The Petition for Rehearing is DENIED and the court having been polled at the *164 request of one of its members, and a majority of the judges who are in regular active service and not disqualified not having voted in favor ( Fed. R. App. P. 35 and 5th Cir. R. 35 ), the Petition for Rehearing En Banc is DENIED.

In the en banc poll, two judges voted in favor of rehearing (Judges Jones and Ho) and twelve judges voted against rehearing (Chief Judge Stewart and Judges Smith, Dennis, Clement, Owen, Elrod, Southwick, Haynes, Graves, Higginson, Costa, and Willett).

JAMES C. HO, Circuit Judge, with whom EDITH H. JONES, Circuit Judge, joins as to Parts I and II, dissenting from denial of rehearing en banc:

The unfortunate trend in modern constitutional law is not only to create rights that appear nowhere in the Constitution, but also to disfavor rights expressly enumerated by our Founders. See , e.g. , Silvester v. Becerra , --- U.S. ----, 138 S.Ct. 945 , --- L.Ed.2d ---- (2018) (Thomas, J., dissenting from denial of certiorari). This case reinforces this regrettable pattern.

There is no more quintessentially American principle than the right of the people to participate in their own governance. The First Amendment protects the freedom of speech, and that freedom emphatically includes the right to speak about who our elected leaders should and should not be. This foundational American liberty includes not only the freedom to engage in one's own political speech, but also the freedom to support like-minded candidates for office.

The First Amendment therefore protects campaign contributions. For example, in Randall v. Sorrell , the Supreme Court invalidated various campaign contribution limits imposed by the State of Vermont. 548 U.S. 230 , 126 S.Ct. 2479 , 165 L.Ed.2d 482 (2006). That included a limit of $300 per election cycle-that is, $150 per election (primary and general), or $215 in 2015 dollars-for state senators representing between 20,000 and 120,000 people. Id. at 236-38 , 126 S.Ct. 2479 (plurality); see also Joint App'x at 21-22, Randall , 548 U.S. 230 (Nos. 04-1528, 04-1530, 04-1697), 2005 WL 3477006 , at *55-56, 79.

This case involves a similarly low contribution limit of $350 per election, in 2015 dollars, for city council members representing fewer than 100,000 people in Austin, Texas. Zimmerman v. City of Austin , 881 F.3d 378 , 387 & n.3 (5th Cir. 2018). For several reasons, we should have granted rehearing en banc and held that the Austin contribution limit violates the First Amendment.

I.

Campaign contributions are not personal gifts-they are donations to support and defray the costs of campaign speech. See , e.g. , FEC v. Mass. Citizens for Life, Inc. , 479 U.S. 238 , 261, 107 S.Ct. 616 , 93 L.Ed.2d 539 (1986) ("[I]ndividuals contribute to a political organization in part because they regard such a contribution as a more effective means of advocacy than spending the money under their own personal direction."); McCormick v. United States , 500 U.S. 257 , 272, 111 S.Ct. 1807 , 114 L.Ed.2d 307 (1991) ("[E]lection campaigns are financed by private contributions or expenditures, as they have been from the beginning of the Nation.").

Accordingly, the Supreme Court has carefully delimited the narrow circumstances in which the government may permissibly interfere with campaign contributions. In fact, the only legitimate government interest for limiting campaign contributions is preventing unlawful quid pro quo corruption or the appearance thereof.

*165 McCutcheon v. FEC , --- U.S. ----, 134 S.Ct. 1434 , 1450, 188 L.Ed.2d 468 (2014) (plurality). And as the Court has made clear, quid pro quo corruption requires "a direct exchange of an official act for money." Id. at 1441 .

The Court has also explicitly rejected other purported justifications for restricting campaign contributions. It has held that amorphous concerns about "improper influence" or "access" are too ambiguous and imprecise to warrant interference with First Amendment rights. Compare Nixon v. Shrink Mo. Gov't PAC , 528 U.S. 377 , 388-89, 120 S.Ct. 897 , 145 L.Ed.2d 886 (2000), with McCutcheon , 134 S.Ct. at 1451 ("The line between quid pro quo corruption and general influence ... must be respected in order to safeguard basic First Amendment rights."), and Citizens United v. FEC , 558 U.S. 310 , 360-61, 130 S.Ct. 876 , 175 L.Ed.2d 753 (2010) ("Ingratiation and access ... are not corruption."). Nor may government regulate contributions "simply to reduce the amount of money in politics, or to restrict the political participation of some in order to enhance the relative influence of others." McCutcheon , 134 S.Ct. at 1441 .

Moreover, the risk of quid pro quo corruption must be established by evidence-courts may not "accept[ ] mere conjecture as adequate to carry a First Amendment burden." Id. at 1452 (emphasis added) (quoting Shrink , 528 U.S. at 392 , 120 S.Ct. 897 ).

