Libertarian National Committee v. FEC

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued November 30, 2018 Decided May 21, 2019 No. 18-5227 LIBERTARIAN NATIONAL COMMITTEE, INC., APPELLANT v. FEDERAL ELECTION COMMISSION, APPELLEE On Certification of Constitutional Questions from the United States District Court for the District of Columbia (No. 1:16-cv-00121) Alan Gura argued the cause and filed the briefs for appellant. Timothy Sandefur and Aditya Dynar were on the brief for amicus curiae Goldwater Institute in support of appellant. Allen Dickerson and Zac Morgan were on the brief for amicus curiae Institute for Free Speech in support of appellant. Jacob S. Siler, Attorney, Federal Election Commission, argued the cause for appellee. With him on the brief were Kevin A. Deeley, Associate General Counsel, and Harry J. Summers, Assistant General Counsel. 2 Paul M. Smith, Tara Malloy, Megan P. McAllen, Fred Wertheimer, and Donald J. Simon were on the brief for amici curiae Campaign Legal Center, et al. in support of appellee. Before: GARLAND, Chief Judge, and HENDERSON, ROGERS, TATEL, GRIFFITH, SRINIVASAN, MILLETT, PILLARD, WILKINS, and KATSAS, Circuit Judges.* Opinion for the Court filed by Circuit Judge TATEL. Opinion concurring in part and dissenting in part filed by Circuit Judge GRIFFITH. Opinion concurring in part, concurring in the judgment in part, and dissenting in part filed by Circuit Judge KATSAS, with whom Circuit Judge HENDERSON joins. TATEL, Circuit Judge: When Joseph Shaber passed away, he left over $235,000 to the Libertarian National Committee (LNC). This case is about when and how the LNC can spend that money. The LNC argues that the Federal Election Campaign Act (FECA), which imposes limits on both donors and recipients of political contributions, violates its First Amendment rights in two ways: first, by imposing any limits on the LNC’s ability to accept Shaber’s contribution, given that he is dead; and second, by permitting donors to triple the size of their contributions, but only if the recipient party spends the money on specified categories of expenses. Scrutinizing each provision in turn, we find no constitutional defects and reject the LNC’s challenges. * Circuit Judge Rao did not participate in this matter. 3 I. Over half a million voters have registered as Libertarians. See Findings of Fact (“CF”) ¶ 3, Libertarian National Committee, Inc. v. Federal Election Commission, 317 F. Supp. 3d 202 (D.D.C. 2018). The LNC, the national committee of the Libertarian Party, has over 130,000 members and about 15,000 active donors. See CF ¶¶ 1, 3. During his lifetime, Joseph Shaber was one of those donors, contributing a total of $3,315 in a series of relatively small donations over some twenty-five years. See CF ¶¶ 109– 10. Unbeknownst to the LNC, Shaber intended to be a donor in death as well. See CF ¶ 115. In 2015, shortly after Shaber had passed away, the LNC learned that Shaber left it the generous sum of $235,575.20. See CF ¶¶ 117, 121. But the LNC had a problem. Under FECA, “no person,” 52 U.S.C. § 30116(a)(1), may make a contribution to a national political party committee above an inflation-adjusted annual limit, see id. § 30116(c)—which, in 2015, capped contributions at $33,400, see CF ¶ 119—and national party committees, in turn, “may not solicit, receive, . . . or spend any funds” donated in excess of that limit, 52 U.S.C. § 30125(a). Furthermore, the Federal Election Commission (the “Commission”), the agency charged with enforcing FECA, interprets “person” to include the dead and their estates. See FEC Advisory Opinion 1999–14 (Council for a Livable World), 1999 WL 521238, at *1 (July 16, 1999) (“[A] testamentary estate is the successor legal entity to the testator and qualifies as a person under the Act . . . .”). Taken together, these restrictions prohibited the LNC from accepting more than $33,400 of Shaber’s donation into the LNC’s general fund in 2015. But there was another way. Just the previous year, in 2014, Congress had amended FECA to permit donors to contribute, 4 over and above their general-purpose contributions, amounts up to three times the base limit into each of three new kinds of “separate, segregated” party-committee accounts. Consolidated and Further Continuing Appropriations Act, 2015, Pub. L. No. 113-235, div. N, § 101, 128 Stat. 2130, 2772–73 (2014) (codified at 52 U.S.C. § 30116(a)(1)(B), (a)(9)). Recipient parties may use these accounts to pay for “presidential nominating convention[s],” party “headquarters buildings,” and “election recounts . . . and other legal proceedings.” 52 U.S.C. § 30116(a)(9). In 2015, then, the LNC could have accepted up to $334,000 from Shaber’s bequest, taking $33,400 into its general fund and $100,200 into each of three segregated funds. The LNC, however, preferred not to tie up the majority of Shaber’s gift in segregated accounts, and the trustee in charge of distributing Shaber’s gift concluded that she had no authority to require the LNC to accept the full bequest into a combination of general- and dedicated-purpose accounts because she “could not impose restrictions on Mr. Shaber’s bequest that Mr. Shaber did not himself place.” CF ¶¶ 126–27. Accordingly, the LNC accepted only $33,400 of Shaber’s donation, see CF ¶ 119, and the trustee asked the Commission for an advisory opinion on what to do with the rest, see 52 U.S.C. § 30108(a) (requiring the Commission to issue written advisory opinions upon request). In that request, the trustee proposed to put the balance of Shaber’s bequest into an escrow account that would disburse the maximum base-limit contribution into the LNC’s general fund each year until the entire gift had been depleted (about seven years in total). See FEC Advisory Opinion 2015–05 (Shaber), 2015 WL 4978865, at *1 (Aug. 11, 2015). The Commission approved this plan, with the caveat that the escrow agreement must prevent the LNC from “exercis[ing] control over the undisbursed funds.” Id. at *3 n.4. 5 In September 2015, the trustee and the LNC signed an agreement under which the remaining $202,175.20 of Shaber’s bequest would be deposited into an escrow account. See CF ¶ 128. Pursuant to the escrow agreement, in January of every year the LNC receives a payment equal to the inflation- adjusted contribution limit. See CF ¶ 128; see also Defendant Federal Election Commission’s Memorandum in Support of its Motion to Dismiss and in Opposition to Plaintiff’s Motion to Certify Facts and Questions, Ex. 27 (“Escrow Agreement”) ¶ 3, Libertarian National Committee, 317 F. Supp. 3d 202 (No. 16- cv-00121), ECF No. 26-31. Although the escrow agreement prohibits the LNC from requesting any money in excess of the contribution limit, it does allow the committee to accept the “entire balance of the Escrow Fund” if it successfully “challenge[s] the legal validity of the [c]ontribution [l]imit in federal court.” Escrow Agreement ¶ 3. The LNC now seeks to do just that. On January 25, 2016, it filed this action challenging both the application of FECA’s contribution limits to Shaber’s bequest and FECA’s new two- tiered limit on contributions to general and segregated accounts. See Complaint ¶¶ 21–34, Libertarian National Committee, 317 F. Supp. 3d 202 (No. 16-cv-00121), ECF No. 1. Proceeding under FECA’s special judicial review provision, the district court then certified factual findings and “non- frivolous constitutional questions” to this en banc court. Holmes v. Federal Election Commission, 875 F.3d 1153, 1157 (D.C. Cir. 2017) (en banc); see also 52 U.S.C. § 30110 (“The district court immediately shall certify all questions of constitutionality of [FECA] to the United States court of appeals for the circuit involved, which shall hear the matter sitting en banc.”). With the benefit of the district court’s findings of fact and certification order, we now consider the three legal questions 6 articulated by the district court. See Order, Libertarian National Committee, 317 F. Supp. 3d 202 (No. 16-cv-00121), ECF No. 34 (“Certification Order”). First: Does imposing annual contribution limits against the bequest of Joseph Shaber violate the First Amendment rights of the Libertarian National Committee? Id. at 2. Second: Do [FECA’s contribution limits], on their face, violate the First Amendment rights of the Libertarian National Committee by restricting the purposes for which the Committee may spend its contributions above [the] general purpose contribution limit to those specialized purposes enumerated in § 30116(a)(9)? Id. Or, put more simply, does FECA’s two-tiered contribution limit, on its face, violate the First Amendment? And third: Do [FECA’s contribution limits] violate the First Amendment rights of the Libertarian National Committee by restricting the purposes for which the Committee may spend that portion of the bequest of Joseph Shaber that exceeds [the] general purpose contribution limit to those specialized purposes enumerated in § 30116(a)(9)? Id. Again, put more simply, does FECA’s two-tiered contribution limit, as applied to Shaber’s bequest, violate the First Amendment? After assuring ourselves of subject-matter jurisdiction, we address each question in turn. 7 II. “[T]he ‘irreducible constitutional minimum’ of standing consists of three elements. The plaintiff must have (1) suffered an injury in fact, (2) that is fairly traceable to the challenged conduct of the defendant, and (3) that is likely to be redressed by a favorable judicial decision.” Spokeo, Inc. v. Robins, 136 S. Ct. 1540, 1547 (2016) (internal citation omitted) (quoting Lujan v. Defenders of Wildlife, 504 U.S. 555, 560 (1992)). The Commission sees three defects in the LNC’s standing. We see none. The Commission first argues that by electing to place the balance of Shaber’s gift into escrow instead of accepting it into segregated accounts, the LNC has inflicted its own injury. See National Family Planning & Reproductive Health Ass’n v. Gonzales, 468 F.3d 826, 831 (D.C. Cir. 2006) (explaining that self-inflicted harm “does not amount to an ‘injury’ cognizable under Article III,” nor is it “fairly traceable to the defendant’s challenged conduct”). Of course the Commission is correct in the most literal sense: the LNC did, indeed, put pen to paper and sign the escrow agreement. But as the district court explained in rejecting the Commission’s self-infliction argument, the LNC’s injury stems not from its inability to accept the entire bequest immediately (which it could have done), but rather from the committee’s “inability to accept [immediately] the entire bequest for general expressive purposes” (which FECA prohibits). Libertarian National Committee, Inc. v. Federal Election Commission, 228 F. Supp. 3d 19, 25 (D.D.C. 2017). The Commission forced the LNC to choose between immediate access to the money and long-term flexibility in spending it; that the committee chose the lesser of two evils hardly transforms FECA’s limitation into a self- imposed restriction. 8 The Commission, however, has a response: because “[m]oney is fungible,” a dollar contributed into a segregated account “is an extra dollar from the . . . general account that becomes available for [the LNC’s] general expressive purposes.” Federal Election Commission’s Motion to Dismiss for Lack of Subject-Matter Jurisdiction (“Motion”) at 14–15. Perhaps so, but the arithmetic just does not work. In 2015, the year the LNC first gained access to Shaber’s $235,575 bequest, it spent only $341 on its 2016 presidential nominating convention and $7,261 on legal proceedings. Therefore, even assuming the LNC could have maxed out its headquarters spending at $100,200 and accepted an additional $33,400 into its general account, some $94,373 of Shaber’s bequest would have remained unused as of December 31, 2015. Contrary to the Commission’s argument, we have no need to examine the LNC’s “2016 budget expectations and expenditures.” Motion at 17. True, the LNC must demonstrate standing “as of the time [its] suit commence[d]” in January 2016, Del Monte Fresh Produce Co. v. United States, 570 F.3d 316, 324 (D.C. Cir. 2009), and expense reports reveal that by the end of 2016, the LNC had incurred enough convention, headquarters, and legal costs to have fully absorbed what remained of Shaber’s donation—assuming the money it spent on those expenses was itself unrestricted and thus fully fungible. But by January 2016, Shaber’s bequest sat locked in an escrow account over which—at the Commission’s direction—the LNC exercised “no control.” FEC Advisory Opinion 2015–05 (Shaber), 2015 WL 4978865, at *3 (Aug. 11, 2015). The relevant date is therefore September 2015, when the LNC committed itself to the escrow arrangement. At that time, although the committee may have projected certain expenses, it lacked perfect information about what costs it would incur and what other donations it might receive in the new year. We cannot rely on hindsight to fault the LNC for its failure of 9 foresight, and in any event, our task is not to assess the committee’s financial planning acumen. Rather, we must determine only whether the LNC suffered a cognizable injury in fact that is fairly traceable to the Commission’s conduct (and, by extension, to FECA). The LNC easily clears that bar. Next, the Commission argues that a favorable judicial determination could not redress the LNC’s injury because this suit, filed in 2016, seeks only injunctive and declaratory relief for harm suffered a year earlier in 2015, when Shaber’s bequest became available. To be sure, our Article III authority does not include the power to turn back time. Nonetheless, much of the money remains tied up in escrow, and we most certainly do have authority to invalidate the challenged portions of FECA— which, per the escrow agreement, would afford the LNC immediate access to the remainder of the bequest for all purposes. See Escrow Agreement ¶ 3. That is redress. Finally, the Commission points out that the LNC “lacks standing to the extent its claims” depend on the allegation that the challenged contribution limits “place the Libertarian Party at a competitive disadvantage vis-à-vis other political parties,” which, the Commission argues, “is akin to the oft-rejected argument that a party is harmed because it is at a fundraising disadvantage to its competitors.” Motion at 20–21. But according to the LNC, “that extent is zero.” Plaintiff’s Opposition to Defendant’s Motion to Dismiss (“Opposition”) at 15. Taking the LNC at its word, we conclude, as did the district court, that the committee has alleged a cognizable harm in its inability to accept immediately “the entire bequest for general expressive purposes.” Libertarian National Committee, Inc., 228 F. Supp. 3d at 25. 10 III. We proceed to the first certified question: whether applying FECA’s annual contribution limits specifically to Shaber’s bequest violates the LNC’s First Amendment rights. A. As the Supreme Court recognized in Buckley v. Valeo— its first and seminal case examining FECA’s constitutionality—contribution limits “operate in an area of the most fundamental First Amendment activities.” 424 U.S. 1, 14 (1976) (per curiam). “There is no right more basic in our democracy,” the Chief Justice explained in his recent plurality opinion in McCutcheon v. Federal Election Commission, “than the right to participate in electing our political leaders.” 