Appellant Hobart Ward Anderson ("Anderson" or "Appellant") appeals the decision of the district court granting summary judgment to Defendants-Appellees Kentucky Board of Elections, the Registry of Finance ("Registry"), the Commonwealth Attorneys as a class, and the Kentucky Attorney General on nine separate claims challenging the constitutionality of various provisions of Kentucky election law. Having concluded that the district court erred in its evaluation of the substantial First Amendment interests asserted by Anderson, we reverse the district court except as to two claims — Anderson's equal protection claim, and his claim that the "trigger" provision is unconstitutional as applied — and remand for proceedings consistent with this order.

Hobart Anderson filed to run as a write-in candidate in Kentucky's 1999 gubernatorial election. Because he and his slated running mate were not eligible to appear on the ballot for the general election, his campaign was not entitled to receive matching funds under Kentucky's Public Financing Campaign Act. Anderson alleges that the statutes he challenges proscribed several of the campaign activities he wanted to conduct, including: distributing within 500 feet of polling places literature instructing voters on how to cast a write-in ballot; soliciting and accepting contributions after the date of the general election; accepting cash contributions by selling items at fundraisers; lending over $50,000 of his own funds to his campaign; and soliciting and accepting contributions within twenty-eight days of the election. Mr. Anderson therefore filed this suit in October 1999 seeking declaratory and injunctive relief pursuant to 42 U.S.C. § 1983. The suit challenges nine separate Kentucky statutes regulating the conduct of elections and campaign finance: (1) KRS § 117.235(3), which prohibits "electioneering" within 500 feet of polling places; (2) § 121A.080(6), a "turn-over" provision requiring campaigns to turn over unexpended funds to the State; (3) § 121.150(16), prohibiting post-election solicitation of contributions; (4) § 121A.050(2), prohibiting cash contributions; (5) § 121.150(13) & (21), providing that a candidate may not loan more than $50,000 of his personal funds to his campaign; (6) §§ 121.150(24) & 121A.030(5), prohibiting solicitation and acceptance of contributions within twenty-eight days before an election; (7) § 121A in general, regulating the public financing of elections; (8) § 121A.080(4)-(5), which allows candidates receiving public financing to exceed the contributions limit when other candidates do so; and (9) § 121A.010(11), which defines "contribution" as including a candidate's personal funds. The district court granted summary judgment for the Defendants on all counts. Anderson timely appealed to this court. The parties do not dispute any facts, and all of the issues presented in the lawsuit are questions of law.

This court reviews the grant of a motion for summary judgment de novo. Faughender v. City of North Olmsted, 927 F.2d 909, 911 (6th Cir.1991). As an initial matter, we must consider whether certain claims in the case are moot. Counsel at oral arguments asked us to take judicial notice of the fact that the Kentucky legislature has defunded the public finance program in the most recent budget, which states in relevant part:

Notwithstanding KRS 118.255(3), 121.150, 121A.015(5), 121A.020, 121A.030, 121A.040, 121A.060, and 121A.080, no funds shall be appropriated to or received into the election campaign fund established by KRS 121A.020, and the Registry of Election Finance shall make no transfer of funds to any slate of candidates from the election campaign fund for any election. Notwithstanding KRS 121.150(24) and 121A.030(5), slates of candidates may accept contributions within the last 28 days immediately preceding a primary or general election, and in addition to the provisions of KRS 121A.020(5), all contributions to slates of candidates made within the last 28 days immediately preceding a primary or general election shall be reported to the Registry of Election Finance within 24 hours of receipt. All other statutes contained in KRS Chapter 121A shall remain in effect for all slates of candidates, except that KRS 121A.080(6) shall not apply, and slated candidates shall be governed instead by KRS 121.180(10), and KRS 121A.030(4) shall not apply, and all slated candidates may receive contributions from permanent committees which, in the aggregate, shall not exceed 25 percent of the contributions received by the slate in any one election up to a maximum of $300,000 in any one election.

Act of March 23, 2003, ch. 156, 2003 Ky. Laws H.B. 269. Accordingly, Appellant will not be subjected to the operation of the public finance system, the 28-day prohibition on contributions prior to elections, or the trigger provision for the duration of Kentucky's current budget cycle. The question then is whether Anderson's claims related to those features of the public financing system continue to raise issues "affect[ing] the rights of the litigants" in this case. North Carolina v. Rice, 404 U.S. 244, 246, 92 S. Ct. 402, 30 L. Ed. 2d 413 (1971).

This court has previously found that Congress may, through appropriations acts, "suspend, amend, or repeal a statute, so long as it does so clearly." Mullis v. United States, 230 F.3d 215, 217 (6th Cir.2000). In Mullis, the plaintiff filed a petition asking the court pursuant to 18 U.S.C. § 925(c) to remove his firearm disability. Section 925(c) permits a district court to review a decision by the Secretary of Treasury to deny an application for relief of firearms disability. However, Congress had passed eight consecutive appropriations bills prohibiting the Secretary of Treasury from expending any funds to review these applications. Id. at 217. The court found that Congress clearly intended to suspend all relief that was otherwise statutorily authorized by defunding such relief. Id. at 218. In Mullis, the issue before us was whether the court's jurisdiction under § 925(c) extended only to reviewing the discretion of the Secretary, or whether the federal court had a basis for independent judicial review of Mullis's application for removal of disability. Id. at 219. The case did not involve a facial challenge, but only addressed Mr. Mullis's particular petition. Because the Secretary had not rendered a decision to review, this court found that the federal courts lacked jurisdiction.

Applying to state lawmakers Mullis's premise that the legislature may "suspend, amend or repeal" a statute by clear operation of appropriation acts, the question arises whether the acts of the Kentucky legislature in defunding the various parts of this campaign funding scheme renders Mr. Anderson's claims moot.1 Importantly, Mr. Anderson's claims are not limited to this budget cycle. Appellant challenges the application of the 28-day prohibition on contributions and the trigger provision as applied to his candidacy in 1999, and challenges the apportionment of public benefits accomplished by the public funding scheme on its face. As is readily apparent, the action of the legislature suspending the operation of certain provisions of the campaign law does not eliminate the case or controversy which existed by the operation of those laws in 1999. Furthermore, because these laws are still on Kentucky's books, the legislature may choose at any time to allocate funds to the public finance program. Indeed, unlike Mullis, in which the legislature evidenced its intent to cease the operation of the law without regard to budgetary considerations, here there is nothing to suggest that the Kentucky legislature will not choose to refund the campaign scheme at the first fiscally feasible moment. The legislature's suspension of the operation of extant laws is not sufficient to deprive this court of jurisdiction, for it is well-settled that "`voluntary cessation of allegedly illegal conduct does not deprive the tribunal of power to hear and determine the case, i.e., does not make the case moot.'" DeFunis v. Odegaard, 416 U.S. 312, 318, 94 S. Ct. 1704, 40 L. Ed. 2d 164 (1974) (quoting United States v. W.T. Grant Co., 345 U.S. 629, 632, 73 S. Ct. 894, 97 L. Ed. 1303 (1953)). Accordingly, because Mr. Anderson's claims involve a live case or controversy, Kentucky's recent budgetary legislation does not deprive this court of jurisdiction.

WHETHER KENTUCKY'S RESTRICTION ON ELECTIONEERING WITHIN 500 FEET OF POLLING PLACES (KRS § 117.235) IS UNCONSTITUTIONALLY OVERBROAD

The Kentucky legislature prohibits campaign activities near polling places on the date of an election: "No person shall, on the day of any election ... do any electioneering at the polling place or within a distance of five hundred (500) feet of a county clerk's office or any entrance to a building in which a voting machine is located...." KRS § 117.235(3). Anderson brings a two-pronged challenge to the statute, alleging first that the 500-foot buffer zone is generally overbroad in violation of the First Amendment, and second that the definition of electioneering is unconstitutionally overbroad because it includes political speech that does not expressly advocate the election or defeat of candidates for public office. We examine these arguments in turn.

Our inquiry is guided by the Supreme Court's decision in Burson v. Freeman, 504 U.S. 191, 112 S. Ct. 1846, 119 L. Ed. 2d 5 (1992), in which the Court examined the constitutionality of a 100-foot "campaign-free zone" surrounding polling places. First, Burson requires the application of exacting scrutiny to restrictions on political speech around polling places. Burson, 504 U.S. at 198, 112 S. Ct. 1846 (noting that to survive review, a statute must be necessary to serve a compelling state interest and narrowly drawn to achieve that end). While exacting scrutiny does apply, the Supreme Court adopted a modified "burden of proof" in cases in which exercise of a First Amendment right threatens to interfere with the act of voting itself. Id. at 209 n. 11, 112 S. Ct. 1846. Under this modified burden, the state must demonstrate that its response is "reasonable and does not significantly impinge on constitutionally protected rights." Id. at 209, 112 S. Ct. 1846 (emphasis in original) (citation omitted).

