not issued findings and conclusions, we found that we were "severely constrained in [our] consideration of the underlying issues raised in the emergency motion." Order, Lair v. Murry, No. 12-35809, 2012 WL 4883247, at *2 (9th Cir. Oct. 9, 2012). We nevertheless ordered that the injunction be "temporarily stayed pending further order of the court." Id. at *2.

The district court issued an Opinion and Order containing its findings of fact and conclusions of law on October 10, 2012. The state filed a reply in support of its motion for a stay on October 11, 2012.

II. STANDARD OF REVIEW

"A stay is not a matter of right.... It is instead `an exercise of judicial discretion' ... [that] `is dependent upon the circumstances of the particular case.'" Nken, 556 U.S. at 433, 129 S.Ct. 1749 (internal citations omitted) (quoting Virginian Ry. Co. v. United States, 272 U.S. 658, 672-73, 47 S.Ct. 222, 71 L.Ed. 463 (1926)). Judicial discretion in exercising a stay is to be guided by the following legal principles, as distilled into a four factor analysis in Nken: "(1) whether the stay applicant has made a strong showing that he is likely to succeed on the merits; (2) whether the applicant will be irreparably injured absent a stay; (3) whether issuance of the stay will substantially injure the other parties interested in the proceeding; and (4) where the public interest lies." Id. at 434, 129 S.Ct. 1749 (citing Hilton v. Braunskill, 481 U.S. 770, 776, 107 S.Ct. 2113, 95 L.Ed.2d 724 (1987)).2 "The party requesting a stay bears the burden of showing that the circumstances justify an exercise of [this Court's] discretion." Id. at 433-34, 129 S.Ct. 1749.

III. DISCUSSION

As discussed in detail below, we find that the State of Montana has satisfied this burden. As the Nken factors illustrate, especially in light of the delicate campaign contribution equilibrium leading up to the imminent election, we should and will exercise our discretion to stay the district court's order pending resolution of the appeal by a merits panel of this court.

A. Strong Showing that Success is Likely on the Merits

The first two Nken factors "are the most critical." Id. at 434, 129 S.Ct. 1749. Regarding the first factor, Nken held that it is not enough that the likelihood of success on the merits is "better than negligible" or that there is a "mere possibility of relief." Id. (internal quotation marks omitted). Since Nken did not specify "the exact degree of likely success that stay petitioners must show, ... courts routinely use different formulations to describe this [factor]." Leiva-Perez v. Holder, 640 F.3d 962, 966 (9th Cir.2011) (per curiam). We have concluded that many of these formulations, including "reasonable probability," "fair prospect," "substantial case on the merits," and "serious legal questions ... raised," are largely interchangeable. Id. at 967-68. All of these formulations indicate that, "at a minimum," a petitioner must show that there is a "substantial case for relief on the merits." Id. at 968. The standard does not require the petitioners to show that "it is more likely than not that they will win on the merits." Id. at 966.

We find that the State of Montana has met its burden to make a strong showing that success on the merits is likely. In 2003, we specifically considered the constitutionality of the Montana statute at question here. Eddleman, 343 F.3d at 1092-96. Our decision in Eddleman stands as a barrier to be overcome, a barrier that works significantly to the State of Montana's advantage. The plaintiffs in this case do not argue that anything has fundamentally changed in Montana political campaigns since our decision in Eddleman that would call into question our conclusions made in 2003. In fact, the evidence presented before the district court in this case appears quite similar to the evidence that was presented in Eddleman. The only change in circumstance pointed to by the plaintiffs is the Supreme Court's decision in Randall. The presumption is that our holding in Eddleman is controlling in this case, see Miller, 335 F.3d at 892-93, and we find that Randall does not overcome this presumption. Randall is not binding authority because there was no opinion of the Court. Further, even if we looked to Justice Breyer's plurality opinion in Randall, it is not clearly irreconcilable with the pre-existing law that we applied in Eddleman. Finally, even if we apply Randall, our limited review suggests that Randall would not compel a result different from Eddleman. This is particularly the case given the points of tension and possible errors that we find on the face of the district court's Opinion and Order. Therefore, taken as a whole, and based upon our limited review, necessitated by the imminent election, we conclude that the State of Montana has made a "substantial case for relief on the merits."

1. Whether Randall has a majority opinion

Marks v. United States held that "[w]hen a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Justices, the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds." 430 U.S. 188, 193, 97 S.Ct. 990, 51 L.Ed.2d 260 (1977) (internal quotation marks omitted). The Supreme Court has acknowledged that in some cases "[t]his test is more easily stated than applied," and that under certain circumstances it may not be "useful to pursue the Marks inquiry to the utmost logical possibility." Nichols v. United States, 511 U.S. 738, 745-46, 114 S.Ct. 1921, 128 L.Ed.2d 745 (1994) (recognizing that where the application of the Marks test to a prior splintered decision "ha[d] so obviously baffled and divided the lower courts that ha[d] considered it," there is reason to reexamine that prior decision).

