In response to the increasingly apparent need to reform the ways in which political campaigns are financed, the city of Albuquerque implemented a campaign-finance reform system in 1974. It adopted limits on campaign expenditures and contributions in municipal elections. In 2001, mayoral candidate Rick Homans brought a challenge under the First and Fourteenth Amendments1 to the mayoral-candidate expenditure restriction; ruling in favor of Homans, the district court permanently enjoined enforcement of this limit. City-council candidate Sander Rue brought a similar suit challenging the expenditure limit for city-council candidates and obtained a favorable summary judgment as well. Both cases are before us on review, and because they present similar issues, we consolidate them for review. Exercising jurisdiction under 28 U.S.C. § 1291, we affirm. The following expresses the opinion of the court as to parts I, II, and III. As to part IV, it concurs in part with the opinion of Judge Tymkovich, which constitutes the majority opinion of the court as to part IV.
I
As part of an overall restructuring of its city government in 1974, Albuquerque amended its city charter and implemented an election code imposing disclosure requirements and limiting expenditures and contributions for municipal elections. More than ninety percent of voters approved these reform measures. As it reads today, the election code provides:
No candidate shall allow or accept contributions or make expenditures in excess of the following for any election:
1. To a candidate for the office of Councillor, contributions or expenditures equal to twice the amount of the annual salary paid by the City of Albuquerque to Councillors as of the date of filing of the Declaration of Candidacy.
2. To a candidate for the office of May- or, contributions or expenditures equal to twice the amount of the annual salary paid by the City of Albuquerque to the Mayor as of the date of filing of the Declaration of Candidacy.
Albuquerque City Charter, art. XIII, sec. 4(d).2 Candidate-expenditure restrictions *903were in effect for each mayoral and city-council election from 1974 to 1995. Limits on the 1997 election were temporarily enjoined pursuant to a court order; however, parties to that litigation stipulated dismissal of the lawsuit, and the spending limits were restored for the 1999 election. For the 2001 elections, the mayoral-campaign expenditure limit was $174,720; city-council candidates were limited to spending a maximum of $17,056. Violation of these limits carries a fine of up to $500 per violation, removal from office, and/or public reprimand.
After filing suit in federal district court on August 10, 2001, plaintiff Rick Homans filed a motion for a preliminary injunction, which was denied on September 1, 2001. Homans v. City of Albuquerque, 160 F.Supp.2d 1266 (D.N.M.2001) (“Homans I”). He then filed an interlocutory appeal on September 4, 2001, seeking an emergency injunction pending appeal. Two days later, a two-member motions panel of this court granted the request and enjoined enforcement of the expenditure limit pending review of the merits. Homans v. City of Albuquerque, 264 F.3d 1240, 1245 (10th Cir.2001) (per curiam) (“Homans II ”). In doing so, the motions panel held that Homans established a likelihood of success on the merits regarding his claim that the expenditure limit violated the First and Fourteenth Amendments under Buckley v. Valeo, 424 U.S. 1, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976) (per curiam). Homans II, 264 F.3d at 1243-44.
On February 13, 2002, the parties filed a joint motion for stipulated admission of evidence and expedited determination on the merits in district court. The district court entered a declaratory judgment in favor of Homans and permanently enjoined enforcement of the provision in August 2002. Homans v. City of Albuquerque, 217 F.Supp.2d 1197 (D.N.M.2002) (“Homans III”). While the court stated its own view that the expenditure limitations restriction survives exacting scrutiny under Buckley, it found the contrary conclusion mandated by the motions-panel’s ruling in Homans II. Homans III, 217 F.Supp.2d at 1206. Albuquerque appeals.
Plaintiff Sander Rue was a duly qualified candidate for District Five City Coun-cillor who also ran in the October 2001 election. In his suit filed in federal district court in September, 2001, he claimed that the city-council campaign-expenditure limitation violates Buckley. The district court granted summary judgment in favor of Rue and permanently enjoined enforcement of the city-council restriction on October 11, 2002, relying on the motions-panel’s ruling in Homans II and the district court’s decision in Homans III. Rue v. City of Albuquerque, Civ. No. 01-1036 JP/LFG (D.N.M. Oct. 11, 2002). Albuquerque appeals.3
II
As to the grant of summary judgment in Rue’s case, we review the district court’s decision de novo, applying the same standards used by the district court and construing the facts in the light most favorable to Albuquerque. See Simms v. Okla. ex rel. Dep’t of Mental Health & Substance Abuse Servs., 165 F.3d 1321, 1326 (10th Cir.1999).