The modified burden of proof is an important component of the Burson analysis, for it stands as the Supreme Court's recognition of the deference due to the states in our federal system of government. The states' ability to conduct elections — particularly for state officers — should not be usurped or interfered with by the federal courts absent a clear violation of the United States Constitution. By modifying the burden, the Burson Court recognized that states are uniquely equipped to manage their own elections, which "vary from year to year, and place to place," making it "difficult to make specific findings about the effects of a voting regulation." Id. This modified burden therefore assures that the state's interest in conducting elections is respected, while assuring that the First Amendment rights are not significantly burdened by overbroad regulations.

Applying this standard, the Burson Court recognized two compelling interests for buffer zones around polling places: 1) the state's duty to protect "the right to vote freely for the candidate of one's choice," id. at 199, 112 S. Ct. 1846 (quoting Reynolds v. Sims, 377 U.S. 533, 555, 84 S. Ct. 1362, 12 L. Ed. 2d 506 (1964)); and 2) the state's interest in preserving "the integrity and reliability of the electoral process itself," id. (quoting Anderson v. Celebrezze, 460 U.S. 780, 788 n. 9, 103 S. Ct. 1564, 75 L. Ed. 2d 547 (1983)). Put more succinctly, the Court recognized the states' interest in preventing voter intimidation and election fraud. Burson, 504 U.S. at 206, 112 S. Ct. 1846.

Having found a compelling interest, the Court then went through the history of measures enacted in the states and in other countries to address voter intimidation and election fraud. It concluded that "widespread and time-tested consensus demonstrates that some restricted zone is necessary in order to serve the States' compelling interests in preventing voter intimidation and election fraud." Id. at 206, 112 S. Ct. 1846. The Court therefore upheld a 100-foot buffer zone surrounding a polling place. However, of particular relevance to the instant inquiry, the Burson Court refused to provide a bright-line rule as to how far the State could regulate speech around a poll. Id. at 210-11, 112 S. Ct. 1846 ("[T]his Court has not employed any `litmus-paper test' that will separate valid from invalid restrictions.") (internal citations omitted). Instead, it offered two indications of how to determine "how far is too far" — that is, when a buffer zone is large enough to significantly impinge on protected First Amendment rights. First, the Court noted that the difference between the 100-foot boundary upheld and the 25-foot boundary recommended by those challenging the law was "a difference only in degree, not a less restrictive alternative in kind." Id. at 210, 112 S. Ct. 1846. Second, the Court clarified that "[a]t some measurable distance from the polls ... governmental regulation of vote solicitation could effectively become an impermissible burden akin to the statute struck down in Mills v. Alabama, 384 U.S. 214, 86 S. Ct. 1434, 16 L. Ed. 2d 484 (1966) [(striking down a prohibition on editorials endorsing candidates run on election day)]." Id. at 210, 86 S. Ct. 1434. Thus, at some undetermined distance, the Burson Court acknowledged, a regulation could be different in kind insofar as it impermissibly burdens First Amendment speech.

Appellant asserts that under the Burson standard, the 500-foot buffer zone fails because it is not narrowly tailored. Appellant's use of "narrowly tailored" to describe the test is not precisely correct, and fails to grant the State the deference required. While Burson does refer to narrow tailoring, it later applies the loosened requirement which gives deference to a state regulation so long as it is reasonable and does not "significantly impinge on constitutionally protected rights." Burson, 504 U.S. at 209, 112 S. Ct. 1846 (emphasis omitted) (citation omitted). Applying the Burson "significant impingement" test, we conclude that the 500-foot buffer zone is facially overbroad.

The most recent Kentucky Elections Laws had their genesis in 1987, when the Acting Kentucky Attorney General appointed a Task Force to investigate election fraud. The Report of the Attorney General's Task Force on Election Fraud begins its statement concerning the problem of electioneering near polling places with the justification of preventing corruption:

Presently, the law allows "electioneering" to occur outside a 50 foot radius from the voting machine. This makes it fairly easy for persons interested in subverting the election process to accost voters going to and from the polls and harass or intimidate them within close proximity of the voting booth itself.

But the Task Force does not stop there, and suggests that the State's rationale also may include suppression of protected speech:

Further, many people find even legitimate "electioneering" such as handing out brochures to be offensive when conducted near the polling place, especially when the resulting effect is to be required to "run a gauntlet" in order to enter the polling place.

After finding the existing 50-foot buffer zone inadequate to address the potential for election fraud and "unnecessary" electioneering, a subsequent Special Commission on Election Reform recommended that the buffer zone be increased to 500 feet. The Commission and the Task Force subsequently agreed on suggesting 1000 feet. The buffer zone was reduced in the Kentucky House to 500 feet, which is the distance ultimately codified in KRS § 117.235.

While the State does provide ample evidence of Kentucky's history of election fraud and corrupt elections practices, glaringly thin is its evidence as to why the legislature, the Task Force, or the Commission ultimately arrived at a distance of 500 feet. Even Hiram Ely III, Director of the Kentucky Task Force on Election Fraud, cannot recall how his Task Force arrived at its proposed recommendation of 1,000 feet. Indeed, when pressed in his deposition, Mr. Ely could not even say whether the Task Force would have found 2,000 feet unreasonable:

Q: [A]ccording to your Task Force report, you recommended that 1,000-foot be where the line be drawn. I would infer from that that 2,000 feet would probably be unreasonable?

A: I don't know. We just came up — we came up with that number through the process I described. We thought it was reasonable for the reasons I have stated. And that's all, really, I can say about it.

What little discussion is in the record, however, suggests that the State sought through the 500 foot barrier to prohibit all electioneering speech on election day. For example, the minutes of Second Meeting of the Special Commission on Election offers the following default rationale for adopting the 500 foot barrier:

A thorough discussion ensued relative to instituting a statewide ban of electioneering on election day. The Commission voted by 11 yes and 4 abstentions that there should be a statewide prohibition of electioneering on an election day if such a ban can be established constitutionally. If this is not possible, the Commission proposes to recommend prohibiting electioneering as defined in KRS 117.235 within a 500 foot radius of the entrance to the polling place which is generally used by the voters.

It thus appears that Kentucky, to the extent possible, sought to eliminate all electioneering on election day. This Burson simply does not permit. See Burson, 504 U.S. at 210, 112 S. Ct. 1846 ("At some measurable distance from the polls, of course, governmental regulation of vote solicitation could effectively become an impermissible burden....")

While unable to provide much in the way of an ex ante justification for the specific distance, Kentucky nonetheless proffers numerous after-the-fact testimonials to support the decision to expand the buffer zone to 500 feet. First, Kentucky cites the affidavit of Darrell Fugate, an admitted vote buyer, for the proposition that he bought votes from distances beyond 100 feet of the polling places. Mr. Fugate's testimony is interesting, but not for the reason cited by the State. Rather, he confirms that the regulation has had a detrimental effect on constitutionally protected speech. ("[T]he area surrounding the polling places is no longer crowded with people electioneering and attempting to hand out cards.") Furthermore, Mr. Fugate confirmed that he stopped buying votes not because of the 500-foot speech ban, but rather because of the increase in the criminal penalty for buying votes. Mr. Fugate admits that "[a]fter the legislature changed the statute to make vote selling and vote buying a felony instead of a misdemeanor, I no longer bought votes." By contrast, he speculates that the 500-foot barrier would have only "hindered" his ability to buy votes. Accordingly, Mr. Fugate's testimony undermines the State's position by demonstrating first, that the 500-foot restriction substantially impinges on protected speech, and second, that the government's interest is better served through stricter penalties than through broad bans on protected speech.

Kentucky next claims that the 500-foot barrier puts the vote buyer far enough away that he cannot discern signals commonly used between vote buyers and corrupt poll workers/insiders to determine whether the voter has cast his vote according to the wish of the would-be vote buyer. Referring to a series of photographs taken at polling places depicting a subject signaling by taking off his hat, Appellees assert that "[a]t 100 feet, one can easily see a visual signal such as taking off a hat. At 500 feet, it is virtually impossible to ascertain whether a signal is being given." This evidence, however, appears to be contradicted by the testimony of Mr. Ely, who stated during deposition that the problem of signaling was essentially solved by making the buffer zone apply to the building, rather than to the polling booth itself:

Q: During the — during the decision regarding the 1,000-foot number, do you recall whether — what were you talking about a minute ago — the visual signaling — whether that was a factor in your-all's consideration?