572 U.S. 185, 191 (2014) (plurality opinion). In fact, political contributions implicate two distinct First Amendment rights: freedom of speech and freedom of association. “When an individual contributes money to a candidate, he exercises both of those rights: The contribution ‘serves as a general expression of support for the candidate and his views’ and ‘serves to affiliate a person with a candidate.’” McCutcheon, 572 U.S. at 203 (plurality opinion) (quoting Buckley, 424 U.S. at 21–22). The recipient, too, has First Amendment interests in accepting campaign contributions. “[V]irtually every means of communicating ideas in today’s mass society requires the expenditure of money,” from “distributi[ng] . . . the humblest handbill,” to “hiring a hall and publicizing” rallies, to purchasing airtime on “television, radio, and other mass media.” Buckley, 424 U.S. at 19. And, of course, just as contributors associate with candidates and parties by making donations, so, too, do recipients associate with contributors by accepting donations. See id. at 18, 22 (explaining that contributions “enable[] like-minded persons to 11 pool their resources in furtherance of common political goals” and that contribution limits therefore restrict “association by persons, groups, candidates, and political parties”). Altogether, then, in the world of political contributions, the First Amendment protects two kinds of rights (speech and association) belonging to two different rights-holders (donors and recipients). As the parties argue this case, however, the First Amendment interests at issue occupy only one box of the rights/rights-holders two-by-two matrix. Because “Shaber’s death ended his expression and association,” and because the LNC “does not associate with the dead,” the committee admits that “[t]his case concerns primarily the LNC’s speech rights with respect to the Shaber bequest.” Appellant’s Br. 34–35. We thus find ourselves in the speech-recipient box. According to the Commission, contribution limits have only minimal bearing on a recipient’s free-speech rights. On the one hand, as the Commission observes, the Court held in Buckley that “restriction[s] on the amount of money a . . . group can spend on political communication during a campaign”— that is, expenditure limits—“necessarily reduce[] the quantity of expression” and therefore receive “the exacting scrutiny applicable to limitations on core First Amendment rights.” Buckley, 424 U.S. at 19, 44–45 (emphasis added). On the other hand, restrictions on the amount of money someone can donate—that is, contribution limits—“merely . . . require candidates and political committees to raise funds from a greater number of persons” “rather than . . . reduce the total amount of money potentially available to promote political expression.” Id. at 22. Therefore, as the Court explained in Buckley and reiterated in McConnell v. Federal Election Commission, “[b]ecause the communicative value of large contributions inheres mainly in their ability to facilitate the speech of their recipients, . . . contribution limits impose 12 serious burdens on free speech only if they are so low as to ‘preven[t] candidates and political committees from amassing the resources necessary for effective advocacy.’” McConnell v. Federal Election Commission, 540 U.S. 93, 135 (2003) (third alteration in original) (quoting Buckley, 424 U.S. at 21); see also Randall v. Sorrell, 548 U.S. 230, 248 (2006) (plurality opinion) (explaining that contribution limits fail “to survive First Amendment scrutiny” if they “prevent candidates from ‘amassing the resources necessary for effective [campaign] advocacy’” or “magnify the advantages of incumbency to the point where they put challengers to a significant disadvantage” (alteration in original) (quoting Buckley, 424 U.S. at 21)). If that is the test, then FECA’s contribution limit as applied to Shaber’s bequest clearly passes. The LNC nowhere claims that it needs Shaber’s money in order to “amass[] the resources necessary for effective advocacy.” Buckley, 424 U.S. at 21. Surely, Shaber’s gift hardly represents a make-or-break sum for the committee’s ability to engage in political communication. We doubt, moreover, that the LNC could make such a showing given that FECA’s current contribution limits are no lower than the ceilings the Court approved in McConnell. With respect to donors’ rights, by contrast, contribution limits tread closer to core First Amendment activity. To be sure, the speech embodied by a political contribution lacks nuance: because a contribution “does not communicate the underlying basis for the [donor’s] support,” “[a]t most, the size of the contribution provides a very rough index of the intensity of the contributor’s support for the candidate.” Buckley, 424 U.S. at 21. That said, the ability to express support through monetary donations provides an “important means of associating with a candidate or committee,” id. at 22—and a particularly important means, at that, for “individuals who do 13 not have ready access to alternative avenues for supporting their preferred politicians,” such as volunteering in person, McCutcheon, 572 U.S. at 205 (plurality opinion). To protect contributors’ heterogeneous First Amendment interests in making political donations, therefore, the Court has announced a single unified test that applies an intermediate level of scrutiny to contribution limits. See Nixon v. Shrink Missouri Government PAC, 528 U.S. 377, 388 (2000) (explaining that “a contribution limitation surviving a claim of associational abridgment would survive a speech challenge as well”). “Closely drawn” scrutiny, as the Court now calls it, requires that “the [government] demonstrate[] a sufficiently important interest and employ[] means closely drawn to avoid unnecessary abridgment” of First Amendment rights. Buckley, 424 U.S. at 25; see also McCutcheon, 572 U.S. at 197 (plurality opinion) (same). But these decisions have left open the question whether closely drawn scrutiny—usually justified as a mechanism to safeguard donors’ rights—also applies to a law limiting a recipient’s right to receive a donation absent a corollary restriction on a contributor’s right to contribute. Because the typical donor is a living human being capable of both speaking and associating, neither the Supreme Court nor we have had occasion to untangle a recipient’s rights from its donors’. But even though Shaber no longer speaks nor associates, Buckley and its progeny hardly foreclose application of closely drawn scrutiny to the contribution limit at issue in this case. We shall therefore assume, without deciding, that closely drawn scrutiny applies to the imposition of contribution limits on Shaber’s bequest. And because we conclude that FECA’s limits survive even that heightened standard of review, we have no need to interrogate that assumption further. 14 B. “In a series of cases over the past 40 years,” the Supreme Court has repeatedly recognized the government’s interest in imposing contribution limits to combat “‘quid pro quo’ corruption [and] its appearance.” McCutcheon, 572 U.S. at 192 (plurality opinion) (emphasis omitted). The risk that candidates might exchange political favors for money is far from hypothetical. As the Court explained in McConnell, “[t]he idea that large contributions to a national party can corrupt or, at the very least, create the appearance of corruption of federal candidates and officeholders is neither novel nor implausible.” 540 U.S. at 144. Indeed, both Buckley and McConnell cited “deeply disturbing examples” of “pernicious practices” in then- recent election cycles. Buckley, 424 U.S. at 27; see also McConnell, 540 U.S. at 122 (noting “disturbing findings of a Senate investigation into campaign practices related to the 1996 federal elections”). Therefore, given the threat posed by actual and apparent corruption to “the integrity of our system of representative democracy,” Buckley, 424 U.S. at 26–27, the Court has long held that “the Government’s interest in preventing quid pro quo corruption or its appearance . . . may properly be labeled ‘compelling,’” McCutcheon, 572 U.S. at 199 (plurality opinion) (emphasis omitted) (quoting Federal Election Commission v. National Conservative Political Action Committee, 470 U.S. 480, 496 (1985)). The risk of quid pro quo corruption does not disappear merely because the transfer of money occurs after a donor’s death. Individuals planning to bequeath a large sum to a political party have two points of leverage during their lifetimes: they may tell the party about their intentions, and they may change their minds at any time. That latter possibility, as the district court found, “creates an incentive for a national party committee to limit the risk that a planned bequest will be revoked” and could cause that party, “its candidates, or its 15 office holders to grant political favors to the individual in the hopes of preventing the individual from revoking his or her promise.” CF ¶ 100 (first quoting Findings of Fact ¶ 92, Libertarian National Committee, Inc. v. Federal Election Commission (LNC I), 930 F. Supp. 2d 154, 186 (D.D.C. 2013), aff’d, No. 13-5094, 2014 WL 590973 (D.C. Cir. Feb. 7, 2014); then quoting Defendant Federal Election Commission’s Proposed Findings of Facts ¶ 80, Libertarian National Committee, 317 F. Supp. 3d 202 (No. 16-cv-00121), ECF No. 26-3) (internal quotation marks omitted). In other words, a donor’s death simply imposes a sequencing constraint on a quid pro quo exchange. Instead of money for votes, the donor requires votes for money—or, to be more precise, political favors now for the promise of money later. And even that constraint evaporates in the case of corrupt donors seeking favors for their survivors. Although an individual’s death terminates his ability to profit personally from a corrupt quo in exchange for his bequeathed quid, the donor’s surviving friends and family remain all too capable of accepting political favors that their deceased benefactor may have pre-arranged for their benefit. What’s more, where the courts have observed a risk of corruption, so too will the electorate. As the Court explained in Buckley, “[o]f almost equal concern as the danger of actual quid pro quo arrangements is the impact of the appearance of corruption stemming from public awareness of the opportunities for abuse inherent in a regime of large individual financial contributions.” 424 U.S. at 27. Voters lack the means to examine the intentions behind suspiciously sizable contributions, a problem that becomes especially acute in the case of a deceased donor who, of course, is forever unavailable to answer inquiries. As a result, the corruptive potential of unregulated contributions, including the unregulated 16 contributions of the dead, inflicts almost as much harm on public faith in electoral integrity as corruption itself. The LNC acknowledges these risks. “Nobody here disputes the theoretical corruption potential of bequests,” declares the committee. Reply Br. 13. And as a result, the LNC has declined, both before the district court and on appeal, to “revisit” the conclusion that bequests “generally warrant[] . . . subjection to FECA’s contribution limits.” Appellant’s Br. 35; see also CF ¶ 93 (“‘[I]t is possible for a bequest to raise valid anti-corruption concerns,’ as the LNC has ‘concede[d].’” (alterations in original) (quoting LNC I, 930 F. Supp. 2d at 166)). It is precisely because the LNC concedes “the theoretical corruption potential of bequests,” Reply Br. 13, that we do not share our dissenting colleague’s concern that “the [Commission] points to nothing substantiating” the same, Op. at 10 (Katsas, J.). The government may, just like any other litigant in any other case, accept an opposing party’s concession. Moreover, among the district court’s findings that the LNC declines to dispute, see Oral Arg. Rec. 32:01–18 (conceding that this court is bound by the district court’s findings of fact unless clearly erroneous), are several that amount to substantial evidence demonstrating the government’s anticorruption interest in regulating bequests. To begin with, contrary to the dissent’s assertion that “bequests are rarely used for political contributions,” Op. at 10 (Katsas, J.), the district court found that since 1978 donors have contributed “more than $3.7 million in bequeathed funds,” not infrequently in five- and six-figure amounts. CF ¶ 102; see also CF ¶¶ 103–08 (listing bequeathed contributions to national political party committees). And that figure is “likely underreported,” as “reporting entities are not required to inform the [Commission] that a particular contribution they received 17 came from a bequest.” CF ¶ 102. In fact, the LNC did not report Shaber’s bequest as such. See CF ¶ 102. Furthermore, the district court found that “nothing prevents a living person from informing the beneficiary of a planned bequest about that bequest,” CF ¶ 94; that “[p]olitical committees ‘could feel pressure to . . . ensure that a (potential) donor is happy with the committee’s actions lest [that donor] revoke the bequest,’” CF ¶ 100 (second and third alterations in original) (quoting LNC I, 930 F. Supp. 2d at 167); and that this pressure could cause a “national party committee, its candidates, or officeholders . . . [to] grant that individual political favors,” CF ¶ 99 (internal quotation marks omitted). Altogether, the district court’s 178 paragraphs of findings amount to much more than “‘mere conjecture,’” Op. at 11 (Katsas, J.) (quoting McCutcheon, 572 U.S. at 210 (plurality opinion)), that bequests pose a threat of quid pro quo corruption. Disclaiming any “categorical challenge to the limitation of all bequests,” the LNC instead asks us to conduct an “as- applied” inquiry “narrowly focused on one particular bequest”: “whether Shaber’s bequest, specifically, warrants government limitation.” Appellant’s Br. 30, 35. It does not, says the LNC, because the bequest was not corrupt and the government therefore has no legitimate interest in its restriction. As to the first half of the LNC’s argument, we have no trouble making the unremarkable assumption that Shaber’s contribution was not, in fact, part of a corrupt quid pro quo exchange. Buckley rested on precisely the same assumption— that “most large contributors do not seek improper influence over a candidate’s position or an officeholder’s action.” Buckley, 424 U.S. at 29. Indeed, the LNC’s observation that contribution limits restrict legitimate as well as corrupt donations is wholly unsurprising. The Court has often “noted that restrictions on direct contributions are preventative, 18 because few if any contributions to candidates will involve quid pro quo arrangements.” Citizens United v. Federal Election Commission, 558 U.S. 310, 357 (2010) (emphasis omitted). But that is precisely the point: it is “difficult to isolate suspect contributions” in the sea of legitimate donations. Buckley, 424 U.S. at 30. As the LNC sees it, because the government’s interest lies in preventing quid pro quo corruption, the government may restrict only corrupt contributions. The government, however, already has those restrictions on the books: they are called bribery laws. But bribery laws “deal with only the most blatant and specific attempts of those with money to influence governmental action,” id. at 28, and if those laws were sufficient to achieve the government’s compelling interest in preventing quid pro quo corruption and its appearance, then Congress would have had no need in the first place to impose contribution limits to combat prior decades’ “deeply disturbing” quid pro quo arrangements, id. at 27. Accordingly, the problem with the LNC’s proposed regime—one under which actually noncorrupt contributions could exceed FECA’s limits—is that corruption is notoriously difficult to ferret out, and “the scope of . . . pernicious practices can never be reliably ascertained.” Id. Because “the First Amendment does not require Congress to ignore the fact that ‘candidates, donors, and parties test the limits of the current law,’” McConnell, 540 U.S. at 144 (quoting Federal Election Commission v. Colorado Republican Federal Campaign Committee, 533 U.S. 431, 457 (2001)), “prophylactic” contribution limits, McCutcheon, 572 U.S. at 221 (plurality opinion), are permissible—even vital— to forestall the worst forms of political corruption. Critically, moreover, even if through some omniscient power courts could separate the innocent contributions from 19 the nefarious, an appearance of corruption would remain. Although “Congress may not regulate contributions simply to reduce the amount of money in politics,” id. at 191 (plurality opinion), it may certainly do more than ask the public to place groundless faith in a bribery-prevention scheme that has failed to thwart corruption in the past. “It is therefore reasonable,” the Court explained in McConnell, “to require that all parties and all candidates follow the same set of rules” in order to prevent “‘both the actual corruption threatened by large financial contributions and the eroding of public confidence in the electoral process through the appearance of corruption.’” 