This standard is fatal to Austin's $350 contribution limit. It is at best "conjectural" that a $351 contribution to help defray the costs of campaign speech would create a genuine risk of an unlawful quid pro quo exchange. Justice Thomas put it well: "I cannot fathom how a $251 contribution could pose a substantial risk of securing a political quid pro quo "-referring to Missouri's $250 contribution limit in elections involving fewer than 100,000 constituents, which adjusted for inflation is $390 in 2015 dollars. Randall , 548 U.S. at 272-73 , 126 S.Ct. 2479 (Thomas, J., concurring) (alterations and quotations marks omitted) (quoting Shrink , 528 U.S. at 425 , 120 S.Ct. 897 (Thomas, J., dissenting) ). His words are equally applicable here: I too cannot fathom how a $390 contribution could pose a substantial risk of securing a political quid pro quo .

The district court should have heeded Justice Thomas's common-sense observation-particularly because the record is devoid of any evidence to the contrary. The district court merely credited the City's assertion that voters in 1997 had a "perception" of "inordinate influence" based on "large contributions, in the $1000-$2500 range"-which is $1,420-$3,545 in 2015 dollars.

There are numerous problems with the City's defense. It credits voter "perception"-which is perilously close to "mere conjecture." It raises amorphous concerns about "inordinate influence"-not quid pro quo corruption. And even ignoring these defects, this "evidence" would not remotely justify a substantially lower contribution limit of $350-less than 25 percent of the "large contributions" that concerned Austin voters.

Not surprisingly, then, when a respected panel of this Court upheld the district court's judgment, it did not rely on any of the dollar values identified by the district court. Instead, the panel invoked Supreme Court precedent: "[I]n Shrink Mo. the Supreme Court upheld Missouri's $275 limit-which, adjusted for inflation, was equivalent to approximately $390 at the time this appeal was filed-on contributions to candidates for any office representing fewer than 100,000 people." 881 F.3d at 387 . In other words, the panel ruled that the difference between the $390 limit in Shrink and the $350 limit challenged *166 here was immaterial for First Amendment purposes. Id . ("Austin's $350 limit ... is not so low by comparison as to raise suspicion.").

But the reliance on Shrink is mistaken for at least two reasons.

To begin with, Austin's $350 limit is more than 10 percent less than the $390 limit at issue in Shrink . As Justice Thomas explained in his concurrence, the Randall plurality treated "the limits in Shrink as a constitutional minimum, or at least as limits below which 'danger signs' are present." 548 U.S. at 269 , 126 S.Ct. 2479 (Thomas, J., concurring).

But there's an even more basic problem here: The Supreme Court did not pass judgment on the constitutionality of the $390 limit in Shrink . 528 U.S. at 382-83 , 120 S.Ct. 897 (describing the inflation-adjusted "$1,075 [limit] for contributions to candidates for statewide office (including state auditor)" as the "particular provision challenged here"); see also Shrink Mo. Gov't PAC v. Adams , 204 F.3d 838 , 840 (8th Cir. 2000) (analyzing on remand "the $525 and $275 limits" because the Supreme Court "reviewed only the statewide limit of $1,075") (emphasis added). Rather, as Randall explained, "the lowest limit this Court has previously upheld [is] the limit of $1,075 per election ... for candidates for Missouri state auditor." 548 U.S. at 251 , 126 S.Ct. 2479 (plurality) (emphasis added) (citing Shrink , 528 U.S. 377 , 120 S.Ct. 897 ).

Thus, in holding the Vermont limit unconstitutional, Randall specifically noted that "Vermont's limit is well below ... $1,075." Id. (emphasis added). So too here: Austin's $350 limit is "well below" $1,075 (or $1,525 in 2015 dollars). Moreover, Randall observed that the "comparable Vermont limit of roughly $200 per election ... is less than one-sixth of Missouri's current inflation-adjusted limit." Id . And again, so too here: Austin's $350 limit is less than one-fourth of the inflation-adjusted $1,525 limit upheld in Shrink .

Because Austin's contribution limit is "substantially lower" than the limits previously upheld by the Supreme Court, there are "danger signs that [Austin's] contribution limit[ ] may fall outside tolerable First Amendment limits." Id . at 253, 126 S.Ct. 2479 . See also id. at 252 , 126 S.Ct. 2479 ("it [is] difficult to treat Shrink 's (then) $1,075 limit as providing affirmative support for the lawfulness of Vermont's far lower levels"); id. at 269 , 126 S.Ct. 2479 (Thomas, J., concurring) (emphasizing plurality's "treatment of the limits in Shrink as a constitutional minimum, or at least as limits below which 'danger signs' are present"). Based on the evidence presented below, and under my reading of Shrink and Randall , it is difficult to see how Austin's $350 limit is "closely drawn" to serve a recognized government interest, as required by the Supreme Court. Randall , 548 U.S. at 253-63 , 126 S.Ct. 2479 (plurality) (citing Buckley v. Valeo , 424 U.S. 1 , 20-22, 36-37, 96 S.Ct. 612 , 46 L.Ed.2d 659 (1976) ).