*904With regard to the final decision in Homans III, we ordinarily review the district court’s legal conclusions de novo, Dang v. UNUM Life Ins. Co. of Am., 175 F.3d 1186, 1189 (10th Cir.1999), and its factual findings for clear error. Fed. R.Civ.P. 52(a). Because this case implicates First Amendment concerns, however, we have “an obligation to make an independent examination of the whole record in order to make sure that the judgment does not constitute a forbidden intrusion on the field of free expression.” Bose Corp. v. Consumers Union of U.S., Inc., 466 U.S. 485, 499, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984) (quotations omitted). We therefore view the evidence objectively rather than in the light most favorable to the City.
Our standard in reviewing Rue’s case is thus more favorable to the City than our standard in Homans’ case. For this reason, with respect to each claim, we evaluate the evidence first in the light most favorable to the City to determine whether summary judgment was proper in Rue. Only if this less stringent standard is satisfied will we conduct an independent and objective examination of the evidence to review whether the permanent injunction in Homans’ case was proper.
Ill
As an initial matter, we determine whether the Homans II decision, in which a motions panel of this court granted an injunction pending appeal against the expenditure limit, has binding effect. See Homans II, 264 F.3d at 1243, 1245. Attempting to leverage this interlocutory de-cisión to its maximum possible effect, Ho-mans argues that the Homans II ruling constitutes the law of the case and restricts us in our merits determination.4 The district court agreed, stating that although it was inclined to conclude that the expenditure limit for mayoral campaigns survives the exacting scrutiny required under Buckley, it was bound by the motions-panel ruling to conclude otherwise. Homans III, 217 F.Supp.2d at 1206. We disagree.
In general, the law of the case doctrine provides that “when a court decides upon a rule of law, that decision should continue to govern the same issues in subsequent stages in the same case.” Arizona v. California, 460 U.S. 605, 618, 103 S.Ct. 1382, 75 L.Ed.2d 318 (1983). Law of the case “is solely a rule of practice and not a limit on the power of the court.” Mason v. Texaco, Inc., 948 F.2d 1546, 1553 (10th Cir.1991) (citing Messinger v. Anderson, 225 U.S. 436, 444, 32 S.Ct. 739, 56 L.Ed. 1152 (1912)). Thus, the doctrine is discretionary rather than mandatory. Kennedy v. Lubar, 273 F.3d 1293, 1299 (10th Cir.2001); Stifel, Nicolaus & Co. v. Woolsey & Co., 81 F.3d 1540, 1544 (10th Cir.1996).
In the instant matter, the two judge panel decision of our court constituted an interlocutory ruling, and its holding was limited to the conclusion that Homans had shown a likelihood of success on the merits of his claim.5 Homans II, 264 F.3d at 1243-44. Courts repeatedly have emphasized that a decision as to the likeli hood of success is tentative in nature and *905not binding at a subsequent trial on the merits. Univ. of Tex. v. Camenisch, 451 U.S. 390, 395, 101 S.Ct. 1830, 68 L.Ed.2d 175 (1981) (“[F]indings of fact and conclusions of law made by a court granting a preliminary injunction are not binding at trial on the merits.”); Southco, Inc. v. Kanebridge Corp., 324 F.3d 190, 195 (3d Cir.2003); A.M. Capen’s Co. v. Am. Trading & Prod. Corp., 202 F.3d 469, 473 (1st Cir.2000). Were the opposite true, an unacceptable conflation of the merits decision and the preliminary inquiry would result. Moreover, if the district court were bound in the manner suggested, then the decision of the motions panel would also bind the appellate panel reviewing the merits. This is not the rule. When reviewing a decision by a prior motions panel, we are “uninhibited by the law of the case doctrine.” Stifel, 81 F.3d at 1544.
Our own circuit precedent expressly rejects the proposition urged by Homans and articulates the rationale for doing so. See id. at 1543-44. As we have explained, “a motions panel’s decision is often tentative because it is based on an abbreviated record and made without the benefit of full briefing and oral argument.” Id. at 1544. In the instant ease, the motions panel did not hear oral argument, and briefing proceeded on an expedited schedule in light of the impending election. Homans filed his motion on September 4, 2001, and the City was required to submit its response the next day; the motion was granted’ on the following day, September 6, 2001. In light of this truncated schedule and the “avowedly preliminary [and] tentative” nature of the emergency-injunction ruling, Council of Alternative Political Parties v. Hooks, 179 F.3d 64, 69 (3d Cir.1999) (quotation omitted), a motions-panel ruling does not establish the law of the case; therefore, neither the district court nor this court is constrained in its review of the merits by the September 6 ruling.6 Accordingly, we reject Homans’ argument and hold that the district court erred in concluding that it was bound by the law of the case doctrine.
IY
We proceed to the merits of Homans’ and Rue’s claims. Although the two cases stand in differing procedural postures, they raise an identical substantive claim: that the campaign-expenditure limits in the Albuquerque City Charter violate the First and Fourteenth Amendments.