540 U.S. at 136, 159 (quoting Federal Election Commission v. National Right to Work Committee, 459 U.S. 197, 208 (1982)). That is not to say as-applied challenges to FECA’s contribution limits are impossible. Because restrictions that strike a permissible balance between governmental and individual interests may nonetheless “impose heavy burdens on First Amendment rights in individual cases,” John Doe No. 1 v. Reed, 561 U.S. 186, 203 (2010) (Alito, J., concurring), people may bring as-applied challenges to demonstrate that, in their unique circumstances, the law in question works too harshly. For example, “a nascent or struggling minor party can bring an as-applied challenge” to a contribution limit that “prevents [the party] from ‘amassing the resources necessary for effective advocacy,’” McConnell, 540 U.S. at 159 (quoting Buckley, 424 U.S. at 21), and, similarly, a group may bring an as-applied challenge to a campaign-contribution disclosure provision that subjects its donors to “‘threats, harassment, or reprisals,’” Citizens United, 558 U.S. at 367 (quoting McConnell, 540 U.S. at 198); see also Doe, 1 v. Federal Election Commission, 920 F.3d 866, 871 (D.C. Cir. 2019) (“Citizens United left open the possibility of an as-applied First Amendment challenge, but only if the donor proved that revealing its identity would probably bring about threats or 20 reprisals.”). But while an individual may demonstrate that, in his particular case, a contribution limit imposes an impermissibly high burden, donors and recipients may not use the guise of an as-applied challenge merely to relitigate the government’s settled interest in enforcing “preventative” limits, Citizens United, 558 U.S. at 357, against all contributions—corrupt and noncorrupt alike. “[A] plaintiff cannot successfully bring an as-applied challenge to a statutory provision based on the same factual and legal arguments the Supreme Court expressly considered when rejecting a facial challenge to that provision.” Republican National Committee v. Federal Election Commission, 698 F. Supp. 2d 150, 157 (D.D.C.) (three-judge panel), aff’d, 561 U.S. 1040 (2010). Unlike the LNC and the dissent, see Op. at 18 (Katsas, J.), we see nothing to the contrary in SpeechNow.org v. Federal Election Commission, 599 F.3d 686 (D.C. Cir. 2010) (en banc). In that case, we sustained an as-applied challenge to a contribution limit on the grounds that “the government ha[d] no anti-corruption interest in limiting contributions to an independent expenditure group,” id. at 695, but we did not do so because of anything special about the government’s anticorruption interest “in that case” in particular, Op. at 18 (Katsas, J.). Instead, we explained that because the Supreme Court had recently held in Citizens United “as a matter of law that independent expenditures do not corrupt or create the appearance of quid pro quo corruption,” neither could contributions to independent expenditure-only groups “corrupt or create the appearance of corruption.” SpeechNow.org, 599 F.3d at 694 (emphasis omitted). In this case, by contrast, the LNC raises no challenge to Buckley nor to the anticorruption interest that case and its successors recognized. See Appellant’s Br. 60 n.13 (“[T]his case does not challenge Buckley.”). 21 The dissent suggests that even if the government has an interest in limiting bequests disclosed during donors’ lifetimes, it lacks a similar interest in regulating the class of bequests kept secret until donors’ deaths. See Op. at 12–14 (Katsas, J.). The trouble, however, is that because the LNC states in no uncertain terms that its “as-applied Shaber challenge . . . does not contest any contribution limit’s general sweep,” Reply Br. 11, we are limited to addressing only the matters raised and litigated by the parties and certified to this court for review, see 52 U.S.C. § 30110—that is, whether “imposing annual contribution limits against the bequest of Joseph Shaber” violates the LNC’s First Amendment rights. Certification Order 2. Indeed, the LNC expressly foreswears any broader challenge. See supra at 17. Perhaps, as the dissent proposes, the Commission might be able to “police” bequest disclosures in the same manner it distinguishes coordinated from independent expenditures. Op. at 14 (Katsas, J.). But there are significant differences, both practical and constitutional, between independent expenditures, coordinated expenditures, and contributions. See supra at 11–12; see also McConnell, 540 U.S. at 221 (explaining that coordinated expenditures “may be regulated as indirect contributions”). For now, then, we simply observe that the task of distinguishing truly uncoordinated from covertly disclosed bequests would seem to require the same sorts of fact-intensive inquiries and give rise to the same sorts of appearance-of-corruption concerns that prophylactic contribution limits are designed to avoid. Without the parties to guide us, we decline to venture into such challenging terrain. See Carducci v. Regan, 714 F.2d 171, 177 (D.C. Cir. 1983) (“The premise of our adversarial system is that appellate courts do not sit as self-directed boards of legal inquiry and research, but essentially as arbiters of legal questions presented and argued by the parties before them.”). 22 We thus return to the LNC’s bottom line: “[W]hat about Shaber?” Reply Br. 14. By the LNC’s logic, the only individuals who must keep their contributions under FECA’s limits are those who intend to violate the bribery laws. That just cannot be what the First Amendment requires. We therefore answer the first certified question in the negative: imposing FECA’s contribution limits on Shaber’s bequest does not violate the LNC’s First Amendment rights. IV. This brings us to the second and third certified questions— a facial and an as-applied challenge—which ask whether it offends the First Amendment that donors may contribute above the base limit only if they make their contributions into segregated, dedicated-purpose accounts. A. The only portion of FECA at issue here is an amendment contained in the Consolidated and Further Continuing Appropriations Act—what we reluctantly assent to calling the “cromnibus” amendment. The LNC assures us, as it must, that it “would not have brought, and the District Court would not have certified, a challenge to the sort of contribution limits that the Supreme Court upheld in McConnell.” Appellant’s Br. 40. Instead, the LNC contends that because the 2014 cromnibus amendment “radically altered FECA’s nature and structure,” id., we must now apply a heightened level of scrutiny. What was constitutional before, the theory goes, is constitutional no longer. Accordingly, we begin by considering precisely what “sort of contribution limits . . . the Supreme Court upheld in McConnell.” Id. A little history will help. 23 In the FECA Amendments of 1976, Congress imposed a $20,000 limit on “contributions” to national party committees. Federal Election Campaign Act Amendments of 1976, Pub. L. No. 94-283, § 112(2), 90 Stat. 475, 487 (codified as amended at 52 U.S.C. § 30116(a)(1)(B)). But not all donations qualified as contributions. Instead, FECA defined “contribution” as a gift “made . . . for the purpose of influencing any election for Federal office,” thus leaving unregulated any money ostensibly donated for the purpose of influencing state and local elections. Federal Election Campaign Act Amendments of 1979, Pub. L. No. 96-187, § 101, 93 Stat. 1339, 1340 (1980) (codified as amended at 52 U.S.C. § 30101(8)). And so “soft money” was born. While FECA subjected contributions for the purpose of influencing federal elections (so-called hard money) to its limits, parties remained free to “raise [soft money] in massive dollops from single contributors.” Shays v. Federal Election Commission, 414 F.3d 76, 81 (D.C. Cir. 2005). “Over time, political parties took increasing advantage of . . . soft money opportunities,” id., causing, as the Senate Committee on Governmental Affairs described it, “a ‘meltdown’ of the campaign finance system,” McConnell, 540 U.S. at 129 (quoting S. Rep. No. 105-167, vol. 4, at 4611 (1998); id., vol. 5, at 7515). Seeking to close the “soft-money loophole,” McConnell, 540 U.S. at 133, Congress enacted the Bipartisan Campaign Reform Act in 2002. See Bipartisan Campaign Reform Act of 2002, Pub. L. No. 107-155, 116 Stat. 81. Through that statute, known as BCRA, Congress took a two-pronged approach to purging federal elections of soft money: it prohibited national political party committees from accepting or “spend[ing] any funds” “not subject to” FECA, and it prohibited (with limited exceptions) state, district, and local party committees from “expend[ing] or disburs[ing] for Federal election activity” any funds raised outside FECA’s limits. Id. § 101 (codified at 52 24 U.S.C. § 30125(a), (b)). Approving these soft-money restrictions in McConnell, the Supreme Court rejected the argument that BCRA imposes an impermissible expenditure limit rather than a permissible contribution limit. According to the Court, BCRA’s soft-money ban, though styled as a restriction on party “spending,” “simply limit[s] the source and individual amount of donations” without “limit[ing] the total amount of money parties can spend.” McConnell, 540 U.S. at 139. “[I]t is irrelevant,” the Court explained, “that Congress chose . . . to regulate contributions on the demand rather than the supply side.” Id. at 138. So what changed? The 2014 cromnibus amendment introduced gradations into the political party contribution limit where none had been before. As previously explained, see supra at 3–4, FECA now permits donors to contribute up to three times the inflation-adjusted base limit into any of three new “separate, segregated account[s] . . . used . . . to defray expenses incurred with respect to” presidential nominating conventions, headquarters buildings, and recounts and other legal proceedings. 52 U.S.C. § 30116(a)(9). Insisting that this case differs meaningfully from Buckley and McConnell, the LNC argues that we must apply strict scrutiny to FECA’s new two-tiered scheme. We disagree. The LNC first contends that because the statute now restricts how certain funds may be “used,” 52 U.S.C. § 30116(a)(9), the cromnibus amendment “transformed” FECA’s contribution limit into an expenditure limit, Appellant’s Br. 41. But McConnell forecloses this argument. That decision teaches that the difference between an expenditure limit and a contribution limit hinges not on the statute’s use of magic words such as “spend” (as in BCRA) or “use” (as in the cromnibus amendment), but rather on a 25 functional test. “The relevant inquiry is whether the mechanism adopted to implement the contribution limit, or to prevent circumvention of that limit, burdens speech in a way that a direct restriction on the contribution itself would not.” McConnell, 540 U.S. at 138–39. That test makes this an easy case. Neither the general- purpose contribution ceiling nor the 300%-higher dedicated- purpose contribution ceiling “in any way limits the total amount of money parties can spend.” Id. at 139. The cromnibus amendment says nothing about how much money political party committees may expend on general purposes, conventions, headquarters, and recounts. Instead, the two- tiered scheme does nothing more than its single-tiered predecessor: it “simply limit[s] the source and individual amount of donations” for each category of expenses. Id. Or, as the Court put it in Buckley, “[t]he overall effect of the Act’s contribution ceilings is merely to require . . . political committees to raise funds from a greater number of persons . . . rather than to reduce the total amount of money potentially available to promote political expression.” 424 U.S. at 21–22. That is a contribution limit through and through. The LNC’s second tack is somewhat more creative, albeit no more successful. Consider, the LNC posits, a contribution from Donor Doe that exceeds the base limit by $1, i.e., a $33,401 donation. Under the cromnibus amendment’s two- tiered contribution limit, the committee may use Doe’s extra dollar to pay for a presidential nominating convention but not a midterm convention, or for a sign on its headquarters but not a billboard on the street. According to the LNC, then, regardless of whether the two-tiered limit imposes a permissible contribution ceiling on donors, with respect to recipients, FECA’s “spending purpose restrictions directly limit how the LNC may express itself” based on the content of 26 its speech. Appellant’s Br. 46; see also Reply Br. 20 (criticizing the Commission’s “obsessive focus on contributors’ interests” as “irrelevant, because the restrictions at issue target the parties’ accounts” and because “[i]t is not the donors who are barred from spending beyond the accounts’ segregated purposes”). For this proposition, the LNC relies on Reed v. Town of Gilbert, in which the Court recently held that laws “defining regulated speech by particular subject matter, . . . function[,] or purpose,” “are subject to strict scrutiny.” 135 S. Ct. 2218, 2227 (2015); see also Appellant’s Br. 47–48 (arguing that “[c]haracterizing FECA’s revised contribution limit as a pure contribution limit does not alter the fact that it ‘target[s] speech based on its communicative content,’ ‘by particular subject matter, and . . . by its function or purpose’” (second and third alterations in original) (citation omitted) (quoting Reed, 135 S. Ct. at 2226–27)). But the LNC misses one crucial element in the “content- based restriction on speech” inquiry: speech. Recall that Buckley drew a clear distinction between spending money (expenditures) and receiving money (contributions). Restrictions on the former regulate speech, as “virtually all meaningful political communications in the modern setting involve the expenditure of money” so that an absolute limit on a political party’s expenditures necessarily restricts its total amount of expression. Buckley, 424 U.S. at 11. Restrictions on the latter, however, are something different. Receiving money facilitates speech, to be sure, but a bank account balance becomes speech only when spent for expressive purposes. This is why the Court has made clear “that contribution limits impose serious burdens on free speech only if they are so low as to ‘preven[t] . . . political committees from amassing the resources necessary for effective advocacy.’” McConnell, 540 U.S. at 135 (quoting Buckley, 424 U.S. at 21). 