Likewise, we have also held that the Marks standard is not always helpful, and should only be applied "where one opinion can be meaningfully regarded as narrower than another and can represent a common denominator of the Court's reasoning." United States v. Rodriguez-Preciado, 399 F.3d 1118, 1140 (9th Cir.) (internal quotation marks omitted) (citing other circuits that have held similarly), amended by 416 F.3d 939 (9th Cir.2005). This standard requires that the narrowest opinion is actually the "logical subset of other, broader opinions," such that it "embod[ies] a position implicitly approved by at least five Justices who support the judgment." Id. (quoting King v. Palmer, 950 F.2d 771, 781 (D.C.Cir.1991) (en banc)); see also United States v. Williams, 435 F.3d 1148, 1157 (9th Cir.2006) (explaining that Marks requires us to find a "legal standard which, when applied, will necessarily produce results with which a majority of the Court from that case would agree"). If there is no such narrow opinion, "the only binding aspect of a splintered decision is its specific result." Rodriguez-Preciado, 399 F.3d at 1140.

Randall is the epitome of a splintered decision. Although six Justices ultimately concurred in the judgment, the case generated six opinions, four of which were required for the six Justices to concur in the judgment. Since the opinions of both Justices Kennedy and Thomas would revisit — or, as preferred by Justices Thomas and Scalia, overrule Buckley, Justice Breyer's plurality decision offers the narrowest rationale in support of the judgment. See Randall, 548 U.S. at 265, 126 S.Ct. 2479 (Kennedy, J., concurring in the judgment) ("Viewed within the legal universe we have ratified and helped create, the result the plurality reaches is correct; given my own skepticism regarding that system and its operation, however, it seems to me appropriate to concur only in the judgment."); id. at 265-66, 126 S.Ct. 2479 (Thomas, J., concurring in the judgment) ("Although I agree with the plurality that [the Vermont contribution limit statute] is unconstitutional, I disagree with its rationale for striking down that statute.... I continue to believe that Buckley provides insufficient protection to political speech, the core of the First Amendment.... [S]tare decisis should pose no bar to overruling Buckley and replacing it with a standard faithful to the First Amendment.").

It cannot be said, however, that Justice Breyer's plurality opinion represents a "common denominator of the Court's reasoning," enjoying the assent of five Justices. Justices Thomas and Scalia would "overrule Buckley and subject both the contribution and expenditure restrictions of [the Vermont statute] to strict scrutiny, which they would fail." Id. at 267, 126 S.Ct. 2479 (Thomas, J., concurring in the judgment). Thus, further consideration of Justice Kennedy's position is irrelevant for our purposes, since at most Justice Breyer's rationale could only garner the assent of four Justices. If Justice Kennedy's position were relevant to this inquiry, however, his "skepticism regarding that system and its operation," coupled with his previously asserted criticism of Buckley, strongly suggests that only three Justices assented to Justice Breyer's rationale. Id. at 265, 126 S.Ct. 2479 (Kennedy, J., concurring in the judgment); see also Shrink Missouri, 528 U.S. at 409-10, 120 S.Ct. 897 (Kennedy, J., dissenting) ("I would overrule Buckley.... The First Amendment ought to be allowed to take its own course without further obstruction from the artificial system we have imposed. It suffices here to say that the law in question does not come even close to passing any serious scrutiny.").

This analysis is consistent with our previous recognition — a holding binding upon this Court, see Miller, 335 F.3d at 892-93 — that no position in Randall garnered the support of more than three Justices. Thalheimer, 645 F.3d at 1127 & n. 5 (explaining that "Justice Breyer's plurality opinion announced the judgment of the Court," so "we follow the plurality opinion as persuasive authority, though not a binding precedent" since "Justice Breyer's plurality opinion was [only] joined by two justices, one in full and one in part" (internal quotation marks omitted)). The only binding aspect of Randall, then, is its judgment, striking down the Vermont contribution limit statute as unconstitutional. Since Randall is otherwise only persuasive, in this context it could not have altered the law as previously dictated by such cases as Buckley and Shrink Missouri, the law we expressly relied upon in Eddleman.