A: I think it was. And I think that was more related to moving it outside the building as opposed to away from the polling place. We used the building as the solution.

This testimony suggests that the major issue in eliminating signaling was making sure that the building was free of electioneers and vote buyers, rather than placing greater geographic distance between the building and the electioneers. Indeed, having solved the problem by fixing the distance to the building rather than the voting booths, Mr. Ely's next statement suggests that something very different from preventing voter intimidation and voter fraud motivated the State to expand the distance outside the building:

I think the — and again, I don't remember how we settled on the number. One of the concerns about it, once you got outside the building, was you didn't want people to have to work their way to the building from the parking lot. You wanted to be able to have somebody park their car and go vote. If they — if they — there would be plenty of electioneers outside 1,000 feet. If they wanted to see them, they could go see them. But if they wanted to simply get out of their car and go vote without having to deal with that, we felt they should be able to do that.

Accommodating the desire of voters to completely avoid contact with anyone handing out legitimate electioneering communications is a far cry from preventing voter intimidation and voter fraud. This theme of completely preventing voter contact with those who have a legitimate electioneering message was repeated by Kentucky in its reliance on the deposition of James Lewis, who ran for county clerk in 1985. Mr. Lewis spoke about the number of campaign workers, and the crowding around polling places. He testified that with 82 people on the ballot, there were at least 150 poll workers. Again, however, it appears that Appellees' interests were often at best aesthetic and at worst suppression of constitutionally protected speech:

You know, it was almost impossible to get in to vote. You had handfuls of cards that people came in, threw down in the floor, threw down in the polling booth, because they weren't interested in those. And I remember, you know, the first election when we had the 500 feet ban, I had, you know, comment after comment from people, that this was the way elections should be, that, you know, they didn't have to run the gauntlet, they didn't have to take all these cards that they didn't want to take, because they didn't want to offend people by not taking their card. It really has changed the appearance of the polling places. You don't have the crowds hanging around on election day.

While Mr. Lewis makes one passing reference to intimidation and vote buying, his testimony is primarily concerned with restraining protected speech, which voters disfavored for reasons of mere convenience. Given these statements, it should not be surprising that Kentucky closes its argument in this section by relying on the "right of every person `to be left [sic] alone'...." Ky. AG. Br. at 21 (quoting Rowan v. United States Post Office Dept., 397 U.S. 728, 90 S. Ct. 1484, 25 L. Ed. 2d 736 (1970)). Thus, the government suggests an interest based upon the fact that "[m]any voters simply do not want to be approached on their way to the voting booth."

Again, this interest is a far cry from the prevention of corruption and intimidation — the only justifications the Supreme Court recognized in Burson to meet the requirements of exacting scrutiny. The Supreme Court has long recognized that "[m]ere legislative preferences or beliefs respecting matters of public convenience may well support regulation directed at other personal activities, but be insufficient to justify such as diminishes the [First Amendment] rights so vital to the maintenance of democratic institutions." Schneider v. State of New Jersey, Town of Irvington, 308 U.S. 147, 161, 60 S. Ct. 146, 84 L. Ed. 155 (1939). Appellees' reliance on Rowan's limited right to be left alone does not tip the jurisprudential scales. The Supreme Court generally has resisted the invitation to extend to public spaces the limited right of individuals to be left alone inside their homes, even when the messages may prove to be offensive to the listener. See generally National Socialist Party of Am. v. Skokie, 432 U.S. 43, 97 S. Ct. 2205, 53 L. Ed. 2d 96 (1977) (granting a stay to an injunction of a Nazi march scheduled to be conducted in a neighborhood that included numerous holocaust survivors, who respectively had voiced their desire to be left alone); see also Cohen v. California, 403 U.S. 15, 91 S. Ct. 1780, 29 L. Ed. 2d 284 (1971) (noting that the risk of offense was not a basis for restricting the ability of a speaker to wear a jacket adorned with a vulgar message in a courthouse, but rather that those offended may overt their eyes). The Kentucky AG points to an exception to the rule that the right to be left alone applies only in the solitude of one's home: the Supreme Court's recent decision in Hill v. Colorado, 530 U.S. 703, 120 S. Ct. 2480, 147 L. Ed. 2d 597 (2000), applying the right to be left alone to buffer zones around abortion clinics. However, this decision addressed a narrower regulation (100-foot buffer zone with 8-foot floating buffers) in the unique context of abortion. As the dissent in Hill noted, the Court's First Amendment jurisprudence in the abortion context is different in kind from its First Amendment jurisprudence outside the abortion context, resembling an "`ad hoc nullification machine' that the Court has set in motion to push aside whatever doctrines of constitutional law stand in the way of that highly favored practice." Hill, 530 U.S. at 741, 120 S. Ct. 2480 (Scalia, J., dissenting) (quoting Madsen v. Women's Health Center, Inc., 512 U.S. 753, 785, 114 S. Ct. 2516, 129 L. Ed. 2d 593 (1994) (Scalia, J., concurring in judgment in part and dissenting in part)). It is therefore dubious at best that this so-called right to be left alone, which under Rowan has peculiar application to the home, should be extended to those approaching polls, especially where the Burson Court did not do so. The Cohen case is instructive regarding the limits of the right to be left alone found in Rowan:

While this Court has recognized that government may properly act in many situations to prohibit intrusion into the privacy of the home of unwelcome views and ideas which cannot be totally banned from the public dialogue, e.g., Rowan v. United States Post Office Dept., 397 U.S. 728, 90 S. Ct. 1484, 25 L. Ed. 2d 736 (1970), we have at the same time consistently stressed that `we are often `captives' outside the sanctuary of the home and subject to objectionable speech.' Id., at 738, 90 S. Ct., at 1491. The ability of government, consonant with the Constitution, to shut off discourse solely to protect others from hearing it is, in other words, dependent upon a showing that substantial privacy interests are being invaded in an essentially intolerable manner.

Cohen, 403 U.S. at 21, 91 S. Ct. 1780. If the right to be left alone provides an insufficient basis for the states to restrict the display of profanity in the courtroom or Nazis marching down residential streets occupied by objecting holocaust survivors, then the State's interest in assuring that voters are not subjected to any unwanted campaign speech alone cannot be a sufficient basis to regulate that clearly protected speech.

Thus, notwithstanding the justifications offered, the evidence provided by Kentucky suggests that the buffer zone was intended to cut off all electioneering speech. This is consistent with the distance of the speech restriction, which, based on the pictures of the polling places in the record, would place many electioneers far beyond the point where they could come into contact with voters. We need only consider how large Kentucky's 500-foot barrier is to recognize the degree to which the restriction impinges on free speech. At first blush, this buffer zone might appear to be five times as large as the 100-foot buffer zone at issue in Burson — an expansion which alone might generate concern. But such a calculation fails to take into account the fact that the buffer zone runs in all directions from the building. Therefore, the buffer zone, unless it is interrupted by private property, covers an area 25 times larger than the area at issue in Burson. The regulation has the potential to silence constitutionally protected speech for 18 acres around a voting booth, and guarantees that those wishing to express their opinions about the election are prohibited from coming within the length of 1 and 2/3 football fields of the polling place. The geographic scope of this regulation alone raises constitutional concerns. See NAACP v. Button, 371 U.S. 415, 438, 83 S. Ct. 328, 9 L. Ed. 2d 405 (1963) ("Broad prophylactic rules in the area of free expression are suspect.").

Burson permits states to create buffer zones around polling places for two purposes only: the prevention of voter intimidation and the prevention of corruption. Kentucky's own witnesses make clear that the extreme geographic distance was not selected for these permissible purposes, but was intended to prevent voters from being bothered by constitutionally protected speech. This Burson does not permit.

At least one other court has found that a 500-foot buffer zone fails the standard established in Burson. See Calchera v. Procarione, 805 F. Supp. 716 (E.D.Wis.1992). Appellees attempt to distinguish Calchera by noting that Kentucky's statute has an exception for electioneering on private property, while the statute at issue in Calchera did not. Appellees therefore argue that at voting places surrounded by private property, electioneering may occur within 500 feet, thereby obviating what would otherwise be 750,000 sq. feet of enforced silence. But the statute's exception is far narrower than Appellees would have this court believe. The private property exception reads:

Nothing contained in this section shall prohibit electioneering conducted within a private residence or establishment other than that in which the polling place is located by persons having an ownership interest in such property.