27 So there lies the solution to the Donor Doe problem. The LNC’s speech occurs when it spends Doe’s money on political expression. That speech remains unencumbered by FECA because, as discussed above, see supra at 24–25, the cromnibus amendment’s two-tiered contribution limit imposes no expenditure limit. True, the LNC may not spend Doe’s additional dollar on a billboard. But it may spend as many dollars from as many non-Does as it wants on billboards, so long as it spends no more than $33,400 from any single donor. The LNC’s speech is thus subject to no restriction, content- based or otherwise. We emphasize that this case implicates only the sort of line-drawing exercises that inhere in a system of federal campaign finance regulation—that is, lines that define in evenhanded terms covered recipients, donors, and contributions. This case, in other words, presents no plausible claim that FECA’s two-tiered contribution limit restricts contributions based on the donor’s identity or viewpoint. And yet, the LNC argues that FECA’s two-tiered contribution limit merits strict scrutiny. Consequently, by the LNC’s logic, FECA would be rife with content-based restrictions on recipients’ speech. For example, the McConnell- approved BCRA prohibits national party committees from “spend[ing] any funds,” 52 U.S.C. § 30125(a)(1), donated in excess of FECA’s limits, which, in turn, apply to contributions made “for the purpose of influencing any election for Federal office,” id. § 30101(8)(A)(i). Likewise, BCRA’s soft-money ban prohibits state party committees from spending non-FECA contributions on “Federal election activity.” Id. § 30125(b). If, as the LNC argues, a limit on contributions made to segregated accounts dedicated to particular “uses” counts as a content- based restriction on speech, then so, too, would restrictions on spending donations “made . . . for the purpose of influencing 28 any election for Federal office” or on expending funds for “Federal election activity.” But that, of course, is not the case: as the Court explained in McConnell, BCRA does not “burden[] speech in a way that a direct restriction on the contribution itself would not.” 540 U.S. at 139. Consequently, the LNC essentially asks us to conclude that Reed’s application of strict scrutiny to laws that “defin[e] regulated speech by particular subject matter, . . . function[,] or purpose,” 135 S. Ct. at 2227, overruled, by implication alone, McConnell’s application of closely drawn scrutiny to FECA’s contribution limits. To put it mildly, we have our doubts. But if the Supreme Court had intended to shake the constitutional foundation of FECA’s contribution-limit architecture, then it is the Supreme Court’s province to say so. See Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U.S. 477, 484 (1989) (“If a precedent of [the Supreme] Court has direct application in a case, yet appears to rest on reasons rejected in some other line of decisions, the Court of Appeals should follow the case which directly controls, leaving to [the Supreme] Court the prerogative of overruling its own decisions.”). Unless and until the Court expressly abrogates McConnell, this “inferior court” lacks authority to “conclude [that the Supreme Court’s] more recent case[]” has, “by implication, overruled an earlier precedent.” Agostini v. Felton, 521 U.S. 203, 237 (1997). B. With no reason to apply strict scrutiny to the cromnibus amendment’s two-tiered contribution limit, we again assume that closely drawn scrutiny supplies the appropriate test. We say “assume” because it remains unclear whether closely drawn scrutiny applies to a recipient’s First Amendment interests alone, see supra at 13, and the LNC declines to invoke the rights of its donors, see supra at 11, 25–26. Nevertheless, 29 because we conclude that the cromnibus amendment’s two- tiered contribution limit survives closely drawn scrutiny, we have no need to determine whether a less stringent standard of review may apply. In applying closely drawn scrutiny, “we must assess the fit between the stated governmental objective and the means selected to achieve that objective.” McCutcheon, 572 U.S. at 199 (plurality opinion). “[I]f a law that restricts political speech does not ‘avoid unnecessary abridgement’ of First Amendment rights, it cannot survive ‘rigorous’” closely drawn review. Id. (internal citation omitted) (quoting Buckley, 424 U.S. at 25, 29). The LNC makes no attempt to challenge the government’s significant anticorruption interest served by limiting the size of contributions to political parties. Indeed, the LNC invokes the district court’s factual finding on this point: “[T]he essential truth,” says the committee, “is that ‘[a]ll contributions to political parties can create the risk of corruption or its appearance regardless of the way that money is ultimately spent . . . .’” Appellant’s Br. 57 (alterations in original) (quoting CF ¶ 36). Rather than contesting the need for contribution limits, the LNC makes a more refined point. “It is one thing to generalize that larger contributions pose a greater risk, and for that reason, impose a simple contribution limit,” argues the committee, but “[r]estricting how a party spends 90% of a contribution, in 30% tranches tied to presidential nominating conventions, buildings, and litigation, cannot be explained on a corruption-fighting rationale.” Id. at 56. In other words, conceding the need for an overall contribution limit, and taking no issue with drawing that line at either $33,400 or $334,000, the LNC questions whether the government can demonstrate an anticorruption interest in treating general- and dedicated-purpose contributions differently. 30 Right out of the gate, the LNC’s argument faces a high hurdle: the cromnibus amendment increased the total amount individuals may contribute to a political party. Before 2014, the LNC could accept only a base-limit sized contribution from any one person; now it may accept ten times that amount. Consequently, the LNC’s argument sounds very much like a grievance with Congress’s decision to raise contribution limits. But so long as contribution limits apply equally to all donors and recipients, “[t]here is . . . no constitutional basis for attacking contribution limits on the ground that they are too high.” Davis v. Federal Election Commission, 554 U.S. 724, 737 (2008). If, as the LNC concedes, the government had a legitimate anticorruption interest in keeping individual contributions below $33,400, then, by simple mathematics, it must also have an interest in keeping contributions below $334,000. We hasten to add a caveat. Although a law does not offend the First Amendment merely because it “conceivably could have restricted even greater amounts of speech in service of [its] stated interests,” a law’s underinclusivity—in this case, the fact that FECA restricts some contributions less than others— nonetheless “can raise ‘doubts about whether the government is in fact pursuing the interest it invokes.’” Williams-Yulee v. Florida Bar, 135 S. Ct. 1656, 1668 (2015) (quoting Brown v. Entertainment Merchants Ass’n, 564 U.S. 786, 802 (2011)). But we see no reason for such skepticism in this case, as allowing donors to make larger contributions into each of the new dedicated-purpose accounts serves Congress’s legitimate interest in relaxing restrictions on First Amendment activity where, as it has concluded here, it can achieve its anticorruption interest with less stringent limits. Take the new, higher limit on contributions to pay for presidential nominating conventions. In April 2014, Congress 31 ended public funding for such conventions, leaving parties on their own. See Gabriella Miller Kids First Research Act, Pub. L. No. 113-94, 128 Stat. 1085 (2014). The cromnibus amendment gives parties a tool for making up for that shortfall, ensuring, as Congress must, that parties remain capable of “amassing the resources necessary for effective advocacy.” Buckley, 424 U.S. at 21. Equally benign are the other two new dedicated-purpose accounts, one for party headquarters and the other for election recounts and “other legal proceedings.” 52 U.S.C. § 30116(a)(9). As the Court explained in McConnell, the donations “that pose the greatest risk of . . . corruption” are “those contributions . . . that can be used to benefit federal candidates directly.” 540 U.S. at 167. Congress could have permissibly concluded that unlike contributions that can be used for, say, television ads, billboards, or yard signs, contributions that fund mortgage payments, utility bills, and lawyers’ fees have a comparatively minimal impact on a party’s ability to persuade voters and win elections. Indeed, congressional leaders supporting the cromnibus amendment emphasized that “many” of the “expenditures made from the [dedicated-purpose] accounts” are “not for the purpose of influencing federal elections.” 160 Cong. Rec. S6814 (daily ed. Dec. 13, 2014) (statement of Sen. Reid); id. at H9286 (daily ed. Dec. 11, 2014) (statement of Rep. Boehner). That makes good sense: headquarters, once built, exist regardless of whether an election is afoot, and recounts, by definition, can occur only after votes have been cast. In fact, before BCRA, the Commission entirely excluded donations for both party headquarters and election recounts from the definition of “contribution.” See 11 C.F.R. § 100.7(b)(12) (2002) (“A gift . . . made to a national committee . . . of a political party is not a contribution if it is specifically designated to defray any cost incurred for construction or purchase of any office facility 32 which is not acquired for the purpose of influencing the election of any candidate in any particular election for Federal office.”); id. § 100.7(b)(20) (“A gift . . . made with respect to a recount of the results of a Federal election, or an election contest concerning a Federal election, is not a contribution . . . .”). We are untroubled in this case by the fact that, as the LNC observes, the cromnibus amendment passed Congress without the sort of robust record of congressional factfinding that accompanied BCRA. In one sense this might be expected; after all, BCRA imposed new contribution limits, so its additional restriction on First Amendment rights required justification. The cromnibus amendment, by contrast, did just the opposite: it relaxed contribution limits. Had BCRA’s extensive legislative history identified some troubling finding related specifically to conventions, headquarters, or legal expenses, we would perhaps harbor more concern about the cromnibus amendment’s relatively stingy congressional record. But we have discovered in that record no basis for any such concern, leaving us without any reason to conclude that the Congress of 2014 committed constitutional error by determining that, a dozen years after BCRA, times and circumstances had sufficiently changed to permit it to deal more generously with expense categories less directly tied to particular candidates or elections. See Wagner v. Federal Election Commission, 793 F.3d 1, 30 (D.C. Cir. 2015) (en banc) (noting that contribution restrictions need not address “speculative” concerns). Our dissenting colleague worries that Congress may have enacted the cromnibus amendment not to better tailor contribution limits to serve the government’s anticorruption interest, but rather to benefit the major parties that do the most spending on segregated-account activities. See Op. at 7–9 (Griffith, J.). But the LNC itself, though displeased that 33 FECA’s two-tiered contribution limit more closely “align[s] with the financial needs and goals of the incumbent parties,” Appellant’s Br. 58 (internal quotation marks omitted), expressly disclaims any argument that “the First Amendment requires a level electoral playing field, free of the advantages that speakers may have owing to their resources,” Opposition at 26; see also id. at 27 (stating that the LNC’s “merits briefing [is] bereft of even a molecule of competitive disadvantage theory” and arguing that “it is absurd for the [Commission] to insist” otherwise). And indeed, the First Amendment requires no such thing. While Congress may not enact contribution limits that “magnify the advantages of incumbency to the point where they put challengers to a significant disadvantage,” Randall, 548 U.S. at 248 (plurality opinion), neither is it “an acceptable governmental objective,” “[n]o matter how desirable it may seem,” “to ‘equaliz[e] the financial resources of candidates,’” McCutcheon, 572 U.S. at 207 (plurality opinion) (second alternation in original) (quoting Arizona Free Enterprise Club’s Freedom Club PAC v. Bennett, 564 U.S. 721, 748, 750 (2011)). Therefore, “if Congress concludes that allowing contributions of a certain amount does not create an undue risk of corruption or the appearance of corruption,” the Court has explained, then “a candidate who wishes to restrict an opponent’s fundraising cannot argue that the Constitution demands that contributions be regulated more strictly.” Davis, 554 U.S. at 737. By the same token, the mere fact that additional fundraising opportunities will benefit some political parties over others does not itself render Congress’s relaxation of contribution limits suspect under the First Amendment. See Federal Election Commission v. Massachusetts Citizens for Life, Inc., 479 U.S. 238, 257 (1986) (“Political ‘free trade’ does not necessarily require that all who participate in the political marketplace do so with exactly equal resources.”). We thus see no reason to “‘doubt[] . . . [that] the government is in fact pursuing the interest it invokes,’” Williams-Yulee, 135 S. Ct. at 34 1668 (quoting Brown, 564 U.S. at 802), to justify FECA’s two- tiered contribution limit: combatting quid pro quo corruption and its appearance. At bottom, the cromnibus amendment represents just another tweak in Congress’s decades-long project to fine-tune FECA’s balance between speech and associational rights, on the one hand, and the government’s anticorruption interest, on the other. That balance, to be sure, remains imperfect. But closely drawn scrutiny “require[s] ‘a fit that is not necessarily perfect, but reasonable; that represents not necessarily the single best disposition but one whose scope is in proportion to the interest served . . . .’” McCutcheon, 572 U.S. at 218 (plurality opinion) (internal quotation marks omitted) (quoting Board of Trustees of the State University of New York v. Fox, 492 U.S. 469, 480 (1989)). And lacking any “‘scalpel to probe’ each possible contribution level,” we “defer[] to the legislature’s” “empirical judgments” about “the precise restriction necessary to carry out the statute’s legitimate objectives.” Randall, 548 U.S. at 248 (plurality opinion) (quoting Buckley, 424 U.S. at 30). Here, Congress drew that line at $33,400 for general- purpose spending and $100,200 for dedicated-purpose spending. The LNC has given us no reason to think that this two-tiered limit would offend the First Amendment. The cromnibus amendment’s limits are closely drawn to the government’s anticorruption interest, and, as compared to the pre-2014 baseline, they certainly avoid unnecessary infringement of associational and speech rights. We therefore answer the second and third certified questions in the negative: FECA’s two-tiered contribution limit, both on its face and as applied to Shaber’s bequest, does not violate the LNC’s First Amendment rights. 