2. Whether Justice Breyer's opinion alters Buckley

Even if Justice Breyer's plurality did represent a majority opinion under Marks, however, Randall is not irreconcilable with the principles of Buckley and Shrink Missouri. In Miller v. Gammie, sitting en banc, we considered the question of "when a three-judge panel may reexamine normally controlling circuit precedent in the face of an intervening United States Supreme Court decision." Miller, 335 F.3d at 892. We held that "where the reasoning or theory of our prior circuit authority is clearly irreconcilable with the reasoning or theory of intervening higher authority, a three-judge panel should consider itself bound by the later and controlling authority, and should reject the prior circuit opinion as having been effectively overruled." Id. at 893. We further held that "the issues decided by the higher court need not be identical in order to be controlling." Id. at 900. We made it clear that this standard applies not only to three-judge panels but also to district courts within this circuit. Id. at 899 (describing prior circuit decisions effectively overruled based on higher intervening authority as "no longer binding on district judges and three-judge panels of this court"); see also Day v. Apoliona, 496 F.3d 1027, 1031 (9th Cir.2007) ("The Miller standard is thus not met, and we (and the district court) are bound by our earlier precedent.").

Since Miller, we have elaborated on this standard. Recently, in In re Flores, we explained that "we are bound by our prior precedent if it can be reasonably harmonized with the intervening authority." In re Flores, 692 F.3d 1021, 1030 (9th Cir.2012). In that case, we explained that under Miller, we were compelled to defer to prior circuit precedent because (1) the "overall analytical framework" of the intervening Supreme Court case was "consistent with our overall analytical approach" in prior circuit precedent, id. at 1030-31, and (2) the specific application of that framework in the intervening Supreme Court case did not mandate a result in the prior case in conflict with the decision rendered by this Court in that case. Id. at 1030-38. As Flores' first consideration suggests, "Miller v. Gammie ... instructs us to focus on the reasoning and analysis in support of a holding, rather than the holding alone." United States v. Lindsey, 634 F.3d 541, 550 (9th Cir.2011); see also Miller, 335 F.3d at 900 (citing Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. Chi. L.Rev. 1175, 1177 (1989)) (favorably discussing Justice Scalia's assertion in a law review essay that lower courts are bound by higher courts' "mode of analysis," not just their holdings).

Although we should consider the intervening authority's reasoning and analysis, as long as we can apply our prior circuit precedent without "running afoul" of the intervening authority, we must do so. United States v. Orm Hieng, 679 F.3d 1131, 1140 (9th Cir.2012). It is not enough for there to be "some tension" between the intervening higher authority and prior circuit precedent, id. at 1140-41, or for the intervening higher authority to "cast doubt" on the prior circuit precedent, United States v. Delgado-Ramos, 635 F.3d 1237, 1239 (9th Cir.2011). The intervening higher precedent must be "clearly inconsistent" with the prior circuit precedent. Orm Hieng, 679 F.3d at 1141. This is a "high standard." Delgado-Ramos, 635 F.3d at 1239.

Applying these principles here, it is obvious that even if Justice Breyer's plurality opinion were binding on this court, Randall is not "clearly irreconcilable" with Eddleman. Miller, 335 F.3d at 893. On its face, Justice Breyer's plurality opinion does not purport to change the state of the law but expressly looked to Buckley and its progeny: "Over the last 30 years, in considering the constitutionality of a host of different campaign finance statutes, this Court has repeatedly adhered to Buckley's constraints...." Randall, 548 U.S. at 242, 126 S.Ct. 2479 (opinion of Breyer, J.); see id. at 246, 126 S.Ct. 2479 ("[W]e begin with Buckley."). Indeed, the Breyer opinion specifically found that "[s]ince Buckley, the Court has consistently upheld contribution limits." Id. at 247, 126 S.Ct. 2479. Although the Court ultimately struck down Vermont's contribution limits, it did so consistent with the principles announced in Buckley.

If anything, Randall's plurality only clarified and reinforced Buckley and its progeny. In Randall, Justice Breyer observed that Buckley "general[ly] approv[ed] of statutes that limit campaign contributions," as long as the statute could demonstrate a "sufficiently important interest." Id. at 246-47, 126 S.Ct. 2479. The importance of Randall, then, was that the plurality affirmed Buckley, while at the same time showing that Buckley was not a rubberstamp. Other courts and scholars have concluded that Randall is an application of Buckley, not a repudiation.3

As Justice Breyer wrote, Buckley requires that contribution limits not "prevent candidates from `amassing the resources necessary for effective [campaign] advocacy.'" Id. at 248, 126 S.Ct. 2479 (quoting Buckley, 424 U.S. at 21, 96 S.Ct. 612) (alteration in Randall). He also emphasized that contribution limits cannot "magnify the advantages of incumbency to the point where they put challengers to a significant disadvantage." Id.; see also Shrink Missouri, 528 U.S. at 403-04, 120 S.Ct. 897 (Breyer, J., concurring). But, as Justice Breyer said, "we have `no scalpel to probe' each possible contribution level," so the Court "cannot determine with any degree of exactitude the precise restriction necessary to carry out the statute's legitimate objectives." Randall, 548 U.S. at 248, 126 S.Ct. 2479 (opinion of Breyer, J.) (quoting Buckley, 424 U.S. at 30, 96 S.Ct. 612). Accordingly, "the legislature is better equipped to make such empirical judgments." Id. Randall reaffirmed Buckley's recognition that such deference to the legislature has a limit, that after Buckley there is a "lower bound." Id. at 248-49, 126 S.Ct. 2479.

