14 FEDERAL ELECTION COMM’N v. TED CRUZ FOR SENATE KAGAN, J., dissenting turn.” App. 351–353. That bears repeating: 81%—an overwhelming perception across all demographic categories, as well as across all party affiliations and political ideologies. See ibid. As the court reviewing the Kentucky version of Section 304 explained: “[T]here is an impression” when a contribution repays a loan after an election that the contributor is simply “lining the candidate’s pocket, as there is no ongoing campaign to which the contribution may be made.” Wilkinson, 876 F. Supp., at 930; see supra, at 12. The majority flyspecks the polling questions: Why didn’t the poll define “political favor”? Did the poll mention that the contributions had to comply with the $2,900 cap? And so forth. See ante, at 17–18. But really—is it likely that such tinkering would have made a real difference? The poll results were so lopsided because the post-election contributions Section 304 targets—ones adding to the candidate’s personal wealth—have so conspicuous a potential to corrupt. The public knows that to be true. The public’s representatives in Congress knew it to be true. Only this Court— somehow—does not. * * * “Democracy works only if the people have faith in those who govern.” Shrink Missouri, 528 U. S., at 390 (internal quotation marks omitted). And the people cannot have faith in representatives who trade official acts for financial gain. Section 304 prevents that kind of corruption, at barely discernable cost to First Amendment freedoms. The provision limits one narrow use of third-party contributions to a campaign, thus “entail[ing] only a marginal restriction” on speech. Buckley, 424 U. S., at 20. And the provision targets a practice posing exceptional risks of quid pro quo deals. Repaying a candidate’s loan after he has won election cannot serve the usual purposes of a contribution: The money comes too late to aid in any of his campaign activities. All the money does is enrich the candidate personally