35 V. The task of crafting campaign finance restrictions is, in many ways, a zero-sum game. Make the regime too restrictive, and you threaten “fundamental First Amendment interests” by burdening citizens’ political expression. Buckley, 424 U.S. at 23. Make the regime too permissive, and you threaten “the integrity of our system of representative democracy” by failing to prevent quid pro quo corruption and its appearance. Id. at 26–27. Balancing these interests has turned out to be a difficult and iterative task. For the reasons given above, we conclude that the current version of FECA—both its application of contribution limits to Shaber’s bequest and its use of a two- tiered contribution limit—has achieved a constitutionally permissible balance. Therefore, although we deny the Commission’s motion to dismiss for lack of standing, we reject each of the LNC’s three constitutional challenges on the merits. So ordered. GRIFFITH, Circuit Judge, concurring in part and dissenting in part: When the government restricts First Amendment freedoms, it “bears the burden of proving the constitutionality of its actions.” McCutcheon v. FEC, 572 U.S. 185, 210 (2014) (plurality opinion) (quoting United States v. Playboy Entm’t Grp., 529 U.S. 803, 816 (2000)). Here, the government has not justified the cromnibus amendments’ two-tiered scheme for contributions to national political parties. I therefore part ways with the majority on the second and third certified questions. The appropriate standard of review is closely drawn scrutiny, as the majority assumes and Judge Katsas explains. See Maj. Op. at 28; Op. at 1-5 (Katsas, J.). Under this standard, the government must “demonstrate[] a sufficiently important interest and employ[] means closely drawn to avoid unnecessary abridgment” of First Amendment freedoms. Buckley v. Valeo, 424 U.S. 1, 25 (1976) (per curiam). The only qualifying interest is combating quid pro quo corruption and its appearance, and we require the government to employ “a means narrowly tailored to achieve the desired objective.” McCutcheon, 572 U.S. at 192, 218 (quoting Bd. of Trs. of State Univ. of N.Y. v. Fox, 492 U.S. 469, 480 (1989)). This standard is “rigorous,” and the government will not prevail if there is “a substantial mismatch between [its] stated objective and the means selected to achieve it.” Id. at 197, 199 (first quoting Buckley, 424 U.S. at 29). The Libertarian National Committee (LNC) would take no issue with a single contribution limit set at $33,400 or $334,000. Maj. Op. at 29. Indeed, a challenge to such a limit would be foreclosed by McConnell v. FEC, 540 U.S. 93 (2003). There, the Supreme Court held that the Federal Election Campaign Act permissibly prohibited a donor from contributing more than $25,000 to a national political party because the government showed that the prohibition substantially advanced, and was properly tailored to, the 2 government’s interests in preventing corruption or its appearance. See McConnell, 540 U.S. at 142-61. But McConnell does not resolve this case, because the two-tiered scheme here differs in important ways from the limit upheld in McConnell. Rather than limiting all contributions above a certain level, the scheme prohibits contributions above the general limit of $33,400 but makes exceptions to that general limit by allowing additional contributions of up to $100,200 to each of three segregated accounts for presidential nominating conventions, party headquarters, and election recounts and litigation. See 52 U.S.C. § 30116(a)(1)(B), (a)(9); Maj. Op. at 3-4.1 This is a new scheme. McConnell did not address the propriety of a regime with these exceptions, the presence of which “can raise doubts about whether the government is in fact pursuing the interest it invokes” or “reveal that a law does not actually advance” that interest. Williams-Yulee v. Fla. Bar, 135 S. Ct. 1656, 1668 (2015) (internal quotation marks omitted). Put differently, the cromnibus amendments introduced a critical feature not present in McConnell: “Congress’ judgment” that contributions of $300,600 to segregated accounts “do not unduly imperil anticorruption interests.” Davis v. FEC, 554 U.S. 724, 741 (2008). Given this judgment by Congress, it is now “hard to imagine how” limiting general contributions to $33,400 “serv[es] anticorruption goals sufficiently to justify the resulting constitutional burden”—unless general and segregated contributions differ in a constitutionally meaningful way. Id. For these reasons, the government cannot justify treating general contributions more restrictively than segregated contributions based on McConnell’s approval of a since-abandoned congressional judgment. Rather, the 1 Like the majority, I use the limits adjusted for inflation as of 2015. Maj. Op. at 3-4. 3 government must show that a new scheme that differentiates between general and segregated contributions is closely drawn to serve anticorruption interests. To do so, the government argues that general and segregated contributions raise different corruption concerns. This is because general-account spending is more likely to be for the purpose of influencing elections and thus raise corruption concerns, while segregated-account spending is less likely to be for the purpose of influencing elections and thus does not raise comparable corruption concerns. See FEC Br. 46-50. The record does not support this distinction. The government relies on identical statements from Senator Reid and Representative Boehner, who both asserted that “many” of the expenditures from segregated accounts are “not for the purpose of influencing Federal elections.” 160 Cong. Rec. S6814 (daily ed. Dec. 13, 2014); id. at H9286 (daily ed. Dec. 11, 2014). But these self-serving assertions by representatives of the major parties do not tell us whether segregated-account spending is any different from general- account spending with respect to influencing elections or raising corruption concerns. Without that information, we simply do not know whether the cromnibus amendments are justified in prohibiting all contributions above the general limit except those made to segregated accounts. And an ambivalent record is not enough to survive closely drawn scrutiny. See McCutcheon, 572 U.S. at 217 (rejecting aggregate contribution limits in part because the government did not provide “any real-world examples” that they served anticorruption interests by preventing donors from circumventing the base limits); McConnell, 540 U.S. at 145-154 (upholding limits on soft- money contributions only after identifying extensive evidence connecting the limits to the government’s legitimate interests); cf. Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622, 666-67 4 (1994) (in applying intermediate scrutiny to a speech restriction, explaining that “we cannot determine” whether Congress drew “reasonable inferences based on substantial evidence” without “a more substantial elaboration in the District Court of the predictive or historical evidence upon which Congress relied, or the introduction of some additional evidence”); Annex Books, Inc. v. City of Indianapolis, 581 F.3d 460, 463 (7th Cir. 2009) (Easterbrook, J.) (“[T]here must be evidence” to carry a First Amendment burden.). The government’s position does not fare any better when we examine the segregated accounts more closely. As the majority points out, the higher limits on contributions to pay for presidential nominating conventions were prompted by the end of public funding for such conventions in 2014. The cromnibus amendments gave parties a “tool for making up for that shortfall.” Maj. Op. at 31. That explanation is understandable, but it does not establish that there are lesser corruption concerns with contributions that help put on nominating conventions. There can be no serious doubt that the nominating conventions of the major parties are closely connected to elections. Contributions to their staging therefore appear to raise the same corruption risks as general contributions, and the record provides no reason to think otherwise. The record is similarly slim as to the segregated accounts for maintaining party headquarters and contesting election results. The majority offers that “Congress could have permissibly concluded that unlike contributions that can be used for, say, television ads, billboards, or yard signs, contributions that fund mortgage payments, utility bills, and lawyers’ fees have a comparatively minimal impact on a party’s ability to persuade voters and win elections.” Id. Perhaps, but that inference lacks record support. The record 5 gives no reason to think that spending on party headquarters or election contests has a different influence on elections than general-account spending, and the majority might just as reasonably have said the opposite: that Congress “could have” determined that elections are significantly influenced by a party headquarters (where parties might host donors and connect them to party leaders and candidates) and election recounts and litigation (which resolve whether an actual candidate wins or loses a particular election). My point is not that either of these potential determinations is more reasonable than the other; my point is that without record support they are “too speculative” to carry a First Amendment burden. McCutcheon, 572 U.S. at 210. Finally, the factual findings made by the district court provide no better support for the government. The district court found that “unrestricted funds are more valuable to national party committees and their candidates than funds that may only be used for particular categories of expenses.” Findings of Fact (“CF”) ¶ 50, Libertarian Nat’l Comm. v. FEC, 317 F. Supp. 3d 202 (D.D.C. 2018). And according to the government, “it is simple common sense that the more a political party values a contribution, the more likely that contribution will be or appear to be part of a quid pro quo corruption scheme,” making it more reasonable for the cromnibus amendments to treat general contributions more restrictively than segregated contributions. FEC Br. 47. The problem for the government, however, is that the district court’s findings simultaneously point in the opposite direction: “A political party may in some circumstances value a contribution with use restrictions more highly than a smaller contribution without such restrictions,” particularly because money is generally fungible and every dollar received through segregated accounts “potentially frees up another dollar in the recipient’s general account for unrestricted spending.” CF ¶¶ 38-39. The record does not 6 clarify whether such a “circumstance” is presented by this case; again, we just don’t know. Moreover, even if a dollar donated to a general account raised more corruption concerns than a dollar given to a segregated account, the government acknowledges that “larger contributions to political parties are generally more likely to lead to actual or apparent quid pro quo arrangements and can do so regardless of how the funds are ultimately used.” CF ¶ 35 (alterations omitted). This further highlights the poor fit of the cromnibus amendments, which treat larger contributions to segregated accounts as if they were less likely to raise corruption concerns than substantially smaller contributions to a general account. In the absence of any corruption-related difference between general and segregated contributions, the government has not carried its burden of showing that the two-tiered scheme is closely drawn to serve anticorruption interests. This conclusion does not rely on a “freestanding underinclusiveness limitation,” as Judge Katsas fears. Op. at 20 (Katsas, J.) (quoting Williams-Yulee, 135 S. Ct. at 1668). Although “the First Amendment imposes no freestanding ‘underinclusiveness limitation,’” underinclusivity still “creates a First Amendment concern when the State regulates one aspect of a problem while declining to regulate a different aspect of the problem that affects its stated interest in a comparable way.” Williams- Yulee, 135 S. Ct. at 1668, 1670 (first quoting R.A.V. v. City of St. Paul, 505 U.S. 377, 387 (1992)). That’s the problem with the two-tiered scheme in this case. On this record, segregated and general contributions affect the government’s anticorruption interests in the same way, yet the scheme restricts general contributions while declining to restrict segregated contributions. Thus, the scheme’s underinclusiveness—its exceptions allowing some contributions above the general limit—shows that the government has not justified prohibiting other contributions 7 from exceeding the general limit. See id. at 1670; see also Reed v. Town of Gilbert, 135 S. Ct. 2218, 2231 (2015) (rejecting a speech restriction as “hopelessly underinclusive” under strict scrutiny because it drew distinctions between prohibited and permissible categories of speech in a way that was not justified by the interests asserted by the government); id. at 2239 (Kagan, J., concurring in the judgment) (rejecting the same restriction under intermediate scrutiny due to its underinclusivity); Edwards v. District of Columbia, 755 F.3d 996, 1007-08 (D.C. Cir. 2014) (rejecting a speech restriction as “fatally underinclusive” under intermediate scrutiny). That is enough to resolve the second and third certified questions in the LNC’s favor, but in closing I note that there are additional reasons to be skeptical of the government’s position. The two-tiered scheme’s exceptions loosen restrictions on the very contributions that are highly sought by major parties but of little use to minor parties. In my view, this further undercuts the government’s position that the scheme pursues the only permissible government interest: combating quid pro quo corruption and its appearance. Under the scheme, a donor may contribute a total of $334,000 to a political party: $33,400 to the general account and $100,200 to each of the three segregated accounts. The major parties benefit from this scheme because they spend substantial sums on activities that can be paid for through segregated accounts: They put on lavish nominating conventions that are spectacles made for a national audience, they maintain expensive headquarters, and they challenge and defend in court the outcomes of numerous elections across the country. Indeed, from December 2014 through December 2016, the Republican Party received more than $23 million for its convention, $26 million for its headquarters, and $5 million for election recounts and litigation; the Democratic Party 8 received more than $12 million for its convention, $3 million for its headquarters, and $6 million for election recounts and litigation. CF ¶¶ 45-46; J.A. 90. The cromnibus amendments enable the major parties