Randall's discussion of "danger signs" and the plurality's subsequent analysis of "five sets of considerations" did not present a new test for analyzing contribution limits; rather, this discussion only explained a mode for determining whether the limits were "narrowly tailored" under Buckley. Id. at 249-262, 126 S.Ct. 2479. Randall stands as a warning to lower courts that Buckley does not license them to approve any contribution limitation that professes an anti-corruption rationale; instead, lower courts must carefully analyze statutes to ensure that they are narrowly tailored. Id. at 249-50, 126 S.Ct. 2479.

We took such a careful approach in Eddleman. As such, the "overall analytical framework" in Eddleman is in harmony with Randall. In Eddleman, we began with Buckley's premise that contribution limits are constitutional as long as they do not prevent candidates from "amassing the resources necessary for effective advocacy." Eddleman, 343 F.3d at 1091 (quoting Buckley, 424 U.S. at 21, 96 S.Ct. 612). We noted that such restrictions "are subject to the `closest scrutiny'" and must be "closely drawn." Id. (quoting Buckley, 424 U.S. at 25, 96 S.Ct. 612). We reviewed the Court's post-Buckley opinions and summarized the principles to be derived therefrom.4

3. Whether Eddleman's analysis is consistent with Randall's analysis

Even if we thought Randall altered Buckley in some way, our decision in Eddleman considered the same issues that were important in Justice Breyer's plurality opinion. The State of Montana has made a strong showing that Randall would not have mandated a different result in Eddleman.

a. The Four "Danger Signs"

Justice Breyer's opinion in Randall identified four "danger signs" to look for in a campaign contribution statute: "(1) The limits are set per election cycle, rather than divided between primary and general elections; (2) the limits apply to contributions from political parties; (3) the limits are the lowest in the Nation; and (4) the limits are below those we have previously upheld." Id. at 268, 126 S.Ct. 2479 (Thomas, J., concurring in the judgment) (listing Justice Breyer's "danger signs"). We considered, in some form, each of these "danger signs" in Eddleman.

First, we found that the Montana contribution limits "apply to `each election in a campaign,' [so,] the amount an individual may contribute to a candidate doubles when the candidate participates in a contested primary." Eddleman, 343 F.3d at 1088. By comparison, Vermont's limits applied to a "two-year general election cycle." Randall, 548 U.S. at 238, 126 S.Ct. 2479.

Second, although Eddleman did not specifically deal with the limit on campaign contributions by political parties, there was no need to do so, because in Montana the aggregate contribution limits for political parties is much higher than the individual and political committee contribution limits. See Eddleman, 343 F.3d at 1094 ("[W]hile decreasing PAC and individual contributions, [Montana's contribution limit statute] simultaneously increased the amount of money political parties may contribute to a candidate, almost doubling the amount that may be contributed in some races."); see also Mont.Code Ann. § 13-37-216(3). In this regard, Montana's statute stands in stark contrast with Vermont's, which applied the same low contribution limit to individuals, PACs, and political parties alike. Randall, 548 U.S. at 238-39, 126 S.Ct. 2479.

Third, we acknowledged that Montana's limits were "some of the lowest in the country," but also observed that this was "unsurprising in light of the fact that Montana is one of the least expensive states in the nation in which to mount a political campaign." Eddleman, 343 F.3d at 1095. As Randall shows, Montana retains some of the lowest contribution limits in the nation, but it is not the lowest, a distinction that belonged to Vermont. See Randall, 548 U.S. at 250-51, 126 S.Ct. 2479.5

Fourth, while Eddleman did not specifically compare Montana's contribution limits with other instances where the Court has upheld a contribution limit as constitutional, we did compare the change in Montana's total campaign spending with other instances where the Court had upheld limits that involved greater decreases in total campaign spending. Eddleman, 343 F.3d at 1094 ("Indeed, the Shrink Missouri Court upheld contributions limits despite a decrease of more than 50% in total spending in Missouri elections, nearly twice the decrease present here.").

We also considered that there are "provision[s] preventing incumbents from using excess funds from one campaign in future campaigns. Such provision[s] keep incumbents from building campaign war chests and gaining a fundraising head start over challengers." Id. at 1095. We stressed that, in the end analysis, it is not the dollar amount that is critical, it is whether a candidate can amass the resources necessary to mount an effective campaign, id., a position in harmony with Randall, see 548 U.S. at 248-49, 126 S.Ct. 2479.