Every challenge to campaign-finance reform provisions must begin with an analysis of the watershed case of Buckley v. Valeo, 424 U.S. 1, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976). Analyzing the 1974 amendments to the Federal Election Campaign Act (“FECA”), the Supreme Court developed a jurisprudential distinction between restrictions on campaign expenditures and restrictions on campaign contributions. Although both types of restrictions limit core political speech and are therefore subject to “exacting scrutiny,” id. at 16, 96 S.Ct. 612, the Court concluded that expenditure limits impose “significantly more severe restrictions on protected freedoms” than limits on contributions. Id. at 22, 96 S.Ct. 612. As the Court explained, a “restriction on the amount of money a person or group can spend on political communication during a campaign necessarily reduces the quantity of expression by restricting the number of issues discussed, the depth of their exploration, and the size of the audience reached.” Id. at 19, 22, 96 S.Ct. 612. For this reason, expenditure limits raise graver constitutional concerns *906and are invalidated more frequently. Id. at 55-56, 59, 96 S.Ct. 612 (upholding contribution limits but invalidating expenditure limits); see also McConnell v. Fed. Election Comm’n, 540 U.S. -, 24, 124 S.Ct. 619, 655, 157 L.Ed.2d 491, (2003) (“In Buckley and subsequent cases, we have subjected restrictions on campaign expenditures to closer scrutiny than limits on campaign contributions.”); Fed. Election Comm’n v. Colo. Republican Fed. Campaign Comm., 533 U.S. 431, 440, 121 S.Ct. 2351, 150 L.Ed.2d 461 (2001) (“Colorado Republican II ”) (“[L]imits on political expenditures deserve closer scrutiny than restrictions on political contributions.”); Nixon v. Shrink Mo. Gov’t PAC, 528 U.S. 377, 387-88, 120 S.Ct. 897, 145 L.Ed.2d 886 (2000) (“Shrink Missouri ”) (concluding that differing standards govern review of contribution limits and expenditure limits, and that contribution limits may be justified when they are “closely drawn” to serve a “sufficiently important interest”); Fed. Election Comm’n v. Mass. Citizens for Life, 479 U.S. 238, 259-60, 107 S.Ct. 616, 93 L.Ed.2d 539 (1986) (“We have consistently held that restrictions on contributions require less compelling justification than restrictions on independent spending”).
Although a less stringent standard of review applies to limits on political contributions, we conclude that the standard for expenditure limits operates identically to strict scrutiny review. To be upheld, therefore, the campaign-expenditure restrictions must be both narrowly tailored, Fed. Election Comm’n v. Nat’l Conserv. Political Action Comm., 470 U.S. 480, 496, 105 S.Ct. 1459, 84 L.Ed.2d 455 (1985) (“NC-PAC ”), and necessary to serve a compelling state interest, Mass. Citizens for Life, 479 U.S. at 251-52, 107 S.Ct. 616. See also Austin v. Mich. Chamber of Commerce, 494 U.S. 652, 657, 110 S.Ct. 1391, 108 L.Ed.2d 652 (1990) (applying strict scrutiny to review state restrictions on corporate political expenditures); Kruse v. City of Cincinnati, 142 F.3d 907, 912-13 (6th Cir.1998) (holding that municipal restrictions on candidate expenditures are subject to strict-scrutiny review).
In conducting strict scrutiny review, it is essential to acknowledge that such scrutiny is not “strict in theory, but fatal in fact.” See, e.g., Grutter v. Bollinger, 539 U.S. 306, 123 S.Ct. 2325, 2338, 156 L.Ed.2d 304 (2003). Despite this repeated admonition by the Supreme Court, appellees insist that Buckley imposes a per se ban on all candidate-expenditure restrictions. Given the Supreme Court’s distaste for “imposing judicial formulas so rigid that they become a straitjacket that disables government from responding to serious problems,” Denver Area Educ. Telecomm. Consortium, Inc. v. FCC, 518 U.S. 727, 741, 116 S.Ct. 2374, 135 L.Ed.2d 888 (1996), we are obliged to disagree.
In Buckley, defenders of FECA’s restrictions on candidate expenditures proffered various rationales to justify the limits, 424 U.S. at 55-57, 96 S.Ct. 612;7 nonetheless, the Court invalidated the expenditure cap, holding that “[njo governmental interest that has been suggested is sufficient to justify the restriction.” Id. at 55, 96 S.Ct. 612 (emphasis added). The Court’s chosen language leaves open the possibility that at least in some circumstances expenditure limits may withstand constitutional scrutiny. See Kruse, 142 F.3d at 920 (Cohn, J., concurring) (stating that “Buckley ... is not a broad pronouncement declaring all campaign expen*907diture limits unconstitutional,” and that it remains possible to develop a factual record that would sustain such restrictions). Concluding that it might be possible to devise a system of campaign-expenditure limits that would survive exacting scrutiny, we evaluate Albuquerque’s attempt to do so.