to raise such sums with individual contributions of up to $334,000. What’s more, those contributions are in effect no different from general contributions. So long as a party has segregated-account expenses, a dollar received in a segregated account frees up a dollar in the general account that otherwise might have been used to defray the segregated-account expenses. Therefore, until a party receives enough money to cover its segregated- account expenses, the two-tiered scheme establishes an effective general contribution limit of $334,000. By contrast, minor parties gain little from this scheme because they do not have much use for segregated-account contributions. The LNC, for example, holds more modest conventions and maintains a less expensive headquarters than the major parties, and the LNC has never spent money on election recounts and is unlikely to do so in the future. See LNC Br. 13-15. In most years, its expenses for these purposes are less than $500,000. See id.; CF ¶¶ 25-29. Lacking further segregated-account expenses, the LNC and similar minor parties do not benefit much from the higher limit for segregated-account contributions. Instead, they seek contributions that can be used for other purposes, and those contributions are limited to $33,400. In this way, the scheme’s exceptions loosen restrictions on those contributions that are useful to major parties but not to minor parties. Of course, this effect is in part attributable to the various levels of support for different parties and the parties’ decisions on how to raise and spend contributions. And as the majority notes, this effect alone does not render the scheme unconstitutional. See Maj. Op. at 33. Even so, it raises further 9 doubts that the scheme is tailored to serve anticorruption interests rather than an impermissible interest, such as disadvantaging minor parties. See Williams-Yulee, 135 S. Ct. at 1668. This concern overlaps with those that motivate comparative-disadvantage cases, see, e.g., Randall v. Sorrell 548 U.S. 230, 248 (2006) (a statute regulating contributions must not “magnify the advantages of incumbency to the point where [it] put[s] challengers to a significant disadvantage”), but it is not an attempt to raise a comparative-disadvantage claim on the LNC’s behalf, Maj. Op. at 32-33. It simply provides further record-based reasons to be skeptical that the two-tiered scheme is tailored to serve anticorruption interests. Because the government has not carried its burden of showing that the scheme is closely drawn to combat corruption or its appearance, I would hold that the scheme violates the First Amendment. Having reached a different decision on the merits, the majority has no occasion to address the appropriate remedy. I therefore do not reach the issue either.2 But on the merits of the second and third certified questions, I respectfully dissent. 2 The appropriate remedy, i.e., the “upshot” of holding that the scheme violates the First Amendment, Op. at 23 (Katsas, J.), is disputed by the parties. The LNC argues that the appropriate remedy is excising the use restrictions while leaving the increased overall limit, allowing a donor to contribute $334,000 for general use. LNC Br. 62-63; accord Amicus Br. of the Goldwater Inst. 8. The government urges the pre-cromnibus status quo, which would allow a donor to contribute $33,400 for general use and nothing more. FEC Br. 54-56. Alternatively, the court could remand this matter for further record development. See Order, Holmes v. FEC, No. 14-5281 (D.C. Cir. Jan. 30, 2015) (en banc) (per curiam); Buckley v. Valeo, 519 F.2d 817, 818 (D.C. Cir. 1975) (en banc) (per curiam); see also Turner, 512 U.S. at 668. KATSAS, Circuit Judge, with whom Circuit Judge HENDERSON joins, concurring in part, concurring in the judgment in part, and dissenting in part: This case involves statutory limits on contributions that individuals may make to political parties. In McConnell v. FEC, 540 U.S. 93 (2003), the Supreme Court held that these contribution limits are not facially unconstitutional. Here, we consider whether the limits are unconstitutional as applied to contributions made through bequests. We also consider whether the limits became unconstitutional when Congress amended them in 2014. I To frame the relevant inquiries, we must first decide the appropriate level of First Amendment scrutiny. The majority reserves this question, ante at 13, 28, but I would decide it. In 1976, the Supreme Court fixed the level of scrutiny for limits on contributions to candidates for federal elective offices. Those limits “may be sustained if the State demonstrates a sufficiently important interest and employs means closely drawn to avoid unnecessary abridgment” of speech and associational freedoms. Buckley v. Valeo, 424 U.S. 1, 25 (1976) (per curiam). Subsequently, the Court has applied this same level of scrutiny to assess the constitutionality of contribution limits imposed on all kinds of donors and recipients, including candidates for federal and state offices; national, state, and local political parties; and political action committees. See, e.g., McCutcheon v. FEC, 572 U.S. 185, 196–99 (2014) (plurality opinion); Ariz. Free Enter. Club’s Freedom Club PAC v. Bennett, 564 U.S. 721, 734–35 (2011); Davis v. FEC, 554 U.S. 724, 736–37 (2008); Randall v. Sorrell, 548 U.S. 230, 246–48 (2006) (plurality opinion); McConnell, 540 U.S. at 134–41; FEC v. Beaumont, 539 U.S. 146, 161–62 (2003); FEC v. Colo. Republican Fed. Campaign Comm., 533 U.S. 431, 446–56 (2001); Nixon v. Shrink Mo. Gov’t PAC, 528 2 U.S. 377, 387–88 (2000). For shorthand, this level of scrutiny is now referred to (rather clumsily) as “closely drawn scrutiny.” Despite this long line of precedent, the Federal Election Commission urges us to lower the bar, at least with respect to bequests. The FEC asks us to consider only whether the challenged contribution limits prevent the Libertarian National Committee, which received the bequest at issue here, from “amassing the resources necessary for effective advocacy.” The FEC plucks that phrase out of Buckley, which observed that contribution limits “could have a severe impact” if they prevented recipients from amassing such resources. 424 U.S. at 21. The FEC reasons that bequests implicate neither the donor’s speech interests nor anyone’s associational interests, and the recipient’s speech interests are impaired only if it is prevented from mounting, in the aggregate, some quantum of “effective” advocacy. This analysis is flawed at every turn. To begin, “effective advocacy” is not a reduced, free-floating level of First Amendment scrutiny. If a contribution limit prevents effective advocacy, then it is insufficiently tailored to satisfy closely drawn scrutiny. See Randall, 548 U.S. at 246–62 (plurality opinion); id. at 267–73 (Thomas, J., concurring in the judgment). But contribution limits may be insufficiently tailored for other reasons, such as “a substantial mismatch between the Government’s stated objective and the means selected to achieve it.” McCutcheon, 572 U.S. at 199 (plurality opinion). And regardless of any tailoring problems, contribution limits are unconstitutional if the asserted government interest is insufficiently important. See, e.g., Davis, 554 U.S. at 740 n.7; SpeechNow.org v. FEC, 599 F.3d 686, 695 (D.C. Cir. 2010) (en banc). 3 Likewise, the Supreme Court has never attempted “to parse distinctions between the speech and association standards of scrutiny for contribution limits.” Shrink Mo. Gov’t, 528 U.S. at 388. Rather, it has fashioned what the majority aptly describes as a “single unified test that applies an intermediate level of scrutiny to contribution limits.” Ante at 13. Thus, in reaffirming the appropriateness of closely drawn scrutiny in McConnell, the Court held it immaterial that the challenged provisions restricted the acceptance of contributions by parties rather than the giving of contributions by donors. See 540 U.S. at 138. Applying closely drawn scrutiny in SpeechNow, this Court held that the challenged contribution limits violated the First Amendment rights of both the donors and the recipient, without hinting at any distinction between the two. See 599 F.3d at 690–96. And the three-judge district court in Republican National Committee v. FEC, 698 F. Supp. 2d 150 (D.D.C.) (Kavanaugh, J.), aff’d, 561 U.S. 1040 (2010) (mem.), applied closely drawn scrutiny to assess contribution limits challenged only by recipients. See id. at 153, 156. Of course, different contribution limits may impact speech and associational interests in different ways, but “we account for [those impacts] in the application, rather than the choice, of the appropriate level of scrutiny.” McConnell, 540 U.S. at 141. The FEC’s plea for less-than-intermediate scrutiny is also radical. For over four decades, various justices have urged that because contribution limits “operate in an area of the most fundamental First Amendment activities,” Buckley, 424 U.S. at 14, they should be subjected to strict rather than closely drawn scrutiny. See, e.g., Shrink Mo. Gov’t, 528 U.S. at 405–10 (Kennedy, J., dissenting); id. at 410–30 (Thomas, J., joined by Scalia, J., dissenting); Colo. Republican Fed. Campaign Comm. v. FEC, 518 U.S. 604, 635–44 (1996) (Colorado I) (Thomas, J., dissenting in part); Buckley, 424 U.S. at 241–46 (Burger, C.J., dissenting in part); id. at 290 (Blackmun, J., 4 dissenting in part). McConnell acknowledged this “significant criticism.” 540 U.S. at 137. And in McCutcheon—the Court’s most recent decision in this area—the plurality sought to minimize the differences between strict and closely drawn scrutiny, see 572 U.S. at 196–99, in the face of a continuing call for strict scrutiny, see id. at 228–32 (Thomas, J., concurring in the judgment). Given this longstanding debate over whether closely drawn scrutiny sets the bar too low, it is quite a stretch to posit that, here, it sets the bar too high. The FEC’s proposal would create anomalies in First Amendment law more generally. Effective speech often requires multiple parties—speakers, listeners, and, in the context of mass markets, patrons. The Supreme Court generally treats the rights of these parties as “reciprocal.” Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, Inc., 425 U.S. 748, 756–57 (1976) (“the protection afforded is to the communication, to its source and to its recipients both”). So, the right of one party to speak implies the right of another party to listen. See id. Likewise, the right of one party to fund speech implies the right of another party to accept the funds. Cf. McConnell, 540 U.S. at 138 (“it is irrelevant that Congress chose … to regulate contributions on the demand rather than the supply side”). It would be odd enough to isolate one from the other in deciding the merits, much less to do so in fixing an appropriate level of scrutiny. Finally, in fixing the level of scrutiny, death should make no difference. Of course, living donors have substantial speech and associational interests in contributing money to political parties of their choice. See, e.g., McCutcheon, 572 U.S. at 191– 92 (plurality opinion); id. at 228 (Thomas, J., concurring in the judgment). Yet a contribution is no less speech and expressive association if the donor makes it through a bequest rather than a lifetime transfer. Either way, the donor intends to support the 5 political views of the party, and an observer would reasonably understand as much. See Buckley, 424 U.S. at 16–17; cf. Spence v. Washington, 418 U.S. 405, 410–11 (1974) (per curiam). Likewise, the speech and associational interests of recipients—in using all available resources to fund political speech—do not vary depending on whether contributions come from living or deceased donors. In sum, the FEC’s attempt to ratchet down the level of scrutiny by separating speech from expressive association, donors from recipients, and the living from the dead is unsupported by precedent and unsound in principle. I would hold what the majority only assumes—that closely drawn scrutiny governs this case. II Under closely drawn scrutiny, limits on political contributions are constitutional “if the State demonstrates a sufficiently important interest and employs means closely drawn to avoid unnecessary abridgment” of speech and associational freedoms. Buckley, 424 U.S. at 25. In this sensitive area, the only sufficiently important government interests are the prevention of quid pro quo corruption—“a direct exchange of an official act for money”—and its appearance. McCutcheon, 572 U.S. at 192 (plurality opinion). Interests in equalizing “electoral opportunities,” and in preventing donors from acquiring “influence over or access to elected officials or political parties,” are insufficient. Id. at 207–08 (quotation marks omitted). Moreover, “the Government bears the burden of proving the constitutionality of its actions,” id. at 210 (quotation marks omitted), consistent with how intermediate scrutiny works in other First Amendment contexts. See, e.g., United States v. Nat’l Treasury Emps. Union, 513 U.S. 454, 475 (1995) (“the 6 Government … must demonstrate that the recited harms are real, not merely conjectural, and that the regulation will in fact alleviate these harms in a direct and material way” (quotation marks omitted)) (speech restrictions on government employees); Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622, 664–68 (1994) (plurality opinion) (same for content-neutral speech restrictions); Edenfield v. Fane, 507 U.S. 761, 770–71 (1993) (same for commercial speech restrictions). In Buckley, the Supreme Court applied these principles to reject a facial challenge to limits on contributions made to candidates for federal elective offices. The Court noted “deeply disturbing examples” of “quid pro quo” corruption, which proved that the government’s asserted interest was “not an illusory one.” 424 U.S. at 26–27. The Court cited “a number of the abuses” discussed in our Buckley opinion, id. at 27 n.28, which explained that the record before Congress was “replete with specific examples of improper attempts to obtain governmental favor in return for large campaign contributions,” 519 F.2d 821, 839 n.37 (D.C. Cir. 1975) (en banc). The Supreme Court further reasoned that, even if most contributors do not improperly seek quid pro quo exchanges, “suspect contributions” are “difficult to isolate.” 424 U.S. at 30. So, to prevent actual and apparent corruption, the government may eliminate the “opportunity for abuse” from large contributions. Id. In McConnell, the Court rejected a facial challenge to limits on contributions to political parties. Given what it described as the “unity of interest” between parties and elected officials, the Court found “neither novel nor implausible” the supposition that large contributions to a party could corrupt its elected officials. 