Thus, all of Randall's "danger signs" were considered in one form or another. Most importantly, and consistent with Randall, our decision in Eddleman "review[ed] the record independently and carefully with an eye toward assessing the statute's `tailoring.'" Id. at 249, 126 S.Ct. 2479. We think Eddleman took ample account of the "danger signs" identified in Randall.

b. The "Five Considerations"

Aside from the four "danger signs," our decision in Eddleman addressed broadly what Justice Breyer called "five sets of considerations." Id. at 261, 126 S.Ct. 2479. The five considerations were: (1) Whether the "contribution limits will significantly restrict the amount of funding available for challengers to run competitive campaigns"; (2) whether "political parties [must] abide by exactly the same low contribution limits that apply to other contributors"; (3) how "volunteer services" are treated; (4) whether "contribution limits are ... adjusted for inflation"; and (5) whether there exists "any special justification that might warrant a contribution limit so low or so restrictive." Id. at 253-61, 126 S.Ct. 2479. In Eddleman, we addressed each of these considerations in some way.

(1). With respect to the first consideration, whether the limits restrict challengers, the Court in Randall considered statistical analyses relevant to discerning "the critical question ... [, i.e., whether] a candidate running against an incumbent officeholder [can] mount an effective challenge." Id. at 255, 126 S.Ct. 2479. The Court noted that it emphasized the competitiveness of races because it was a proxy for the relative ability of a challenger to overcome the advantages of incumbency. Id.

In Eddleman, we recognized the importance of considering "all dollars likely to be forthcoming in a campaign, rather than the isolated contribution, and ... consider[ed] factors such as whether the candidate can look elsewhere for money, the percentage of contributions that are affected, the total cost of a campaign, and how much money each candidate would lose." Eddleman, 343 F.3d at 1094 (internal citations omitted). We repeatedly emphasized that the mere fact that a candidate could have raised more money without the limits was not the relevant inquiry; rather, the issue was whether the limit prevented a campaign from being effective. Id. at 1095 ("[A]part from bald, conclusory allegations that their campaigns would have been more effective had they been able to raise more money, none of the witnesses offered any specifics as to why their campaigns were not effective.") (internal quotation marks omitted). We found that "Montana candidates remain able to mount effective campaigns." Id. (describing candidates who claimed the limits prevented effective campaigns but some of which raised more money after the limits were in place and another who won with a large surplus of campaign funds). Additionally, even though the contribution limits restrict the total amount of funds raised, candidates were still able to raise funds "well within the range of money needed to run an effective ... campaign." Id. at 1094-95.

Specific to the Court's concern with challengers to incumbency, we discussed provisions that increased the ability of challengers to overcome the effects of incumbency. First, we pointed out that "§ 13-37-216 also contains a provision preventing incumbents from using excess funds from one campaign in future campaigns." Id. at 1095. Second, we found that "the average gap between the total amount of money raised by incumbents and challengers for all legislative races was only $65.00 per race," so there was almost no difference between incumbents and challengers in the amount of money they raised. Id. Third, relying on Buckley and Shrink Missouri, we suggested that there was no evidence that Montana's limitations allowed incumbents to leverage their incumbency unfairly against their challengers. Id. at 1095-96.

The district court did not look to our opinion in Eddleman. Instead, it conducted its own inquiry. For example, it compared the Vermont limits for state senate and house with those of Montana and concluded that Montana's were lower. Opinion and Order at 29. We think the district court did not account for one key difference between Vermont and Montana. While Vermont's contribution limits apply to a "two-year general election cycle," Randall, 548 U.S. at 238-39, 126 S.Ct. 2479, Montana's limits apply to "each election," Mont.Code Ann. § 13-37-216(1)(a), meaning that if there is a contested primary, the district court has understated Montana's limits by half. See Eddleman, 343 F.3d at 1088 ("[T]he amount an individual may contribute to a candidate doubles when the candidate participates in a contested primary."). In other words, if there is a primary, Montana's limit for the state legislature is $320, which is greater than Vermont's limit for state senate ($300) and much higher than its limit for state house ($200).

