The U.S. Supreme Court’s decision in Citizens United v. FEC is one of its most unpopular in history. It is also one of the most misunderstood. Critics tend to see in Citizens United a cliff-jump into an abyss of unlimited political money, a descent that was well underway sometime before. In reality, the case represented a turning point in a different sense: It sounded the death knell for traditional, television-based campaigning and presaged a world of unregulated and unregulatable online political communication.
Contrary to popular conception, Citizens United was not a case about corporate campaign advertising. One may be forgiven for thinking it was, given that the Supreme Court resolved the case that way. But the case itself involved a movie—called, in fact, Hillary: The Movie—that a nonprofit corporation placed “on demand” for viewers to download. It was not a manipulative thirty-second ad run by a shady corporation. It did not invade the privacy of the home, washing over captive eyeballs and taking advantage of attention paid to other programming. Hillary: The Movie was an ideologically motivated hatchet job on a presidential candidate, to be sure, but it was one that viewers needed to seek out if they wished to watch it. It was not imposed on them.
Because the Federal Election Commission (FEC) regulated the movie just as it would regulate any corporate-sponsored, election-related television advertisement, it implicated the constitutionality of bans on corporate campaign advertising. The 2002 McCain-Feingold Law (or Bipartisan Campaign Reform Act) banned the use of corporate treasury funds for “broadcast, cable, and satellite communications” that referred to clearly identified Federal candidates within thirty days of a primary or sixty days of a general election. The Court, or the FEC for that matter, could have used the case as an opportunity to carve out an exception for downloadable movies, but it didn’t. An on-demand movie, just like a thirty-second ad, was deemed covered by the law, or at least it was covered until the Court decided that the regulation, in all of its manifestations, violated the First Amendment.
In the sturm und drang that followed the Court’s decision—cries of “corporations are not people” or “democracy for sale” and the like—little attention was paid to issues of evolving communication technology actually at the crux of the case, and that is a shame. In the six years since Citizens United, with technological change in campaign tactics proceeding at a rapid pace, we still have not come to grips with the legal and political implications of the impending decline of traditional forms of televised campaign communication.
Were we to grapple with those implications we would confront several uncomfortable truths. First, conventional and familiar assessments of the campaign finance “problem” are heavily dependent on television as the chief mode of campaigning. Second, regulation of campaign activity becomes increasingly difficult once campaigning moves online and traditional legal categories, such as who constitutes the “media”, become ever fuzzier. Third, a new set of challenges emerges both from the newfound precision with which voters can be targeted and the concentrated market power of certain platforms, such as Google and Facebook. Fourth, future campaign regulation, as a consequence, will not be the exclusive or even dominant province of government actors: The terms of service for platforms chiefly used for political communication will become more important than formal law in this regard.
This final development presents the danger that ill-fitted standards adopted for advertising in the commercial marketplace will merely be replicated for the political sphere. But it also presents opportunities to adapt these standards for online campaign communication to address some of the well-known pathologies of the campaign finance system.
Neither campaign finance reformers nor their chief critics will admit it, but much of the way we think about campaign finance is predicated on television—specifically the thirty-second television ad—existing as the primary mode of political campaigning. The constitutional concerns, relevant statutes, and regulatory activity in the context of campaign spending derive from a concern about the influence that uncontrolled TV ad spending will have on voters and candidates.
The constitutional debate has often focused on whether uncontrolled spending is “corrupting”, either of the candidates who benefit or of “the process”, when rich interests have disproportionate effects on electoral outcomes. But suppose “big money” was being spent on get-out-the-vote drives, door-knocking, lawn signs, or movies instead of TV ads. Would the same concerns arise? The notion of undue influence at the heart of the state’s anti-corruption interest implies not just the fact that a large amount of money is being spent, but that it is being spent on a medium seen as overly manipulative of voter decision-making. Consider, for example, the way Justice John Paul Stevens described the campaign finance problem in his 2005 dissent in a Vermont case: “Just as a driver need not use a Hummer to reach her destination, so a candidate need not flood the airways with ceaseless sound-bites of trivial information in order to provide voters with reasons to support her.”
Although the argument lurks below the surface of both the constitutional and policy discussions, television advertising has been seen as both special and especially dangerous. Even in the age of cable, when spectrum space has been less limited than it was when only three networks existed, the notion of the public airwaves as a public trust has made them especially regulable. So too, the “captivity” of a TV audience has justified heightened concern that advertising might be especially (and perhaps unfairly) effective when it coopts viewers who tuned in for other purposes.