A
To satisfy strict scrutiny review, Albuquerque must establish that its candidate-expenditure restrictions are necessary to further a compelling state interest. Albuquerque sets forth the following rationales to justify the limits: (1) deterrence of corruption and enhancement of public confidence in the electoral process; (2) preservation of officeholders’ ability to perform their duties without devoting excessive time to fundraising; and (3) encouragement of electoral competition. We address each rationale.
1
Speaking nearly five decades ago, Justice Frankfurter made the following assessment of the corruption and public confidence issue:
We all know ... that one of the great political evils of the time is the apparent hold on political parties which business interests and certain organizations seek and sometimes obtain by reason of liberal campaign contributions. Many believe that when an individual or association of individuals makes large contributions for the purpose of aiding candidates of political parties in winning elections, they expect, and sometimes demand, and occasionally, at least, receive, consideration by the beneficiaries of their contributions which not infrequently is harmful to the general public interest.
United States v. United Auto. Workers, 352 U.S. 567, 576, 77 S.Ct. 529, 1 L.Ed.2d 563 (1957) (quotation omitted). Some time later, the Supreme Court explained:
Leave the perception of impropriety unanswered, and the cynical assumption that large donors call the tune could jeopardize the willingness of voters to take part in democratic governance. Democracy works only if the people have faith in those who govern, and that faith is bound to be shattered when high officials and their appointees engage in activities which arouse suspicions of malfeasance and corruption.
Shrink Missouri, 528 U.S. at 390, 120 S.Ct. 897 (quotation omitted). Most recently, the Supreme Court reiterated the importance of preventing corruption or its appearance in the context of political contribution limits. See McConnell, 540 U.S. at 33, 124 S.Ct. at 660 (“Our cases have made clear that the prevention of corruption or its appearance constitutes a sufficiently important interest to justify political contribution limits.”). Despite this precedent, Homans argues to no avail that as a matter of law Buckley mandates that expenditure limits can never be justified by the anti-corruption rationale.
It is well-established that the deterrence of corruption constitutes a compelling state interest. See Austin, 494 U.S. at 657-60, 110 S.Ct. 1391; NC-PAc, 470 U.S. at 496-97, 105 S.Ct. 1459. The question remains whether Albuquerque’s expenditure limits are necessary to serve this end. See Burson v. Freeman, 504 U.S. 191, 198, 112 S.Ct. 1846, 119 L.Ed.2d 5 (1992) (noting that exacting scrutiny requires that the regulation be necessary to serve a compelling state interest). In the particular circumstances of Buckley, the Court rejected the anti-corruption rationale in reviewing FECA’s campaign-expenditure limits, concluding that the interest in preventing cor*908ruption was served adequately in that case by the federal contribution limits and disclosure provisions. Rejecting the corollary argument that expenditure limits were necessary to prevent circumvention of permissible campaign-finance provisions, the Court relied on the factual conclusion that “[tjhere is no indication that the substantial criminal penalties for violating the contribution ceilings combined with the political repercussions of such violations will be insufficient.” 424 U.S. at 26, 55-56, 96 S.Ct. 612. Attaching significance to the fact that FECA permitted successful candidates to retain contributions in excess of expenditure limits and to use these funds for any lawful purpose, the Buckley Court explained: “This provision undercuts whatever marginal role the expenditure limits might otherwise play in enforcing contribution ceilings.” Id. Contrary to ap-pellees’ contention, Buckley does not preclude the use of expenditure limits to further a state’s anti-corruption interest in all circumstances. Cf. Kruse, 142 F.3d at 915 (acknowledging that Buckley may be interpreted to leave open the possibility that the anti-corruption rationale may, under some circumstances, justify candidate-expenditure caps).
Nor are we persuaded that subsequent case law prohibits the use of expenditure caps to deter corruption as a matter of law. In Kruse, the Sixth Circuit reached the opposite conclusion and held that NC-PAC and Colo. Republican Federal Campaign Comm. v. Federal Election Comm’n, 518 U.S. 604, 116 S.Ct. 2309, 135 L.Ed.2d 795 (1996) (“Colorado Republican I ”) “make eminently clear that spending limits ... are unconstitutional not simply because of the presence of contribution limits but because they are not narrowly tailored to serve this interest.” 142 F.3d at 915. We conclude that this view reads too much into NC-PAC and Colorado Republican I.