Additionally, we are concerned that the evidence the district court received and credited — which because of our time constraints, the parties have not briefed and we have not examined as thoroughly as we ordinarily would like — does not adequately account for the revenues actually available to candidates. For example, Montana only requires that the identity of donors contributing $35 or more, and their aggregate amount of contributions, be disclosed. Mont.Code Ann. § 13-37-229(2)-(3). While a candidate is required to disclose an "itemized account of proceeds that total less than $35 from a person," the donor's identity is not disclosed and therefore does not count against an individual's aggregate contribution limit. Mont.Code Ann. § 13-37-229(8). Thus, it is likely that Montana's limits understate the actual contributions made to the candidates. These are matters that, undoubtedly, would benefit from briefing and oral argument but raise serious concerns in our minds whether there is sufficient evidence to overrule Eddleman.6

(2). With respect to the second consideration, the limits on political parties, the Court was concerned that Vermont's statute required "that political parties abide by exactly the same low contribution limits that apply to other contributors." Randall, 548 U.S. at 256, 126 S.Ct. 2479. The cumulative restrictions imposed by the Vermont statute "severely inhibit[ed] collective political activity by preventing a political party from using contributions by small donors to provide meaningful assistance to any individual candidate," including a party's ability to engage in "coordinated spending on advertising, candidate events, voter lists, mass mailings, even yard signs." Randall, 548 U.S. at 256-58, 126 S.Ct. 2479.

In contrast to Vermont's statute, we noted, in Eddleman, that in Montana political parties were not subject to the same low contribution limit as individuals. Eddleman, 343 F.3d at 1094 (discussing the increase in amount that can be contributed by political parties, "almost doubling the amount that may be contributed in some races").

Despite the obvious differences between Vermont and Montana, the district court concluded that the Montana statute was inconsistent with this factor because "political committees [were held] to the same contribution limits as individuals" and this "inhibit[s] the associational rights of political committees and, consequently, a full and robust exchange of views." Opinion and Order at 32 (internal quotation marks omitted). Instead of addressing Randall's concern with limits on political parties, the district court focused on limits on political committees under § 13-37-216. Political committees are not political parties. Political committees — including PACs and local party affiliates — are subject to the same limits as individuals. Mont.Code Ann. § 13-37-216(3). "Political party organizations," however, are exempted from this restriction under the statute and subject to a much higher cap. For example, individuals and political committees may not contribute more than $630 to a gubernatorial candidate, but a political party organization can contribute up to $22,600. Id. § 13-37-216(1)(a), (3)(a), as adjusted by Admin. R. Mont. § 44.10.338(1)(a), (2)(a).

Furthermore, the district court's opinion fails to acknowledge that even political committees remain free to spend as much money as they desire promoting a candidate. See Citizens United v. Fed. Election Comm'n, 558 U.S. 310, 130 S.Ct. 876, 175 L.Ed.2d 753 (2010); see also Am. Tradition P'ship, Inc. v. Bullock, ___ U.S. ___, 132 S.Ct. 2490, 183 L.Ed.2d 448 (2012). As we pointed out, the PACs have many other ways "to convey their support." Eddleman, 343 F.3d at 1094. They just cannot give the money directly to the candidate. Thus, the district court's analysis on this point is inapposite; under the Montana statute political committees remain free to participate in a "full and robust exchange of views."

(3). The third consideration is the treatment of volunteer services. Montana's scheme, however, is far more permissive than Vermont's statute. In Randall, Vermont counted expenses incurred during the provision of volunteer services as contributions. Randall, 548 U.S. at 259-60, 126 S.Ct. 2479. As we explained, "the [Montana] statute in no way prevents PACs[, and individuals,] from affiliating with their chosen candidates in ways other than direct contributions, such as donating money to a candidate's political party, volunteering individual members' services, sending direct mail to their supporters, or taking out independent newspaper, radio, or television ads to convey their support." Eddleman, 343 F.3d at 1094. Moreover, we noted that nothing prevents "individuals and PACs [from] ... engag[ing] in independent political expression, to associate actively through volunteering their services, and to assist in a limited but nonetheless substantial extent in supporting the candidates and committees with financial resources." Id. at 1096.

The district court concluded that Montana treats volunteer services in the same manner as Vermont, "not exclud[ing] the expenses ... volunteers incur, such as travel expenses, in the course of campaign activities." Opinion and Order at 34 (internal quotation marks omitted). This conclusion appears to be error. Testimony provided by the plaintiff's own witnesses — as well as a stipulation of the parties — established that expenses incurred by volunteers are not considered contributions under Montana law. Tr. at 50-54, 74-76, 154-56 (Sept. 12, 2012). Even more importantly, other testimony established that an individual, political party, or political committee can actually hire staff for a candidate, and that would not be considered a contribution. Id.

(4). The fourth consideration is whether the limits are adjusted for inflation. Vermont's limits were not. Randall, 548 U.S. at 261, 126 S.Ct. 2479. As we noted in Eddleman, the Montana contribution limits are regularly adjusted for inflation. Eddleman, 343 F.3d at 1089.