The importance of TV to our thinking about campaign finance extends beyond the constitutional debate to the statutes and regulations on the books. The campaign finance laws (at least before Citizens United) regulated spending on all types of activities, but television was a particular cause of concern. The Bipartisan Campaign Reform Act’s express advocacy restriction regulated corporate-sponsored “broadcast, cable, and satellite communications.” In addition, its “stand by your ad” requirement (for example, “I am Donald Trump and I approve this message”) was more elaborate for television ads.
In general, the FEC (and Congress) have left the internet alone when it comes to campaign regulation. Disclosures are required for candidate websites, email communications to over 500 people, and ads placed on someone else’s website for a fee. Besides those few exceptions, though, online communications are uniquely unregulated. Indeed, when the current Chair of the FEC merely hinted at updating regulations to account for the transition to online campaigning, her opponents both within and beyond the FEC publicly criticized her.
The concern about regulating online campaign communication is well placed. Certain critical concepts in campaign finance law become especially contested in the cyberspace setting. Moreover, any domestic election authority, from whichever country, will find it difficult to enforce regulations on the web—which is worldwide, after all.
One concept that must be reconfigured for the web would be the legal definition of the media. In general, the media gets a pass in campaign finance law. Otherwise, the government could regulate how much money the New York Times Corporation or NBC, for example, could spend on its coverage of a presidential campaign. If the First Amendment means anything, it would probably cover the right of the institutional press to report and comment on political campaigns.
In the age of the internet, though, who (or what) counts as the media? Any person or group can start a blog, tweet a comment, or post a YouTube video. It is difficult to see how or why the “media” should be treated differently, even assuming we could find a way to bifurcate the web into the media and everyone else.
The difficulty in categorizing different actors in the internet campaign system is emblematic of the challenges in identifying the source for any communication on the web. Although Federal law prohibits foreign nationals from spending money in connection with a Federal election, for example, neither the FEC nor any other body of the Federal government could take down a website put up by a foreign government, corporation, or individual. To be sure, American campaigns are sufficiently expensive and complex that we should not expect a rush of foreign money to move in now that the internet spigot has opened up. But the point here is that enforcement of virtually any rule that depends on establishing the identity of the actor becomes very difficult to implement in the context of online communication.
That said, the internet’s inevitable lurch toward a libertarian regulatory environment does not translate into either a more egalitarian politics or a campaign environment in which money plays less of a role. New technologies and platforms give a potential voice to many who could not have had the resources or ability to broadcast their views when the costs of entry into television were quite high. But amid the cacophony of online voices, well-financed interests will still have a larger megaphone than the “poorly financed causes of little people” (to quote a famous 1943 Supreme Court case, Martin v. Struthers).
Some of the dangers of unequal online political influence are familiar and analogous to those in the TV-dominated environment, while others are unique to the communication possibilities that the internet enables. The hopeful utopians of the internet would cheer the liberation of campaigns from limited spectrum space and the concomitant pluralism of voices enabled by the multiple avenues for online political communication. And if one merely pays attention to the capacity of any given individual to make information or commentary available to the world, then the move to internet-based campaigning looks like a boon to democracy. It enables videos like those produced by “Obamagirl”, JibJab, conservative activist James O’Keefe, and the guerilla videographer who brought down George Allen’s political career.
Although campaigns can now employ any number of tools and apps through many different types of devices and platforms, there are still a few channels—namely Google and Facebook—that exert considerable market power. The power and reach of these few firms translates into electoral clout that can be used for good or ill, just like the broadcasters of yesteryear. Just as Walter Cronkite, during the heyday of the evening news, had the power to shape opinion over a sizable share of the American audience, so too these platforms could use their market power, should they so desire, to subtly affect electoral outcomes. As a result, familiar concerns (and some novel ones) are lodged about the institutional or corporate biases of these media and whether they are behaving responsibly and fairly.
One set of concerns revolves around the power that these platforms might abuse to shift election outcomes in their favor. Harvard Law Professor Jonathan Zittrain warns of the platforms’ manipulation of voters through what he calls “digital gerrymandering”, which refers to the distribution of information by a portal in order to serve its own ideological agenda. Zittrain cites two examples, one where Facebook ran an experiment that boosted voter turnout and the other from Google, which blacked out its homepage to generate opposition to the Stop Online Piracy Act. In a similar vein, Psychologist Robert Epstein, based on a series of experiments he has conducted, warns how Google’s search algorithm either intentionally or unintentionally could manipulate the outcome of elections by ranking certain candidates higher in search results.