KRS § 117.235(3) (emphasis added). This exception is substantially narrowed in two ways by its own terms. First, the word "within" makes the exception virtually non-existent by prohibiting any political speech outside the interior confines of the actual house or business. Based upon the plain language of the statute, an individual who owns a house within 500 feet of a polling place may not display a political yard sign, or stand on his lawn or near the edge of his property distributing literature, because this speech would not be "within" a private residence or establishment. Indeed, the only place that the property owner would be free to speak would be physically inside his own home. It strains credulity to assert that a statute is narrowly tailored because it exempts residential living rooms. Second, the statute only exempts speakers who have an ownership interest in the property. Therefore, the owner of a home or business could not invite a campaign worker into her home or establishment to speak about the election without running afoul of the 500-foot restriction on campaign speech. Contrary to the State's assertion, this wafer-thin exception does not cure the overbroad speech regulation.

Furthermore, while it is laudable (and requisite) that the Kentucky statute exempts private property, this carve-out does very little for voting places surrounded by public parking lots and roads. The problem of buffer zones surrounding urban voting places was recognized in Louisiana v. Schirmer, 646 So. 2d 890, 901 (La.1994), in which the court noted that buffer zones raise serious concerns

in a crowded urban context, where, because of greater population density, polling places tend to be more closely situated. In such circumstances the 600 foot [buffer zone] radius may often include a large number of surrounding streets, alleyways, and neutral grounds, with the application of the statute as written stifling political speech in traditional public fora.

Id. Accordingly, for urban voting places in Kentucky, the 500-foot barrier does create 750,000 sq. feet of silence, and threatens to stifle speech in public fora. The fact that the State recognizes that it may not trample free speech within private property does not relieve it of its obligations under Burson to avoid significantly impinging on First Amendment rights on public property.

To contrast Calchera, Appellees point this court to Schirmer v. Edwards, 2 F.3d 117 (5th Cir.1993), in which the Fifth Circuit upheld Louisiana's 600-foot buffer zone around polling places. But see Louisiana v. Schirmer, 646 So. 2d 890 (finding that a 600-foot buffer zone applied to all political speech was overbroad). The Fifth Circuit found that Louisiana had a compelling interest in keeping poll workers — who the State demonstrated were used to intimidate voters — from impeding the voting process, and suggested that the difference between Louisiana's previous 300-foot barrier and the 600-foot barrier is only a difference in degree. Id. at 122. By contrast, the evidence presented by Kentucky's own witnesses suggests that the 500-foot barrier is different in kind: it is designed to prevent voters from having contact with any speech whatsoever immediately prior to voting. This overbroad restriction significantly impinges on protected speech, and fails the test established by Burson.

Overbreadth of "Electioneering"

Appellant also challenges the 500-foot barrier by claiming that its restriction on "electioneering" is overbroad, and captures more constitutionally protected speech than is necessary to promote the State's interest. The 500-foot buffer zone prescribed by Kentucky law applies to electioneering, which is defined to include "the displaying of signs, the distribution of campaign literature, cards, or handbills, the soliciting of signatures to any petition, or the solicitation of votes for or against any candidate or question on the ballot in any manner, but shall not include exit polling." KRS § 117.235(3). Counsel for the Kentucky State Board of Elections informed Mr. Anderson that distributing instructions to voters on how to cast a write-in votes "would be considered `electioneering' and, therefore would be subject to the restrictions of KRS § 117.235 (e.g., [sic] this practice would be precluded at the polling place or within five-hundred (500) feet of the voting place, etc.)."

Mr. Anderson challenges the buffer zone as overbroad based on the content proscribed by the definition of electioneering. Specifically, Anderson suggests that buffer zones should be applied only to "express advocacy," and not to "issue advocacy."

To understand the difference between express advocacy and issue advocacy, we must look to the seminal case of Buckley v. Valeo, 424 U.S. 1, 96 S. Ct. 612, 46 L. Ed. 2d 659 (1976). In Buckley, the Supreme Court interpreted a provision of the Federal Election Campaign Act ("FECA") that capped the amount of independent expenditures an individual could make "relative to a clearly identified candidate...." Id. at 39, 96 S. Ct. 612. The statute therefore treated expenditures that were made by individuals other than the candidate or the candidate's campaign as subject to contribution limitations if the expenditure satisfied the vague requirement of being "relative to a clearly identified candidate."2 In interpreting this section, however, the Court was confronted with a substantial statutory vagueness and overbreadth issue. See McConnell v. Federal Election Comm'n, 540 U.S. 93, 124 S. Ct. 619, 687, 157 L. Ed. 2d 491 (2003). If the Court did not circumscribe the term "relative to," the regulation could apply to broad categories of issue-related speech, which may or may not have any relation to the election or defeat of specific candidates. In order to avoid overbreadth, the Court utilized a bright-line rule, and found that "relative to" referred only to expenditures using terms of express advocacy, which it defined as words such as vote for, elect, support, cast your ballot for, Smith for Congress, vote against, defeat, and reject. Id. at 83-84, 96 S. Ct. 612; see also Buckley, 424 U.S. at 44 & n. 52, 96 S. Ct. 612. By offering a narrowing construction, the Supreme Court interpreted the statute so as to avoid sweeping in more protected speech than is necessary to prevent corruption.

Anderson asserts that the rationale behind the express advocacy distinction applies with equal force in the Burson context. Kentucky responds by noting that Burson, which was decided after Buckley, did not make any distinction between issue and express advocacy. But it would be extreme indeed to infer a rejection of the theory from silence, especially where there is no evidence suggesting that the argument was ever raised before the Burson Court.

Kentucky begins its opposition to applying an issue advocacy/express advocacy distinction by stating that "[n]o court has ever applied these concepts to the protection of the integrity of the polling place." For this proposition, Kentucky again cites Schirmer v. Edwards, in which the Fifth Circuit upheld a complete ban on politicking within a 600-foot zone of a polling place. Appellees interestingly fail to address the subsequent decision of the Louisiana Supreme Court in that very case, which did apply something like an issue advocacy/express advocacy distinction. In Louisiana v. Schirmer, the court found that a complete ban on all political speech (that is, both issue advocacy and express advocacy) within 100 feet of a polling place would likely pass muster. However, the court struck down the same provision when it was extended out to 600 feet, because at such a great distance it significantly impinged upon First Amendment rights by sweeping in adjacent alleys, sidewalks, and other property traditionally open to political discourse. See 646 So.2d at 901.

We must also take note of the Supreme Court's recent decision in McConnell v. Federal Election Commission. In McConnell, the Supreme Court revisited the express advocacy/issue advocacy line first drawn in Buckley. See McConnell, 124 S.Ct. at 687-88. The McConnell Court stated that:

a plain reading of Buckley makes clear that the express advocacy limitation, in both the expenditure and disclosure contexts, was the product of statutory interpretation rather than a constitutional command. In narrowly reading the FECA provisions in Buckley to avoid problems of vagueness and overbreadth, we nowhere suggested that a statute that was neither vague nor overbroad would be required to toe the same express advocacy line.

Id. at 688. Because the Court found that the Bipartisan Campaign Reform Act's definition of "electioneering communication" raised "none of the vagueness concerns that drove our analysis in Buckley," the Court found that the express advocacy distinction was not necessary. Id. at 689. In eschewing the express advocacy distinction, the Court also relied upon substantial evidence that the line between express and issue advocacy had become "functionally meaningless" as applied to the Federal Election Campaign Act. Id. at 689. Accordingly, while the McConnell Court disavowed the theory that "the First Amendment erects a rigid barrier between express advocacy and so-called issue advocacy," it nonetheless left intact the ability of courts to make distinctions between express advocacy and issue advocacy, where such distinctions are necessary to cure vagueness and overbreadth in statutes which regulate more speech than that for which the legislature has established a significant governmental interest. And McConnell in no way alters the basic principle that the government may not regulate a broader class of speech than is necessary to achieve its significant interest.

Unlike the statute at issue in McConnell, Kentucky's statute is vague, sweeping in, inter alia, "the displaying of signs, the distribution of campaign literature, cards, or handbills...." KRS 117.235(3). While this language could be interpreted as limited to express advocacy, the Kentucky State Board of Elections has chosen a broader — indeed an overbroad — interpretation of the statute in finding that instructions on how to cast an absentee ballot constitute electioneering. Also unlike McConnell, the record here is devoid of evidence that such a broad definition is necessary to achieve the State's interest in preventing corruption — or, to use McConnell's words, that an express advocacy line would be "functionally meaningless" as applied to electioneering proximate to voting places.