540 U.S. at 144–45. The Court also discussed at length the supporting evidence: the major political parties annually had been raising hundreds of millions of dollars in 7 previously unregulated soft money, id. at 124; these contributions often were solicited by, and used to help, individual candidates, id. at 146; wealthy donors made large contributions to both major parties, id. at 148; and these contributions impacted a wide range of legislation, id. at 150. III A This case presents a challenge to limits on contributions to political parties made through bequests. In a prior case, the LNC unsuccessfully sought to enjoin application of the contribution limits to all bequests. Libertarian Nat’l Comm., Inc. v. FEC, 930 F. Supp. 2d 154 (D.D.C. 2013) (LNC I). Here, the LNC seeks to enjoin application of the limits only to a bequest made by Joseph Shaber. The facts surrounding this bequest are undisputed. Shaber neither coordinated with the LNC regarding his decision to include the party in his will nor even informed the party of that decision. Libertarian Nat’l Comm., Inc. v. FEC, 317 F. Supp. 3d 202, 249 (D.D.C. 2018) (LNC II). “Aside from pursuing its ideological and political mission, the LNC has provided nothing of value to Mr. Shaber, or to anyone else, in exchange for his bequest.” Id. at 251. The bequest imposed no conditions and made no requests, but instead provided for the LNC to take “outright” a contribution ultimately valued at about $235,000. Id. at 250 (quotation marks omitted). Over the course of his lifetime, Shaber donated a total of $3,315 to the LNC, made in 46 separate gifts spread out over 24 years. Id. at 248–49. Besides making these contributions, Shaber had no other relationship with the LNC. Id. at 251. In its prior cases on contribution limits, the Supreme Court considered no issues specific to bequests. Because the LNC 8 does not rest its claim on “the same factual and legal arguments the Supreme Court expressly considered” in Buckley and McConnell, those precedents do not foreclose the LNC’s as- applied challenge here. Republican Nat’l Comm., 698 F. Supp. 2d at 157 (“McConnell permits as-applied challenges”); see also Doe v. Reed, 561 U.S. 186, 201 (2010) (“upholding the law against a broad-based challenge does not foreclose a litigant’s success in a narrower one”). Indeed, the Supreme Court has sustained an as-applied challenge to corporate- expenditure limits previously held facially constitutional, FEC v. Wis. Right to Life, Inc., 551 U.S. 449, 476–82 (2007) (WRTL) (plurality opinion), and this Court has sustained an as-applied challenge to contribution limits previously held facially constitutional, SpeechNow, 599 F.3d at 692–96. Moreover, because the LNC’s challenge raises issues not addressed in Buckley or McConnell, the government retains its burden of proof under heightened scrutiny. See WRTL, 551 U.S. at 464– 65 (plurality opinion). Of course, we must determine which facts, if any, distinguish this case from Buckley and McConnell, and the breadth of our reasoning will impact the law going forward. See Citizens United v. FEC, 558 U.S. 310, 331 (2010) (“no general categorical line bars a court from making broader pronouncements of invalidity in properly ‘as applied’ cases” (quotation marks omitted)). But regardless of the breadth of our reasoning, the LNC’s first claim seeks relief only as to Shaber’s individual bequest. Under these rules for assessing as-applied challenges, I would hold that the challenged contribution limits are unconstitutional as applied to any of three nested categories: bequests, uncoordinated bequests, and Shaber’s bequest. I will address the categories from broadest to narrowest. 9 1 “The quantum of empirical evidence needed to satisfy heightened judicial scrutiny of legislative judgments will vary up or down with the novelty and plausibility of the justification raised.” McConnell, 540 U.S. at 144 (quotation marks omitted). Here, that means requiring more evidence rather than less, for there are strong reasons to think that bequests—in contrast to contributions from living donors—do not pose a significant risk of actual or apparent quid pro quo corruption. For one thing, politics operates on notoriously “short timeframes,” Citizens United, 558 U.S. at 334, so gifts deferred until death—perhaps many election cycles down the road— will have relatively little value to political parties or their candidates. For another, there is no easy means for deceased donors or their beneficiaries to enforce any corrupt bargains. In the context of contributions from living donors, such bargains are managed through winks and nods over time, as money flows one way and political favors flow the other. See McConnell, 540 U.S. at 147 (quoting lobbyist’s testimony that “overt words are rarely exchanged about contributions, but people do have understandings”). Bequests cannot work like that, because the money flows only once, and at death. So, if a corrupt donor seeks political favors during his lifetime, when the bequest is nothing more than a revocable promise, the recipient will have no way to prevent the donor from accepting the favors but then reneging on the promise. Or, if the donor seeks favors for survivors, he will have no way to ensure delivery after death makes the bequest irrevocable and removes him from the picture. Either way, inherent constraints limit the feasibility of any contemplated exchange. Bequests are thus generally “less susceptible … to misuse,” Beaumont, 539 U.S. at 160, than contributions from living donors. 10 The evidence confirms this point. To justify its concerns about possible corruption through bequests, the FEC could have pointed to anything in any of four records: the legislative record of a select committee established by Congress to investigate fundraising for the 1972 presidential election, see Buckley, 519 F.2d at 839 n.35; the 100,000-page record compiled for the three-judge district court in McConnell, see 251 F. Supp. 2d 176, 209 (D.D.C. 2003); the district-court record in LNC I, where all bequests were at issue; or the district-court record in this case. Yet, despite the massive records in Buckley and McConnell, and the two records made in the bequest-specific LNC cases, the FEC points to nothing substantiating its concerns. In fact, these records undercut its position in three critical respects. First, bequests are rarely used for political contributions. From 1978 through August 2017, bequests accounted for only about $3.7 million in contributions to federal candidates, political parties, and all other entities required to file reports with the FEC. LNC II, 317 F. Supp. 3d at 247. To put that number in perspective, the same group of recipients spent $7 billion in the 2012 election cycle alone, McCutcheon, 572 U.S. at 219 (plurality opinion), and the major political parties spent nearly $1.2 billion in 2000 alone, see McConnell, 540 U.S. at 124. Of course, bequests to political parties might increase if the relevant contribution limits were invalidated. But, from 1978 to 2002, donors could have made unlimited soft-money bequests to political parties. See id. at 122–24. And if McConnell correctly understood the “unity of interest” between political parties and elected officials, such bequests would have been almost as enticing as ones made directly to the officials. See id. at 144–45. In sum, despite decades of little or no relevant regulation, contributions through bequests have remained a drop in the proverbial bucket. 11 Second, and perhaps most striking, the FEC does not point to even a single quid pro quo exchange—at any time in American history—allegedly effected through a bequest. Nor do the careful, extensive findings made by the district courts in the LNC cases. See LNC I, 930 F. Supp. 2d at 171–90; LNC II, 317 F. Supp. 3d at 225–57. In developing the records for those cases, all the FEC could muster up was more evidence of corruption involving contributions from living donors. See id. at 236–42. In striking down limits on independent expenditures by corporations, the Supreme Court stressed that “[t]he McConnell record was over 100,000 pages long, yet it does not have any direct examples of votes being exchanged for expenditures.” Citizens United, 558 U.S. at 360 (cleaned up). The FEC’s failure of proof here is no less dramatic. Third, there is no evidence of testators trying to play both sides. In McConnell, the Court found it “[p]articularly telling” that wealthy individuals “gave substantial sums to both major national parties, leaving room for no other conclusion but that these donors were seeking influence, or avoiding retaliation, rather than promoting any particular ideology.” 540 U.S. at 148. The FEC alleges nothing comparable as to bequests. This should hardly be surprising, for the possibility of a corrupt donor securing political favors, not by giving large sums to both parties during his lifetime, but by simultaneously remembering both parties in his will, seems almost fantastic. Against this evidence (or lack thereof), and despite the practical problems with effectuating any quid pro quo through a bequest, the majority posits that a corrupt bequest might be possible—in theory—if the donor and the party worked out the exchange in advance. Ante at 14–15. With respect, I find that possibility insufficient to discharge the FEC’s significant burden of proof under closely drawn scrutiny. The Supreme Court has “‘never accepted mere conjecture as adequate to 12 carry a First Amendment burden,’” McCutcheon, 572 U.S. at 210 (plurality opinion) (quoting Shrink Mo. Gov’t, 528 U.S. at 392), and so neither should we. 2 In any event, contribution limits are unconstitutional as applied to uncoordinated bequests. To reiterate, the majority posits that bequests could be corrupt if the testator bargained with the intended beneficiary before his death. Ante at 14–15; see also LNC I, 930 F. Supp. 2d at 166 (“making one’s bequest known before death could be treated just as a contribution is”). But this cannot happen if the testator does not even tell the recipient about the planned bequest during his lifetime. In that circumstance, a quid pro quo exchange is impossible. The only response is that coordinated and uncoordinated bequests may be difficult to distinguish, so both must be regulated together. But this reasoning runs counter to perhaps the most fundamental distinction in campaign-finance law— between contributions and independent expenditures. In Buckley, the Court invalidated a limit on the expenditures that any person could make “relative to a clearly identified candidate.” See 424 U.S. at 39–51 (quotation marks omitted). The government defended the expenditure limit as necessary to prevent evasion of the limits on contributions to candidates. But the governing statute already treated “controlled or coordinated expenditures” as “contributions rather than expenditures.” Id. at 46 & n.53. And the Court held that this distinction between coordinated and independent spending also marked a critical constitutional line. Thus, the treatment of “prearranged or coordinated expenditures” as contributions permissibly addressed the government’s concern about evading contribution limits. Id. at 47. But the limit on independent expenditures did not. As the Court explained: 13 “The absence of prearrangement and coordination of an expenditure with the candidate or his agent not only undermines the value of the expenditure to the candidate, but also alleviates the danger that expenditures will be given as a quid pro quo for improper commitments from the candidate.” Id. Later decisions have reinforced this “fundamental constitutional difference” between independent expenditures, which are fully protected, and coordinated expenditures, which may be and are regulated as contributions. FEC v. Nat’l Conservative Political Action Comm., 470 U.S. 480, 497 (1985); see, e.g., McConnell, 540 U.S. at 202–03, 219–22; Colorado I, 518 U.S. at 613–16 (plurality opinion); FEC v. Mass. Citizens for Life, Inc., 479 U.S. 238, 251–63 (1986). Most recently, in Citizens United, the Court applied this reasoning to invalidate limits on independent expenditures by corporations and unions. 558 U.S. at 356–60, 365–66. In SpeechNow, this Court recognized that the protection for independent expenditures also constrains the government’s ability to regulate contributions. We held that contribution limits are unconstitutional as applied to recipients that engage only in independent expenditures. We noted that, after Citizens United, “the government has no anti-corruption interest in limiting independent expenditures.” 599 F.3d at 693. Then, we reasoned: “In light of the Court’s holding as a matter of law that independent expenditures do not corrupt or create the appearance of quid pro quo corruption, contributions to groups that make only independent expenditures also cannot corrupt or create the appearance of corruption.” Id. at 694. Because no legitimate government interest was implicated, even a modest impairment of speech and associational rights would be unconstitutional. See id. at 695 (“something … outweighs nothing every time” (quotation marks omitted)). 14 The line between coordinated and uncoordinated spending thus runs throughout campaign-finance law, and the FEC routinely must police it. Congress has long defined an expenditure “independent” of a candidate as one that, in pertinent part, was “not made in concert or cooperation with or at the request or suggestion of such candidate, the candidate’s authorized political committee, or their agents, or a political party committee or its agents.” 52 U.S.C. § 30101(17)(B); see also id. § 30116(a)(7)(B)(i) (treating expenditures not independent of a candidate as “a contribution to such candidate”); McConnell, 540 U.S. at 221–22 & n.99. A parallel definition now distinguishes expenditures “independent” of political parties from contributions to those parties. See id. at 219–20 & n.97. The Supreme Court has held that this definition is not impermissibly vague, id. at 222–23; the FEC has promulgated a swath of regulations implementing it, see generally 11 C.F.R. pt. 109; and the Commission or the courts frequently apply it to determine whether disputed expenditures were in fact independent, see, e.g., Colorado I, 518 U.S. at 619–23 (plurality opinion); AFL-CIO v. FEC, 333 F.3d 168, 171 (D.C. Cir. 2003). Likewise, other decisions assess whether specific entities make only independent expenditures and thus have a First Amendment right to receive unrestricted contributions under SpeechNow. See, e.g., Vt. Right to Life Comm., Inc. v. Sorrell, 758 F.3d 118, 140–41 (2d Cir. 2014). Armed with extensive disclosure requirements and enforcement powers, the FEC routinely determines whether disputed expenditures were coordinated or independent. The FEC offers no reason why it cannot make the same determination as to bequests. Because coordinated and uncoordinated bequests can be manageably distinguished, and because uncoordinated bequests are not even alleged to present any corruption risk, the contribution limits are unconstitutional at least as applied to them. 15 3 Finally, the contribution limits are unconstitutional as applied to Shaber’s individual bequest. Not only was his bequest uncoordinated, but several additional facts make the LNC’s challenge even stronger. First, far from coordinating with the LNC, Shaber never even told the LNC of the bequest before his death. LNC II, 317 F. Supp. 3d at 249. With the LNC unaware that a testamentary quid might be forthcoming, there could be no quid pro quo agreement—nor even any debate about whether to infer such an agreement based on winks, implicit understandings, or other ambiguous circumstances. Second, the bequest came with no strings attached. LNC II, 317 F. Supp. 3d at 250. It neither demanded nor even asked that the LNC do anything in return. The district court noted that, in one other instance, a trustee had requested that the LNC use the bequest to help defeat specific candidates. See id. at 248. There would be nothing wrong with such an agreement, for that quo would not involve any “official act” of the government. See McCutcheon, 572 U.S. at 192 (plurality opinion). But, here, Shaber never sought any quo at all. Third, the LNC “provided nothing of value” in exchange for the bequest, except perhaps for continuing to “pursu[e] its ideological and political mission.” LNC II, 317 F. Supp. 3d at 251. In LNC I, the FEC expressed concern that a political party could grant “preferential access” to testators who (unlike Shaber) tell the party of the intended gift during their lifetime. 