Strategic biasing of the electorate toward the platforms’ favored candidates or parties, while cause for concern, is less likely than their use by well-financed campaigns to achieve the same ends. Indeed, that is the service for sale to the campaigns and interest groups. And while television advertising has its share of problems when it comes to “dark money” expenditures, the internet is potentially even more fertile ground for anonymous, unaccountable spending. As difficult as it may be to track down the source of funds for a televised ad purchased by some group akin to “Americans for America”, a website video could come from any source anywhere in the world.
Although online campaigning can enable formerly voiceless dissidents to have out-of-proportion effects on elections, the cheapness and openness of the medium enables all kinds of outside groups to have greater impact—and the richer and better organized the group, the better able it will likely be to use these new media, just like all others. As is true with direct democracy and any number of other populist reforms, greater opportunities for participation often translate into a decline in the role of traditional mediating institutions such as political parties. They cede control of the campaign conversation to other well-funded groups that do not have the same interests in forming a governing majority and pursuing policy objectives. The internet, under this view and given current politics, provides the perfect breeding ground for the most polarizing, least accountable, and most finely targeted forms of political communication.
The microtargeting enabled by new media is really what sets it apart. Whether one views this merely as a difference in degree (the two Obama campaigns perfected cable TV targeting, for instance, and direct mail and robocalling have always been personalized to some extent) or a difference in kind is less important than appreciating that targeting is now the coin of the campaign realm. The new forms of communication will enable campaign appeals personalized to an unprecedented degree. Microtargeting will be facilitated by the elaborate individual-level data that parties, interest groups, advertising agencies, and (especially) the platforms themselves have on individual voters, and by the widespread use of experimental methods to refine appeals to particular classes of voters.
The challenges posed by individualized politics derives, like the ones discussed earlier, from the echo-chamber effects magnified by online forms of communication and social media. Narrow appeals, by their nature, need not account for the diversity of reactions from a broader electorate. Those who worry that a candidate will send diametrically opposing communications to cater to different voters—a pro-Keystone pipeline ad for oil lovers, and an anti-pipeline ad to environmentalists, for example—are probably minimizing the capacity of the internet to check such inconsistencies. But different groups (for example, SuperPACs, 501(c)(4)s, parties, candidates) may send different types of messages to voters and even the same source will segment the electorate to the greatest extent possible to tailor messages to emphasize those issues or sentiments most likely to persuade, mobilize, or, in some cases, even to alienate.
This description of the Brave New World of online campaigning might lead to resignation to the future of anarchic electoral politics, with ideological echo chambers, massive anonymous spending, and microtargeting of unprecedented precision. Indeed, that is a possible future, but it is not the only or the inevitable one. This transitional moment in the dominant forms of political communication requires creative thinking about how to address well-known pathologies and newfound challenges of the campaign finance system.
To begin with, the principal regulator of political communication will not be a government agency but rather the internet portals themselves. For reasons already explained, the task of “regulating campaigning on the internet” is simply too large in scope and conceptually confusing to craft the kind of regulations that would be necessary in the area of political campaigns most sensitive to First-Amendment concerns. If regulation is going to occur, it will grow out of both the business necessity and sense of public obligation of the main portals.
However, let me dispel one popular red herring at the outset of this more normative discussion. Regulation will and does happen. The idea that political speech will be completely unconstrained on YouTube, Google, and Facebook is untrue now, and will continue to be false even as they adapt to their more important role in channeling campaign speech. The relevant question is not whether these portals will regulate, but rather whether the same regulations adopted for the commercial sphere will be replicated in toto for election-related activity.
The standards the portals currently apply to online speech read like a syllabus for a First Amendment course, but if adopted by the government, they all would violate the Constitution. Google, Twitter, Facebook, and YouTube all have standards for threats, harassment, hate speech, sexual content, incitement, violence, impersonation, and breaches of privacy. They also have a range of well-known and oft-debated protections for intellectual property.
In addition to regulations that apply to all speech on their platforms, each also has additional regulations for advertisers. They restrict advertising for alcohol, tobacco, and gambling. Google even prevents ads that promote prescription or over-the-counter medication, pregnancy or fertility related products, and a host of other health care-related content. Facebook, for its part, prevents ads that even “imply nudity”, show “shocking, sensational, disrespectful, or excessively violent content”, or contain “content that asserts or implies personal attributes” such as race or sexual orientation. It even edges into political territory with standards that ban “deceptive, false or misleading content” as well as “content that exploits controversial political or social issues for commercial purposes.”