Accordingly, because Kentucky's statute is vague and because the State has failed to provide any evidentiary support for regulating both express and issue advocacy, we find that this Court should apply a limiting construction. The reasoning of Buckley, McConnell, Schirmer, and Burson suggests that a prophylactic restriction which extends to issue advocacy — that is, protected speech which does not directly seek to elect or oppose specific candidates — cannot be maintained unless the state demonstrates that the limitation was necessary to prevent intimidation and election fraud. Because Kentucky has failed to demonstrate that interest here, we apply a narrowing construction to the term "electioneering," and find that it may permissibly apply only to speech which expressly advocates the election or defeat of a clearly identified candidate or ballot measure.

The question of whether Kentucky may regulate issue advocacy proximate to the polls raises an issue of broad constitutional import. While Mr. Anderson's particular speech — i.e., providing instructions on how to vote for write-in candidates — at first glance looks like a relatively narrow class of speech, his legal challenge to a definition of electioneering which includes issue advocacy raises constitutional concerns about a broad class of speech. The Kentucky State Board of Elections fails to explain why providing instructions on how to cast a write-in vote would constitute "electioneering" for the purposes of the statute. Thus, as best we can tell, the Kentucky law as interpreted by the Board of Elections would forbid an individual to remind voters to fill in the ovals completely on optical scan ballots. Given the Board's decision, it would also appear that individuals would be prohibited from displaying signs or distributing leaflets which fall into core issue advocacy: that is, promoting issues rather than specific candidates. If "electioneering" includes Mr. Anderson's instructing voters on how to cast a write-in vote, does it also include, for example, parents urging voters to "support our schools"? All issue-related speech is chilled by the Board's interpretation of "electioneering." However, the State has failed to provide evidence to support a finding either that a regulation so broad is necessary to prevent corruption and voter intimidation, or that the regulation does not significantly impinge on the rights protected by the First Amendment.

The Kentucky Attorney General objects that a narrowing construction of the term "electioneering" would have dire consequences, and points to Ellis v. Meeks, 957 S.W.2d 213 (Ky.1997), for support. In Ellis, a candidate was found to have violated KRS 117.235 by, inter alia, bringing chicken to poll workers and making it available to voters at polling places. Appellees suggests that under the Appellant's theory, a candidate could engage in "all manner of improper conduct that may or may not constitute `express advocacy'" within 500 feet of a polling place. They offer a parade of horribles, suggesting that such a rule would allow "a candidate [to] stand within 500 feet of the polls and hand out $100 bills or half pints of whiskey to the voters as they enter the polling place." As the Appellant rightly responds, such actions would be conduct — not issue advocacy — and therefore would be regulable whether or not they occurred within 500 feet of a polling place. Indeed, many such actions would be covered by Kentucky's vote-buying statute. See KRS § 119.205 ("Any person who makes or offers to make an expenditure to any person, either to vote or withhold his vote, or to vote for or against any candidate or public question at an election shall be guilty of a Class D felony.")

Accordingly, because the statute is overbroad in that it prohibits speech over too much geography, and because, absent a narrowing construction it prohibits more speech than is necessary to meet the State's protected interest, the decision of the district court must be reversed.

WHETHER KENTUCKY'S DEFINITION OF "CONTRIBUTION" (KRS § 121A.010(11)) IS UNCONSTITUTIONAL ON ITS FACE

In addressing Appellant's numerous challenges to Kentucky's campaign finance statutes, we begin with a challenge to one of the fundamental terms in Kentucky's regulatory scheme: "contribution."

Kentucky's campaign finance statute defines "contribution" to include any "[p]ayment, distribution, loan, deposit, or gift of money or other thing of value...." KRS § 121A.010(11). This definition is broad, and on its face includes disbursements by candidates to their own campaigns. Indeed, other segments of Kentucky's statutory scheme rely upon this broad definition of contribution to indirectly regulate these internal disbursements. In Gable v. Patton, 142 F.3d 940 (6th Cir.1998), this court addressed a challenge to KRS § 121A.030(5), which, with limited exceptions, prohibits a candidate from receiving a contribution within 28 days of an election. Because "contribution" included disbursements made by the candidate to his own campaign, the 28-day ban restricted the ability of a candidate to spend any money not already in the campaign coffers during the final 28 days. Gable, 142 F.3d at 944. In reviewing the statute, this court looked to Buckley, and noted that the Supreme Court "explicitly rejected a lower court's attempt to characterize `the personal funds expended by the candidate on his own behalf as a contribution rather than an expenditure.'" Id. at 952 (citing Buckley, 424 U.S. at 52-53 n. 58, 96 S. Ct. 612). This court therefore struck down the 28-day prohibition as applied to "internal" contributions, because the regulation, while using the term contribution, operated as an impermissible limitation on candidate expenditures. This distinction is of particular import because of the dichotomy Buckley recognized between contributions and expenditures. Candidate expenditures may not ordinarily be capped (except as a condition of participation in public funding program), because such ceilings "impose[] a substantial restraint on the ability of persons to engage in protected First Amendment expression." Buckley, 424 U.S. at 52, 96 S. Ct. 612.

This, however, is not the only place in which the regulatory scheme relies upon a definition of "contribution" to include candidate contributions. Kentucky offers 2-for-1 matching funds for those who participate in their public funding program. Participating candidates must agree to limit their expenditures to $1.8 million. The statute also contains what is referred to as a trigger provision, which states that if a non-participating candidate receives more than $1.8 million in contributions — where contribution includes disbursements by the candidate to his own campaign — then the participating candidate is released from the expenditure ceiling, and may again receive 2-for-1 matching dollars.

Appellant challenges the definition of "contribution" here, arguing that the trigger deterred him from making expenditures. While the challenge is focused on how the term "contribution" is used in the context of the trigger, the broad definition is not found in the section of the code related to the trigger, but rather is given at KRS § 121A.010(11)(a) and applied throughout the regulatory scheme. Accordingly, for this court to apply a facial challenge to the term "contribution" only within the context of the trigger would require us to ignore the very structure of the statute. Therefore, we must examine whether the term "contribution," defined to include candidates' disbursements to their own campaigns, is facially constitutional. We conclude that it is not.

As this court noted in Gable, Buckley drew a line in the sand, and prohibited the government from restricting a candidate's ability to make expenditures on his own behalf. By defining contribution to include contributions by the candidate to his own committee, the statute runs the risk of limiting expenditures. While Kentucky avoids the brunt of this problem by exempting candidate contributions to their own campaigns from otherwise applicable contribution limits, the problem still exists in those sections of the code where candidate contributions to their own campaigns are not exempted from regulation. Thus, by failing to exempt candidate contributions to their own campaigns from the trigger provision, Kentucky applies an indirect regulation on expenditures.

As tempting as it might be to offer a narrowing construction here, the language of the statute does not permit it. Unlike the definition of electioneering communication, which could be plausibly and facially read to apply only to express advocacy, the definition of contribution is not susceptible of a limited reading, and to save it would require this court to take out its blue pencil to add an exception — for candidates' contributions to their own campaign accounts — to the definition. This kind of modification is reserved for the legislature. It is for this court to decide whether the definition of contribution, written as it is, meets the requirements of the Constitution. Because Kentucky's definition of contribution includes what are candidate expenditures, and because this definition is relied upon by other segments of the regulatory scheme to indirectly limit candidate expenditures, the statute infringes upon constitutionally protected speech. Accordingly, the district court erred, and the definition of contribution in KRS § 121A.010 must be struck down.

WHETHER KENTUCKY'S REQUIREMENT THAT CAMPAIGNS "TURN OVER" UNEXPENDED CONTRIBUTIONS AND FUND TRANSFERS TO THE COMMONWEALTH (KRS § 121A.080(6)) IS CONSTITUTIONAL ON ITS FACE AND AS APPLIED TO A WRITE-IN SLATE OF CANDIDATES

Kentucky law requires disgorgement of campaign funds remaining in a campaign account after an election. Specifically, the relevant statute provides that:

[t]he unexpended balance of contributions and fund transfers in a candidate campaign account of a slate of candidates which remains after all financial obligations of the particular election for which the account is established have been satisfied shall be forwarded to the registry for deposit in the fund when the account is closed.

KRS § 121A.080(6). Appellant does not dispute that the statute may be applied to participants in the public financing program. Rather, Appellant argues that the statute (both facially and as applied) violates the First Amendment and the Takings Clause of the Fifth Amendment by implicating those campaign slates that do not participate in the public finance system.