In NC-PAC, the Court invalidated restrictions on political action committees’ independent expenditures when the spending was not coordinated with a campaign; the court did not address the permissibility of restrictions on candidate expenditures. 470 U.S. at 496-98, 105 S.Ct. 1459. Even were NC-PAC’s reasoning to apply to restrictions on candidate expenditures, its holding would be limited to the facts in that case. The Court’s explicit holding was that “[o]n this record, ... an exchange of political favors for uncoordinated expenditures remains a hypothetical possibility and nothing more.” Id. at 498, 105 S.Ct. 1459 (emphasis added). Similarly fact-bound is the holding in Colorado Republican I, in which the Supreme Court invalidates limits on spending by political parties, which are treated as “independent expenditures.” 518 U.S. at 613-19, 116 S.Ct. 2309. The Court based its conclusion on the absence of coordination between the candidate and the source of the expenditure, which prevented the Court from assuming, “absent convincing evidence to the contrary, that a limitation on political parties’ independent expenditures is necessary to combat a substantial danger of corruption of the electoral system.” Id. at 617-18, 116 S.Ct. 2309 (emphasis added). Both NC-PAC and Colorado Republican I thus explicitly leave open the possibility that in certain circumstances, a factual record may establish “convincing evidence” and thus justify the need for expenditure limits to reduce corruption or the appearance of corruption.
Unfortunately for Albuquerque, we conclude that the record in this case does not aggregate to such “convincing evidence.” As a consequence, we are bound to reject the contention that the City’s expenditure limits are necessary to deter corruption. The City submits the following as evidence to support its position: (1) evidence demonstrating the ease with which contribu*909tion limits are circumvented; (2) voter turnout statistics; (3) voter surveys; and (4) anecdotal evidence of corruption.
As to the first category of evidence, the City focuses on the use of tactics such as “bundling” at the federal level, arguing that expenditure limits are necessary to prevent circumvention of contribution limits. Assuming that such evidence could demonstrate that expenditure limits are necessary to reduce corruption,8 we are not persuaded that bundling practices at the federal level are comparable to those at the local level. Although the City is entitled to rely on evidence from other jurisdictions to justify campaign-finance reform measures, Shrink Missouri, 528 U.S. at 394 & n. 6, 120 S.Ct. 897 (citing Renton v. Playtime Theatres, Inc. 475 U.S. 41, 51-52, 106 S.Ct. 925, 89 L.Ed.2d 29 (1986)), it may only do so if the evidence relied upon is “reasonably believed to be relevant to the problem that the city addresses.” Id. The City does not proffer any argument, much less evidence, suggesting that local bundling practices are analogous to federal ones.
But the City’s evidence of the need to deter corruption is not limited to documenting bundling at the federal level — the City introduces additional, independent evidence of the public appearance of corruption. Submitting statistics on Albuquerque voter turnout, the City argues that, contrary to common claims that expenditure limits suppress voter turnout, Albuquerque voters are more likely to vote when expenditure limits are imposed. Recognizing a positive relationship between voter turnout and the public perception of corruption is not unprecedented; as earlier noted, the Supreme Court has articulated, “Leave the perception of impropriety unanswered, and the cynical assumption that large donors call the tune could jeopardize the willingness of voters to take part in democratic governance.” Shrink Missouri, 528 U.S. at 390, 120 S.Ct. 897.9 To establish the relationship in the instant case, the City submits that in Albuquerque elections from 1974 until 2001, an average of 40.3% of registered voters participated in all city elections for which the spending limits were in place.10 This figure is particularly impressive when *910compared to the 25 to 35% turnout rate for city elections nationally. This comparison figure, however, is based on the percentage of the voting-age population, not the percentage of registered voters as it was in the Albuquerque figure; indeed, a calculation of Albuquerque voter turnout as a percentage of the voting age population reveals a turnout rate of only 28.3% between 1984 and 1989. While these figures may cast doubt on the suggestion that spending limits inhibit turnout rates, they assuredly do not establish the affirmative of the proposition: that Albuquerque’s spending limit actually increases turnout rates, as the City’s own expert readily concedes.11
Telephone survey results12 are submitted by Albuquerque in support of its argument that expenditure limits are necessary to combat the public perception of corruption.13 This survey evidence does suggest that Albuquerque voters have more confidence in the integrity of local elections than federal elections, which have no spending limits. On the other hand, appel-lees’ evidence also suggests that voters generally trust local government more than state and federal government, regardless of spending limits. Albuquerque’s survey results demonstrate that voters think that the removal of spending limits increases the potential for corruption. By contrast, appellees’ evidence suggests that the amount of spending in an election does not affect voter cynicism when other variables are controlled.
Anecdotal evidence supports Albuquerque’s contention that special interests are perceived to exercise an undue influence in elections. For example, a New Mexico State Senator described the local influence of contributions from certain industries. Specifically, she cited the influence of the alcohol industry in delaying legislation to prohibit “drive-up windows” for the purchase of alcohol, and the influence of the gambling industry in obtaining favorable revenue-sharing compacts with the state. An expert in the field of campaign-finance, Larry Makinson, testified to a correlation between legislative voting and campaign-contributions at the federal level,14 citing as an example the Tauzin-Dingell bill.