930 F. Supp. 2d at 186. However, “[i]ngratiation and access … are not corruption.” Citizens United, 558 U.S. at 360. And, here, Shaber did not seek even that. 16 Fourth, Shaber made only modest contributions to the LNC during his lifetime. As the district court explained, Shaber’s total lifetime donation of $3,315, made in 46 separate contributions spread out over 24 years, “is a drop in the bucket relative to current law’s annual limit of $33,900 for individuals to contribute for any purpose to national political party committees, and an even smaller drop relative to the limit of $339,000 that individuals may contribute for either general or specialized purposes.” LNC II, 317 F. Supp. 3d at 216. Likewise, Shaber’s contribution history did not qualify him for any of the benefits that the LNC affords to its major donors. See id. at 242. So, there is no reason to think that the LNC might have even identified Shaber as someone likely to make a large bequest, much less used that possibility to engineer a secret quid pro quo before his death. Finally, besides making his modest gifts, Shaber had no other relationship with the LNC during his lifetime, LNC II, 317 F. Supp. 3d at 251, thus making the prospect of corruption even more unlikely. B The majority views the LNC’s as-applied claim as resting on nothing more than a factual contention that Shaber’s individual bequest was not corrupt. Ante at 16–17. It then rejects the claim as inconsistent with Buckley’s holding that, because corrupt and legitimate contributions are hard to distinguish, “prophylactic” limits may be applied to both. Ante at 17–19 (quotation marks omitted). But there is more to the LNC’s claim. As noted above, the fact that the LNC sought relief only as to Shaber’s bequest did not prevent it from making substantive arguments that sweep more broadly. See Bucklew v. Precythe, 139 S. Ct. 1112, 1127–28 (2019); Citizens United, 558 U.S. at 17 331. Although the LNC asks us to assess Shaber’s bequest based on a totality of the circumstances, it also makes broader arguments keyed to the general nature of bequests and uncoordinated bequests. See, e.g., LNC Opening Br. at 37 (“[B]arring supernatural intervention, the potential for quid pro quo activity is rather more limited than in the case of a living donor, as are prospects for its enforcement. Regardless of what the LNC might do for Shaber now, he will give it nothing more or less than his bequest.”); LNC Reply Br. at 14 (“Bequests are different. Until death, they are merely a revocable promise. After death, they are irrevocable, and cannot be policed by the dead for quid pro quo compliance.”). In my judgment, that was enough to preserve the broader arguments—and, as to them, to trigger the FEC’s burden of proof under closely drawn scrutiny. The FEC did not misapprehend this point; to the contrary, it argued both that Buckley forecloses as-applied challenges based on the facts of individual cases, FEC Br. at 25–28, and that bequests as a category raise the same corruption concerns as other kinds of political contributions, id. at 29–32. On the merits, the LNC’s substantive arguments do not threaten the general justification for prophylactic contribution limits. As explained above, contributions made through bequests may be safely distinguished as a category—just like contributions to groups that make only independent expenditures. See SpeechNow, 599 F.3d at 692–96. The same is true for the narrower category of contributions made through uncoordinated bequests. And to the extent that additional facts strengthen the LNC’s challenge, there is nothing inappropriate about considering them. Successful as-applied challenges often turn on the facts of individual cases. See, e.g., WRTL, 551 U.S. at 469–81 (plurality opinion) (expenditure limit impermissibly extended beyond functional equivalent of express advocacy); Brown v. Socialist Workers ’74 Campaign Comm., 459 U.S. 87, 88 (1982) (disclosure requirement 18 impermissibly subjected party to threats or harassment). Likewise, case-specific facts would be necessary to determine whether contribution limits prevent individual recipients from “amassing the resources necessary for effective advocacy”—a type of as-applied challenge that McConnell repeatedly invited. 540 U.S. at 159 (quotation marks omitted); see id. at 173. The majority also suggests that as-applied challenges to contribution limits may be appropriate in cases where the burdens imposed on speakers are particularly harsh, but not in cases where the relevant government interests are particularly weak. Ante at 19–20. There is no conceptual reason why that should be so, for closely drawn scrutiny requires proof both that an important government interest is implicated and that the challenged restriction does not infringe speech or associational interests unnecessarily. SpeechNow confirms this point. There, in striking down contribution limits as applied to recipients that make only independent expenditures, we rested squarely on the premise that “the government ha[d] no anti- corruption interest” in that case, without reaching the question of how severely the challenged limits infringed speech and associational interests. 599 F.3d at 694–95. Finally, it is worth remembering that Buckley and McConnell are themselves exceptions to an overarching First Amendment principle. “Broad prophylactic rules in the area of free expression are suspect,” and “[p]recision of regulation must be the touchstone” in this area. NAACP v. Button, 371 U.S. 415, 438 (1963). Buckley and McConnell qualify that principle, by approving “prophylactic” restrictions extending to some non-corrupt contributions. McCutcheon, 572 U.S. at 221 (plurality opinion). But the “prophylaxis” must also have limits. See id. Under closely drawn scrutiny, it cannot properly be extended to bequests that, as a group and individually, may reliably be determined to be legitimate. 19 IV Beyond any question about bequests, the LNC challenges the contribution limits as amended in 2014. The LNC contends that the current limits are unconstitutional, both on their face and as applied. On this point, the LNC does not highlight any facts about Shaber’s individual contribution, but instead attacks the statutory scheme itself. The provisions at issue are structured as one old rule subject to three new exceptions. The rule is that no person may contribute over $25,000 per year to a national political party, 52 U.S.C. § 30116(a)(1)(B), subject to adjustment for inflation, id. § 30116(c). It is contained in the Federal Election Campaign Act of 1971 (FECA), as amended by the Bipartisan Campaign Finance Reform Act of 2002 (BCRA), and it was upheld by McConnell. See Pub. L. No. 107-155, § 307(a)(2), (d), 116 Stat. 81, 102–03; 540 U.S. at 142–61. The exceptions permit individuals to make additional annual contributions of up to $75,000 for presidential nominating conventions, $75,000 for party headquarters, and $75,000 for recounts and other legal proceedings, all subject to the same inflation adjustment. 52 U.S.C. § 30116(a)(1)(B), (a)(9). They were created by a 2014 amendment to FECA. Pub. L. No. 113-235, div. N, § 101, 128 Stat. 2130, 2772–73. The LNC’s challenge to this scheme mixes attacks on the new exceptions, attacks on the old rule, and attacks on how the two treat different categories of speech differently. The LNC also combines arguments based on overbreadth and underbreadth. But once these various arguments are unpacked, none of them succeeds. Most obviously, the new contribution limits do not themselves restrict too much speech. On this point, McConnell controls. If a prohibition on contributing more than $25,000 to a political party for any purpose does not restrict too much 20 speech, then neither do exceptions that permit additional contributions of up to three times that amount. The majority correctly concludes that this much is a matter of “simple mathematics,” ante at 30, and Judge Griffith agrees, ante at 1. The LNC further attacks the statutory distinction between contributions for nominating conventions, headquarters, and legal proceedings (now governed by the higher 2014 limits) and contributions for all other purposes (still governed by the lower BCRA limit). It contends that there is no anti-corruption justification for treating these categories differently. The majority concludes that there are such justifications, ante at 30– 32, while Judge Griffith concludes that there may not be, ante at 3–7. In my view, Judge Griffith has the better of this argument, so I would join his dissent if the First Amendment required proof of a corruption-based justification for the differential treatment of these speech categories. But I do not think that such proof is necessary in this case. As a general matter, “the First Amendment imposes no freestanding ‘underinclusiveness limitation.’” Williams-Yulee v. Fla. Bar, 135 S. Ct. 1656, 1668 (2015) (quoting R.A.V. v. City of St. Paul, 505 U.S. 377, 387 (1992)). So, for example, if a state may prohibit obscenity across the board, then it may prohibit obscene telephone calls but not obscene telegrams— even if the two raise comparable concerns. See R.A.V., 505 U.S. at 387. Otherwise, laws might “violate[] the First Amendment by abridging too little speech”—which is highly “counterintuitive.” Williams-Yulee, 135 S. Ct. at 1668. In my view, that principle governs this case. Under closely drawn scrutiny, Congress needed an anti-corruption justification both to impose BCRA’s original contribution limit and to limit the additional categories of spending permitted by the 2014 amendment. As noted above, McConnell found 21 sufficient justification for the former, and the latter follows from it. But Congress did not need a further, corruption-related justification to restrict contributions for nominating conventions, headquarters, and legal expenses less severely than it restricts other contributions. Rather, Congress could have chosen to restrict those contributions less severely for other reasons, such as a desire to make up for the loss of public funds for nominating conventions, or simply to permit more speech rather than less. The First Amendment demands a strong anti-corruption justification when Congress chooses to restrict campaign contributions, not when it chooses to loosen the restrictions. There are two important qualifications to this analysis, but neither affects the bottom line here. First, distinctions among categories of speech may violate the First Amendment if they are based on content. See R.A.V., 505 U.S. at 387 (“the First Amendment imposes not an ‘underinclusiveness’ limitation but a ‘content discrimination’ limitation”). Here, the LNC contends that the more favorable treatment of contributions for nominating conventions, headquarters, and legal expenses is content-based, because it targets speech based on its “function or purpose.” Reed v. Town of Gilbert, 135 S. Ct. 2218, 2227 (2015). Moreover, if a distinction between “political” and other speech is content- based, see id. at 2224–30, then so are the distinctions among the types of political-speech contributions at issue here. Whatever the force of this argument in the abstract, it cannot carry the day. Reed did not involve campaign contribution limits, which the Supreme Court has long treated as content-neutral restrictions subject to intermediate scrutiny. So, while I disagree with the majority’s suggestion that Reed is inapposite because this case does not involve speech 22 restrictions, ante at 26, I agree with its ultimate conclusion, ante at 27–28, that a lower court cannot follow the implications of Reed as against the holdings of the campaign-finance cases. See Agostini v. Felton, 521 U.S. 203, 237 (1997). Second, underinclusiveness can raise First Amendment concerns for another reason, by suggesting that the government is not pursuing its asserted interests or that the challenged speech restriction will not substantially advance them. See Williams-Yulee, 135 S. Ct. at 1668. The majority concludes that the 2014 scheme does not raise these concerns, ante at 30– 32, while Judge Griffith concludes that it does, ante at 3–9. Were we free to engage this question, I would agree with Judge Griffith. But I believe that McConnell forecloses the debate. An underinclusiveness argument along these lines uses speech-enabling exceptions to attack a speech-restricting rule. If the government allows the sale of violent movies, that casts doubt on its asserted need to restrict the sale of violent video games. Brown v. Entm’t Merchs. Ass’n, 564 U.S. 786, 801–02 (2011). If the government permits newspapers to be distributed through newsracks, that casts doubt on its asserted need to prohibit commercial publications from being similarly distributed. City of Cincinnati v. Discovery Network, Inc., 507 U.S. 410, 416–28 (1993). If the government permits electronic media to release names of juvenile offenders, that casts doubt on its asserted need to prohibit newspapers from doing so. Smith v. Daily Mail Publ’g Co., 443 U.S. 97, 104–05 (1979). Here, the analogous argument amounts to a direct attack on BCRA itself: If Congress permits annual contributions to political parties of $225,000 (or $300,600, adjusted for inflation) for three specified categories of activity, that casts doubt on its asserted need to prohibit all other annual contributions over $25,000 (or $33,400, adjusted for inflation). 23 As Judge Griffith explains, the argument is compelling: money is fungible, the exceptions dwarf the rule, and there is no plausible anti-corruption rationale to explain the disparate treatment. Nonetheless, McConnell held that BCRA’s $25,000 contribution limit substantially advances, and is narrowly tailored to, the important government interest in combatting actual or apparent quid pro quo corruption. If we may not revisit that conclusion based on intervening Supreme Court decisions that undermine McConnell’s reasoning, see Agostini, 521 U.S. at 237, then neither may we revisit it based on intervening statutes that do likewise. On this point, any course correction must come from the Supreme Court itself. Judge Griffith concludes that McConnell is not binding on this point because it did not involve a “regime” with the three new exceptions. Ante at 2. True enough, but the upshot of his argument is that “limiting general contributions to $33,400” is now unconstitutional. Id. And that general limit, created by section 307(a)(2) of BCRA, and currently codified at 52 U.S.C. § 30116(a)(1)(B), is precisely the one that McConnell upheld. * * * * I join Part II of the majority opinion, which holds that the LNC has standing to raise its various challenges. For the reasons given above, I respectfully dissent from Part III of the opinion, and I concur in the judgment as to Part IV.