In the trial court, Kentucky conceded that "if read as applicable to both participating and non-participating candidates, KRS § 121A.080(6) `would create a host of constitutional problems.'" In order to avoid these problems, Kentucky urged the trial court to apply a form of the Canon of Constitutional Avoidance, which states that if there are two permissible statutory interpretations, a court should choose the interpretation that will save the statute. See, e.g., Rust v. Sullivan, 500 U.S. 173, 190-91, 111 S. Ct. 1759, 114 L. Ed. 2d 233 (1991). The district court, relying on Kentucky's alternative reading, upheld the statute.

Appellees' alternative reading relies on the fact that the statute requires disgorgement of "[t]he unexpended balance of contributions and fund transfers in a candidate campaign account of a slate of candidates." Registry Br. at 10 (quoting KRS § 121A.080(6) (emphasis added)). Emphasizing the phrase "and fund transfers," Appellees assert that "[b]y its plain meaning, the statute contemplates an account with commingled funds, which include publicly financed, matching dollars." Id. at 10-11.

In order to succeed on the Constitutional Avoidance theory, the reading proffered must be permissible — that is, it cannot be contrary to the plain language of the statute. We think it is a strained reading indeed to find that the statute applies only to those candidate slates that have both unexpended contributions AND fund transfers. Rather, the natural and grammatical reading of the sentence is that BOTH types of funds are subject to disgorgement — whether the candidate has one or both types of funds in his account.

In addition to the natural and grammatical reading of the language, it is useful to look at the statute as a whole. Kentucky demonstrated that when it wished to specify participating or non-participating slates of candidates, it was capable of doing so with clarity. Indeed, just a few short lines above this subsection, the statute ably names both participating and non-participating slates with reasonable clarity:

If the registry makes a finding of fact, after a public hearing of which all slates of candidates for Governor and Lieutenant Governor shall be notified, that in the course of a primary election, runoff primary election, or regular election campaign a slate of candidates for Governor and Lieutenant Governor that has not accepted the provisions of this chapter has received contributions or made expenditures in excess of the expenditure limit as provided in KRS 121A.030(1), the registry shall certify that those slates of candidates that have elected to become eligible for fund transfers or may in the future elect to become eligible for fund transfers shall be released from expenditure limitations.... KRS § 121A.080(4)(a) (emphasis added). Thus, if the legislature intended the disgorgement provision to apply only to participating slates of candidates, reasonable rules of construction force us to presume that it would have specified "participating" slates as it did just a few sentences earlier. Because Kentucky's suggested reading of the statute does not comport with the plain language of the statute, the narrow but implausible reading proffered by the State cannot be relied upon under the Canon of Constitutional Avoidance.

Appellees alternatively assert that their interpretation of the statute is entitled to Chevron deference. Appellees, however, cite no case for the proposition that Chevron deference applies to state agency determinations. Chevron deference is predicated on the idea that legislative gaps serve as delegations from Congress to administrative agencies, whose determinations are given controlling weight unless they are "arbitrary, capricious, or manifestly contrary to the statute." Chevron U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 844, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984). In order to demonstrate that such deference is due to the Kentucky Registry of Election Finance, the agency must, at the very least, establish under Kentucky law that the legislature intends ambiguities or gaps to be treated as delegations to administrative agencies. The agency makes no attempt to do so, and accordingly any claim to Chevron deference must fail.

The Appellees then argue that the claim is not ripe, because the Registry has never sought to enforce the provision against Mr. Anderson or any non-participating slate of candidates. Appellees further assert that they would be precluded from enforcing the provision based upon their previous litigation posture, in which they argued that the provision does not apply to non-participating slates. Kentucky would appear to be correct insofar as Anderson brings the challenge as applied. However, Anderson also brings a facial challenge to the overbreadth of the act. Anderson rightly notes that facial challenges have been permitted even where the agency interpreted a provision in such a way as to preclude enforcement. Thus, in Virginia Soc'y for Human Life, Inc. v. FEC, 263 F.3d 379, 388-90 (4th Cir.2001) ("VSHL"), in a case involving a facially applicable regulation which had not been applied to its challengers, the Fourth Circuit found the case ripe because the agency had not promulgated a rule exempting the parties from the regulation. Id. Similarly, here there is no evidence that the Registry has issued such a regulation, and, notwithstanding its harkening to previous litigation posture, there is nothing to prevent Appellee from promulgating a policy tomorrow applying the statute to non-participating slates in future elections. Therefore, because a party may challenge a statute based upon the "assumption that the statute's very existence may cause others not before the court to refrain from constitutionally protected speech or expression[,]" the facial challenge is ripe for adjudication. Broadrick v. Oklahoma, 413 U.S. 601, 612, 93 S. Ct. 2908, 37 L. Ed. 2d 830 (1973).

Having disposed of Appellees' alternative arguments, we are left with a statute that Appellees concede suffers from a host of constitutional problems. It is clear that the disgorgement constitutes a per se taking for public use. See Brown v. Legal Foundation of Washington, 538 U.S. 216, 123 S. Ct. 1406, 1418-1419, 155 L. Ed. 2d 376 (2003) (finding that a state's act of taking funds out of an IOLTA account would constitute a per se, rather than a regulatory taking). The Fifth Amendment to the United States Constitution therefore requires just compensation, which the State does not provide to those who do not participate in the public funding program. Therefore, the district court erred in finding the disgorgement provision facially constitutional.

WHETHER KENTUCKY'S BAN ON ALL POST-ELECTION CONTRIBUTIONS (KRS § 121.150(16)) IS CONSTITUTIONAL ON ITS FACE

KRS § 121.150(16) prohibits a candidate or his agents from soliciting or accepting contributions after the date of the general election. Relying in large part upon Buckley, the McConnell Court recently restated that contribution limits must be "closely drawn to match a sufficiently important [governmental] interest." McConnell, 124 S.Ct. at 656 (internal quotations and citations omitted).

Kentucky asserts that the statute advances its interest in avoiding the appearance of quid pro quo corruption. Because Buckley recognized that avoiding corruption and the appearance of corruption is sufficient to justify limitations on contributions, there is little doubt that Kentucky has met its burden of demonstrating a sufficiently important interest. See Buckley, 424 U.S. at 26-27, 96 S. Ct. 612.

The question then is whether the regulation is closely drawn to avoid an unnecessary abridgment of associational freedoms. On this point, the district court relied on Alaska v. Alaska Civil Liberties Union, 978 P.2d 597 (Alaska 1999), for the proposition that:

[p]ost-election time limits ... far more clearly address corruption and its appearance because the election has resolved the critical contingency of which candidate will hold office. We think this latter impact on associational rights is narrowly tailored to further compelling state interests.

Id. at 630. The reasoning of the court and of Appellees who rely on this argument appears to be that knowledge of the identity of the successful candidate creates a unique risk of corruption after an election — a unique risk that makes an absolute ban on post-election contributions narrowly tailored to prevent that appearance of corruption. Of course, this unique risk is a double-edged sword. While it may be that post-election contributions are more susceptible to the impression or appearance of corruption when those contributions are made to the winning candidate, the appearance of corruption all but disappears when that same contribution is made to a losing candidate.

Furthermore, the unique risk approach appears to be an attempt to apply the state interest analysis twice, and to skip over the closely drawn (or tailored) component of the Buckley analysis. Thus, to accept Appellees' argument is to find that the State has a sufficient state interest in preventing corruption, and the State's actions are narrowly tailored because ex post election contributions are especially corrupting. But it is not the ban on making contributions after the election that effectively prevents this appearance of corruptions; it is Kentucky's $1,000 contribution limit. While a contribution to a winning candidate may be considered a "sure bet" after the election, the contribution cap prevents the late-comers from being perceived as obtaining or from actually obtaining any more influence than that which a pre-election contributor could muster from a maximum contribution.

In further support of the unique problem of post-election corruption theory, both the district court and Appellees cite to Ferre v. Florida ex rel Reno, 478 So. 2d 1077 (Fla.Dist.Ct.App.1985), in which that court warned that:

if post-election contributions were allowed, a candidate could make large expenditures, secure in the knowledge that immediately after the election members of some currently unpopular group whose support for the candidate had not theretofore been disclosed would send in contributions to pay off the deficit. The result, of course, would be that voters may have been deceived into voting for a candidate with allegiances, or at least bedfellows, with which they violently disagree.