While recognizing the possibility that the government may provide sufficient evidence of the need for candidate-expenditure caps to prevent corruption, we nonetheless conclude that the evidence before us, even viewed in the light most favorable *911to Albuquerque, is no more compelling than the evidence the Court effectively rejected in Buckley, 424 U.S. at 29, 96 S.Ct. 612. In Buckley, FECA’s defenders also submitted survey evidence suggesting that the public perceived undue influence by special interests and anecdotal evidence of corruption surrounding the 1972 election; indeed, the Buckley evidence was of a far greater magnitude than that presented here and smelled of actual quid pro quo. See id. at 27 n. 28, 96 S.Ct. 612 (referencing Buckley v. Valeo, 519 F.2d 821, 839-40, & nn. 36-38 (D.C.Cir.1975) for evidence documenting the undue influence of the dairy industry, illegal corporate contributions, and promise of diplomatic posts in exchange for hefty campaign contributions). While undoubtedly troubling, Albuquerque’s evidence does not approach the palpable sense of corruption prompting the federal amendments in 1974. Because the Buckley evidence was held insufficient to demonstrate that FECA’s candidate-expenditure limits were necessary to serve the compelling state interest in deterring corruption, we are compelled to conclude that Rue is entitled to summary judgment. A fortiori, we hold that the evidence in Homans’ case, which need not be viewed in the light most favorable to the City, fails to sustain the City’s burden.
2
Thus far, the Supreme Court has recognized the existence of but one state interest sufficiently compelling to justify campaign-finance regulation: the anti-corruption rationale. See NC-PAC, 470 U.S. at 496-97, 105 S.Ct. 1459. This initial recognition does not, of course, foreclose the possibility that other compelling state interests may be identified in future cases, and in the case on review Albuquerque submits evidence of additional interests that it claims justify campaign expenditure caps. The City argues that the caps are necessary to serve the compelling state interest of preserving officeholders’ time and enabling them to perform official duties. Claiming that “when campaign spending is unlimited, as is true for congressional elections, fundraising becomes a full-time job for candidates and officeholders fearful of being outmatched by an opponent’s spending,” the city submits evidence documenting the fundrais-ing burdens of candidates and officeholders. (Homans Appellant’s Br. at 52; Rue Appellant’s Br. at 54.)
In Kruse, the Sixth Circuit concluded that the need to preserve officeholders’ time and ability to perform official duties is merely a restatement of the containing-skyrocketing-campaign-costs rationale rejected in Buckley:
The need to spend a large amount of time fundraising is a direct outgrowth of the high cost of campaigns. However, because the government cannot constitutionally limit the cost of campaigns, the need to spend time raising money, which admittedly detracts an officeholder from doing her job, cannot serve as a basis for limiting campaign spending.
142 F.3d at 916-17. We view the Kruse court’s conflation of the two rationales as inaccurate and conclude that the preservation of officeholders’ time is wholly separate. In Buckley, the Court rejected the proffered rationale of containing the skyrocketing costs of campaigns as follows: “[T]he mere growth in the cost of federal election campaigns in and of itself provides no basis for governmental restrictions on the quantity of campaign spending and the resulting limitation on the scope of federal campaigns.” 424 U.S. at 57, 96 S.Ct. 612 (emphasis added). Albuquerque does not merely rely on the skyrocketing costs of campaigns in and of themselves; rather, the City cites an additional reason why increasingly expensive campaigns *912hurt the electoral system: the woes of poor challengers aside, such campaigns distract officeholders from performing their official duties. Buckley makes no mention of this rationale and thus does not necessarily preclude the recognition of this as a compelling state interest.
To show that its campaign-expenditure limits are necessary to further this interest, Albuquerque submits excerpts from the book Speaking Freely containing interviews of retired congressional representatives that present a disturbing view into the fundraising pressures imposed on federal candidates. Additionally, the City cites statements made by local officials describing the time pressures imposed by fundraising in the absence of expenditure limits. Dede Feldman, a New Mexico State Senator who also ran for Albuquerque City Council in 1995, compares the differences in state campaigns versus Albuquerque campaigns:
Campaigning with and without spending limits is very different.... Because there was unlimited spending and I had to raise more money in the Senate races, I spent a lot more time fund-raising and I tried to raise money from larger contributions .... Doing so much fund-raising was incredibly time-consuming and cut into my other campaigning and my regular job. The fund-raising had to be done during the day and I therefore had less time to do my regular job. In addition, I had less time to engage in direct contact with the voters by going door-to-door.
(2 Rue R. Doc. 47 at 708-09.) In discussing the fundraising burden imposed during the recent election for which spending limits were enjoined, former mayor Jim Baca comments, “As a result of this new money chase in this year’s mayoral election in Albuquerque, I am now forced to spend three hours every day making fundraising phone calls. I have never before had to do this in my political career.” (2 Homans Doc. 24 at 512.) On the other hand, other evidence in the record suggests that the strain on Albuquerque officials’ time in the absence of spending limits is not significantly greater than when a spending limit is in place. For example, Michael Guerrero, who entered the 1999 City Council race while the spending limit was in place, testified that he spent between ten to fifteen hours per week raising funds.