II.

A majority of this Court has decided not to rehear this case en banc. But that decision need not foreclose a future challenge to Austin's contribution limit. Indeed, although I would have held unconstitutional Austin's limit based solely on the record in this case, there is additional evidence and argument that Mr. Zimmerman could have marshaled-but did not-that would have brought the unconstitutionality of the Austin contribution limit into even sharper relief.

In his effort to distinguish Shrink , Mr. Zimmerman adjusted for both inflation and population size. But he did not additionally *167 adjust for what I will call locality considerations-such as media market costs and other cultural factors-that affect the cost of campaigning in a particular area. It would not be surprising if the cost of reaching voters were significantly greater in Austin than in Missouri. Accordingly, it may well be that a $350 contribution limit is substantially more disruptive to effective campaign advocacy in Austin than in Missouri. See Randall , 548 U.S. at 248 , 126 S.Ct. 2479 ("Following Buckley , we must determine whether [Vermont's] contribution limits prevent candidates from 'amassing the resources necessary for effective [campaign] advocacy.' ") (second alteration in original) (quoting Buckley , 424 U.S. at 21 , 96 S.Ct. 612 ).

Nothing in Supreme Court precedent precludes such locality considerations in assessing the constitutionality of campaign contribution limits. To the contrary, the parties in Randall well understood the relevance of such considerations. 1 And our sister circuits have too. 2

Because Mr. Zimmerman neither presented this legal theory here nor offered any evidence to support it, the panel decision should not foreclose another Austin *168 citizen from presenting evidence and argument regarding such locality considerations in a future challenge to the Austin contribution limit. See De La Paz v. Coy , 786 F.3d 367 , 373 (5th Cir. 2015) ("[A]ccording to black letter law, 'a question not raised by counsel or discussed in the opinion of the court' has not 'been decided merely because it existed in the record and might have been raised and considered.' ") (quoting United States v. Mitchell , 271 U.S. 9 , 14, 46 S.Ct. 418 , 70 L.Ed. 799 (1926), and citing Henry Campbell Black, Handbook on the Law of Judicial Precedents, or, The Science of Case Law 37 (1912) ). Nor should it foreclose a challenge to Austin's contribution limit for mayoral races, which was not at issue in this case. 881 F.3d at 384 n.1.

III.

The Austin contribution limit is invalid under current Supreme Court precedent. Moreover, there are more fundamental problems with such laws: Contribution limits such as Austin's are simultaneously over- and under-inclusive-defects that have been held fatal in other First Amendment contexts.

First, as to over-inclusiveness: As the Supreme Court has recognized, the First Amendment imposes such a formidable barrier to government interference with speech that it not only forbids the government from imposing a regulation that affects both protected and unprotected speech-it even forbids government from regulating unprotected activities alone, if the regulation also threatens to chill protected speech. See , e.g. , Bates v. State Bar of Ariz. , 433 U.S. 350 , 380, 97 S.Ct. 2691 , 53 L.Ed.2d 810 (1977) ("The reason for the special rule in First Amendment cases is apparent: An overbroad statute might serve to chill protected speech. First Amendment interests are fragile interests, and a person who contemplates protected activity might be discouraged by the in terrorem effect of the statute."); Dombrowski v. Pfister , 380 U.S. 479 , 494, 85 S.Ct. 1116 , 14 L.Ed.2d 22 (1965) (holding unconstitutional an "overly broad statute" because it "creates a 'danger zone' within which protected expression may be inhibited").

In other words, the First Amendment prophylactically protects speech from government intrusion. Yet campaign contribution limits turn this principle on its head: They prophylactically prohibit protected speech, in hopes of targeting the "appearance" of unprotected activity in the form of quid pro quo corruption.

By design, contribution limits categorically bar all contributions over a certain threshold, irrespective of the purpose or motivation of the donor. But this is dramatically over-inclusive. Many contributions have nothing to do with the appearance of-let alone any actual- quid pro quo corruption. Countless Americans contribute for no other reason than to "support candidates who share their beliefs and interests." McCutcheon , 134 S.Ct. at 1441 . Because the candidate and the donor share common beliefs, the candidate is already "expected to be responsive to those concerns," without any inkling of a quid pro quo agreement. Id. Indeed, many Americans contribute without ever even communicating with the candidate-for example, a donor might simply be inspired by the candidate's prior record of public service, proposed future action, or a particular speech or debate performance. Such contributions are far from corrupt-to quote McCutcheon , they "embody a central feature of democracy." Id. The Court nevertheless allows their criminalization. This is textbook over-inclusiveness.