478 So.2d at 1080 n. 10. While disclosure does have a signaling effect, see Buckley, 424 U.S. at 67, 96 S. Ct. 612 (noting that "[t]he sources of a candidate's financial support also alert the voter to the interests to which a candidate is most likely to be responsive"), Ferre overstates the concern, at least as it might apply here. Because Kentucky has a statute capping the amount that any contributor may give to a candidate, concerns about a donor "purchasing" undue influence are substantially mitigated.

Accordingly, because the post-election restriction is not closely drawn, but rather impinges on associational rights even where there is little risk of corruption following an election, the decision of the district court is reversed on this question.

WHETHER KENTUCKY'S BAN ON CASH CONTRIBUTIONS (KRS § 121A.050(2)) IS CONSTITUTIONAL AS APPLIED

Pursuant to KRS § 121A.050(2), Kentucky prohibits any cash contributions to gubernatorial candidates. In an Advisory Opinion issued to Anderson, the Registry clarified that the restriction applied to candidates who do not participate in the public funding scheme, and that campaign paraphernalia purchased from the campaign constitutes a contribution for purposes of the act. Accordingly, candidates in Kentucky are prohibited from selling a campaign button or a bumper sticker for one dollar — or less — without receiving the purchase/contribution price for the item in the form of an negotiable instrument which identifies both the donor and recipient.

The district court recognized, and the Registry concedes, that this cash prohibition is essentially a disclosure requirement. Relying on Buckley's determination that the disclosure requirements of the federal act were constitutional, the district court upheld Kentucky's provision.

Buckley requires that disclosure requirements be subjected to "exacting scrutiny" because of the "significant encroachment on First Amendment rights ... that compelled disclosure imposes...." See Buckley, 424 U.S. at 64, 96 S. Ct. 612. Placing significant weight on a recent decision of this court which held that a provision of Akron's prohibition of cash contributions in excess of $25 did not impose an undue burden on the right of association, see Frank v. City of Akron, 290 F.3d 813 (6th Cir.2002), Appellees argue that Kentucky's disclosure requirement in the form of a prohibition on all cash contributions survives this exacting scrutiny. The Frank Court found that the Act's requirement that modes of payment identify the contributor did not burden the right to associate, and further declared that "[t]he provision serves the significant governmental interest of accountability by forcing contributions to be traceable." Id. at 819.

While Appellees suggest that Frank disposes of this question, such a finding would ignore the difference between the $25 dollar limitation at issue there, and the first penny requirement at issue here. While this may at first glance appear to be a difference in degree and not in kind, it is clear that Kentucky's regulation cannot survive exacting scrutiny because it is not "closely drawn" to avoid unnecessary abridgment of associational freedoms. Buckley, 424 U.S. at 25, 96 S. Ct. 612. The State's ban on all cash contributions effectively forecloses speech by a large body of individuals who will be chilled from making a de minimis contribution. While the parties do not flesh this argument out, we think that as a matter of common sense, a contributor will be substantially less willing — or able — to use a negotiable instrument to make a purchase or contribution of, say, $5 or less than to use such a financial instrument to make a purchase of $25 or more. See, e.g., Fior D'Italia, Inc. v. United States, 242 F.3d 844, 846 (9th Cir.2001), rev'd on other grounds, United States v. Fior D'Italia, Inc., 536 U.S. 238, 122 S. Ct. 2117, 153 L. Ed. 2d 280 (2002) (noting that experience demonstrates that consumer spending habits differ based on whether the transaction is conducted in credit or cash). Thus, to the extent that Kentucky's interest in preventing corruption descends to the smallest contributions, the State could have easily promoted disclosure by requiring the campaigns to take the name of the person making the contribution.3

The Buckley decision itself is instructive of how a statute may be more closely tailored to the end of providing disclosure. The federal statute at issue in Buckley established a two-tier disclosure system, in which campaigns were required to record the names and addresses of those who contributed more than $10, and were required to report to the FEC contributions exceeding $100 in a calendar year. Buckley, 424 U.S. at 82, 96 S. Ct. 612.4 By relying on reporting for even small amounts, rather than a ban on the use of cash, the federal statute provided for disclosure without placing undue burdens on those who wish to associate.

Because the Kentucky statute is not closely drawn to avoid unnecessary abridgment of associational freedoms, we must reverse the district court's judgment finding the prohibition on cash contributions constitutional.

WHETHER KENTUCKY'S PROHIBITION ON A CANDIDATE'S LOANING HIS CAMPAIGN MORE THAN $50,000 (KRS § 121.150(13) & (21)) IS CONSTITUTIONAL ON ITS FACE AND AS APPLIED

Kentucky prohibits candidates from loaning their respective campaigns more than $50,000 in a given election. KRS § 121.150(13). Anderson challenged this provision as being akin to an unconstitutional restriction on candidate expenditures.

We conclude that loans are candidate expenditures, unless and until they are repaid. Kentucky law defines the term "contribution" to include any loan given to a committee. KRS § 121A.010(11)(a)1. Under Buckley, a contribution made by the candidate on his own behalf is an expenditure. See, e.g., Gable, 142 F.3d at 951 (treating the limitation on internal contributions during the 28-day window as a limitation on candidate expenditures). As a matter of campaign finance law, therefore, limitations on candidate loans are limitations on campaign expenditures, and limitations on campaign expenditures are prohibited by Buckley.

The district court rejected the view that loans to campaigns should be treated as expenditures. First, the district court cited to the concern that repayment after the election creates unique risks of quid pro quo corruption. As we have already found, however, the risk of quid pro quo is substantially mitigated by individual contribution limits. See supra, section VI. Furthermore, the risk of quid pro quo is virtually non-existent where the contribution is made to a losing candidate who seeks to recoup some of his debt. Finally, if the risks of after-election repayment are substantial enough to justify regulation, the court must determine whether these particular regulations, that is, the loan limitation and the flat prohibition on post-election contributions combined, are closely drawn to address the risk of corruption.

The district court then adopted the reasoning of Wilkinson v. Jones, 876 F. Supp. 916 (W.D.Ky.1995), which previously upheld the $50,000 loan ban. The Wilkinson Court found two distinct compelling interests: combating the appearance that a "contributor is lining the candidate's pockets" by contributing after the election, and removing "the appearance that heavily indebted candidates are easy bedfellows for quid pro quo contributors." Wilkinson, 876 F.Supp. at 930-31.

Wilkinson's first compelling interest — combating the appearance of a contributor's lining a candidate's pocket — is simply a variation on the post-election contribution issue raised above. We note that while a state does have the ability to regulate in order to combat the appearance of corruption, that appearance must be reasonable. To hold otherwise would be to give unreasonable perceptions of corruption the equivalent of a heckler's veto. The perception that contributions to campaigns, which in turn repay loans to candidates, in some way "line the pockets" of candidates is simply not reasonable, and should not be used as a basis for regulation.

Wilkinson's second compelling interest — removing the appearance that heavily indebted candidates are easy bedfellows for quid pro quo corruption — is once again addressed by the contribution limits. If a $1000 contribution has been found by the Kentucky legislature to be sufficiently low to avoid the appearance or fact of corruption, then a $1000 contribution to a campaign that is indebted to the candidate should also be found to be non-corrupting. While the candidate may have a greater vested interest in assuring that an unsecured loan from his personal coffers gets repaid, the application of Kentucky's contribution limits means that no individual donor can give enough to be considered corrupting or "apparently" corrupting.

Finally, restrictions on loans are particularly onerous because they limit when a party can speak (or how much he can say at a given time). The exigencies of a campaign may require that a candidate spend more early to raise name recognition, or to address an issue of public concern prior to contributions arriving. Indeed, a candidate may need to speak early in order to establish her position and garner contributions. The $50,000 ceiling on loans, while not insubstantial, does significantly impinge upon a candidate's ability to deliver and to time his or her message. We must therefore reverse the district court's judgment concerning the constitutionality of campaign loan limits.

WHETHER KENTUCKY'S PROHIBITION OF CONTRIBUTIONS DURING THE FINAL 28 DAYS BEFORE THE REGULAR ELECTION (KRS § § 121.150(24) & 121A.030(5)) IS CONSTITUTIONAL AS APPLIED

Pursuant to KRS §§ 121.150(24) & 121A.030(5), a candidate is prohibited from receiving contributions in the final 28 days preceding an election, whether the candidate is participating in the public finance scheme or not. Mr. Anderson challenges the applicability of this provision to write-in candidates.