Limited to the record before us, we cannot conclude that Albuquerque has submitted sufficient evidence to demonstrate that candidate-expenditure caps are necessary to further the state’s interest in protecting officials’ abilities to conduct their jobs. While it may seem so to candidates at the time, we are not convinced that the burdens imposed on Albuquerque officeholders’ time amounts to a problem of constitutional proportions. The claim that fundraising prevents officeholders from engaging in alternative campaign tactics such as individual door-to-door contact is interesting. Albuquerque does not articulate any reason why there is a compelling state interest in channeling campaign resources to favor individual voter contact rather than fundraising tactics. There is no indication in the record that, for example, the added burden of fundraising events and phone calls is more demanding on an officeholder’s time than the burden of individual voter contact.
Nor does the record persuade us that individual voter contact is a fundamentally superior campaign strategy because fund-raising efforts compromise an officeholder’s ability to communicate with the public. Given the individual contribution-limits in Albuquerque, officeholders are likely to communicate with a broad swath of potential fundraisers in much the same way they would communicate with the public *913through door-to-door contact; records of past campaign contributors do not suggest that the target for contribution solicitations differs significantly from the general Albuquerque public at large. On the contrary, candidate Guerrero states that time spent on fundraising cannot be distinguished from time spent on other campaign tactics because “a lot of times when you’re fund-raising, you’re also campaigning.” (3 Rue R. Doc. 52 at 760.) We take this to mean that the message disseminated in fundraising efforts often coincides with the message disseminated through voter contact — they are generally one and the same; the City makes no effort to rebut this assumption. For these reasons, we hold that Albuquerque’s evidence, even when viewed in the light most favorable to the City, fails to demonstrate that expenditure limits are necessary to further a compelling state interest in preserving officeholders’ time.
3
Finally, the City argues that campaign-spending limits are necessary to further the state interest in promoting electoral competition. One expert in the field of campaign-finance reform explains:
Electoral competition is another central component of democratic governance. In many respects, the ultimate weapon of public accountability in a democratic system is the ability of citizens to remove political actors through elections. And, electoral competition is the mechanism that keeps accountability viable.... High levels of campaign spending poses a threat to such competition because large incumbent war chests tend to discourage serious challengers.
(2 Rue R. Doc. 43 at 674.) Homans argues that this merely rehashes the equalization-of-candidate-resources rationale rejected in Buckley.15 See Buckley, 424 U.S. at 56, 96 S.Ct. 612 (rejecting the argument that equalizing candidates’ resources constitutes a compelling state interest). In Buckley, however, the Court spoke solely to the equalization of candidate resources; it did not address the possible rationale of improving electoral competition. Id. at 48-49, 56-57, 96 S.Ct. 612. We are persuaded that the improvement of electoral competition constitutes an interest distinct from the equalization-of-candidate-resources rationale rejected in Buckley. Thus, nothing precludes this court from recognizing robust electoral competition as a state interest sufficiently compelling to justify the expenditure limits.
We do not resolve this question because there is insufficient evidence in the record, even when viewed in the light most favorable to the City, to establish that spending limits actually enhance electoral competition. While the City’s statistical evidence does tend to undercut the doomsday prediction that spending limits discourage competition by insulating incumbents, it falls short of proving the contrary — that spending limits actually improve electoral competition — as the City’s own expert admits. Even were we to assume that enhancement of electoral competition constitutes a compelling state interest, there is insufficient evidence in this record to show that expenditure limits serve this end.
B
We conclude that Albuquerque’s evidence, even when viewed in the light most favorable to the City, fails to establish that its candidate-expenditure limits are necessary to serve a compelling state interest. Thus, summary judgment in favor of Rue was proper. It follows that the City’s evidence in Homans III, which need not *914be viewed in the light most favorable to Albuquerque, fails to sustain its burden, and the permanent injunction was appropriately granted. Given these holdings, we have no occasion to determine whether the expenditure limits are narrowly tailored. It is clear from the record, and from the many other cases dealing with the problem, that there is an increasing drive, and need, for campaign finance reform. We do not intend by our holding— that Albuquerque has failed in the instant case to demonstrate a compelling state interest for its expenditure provisions — to discourage future efforts in reforming our electoral system; we merely hold that on the record before us, Albuquerque has failed to justify its expenditure limits.
Other jurisdictions have attempted alternative measures to eradicate corruption: limiting the size of campaign contributions, improvement of electoral competition, enhancement of voter participation, and preservation of candidates’ scarce time resources, and they have done so within constitutional bounds. Notably, the Court has stated that public financing measures including expenditure limits may be implemented to achieve these ends without running afoul of the First Amendment. See Buckley, 424 U.S. at 57 n. 65, 96 S.Ct. 612 (“Congress may engage in public financing of election campaigns and may condition acceptance of public funds on an agreement by the candidate to abide by specified expenditure limitations.”). As for the prescription before us — a rigid limitation of campaign expenditures — Albuquerque has failed to submit sufficient evidence of a compelling state interest .justifying such limits.