Campaign contribution limits are also impermissibly under-inclusive. In other *169 contexts, the Supreme Court has held that the First Amendment forbids laws that infringe on the freedom of speech-even where the government's interest is compelling-if the law is under-inclusive and therefore fails to further a recognized government interest. See , e.g. , The Florida Star v. B.J.F. , 491 U.S. 524 , 540, 109 S.Ct. 2603 , 105 L.Ed.2d 443 (1989) ("[T]he facial underinclusiveness of [the statute] raises serious doubts about whether Florida is, in fact, serving, with this statute, the significant interests which appellee invokes in support of affirmance."); Citizens United , 558 U.S. at 362 , 130 S.Ct. 876 ("[T]he statute is both underinclusive and overinclusive. ... [I]f Congress had been seeking to protect dissenting shareholders, it would not have banned corporate speech in only certain media within 30 or 60 days before an election. A dissenting shareholder's interests would be implicated by speech in any media at any time.").

Take Buckley , for example. The Court held that citizens have a First Amendment right to spend money on their own political speech to support a political campaign-also known as independent expenditures-despite the obvious risk that such independent expenditures may pose the same potential for quid pro quo corruption as direct campaign contributions. 424 U.S. at 45 , 96 S.Ct. 612 (invalidating limits on independent expenditures, while upholding campaign contribution limits, even "assuming, arguendo , that large independent expenditures pose the same dangers of actual or apparent quid pro quo arrangements as do large contributions").

This raises an obvious question: If the government cannot regulate independent expenditures, what government interest is served by regulating only campaign contributions? As any proponent of campaign finance regulation will tell you, a donor with suspect intentions can circumvent campaign contribution limits-and achieve his nefarious goals-simply by making independent expenditures instead. So either the government regulates everything-or there's no point in regulating any of it.

Indeed, that is what the Court said in Buckley itself. There, the Court invalidated a rule that restricted independent expenditures that expressly advocated for a candidate, on the ground that it would be pointlessly underinclusive: Donors could simply make independent expenditures that avoid express advocacy but still benefit the candidate. As the Court observed, it "would naively underestimate the ingenuity and resourcefulness of persons and groups desiring to buy influence to believe that they would have much difficulty devising expenditures that skirted the restriction on express advocacy of election or defeat but nevertheless benefited the candidate's campaign." Id. Accordingly, the Court held that " no substantial societal interest would be served " by such a restriction because it still "permitted unscrupulous persons and organizations to expend unlimited sums of money in order to obtain improper influence over candidates for elective office." Id. (emphasis added).

Limits on campaign contributions are even more under-inclusive-especially considering that, as the Supreme Court has made clear, donors have the right under the First Amendment to make any independent expenditures they desire.

I finish where I began: Campaign speech is core political speech under the First Amendment. Yet current Supreme Court jurisprudence disfavors it. Contribution limits such as Austin's are both over-inclusive and under-inclusive-defects the Court has found unacceptable in other First Amendment contexts.

* * *

*170 Under our Constitution, the people are not subjects, but citizens. As citizens, we enjoy the fundamental right to express our opinions on who does and does not belong in elected office.

To be sure, many Americans of good faith bemoan the amount of money spent on campaign contributions and political speech. But if you don't like big money in politics, then you should oppose big government in our lives. Because the former is a necessary consequence of the latter. When government grows larger, when regulators pick more and more economic winners and losers, participation in the political process ceases to be merely a citizen's prerogative-it becomes a human necessity. This is the inevitable result of a government that would be unrecognizable to our Founders. See , e.g. , NFIB v. Sebelius , 567 U.S. 519 , 132 S.Ct. 2566 , 183 L.Ed.2d 450 (2012).

So if there is too much money in politics, it's because there's too much government. The size and scope of government makes such spending essential. See , e.g. , EMILY's List v. FEC , 581 F.3d 1 , 33 (D.C. Cir. 2009) (Brown, J., concurring) ("The more power is at stake, the more money will be used to shield, deflect, or co-opt it. So long as the government can take and redistribute a man's livelihood, there will always be money in politics.").

But whatever size government we choose, the Constitution requires that it comply with our cherished First Amendment right to speak and to participate in our own governance. If we're going to ask taxpayers to devote a substantial percentage of their hard-earned income to fund the innumerable activities of federal, state, and local government, we should at the very least allow citizens to spend a fraction of that amount to speak out about how the government should spend their money. I respectfully dissent.