The district court recognized that Montana adjusts its limits for inflation, but suggested that the Consumer Price Index ("CPI") is a flawed method of accounting for inflation. Opinion and Order at 35-36. The district court made that determination on the basis of near anecdotal testimony that the cost of pencils, yard signs, postage, and fuel have increased faster than the CPI. Id. at 13. The district court also noted that the CPI does not account for certain inputs that an effective campaign requires. Id. at 35.

This is too thin a reed to cling to in order to overturn our decision in Eddleman. We do not doubt that the CPI fails to capture all changes in campaign costs. It is, however, a well-recognized mechanism for adjusting for inflation, and we have no indication that the Supreme Court intended that states do anything else to "index limits." Randall, 548 U.S. at 261, 126 S.Ct. 2479. We continue to believe that Montana's statute will survive the Court's analysis in Randall. If we were to examine the district court's findings, its methodology would raise a number of questions. For example, the district court apparently did not consider whether pencils, yard signs, postage, and fuel fall within the underlying basket of goods used to calculate the CPI, nor did it question whether other campaign costs — such as office space — may have gone down during the same period. Further, even as we acknowledge that campaign costs have gone up over time, so have contribution limits risen since their inception in 1994, yet the district court made no attempt to compare the overall increase in the contribution limits with the overall increases in campaign inputs that were the subject of testimony at trial.

(5). The fifth and final consideration is a catchall: Whether there are any "special justification[s]" for the limits that "bring about ... serious associational and expressive problems." Randall, 548 U.S. at 261, 126 S.Ct. 2479. We identified at least one justification for why Montana's contribution limits are among the lowest in the nation: "[T]he State of Montana remains one of the least expensive states in the nation in which to run a political campaign." Eddleman, 343 F.3d at 1094. Thus, unlike Randall, where Vermont's justification was based solely upon the prevention of corruption, Montana specifically justified the low limits based on the relative inexpense of campaigning in Montana, a state where, for many offices, "campaign[ing] primarily [takes place] door-to-door, and only occasionally [through] advertis[ing] on radio and television." Id.

Most importantly, in Eddleman, after considering all of the factors deemed important by Justice Breyer's plurality opinion in Randall, we held that the Montana contribution limit does not prevent candidates from amassing the resources necessary to run an effective, competitive campaign. Id. at 1094-95, 1098. We cannot conclude that Randall is, in any material way, inconsistent with our analysis in Eddleman. Therefore, under Miller, we remain bound by Eddleman.7

* * * * * *

Given the procedural posture of the state's motion, we have tried to be careful not to prejudge whether any of the district court's findings of fact are clearly erroneous or its conclusions errors of law. That is the province of a merits panel of this court, to be decided on consideration of the appeal of the permanent injunction after full briefing. Based on our emergency review, however, we have noted that there appear to be sufficient problems with the district court's findings of fact and conclusions of law such that the State of Montana has met its burden of making a substantial case for relief on the merits. This showing is sufficient for us, under this factor, to exercise our discretion to stay the district court's permanent injunction pending appeal. Moreover, given the imminent nature of the election, we find it important not to disturb long-established expectations that might have unintended consequences, particularly in light of our previous holding in Eddleman that this selfsame statute is constitutional, without first allowing a merits panel the benefit of thoroughly examining the Montana statute in light of Randall.8 See Purcell v. Gonzalez, 549 U.S. 1, 5-6, 127 S.Ct. 5, 166 L.Ed.2d 1 (2006) ("Given the imminence of the election and the inadequate time to resolve the factual disputes, our action today shall of necessity allow the election to proceed without an injunction...."). We conclude that the state is likely to succeed in its appeal.

B. Irreparable Injury to the Party Requesting Stay

Nken held that the second stay factor, "whether the applicant will be irreparably injured absent a stay," requires more than "some possibility of irreparable injury." Nken, 556 U.S. at 434-35, 129 S.Ct. 1749 (internal quotation marks omitted). But, in contrast to the first factor, we have interpreted Nken as requiring the applicant to show under the second factor that there is a probability of irreparable injury if the stay is not granted. Leiva-Perez, 640 F.3d at 968 (explaining that while the first factor asks "whether the stay petitioner has made a strong argument on which he could win," the second factor asks us to "anticipate what would happen as a practical matter following the denial of a stay"). In analyzing whether there is a probability of irreparable injury, we also focus on the individualized nature of irreparable harm and not whether it is "categorically irreparable." Id. at 969 (quoting Nken, 556 U.S. at 435, 129 S.Ct. 1749).