For the foregoing reasons, we AFFIRM both decisions of the district court.
. Protections afforded under the First Amendment have been incorporated into the Fourteenth Amendment to apply against the states. Everson v. Bd. of Educ., 330 U.S. 1, 13-15, 67 S.Ct. 504, 91 L.Ed. 711 (1947).
. Albuquerque’s city charter also caps individual contributions at five percent of the annual salary of the office. Id. at sec. 4(e). The propriety of the contribution limits, which theoretically would appear to allow twenty individual contributors to fund one hundred *903percent of a campaign at the full expenditure limitation level, is not before us.
. Although the elections have passed, this does not render the cases moot because the issues raised are “capable of repetition, yet evading review.” See First Nat'l Bank of Boston v. Bellotti, 435 U.S. 765, 774, 98 S.Ct. 1407, 55 L.Ed.2d 707 (1978); Suster v. Marshall, 149 F.3d 523, 527 (6th Cir.1998).
. In similar fashion., Rue argues that the Ho-mans II order binds review of the merits of the city-council limits.
. To the extent that any language in Homans II can be read as an assessment of the actual merits of Homans’ claim, as opposed to his likelihood of success on the merits, such language is dicta. Dicta is not subject to the law of the case doctrine. See In re Meridian Reserve, Inc. v. Bonnett Res. Corp., 87 F.3d 406, 410 (10th Cir.1996).
. From this it follows that neither the district court considering Rue's claim nor this court in reviewing that decision is constrained by the Homans II ruling.
. The three rationales proffered were as follows: (1) to deter corruption and the appearance of corruption; (2) to equalize candidates’ resources; and (3) to contain the skyrocketing costs of political campaigns.
. Notably, evidence of circumvention of contribution limits, standing alone, could not sustain the more onerous burdens imposed by expenditure limits:
The discussion in [the earlier section of the Buckley Opinion] explains why the Act’s expenditure limitations impose far greater restraints on the freedom of speech and association than do its contribution limitations. The markedly greater burden on basic freedoms caused by [FECA’s independent expenditure provision] thus cannot be sustained simply by invoking the interest in maximizing the effectiveness of the less intrusive contribution limitations.
Buckley, 424 U.S. at 44, 96 S.Ct. 612; see also Kruse, 142 F.3d at 915-16 (holding that Buckley expressly rejects the argument that spending caps are justified by the need to enforce contribution limits).
. This statement suggests that turnout rates correlate with expenditure limits because expenditure limits decrease the perception of corruption, which in turn increases turnout. Albuquerque, however, suggests another reason for the relationship between turnout rates and expenditure limits — one which does not implicate the anti-corruption rationale. It proffers the somewhat counterintuitive claim that expenditure limits improve the public's knowledge base, which in turn increases turnout. Implicitly, then, Albuquerque suggests that expenditure limits are necessary not only to further an interest in deterring corruption, but also to further an interest in enhancing the electorate’s knowledge.
. The data for the two elections for which spending limits were enjoined are as follows: in the 1997 election, only 33% of registered voters turned out; in the 2001 election, however, an impressive 42.4% voted.
. Kruse implicitly rejected a similar argument that turnout rates might be used to demonstrate the need to further an interest other than anti-corruption. 142 F.3d at 916. Because we are unpersuaded that turnout rates in fact correlate with expenditure limits, we have no occasion to reach this issue.
. Survey results are an acceptable form of evidence to demonstrate the need for campaign-finance reform measures. Shrink Missouri, 528 U.S. at 394, 120 S.Ct. 897; Mont. Right to Life Ass’n v. Eddleman, 306 F.3d 874, 882 (9th Cir.2002); Daggett v. Comm’n on Governmental Ethics and Elections, 205 F.3d 445, 457-58 (1st Cir.2000).
. The City's evidence is the result of a telephone survey conducted in August 1998 by Lake Snell Perry & Associates of 400 registered voters who reside in Albuquerque. These results have a margin of sampling error of +/-4.9%.
Appellees’ argument that majoritarian views cannot dictate the bounds of First Amendment protections is misplaced. Such barometers of public sentiment are relevant in the campaign-finance context, to the extent that they show the need to remedy a public perception of corruption. See, e.g., Shrink Missouri, 528 U.S. at 394, 120 S.Ct. 897.
.Neither the utilization of candidate issue questionnaires as a tool to achieve such correlation nor the propriety of such questionnaires in the fundraising and lobbying process was presented to the district court and we do not reach the issue.
. Notably, Rue does not raise this argument.