The district court and Appellees assert that this issue is controlled by Gable v. Patton, 142 F.3d 940 (6th Cir.1998). In Gable, this court examined whether non-participating candidates could be subjected to the 28-day restriction. While the Gable court struck down the restrictions on internal contributions (self-financing) during this time period, the court nonetheless upheld the restriction on external contributions during the final 28 days, even as applied to candidates who do not participate in the public finance scheme. Id. at 951. First, the court recognized that the 28-day requirement is an important part of Kentucky's campaign finance scheme. Id. at 950. The court relied upon Kentucky's statements that the 28-day restriction is necessary in order to ensure that all contributions are made before the final pre-election reporting date, so that "if a non-participating slate has exceeded the $1.8 million threshold, the Registry can detect it in time to activate the Trigger." Id. at 949-50. Under the Kentucky campaign finance law, contributions to a participating candidate's campaign are capped at $600,000 (which, with matching public funds equals $1.8 million). The lifting of this contribution cap is "triggered" by a non-participating candidate's receiving an excess of $1.8 million in contributions. Id. at 947. Kentucky argued in Gable that the 28-day window is necessary for the effective operation of the trigger; the trigger provision is an incentive for candidates to participate in the public financing scheme; and the purpose of the public financing scheme is to prevent actual and apparent corruption. Therefore, the State contended application of the 28-day window to non-participating candidates is justified by Kentucky's interest in preventing corruption. This court agreed. Id. at 951.

Here, Anderson challenges the application of the 28-day window to write-in candidates. He attempts to distinguish Gable because write-in candidates are ineligible to participate in the funding scheme, and therefore the 28-day window cannot advance the State's purpose in inducing participation in the public funding program. The district court found this attempt to distinguish the holding in Gable to be a "distinction without a difference." But the district court paints with too broad a brush. The Gable court found that the 28-day window was permissible as applied to non-participating candidates because it advanced Kentucky's interest in combating corruption. Gable, 142 F.3d at 951. The Gable court did not specifically address those candidates who are non-participating because they are ineligible for funding under the act — that is, write-in candidates. While the opinion itself may appear at first blush to be broad enough to encompass both voluntary and involuntary non-participants, Gable's reasoning supports the application of the 28-day window only to voluntary non-participants.

Under the ratio decidendi of Gable, the 28-day window contributes to Kentucky's scheme to combat corruption, but only insofar as it supports the trigger, which in turn channels individuals into the corruption-reducing public finance scheme. Under KRS § 121A, however, write-in candidates are not eligible to participate in that scheme, and therefore cannot be channeled into the public finance system. Therefore applying the 28-day window to write-in candidates simply cannot be intended to combat corruption by channeling write-in candidates into the public finance scheme.

The only remaining question then is whether exempting write-in candidates from the 28-day requirement would impair the incentives for other candidates to participate in the public finance scheme, and thereby undermine the Kentucky statutory framework designed to combat corruption. The public finance program caps contributions to a candidate's campaign at $600,000, and provides matching funds for those dollars on a two-for-one basis. The contribution cap will be lifted at any time (until the final report date which is 28 days before the election) that the non-participating candidate collects more than $1.8 million. If the 28-day window were not applicable to write-in candidates, a candidate eligible for public funding would have an incentive not to participate in the public finance program only if she had reason to believe that a write-in candidate would fail to collect more than $1.8 million before the 28-day window closed, but would substantially outstrip her in contributions in the final 28 days. This is a highly unlikely scenario at best. To the extent that a write-in candidate could raise such last-minute money, those funds would most likely be internal contributions — that is, contributions made by the candidate himself — and this court's decision in Gable clarified that internal contributions are not subject to the 28-day window. Accordingly, we see no reasonable likelihood that exempting write-in candidates would have a negative impact on candidate participation in the program. The 28-day window, as applied to write-in candidates, therefore does not serve the government's purported purpose in combating corruption.

But the 28-day window does unnecessarily abridge associational freedom. As the Supreme Court has noted in McIntyre v. Ohio Elections Com'n, the fact that speech occurs during the heat of an election "only strengthens the protection afforded" under the First Amendment. McIntyre v. Ohio Elections Comm'n, 514 U.S. 334, 347, 115 S. Ct. 1511, 131 L. Ed. 2d 426 (1995). Because write-in candidates are less likely to have substantial name recognition or large bases of support, restricting their ability to receive contributions during the vital 28 days preceding an election constitutes a substantial burden on the associational rights of both the candidate and her would-be supporters. The district court therefore erred in holding that the application of the 28-day window to write-in candidates is constitutional.

WHETHER KENTUCKY'S PUBLIC FINANCE OPTION (KRS 121A) DOES NOT PROVIDE BENEFITS ON A DISCRIMINATORY BASIS IN VIOLATION OF THE CONSTITUTION

Under the provisions of KRS 121A, a write-in candidate is not entitled to matching funds. Anderson challenges this provision as a violation of the Equal Protection Clause. As the district court noted below, Equal Protection challenges to public funding schemes are not unique, as demonstrated by the fact that the Buckley Court entertained an Equal Protection challenge to access to the federal public finance system.

The language in Buckley concerning the inability of a minor candidate to succeed in challenging a public funding scheme on Equal Protection grounds is sweeping, and suggests a low probability of success for such a claim. For example, the Buckley Court found that "Congress' interest in not funding hopeless candidacies with large sums of public money ... necessarily justifies the withholding of public assistance from candidates without significant public support." Buckley, 424 U.S. at 96, 96 S. Ct. 612. The Court then found that eligibility requirements also served the public interest "against providing artificial incentives to `splintered parties and unrestrained factionalism.'" Id. (citation omitted).

The Buckley Court concluded that Equal Protection claims seeking access to public financing are at base claims of the "denial of the enhancement of opportunity to communicate with the electorate that the formulae afford eligible candidates." Id. at 95, 96 S. Ct. 612. However, because the Court found that ineligible candidates are not subject to expenditure limitations applied to those in the public finance system, the Court found that those denied access were not "unfairly or unnecessarily burdened." Id. at 95-96, 96 S. Ct. 612.

In applying Buckley to Kentucky's categorical exclusion of write-in candidates from public-funding eligibility, the district court was "not unmindful that the instant action presents a significant twist on the Buckley equal protection analysis, a distinction that may, indeed, be said to inch the Commonwealth's scheme in the direction of invidious discrimination." Specifically, because of the operation of the "trigger" (which releases participants from spending caps once non-participating candidates have raised $1.8 million), participants in the public financing scheme are not necessarily burdened by an expenditure cap any more than are those candidates who are ineligible to participate in the program. Accordingly, Buckley's rationale that ineligible candidates are not unfairly burdened because they are not subject to the additional expenditure restrictions concomitant with the public finance program is inapplicable. The district court nonetheless and without discussion found that the difference was not significant enough to require a different result from that which occurred in Buckley, and therefore upheld the program.

While this is a close question given that under Kentucky's scheme, write-in candidates are given all the burdens and participating candidates all the benefits, we conclude that the district court did not err in finding that the program can survive on Equal Protection grounds post-Buckley. Specifically, Kentucky's interest in maintaining and managing scarce resources and, as Buckley put it, in "not funding hopeless candidacies with ... public money," is a significant government interest that justifies access requirements for the fund. See Buckley, 424 U.S. at 96, 96 S. Ct. 612. Accordingly, while Buckley is certainly distinguishable insofar as the program here does not balance the burdens among participants and non-participants, the government interest in maintaining scarce resources allows it to treat differently situated candidates differently without running afoul of the Fourteenth Amendment's Equal Protection clause. We therefore affirm the decision of the district court on this question.

WHETHER KENTUCKY'S TRIGGER PROVISIONS ARE CONSTITUTIONAL AS APPLIED

As previously described, Kentucky's trigger provision lifts the contribution cap for candidates participating in the public finance system, and eliminates the 28-day window restricting contributions for all candidates. KRS § 121A.030(5)(a).

Anderson challenges the trigger as excessively coercive. The Gable Court found that the trigger is not excessively coercive, and this panel is bound by that decision. See Gable, 142 F.3d at 949. Anderson also challenges the trigger as applied, suggesting that it is particularly onerous to write-in candidates who are not permitted to participate in the public finance system. Because this argument is essentially a recasting of the Equal Protection claim made in question X, it must fail for the same reason. We therefore affirm the judgment of the district court upholding the constitutionality of the trigger provision as applied.

For the foregoing reasons, the decision of the district court is REVERSED as to those claims addressed by this court in sections III-IX, and AFFIRMED as to those claims addressed in sections X-XI. The case is remanded for further proceedings consistent with this order.