The State of Montana has made a showing that there is a probability of irreparable injury if a stay of the district court's permanent injunction is not granted. Since 1994, a clear framework has been in place allowing candidates to plan for campaigns. In 2003, we held this statute constitutional, in the face of a virtually indistinguishable attack. This has created a background upon which the candidates in the current election have formed their campaign strategies and expectations. Absentee voting has already begun in Montana and the general election is imminent. Allowing the permanent injunction to remain in place before a merits panel of this court can ultimately rule on the constitutionality of the Montana contribution limit statute could throw a previously stable system into chaos. In a state that has operated with some of the most restrictive campaign limits in the country, there would suddenly be no limits whatsoever. In fact, there is some evidence from media reports that before the temporary stay was issued in this case, individuals had already begun to ask for unlimited donations. In light of the fact that the State of Montana has made a substantial case for relief on the merits, this calls into question the fairness and integrity of elections in Montana. Not all candidates might feel comfortable taking unlimited donations in the wake of conflicting judicial decisions. Furthermore, once the election is over, it cannot be reversed, and any consequences flowing from the disruption in equilibrium in the campaign contribution laws would also be irreversible. Regardless, because of the likely disruption to the election and the untold, irreversible consequences that might result, the State of Montana has satisfied its burden of showing a probability that irreparable harm will occur.

C. Substantial Injury to Other Parties Interested in the Proceeding

Finally, Nken explained that the last two factors of the test require us to weigh the public interest against the harm to the opposing party. Nken, 556 U.S. at 435, 129 S.Ct. 1749; see also Leiva-Perez, 640 F.3d at 964-66 (holding that the stay inquiry is "flexible" and involves an equitable balancing of the stay factors). In doing so, we again consider "the particulars of each individual case." Leiva-Perez, 640 F.3d at 970 (citing Nken, 556 U.S. at 436, 129 S.Ct. 1749).

We find that the other interested parties are not likely to be harmed. We well understand that "political speech [is] the core of the First Amendment," Randall, 548 U.S. at 266, 126 S.Ct. 2479, but for 36 years the Court has held that states may restrict political contributions as "only a marginal restriction upon the contributor's ability to engage in free communication." Buckley, 424 U.S. at 20-21, 96 S.Ct. 612. Montana largely restricts cash contributions to candidates; it thus leaves interested parties with a number of other options for engaging in political speech, from volunteering — or even paying for the provision of volunteer services to candidates — to engaging in independent activities to support a candidate. See, e.g., Bullock, 132 S.Ct. at 2491. Additionally, we have already carefully analyzed the Montana contribution limit statute and found it to be constitutional. Eddleman, 343 F.3d at 1098. All interested parties, who have operated under Montana's contribution limit statute since 1994, had clear notice of its constitutionality since 2003. Any harm that might be felt would, at most, be minimal and vastly outweighed by the public interest.

D. Public Interest

Finally, we find that the public interest is closely aligned with the irreparable harm shown by the State of Montana. The people comprising the State of Montana have a deep interest in fair elections. See United States v. Gradwell, 243 U.S. 476, 480, 37 S.Ct. 407, 61 L.Ed. 857 (1917) ("[T]he people of the United States ... have an interest in and a right to honest and fair elections...."). The Montana contribution limit statute has long stood, not only to prevent corruption, see Eddleman, 343 F.3d at 1092-93, but also to create a background of fairness to allow candidates to plan their campaigns and implement their strategies upon the foundation of well-laid and understood groundrules. Given the deep public interest in honest and fair elections and the numerous available options for the interested parties to continue to vigorously participate in the election, the balance of interests falls resoundingly in favor of the public interest.

IV. CONCLUSION

The State of Montana has satisfied the standards for a stay pending appeal. Given the formidable obstacle presented by our decision in Eddleman, the fact that Randall does not compel a different result in Eddleman, and the tensions and possible errors in the district court's application of Randall, the State of Montana has made a strong showing that it is likely to succeed on appeal. Furthermore, because the fairness of the imminent election would be put in danger by our failure to stay the permanent injunction, the State of Montana and the public interest would be irreparably harmed, and that harm vastly outweighs any minimal harm that might come to the interested parties who have operated under the established Montana contribution limits for almost two decades. We therefore find it necessary to exercise our judicial discretion, and we will stay the district court's permanent injunction pending resolution of the appeal by a merits panel of this court. The State of Montana's motion for stay pending appeal is GRANTED.

APPENDIX A

Montana Code Annotated § 13-37-216

13-37-216. Limitations on contributions — adjustment

APPENDIX B

Admin. R. Mont. § 44.10.338

44.10.338 LIMITATIONS ON INDIVIDUAL AND POLITICAL PARTY CONTRIBUTIONS

(1) Pursuant to the operation specified in 13-37-216, MCA, limits on total combined contributions from individuals to candidates are as follows:

(2) Pursuant to the operation specified in 13-37-216, MCA, limits on total combined contributions from political party committees to candidates are as follows:

(3) Pursuant to 13-37-218, MCA, in-kind contributions must be included in computing these limitation totals.