If the portals move more explicitly into the business of regulating political campaigns, what might be the contours of such a regime? To begin with, I should be clear that such standards are only for the principal portals and only for paid content. No one would think that the websites of the RNC, DNC, Drudge Report, or the Sierra Club, for example, should abide by the same standards as Google and Facebook. Moreover, just as it would be impossible for the government to regulate campaigning on the internet, it would be impossible for the portals, if they are to retain their character as forums for social interaction, to restrict online conversations between individuals. This presents a big loophole for any standards, given that the most effective way for candidates to launch campaigns is to have their supporters do it for them, for free, without the official fingerprints of the campaign. But the principal concern animating regulation of internet campaigning has to do with the purchased manipulation of opinion—the replication of economic power in the electoral arena through the skillful targeting and persuasion of voters enabled by new forms of technology. Individual communication may raise concerns, especially as to tone and fairness, but authentic grassroots organizing and individual discussion of politics should be encouraged and protected online.
The types of potential regulations fall into several categories: disclosure, restrictions on sources and amounts of spending, misrepresentation, tone, and fairness. Given the business constraints placed on the advertising operations of the portals, the most likely domain of new regulations will be disclosure and perhaps some loose constraint regarding fairness (namely, that the portals themselves should allow advertising from all comers regardless of party). But it is also possible that they might experiment with regulations in these other areas, as might new portals should they come to rival Google, Facebook, and Twitter.
Courts have deemed disclosure as the safest constitutional form of regulation. Many forget that even the Citizens United decision itself upheld statutory requirements of forced disclosure of corporate and union spending. Indeed, when it comes to disclosure of contributions to candidates, parties, and PACs, and expenditures by those entities, disclosure on the FEC website is robust and exceedingly specific.
However, as any watcher of television ads knows, the disclaimers of spending accompanying advertisements sometimes take the form of “Paid for by Americans for America.” The actual identity of the spender is often cloaked behind a pleasant sounding, patriotic name, and the main donors to such organizations are often difficult to discover. Such spending is, in a literal sense, unaccountable; neither the media nor the voters can hold the speakers to account for substance or tone.
The internet portals could require more robust disclosure than the law currently mandates. They could ensure that the principal donors to any organization paying for ad space must be available to anyone who clicks on the ad or provide for greater detail in the ad itself. Moreover, because they do not need to worry about concerns of vagueness and overbreadth, they could (if they wanted to, and I am not endorsing this) require disclosure for such paid communications even if they occur far out in time from an election and for communications that might have an effect on elections even though they do not expressly advocate for the election or defeat of candidates.
Campaign finance law not only regulates what can be said, and how, but also who can say it. Different restrictions apply to candidates, parties, and a variety of outside groups. The Federal Election Campaign Act (FECA) limited individual expenditures, until the Supreme Court struck down those provisions in 1976 as infringing on the First Amendment. Until Citizens United, of course, corporations and unions were banned by the Bipartisan Campaign Reform Act (BCRA) from using their treasury funds to pay for ads that referred to candidates. Even after that case, however, the law continues to ban foreign nationals from spending money in connection with any federal, state, or local election. Both campaign finance law and tax law also place certain other restrictions, such as disclosure obligations, on different types of entities.
If the portals require advertisers to reveal who is actually paying for the ad, they would be able to enforce restrictions on the sources and amounts of expenditures. Although it would not be in their economic self-interest, they could re-enact, in a sense, the original FECA limits on how much individuals could spend or the BCRA’s absolute prohibitions on expenditures of corporate and union treasury money. They could go even further, for example, and treat candidate and party ads differently than those of SuperPACs and other outside groups. (They might want to do that in order to privilege, for example, the speech of accountable actors in the system over less accountable actors.) At a minimum, they could enforce the existing prohibition on foreign spending in U.S. elections.
None of this seems especially likely, except perhaps enforcing a ban on advertising by foreign nationals. However, it gives a sense of how the portals, on their own, could achieve the aims of statutes the Court has ruled unconstitutional. As explained later, it also might make sense for the portals to enable different levels of targeting for different actors, based on the theory that individual-level campaign communication should be most available to the candidates running for the office.
Disclosure requirements or limits on the sources and amounts of campaign expenditures are the most familiar tools of campaign finance regulation. However, the government regulates campaign communication even apart from its financing. The content of advertisements can also run into certain restrictions based on its misrepresentation, unfairness or tone. For commercial ads, the platforms (and for that matter, the Federal Trade Commission) have rules against misrepresentation of products and services. Many of the same policies, on their face, might apply to political ads. Google, for example, prohibits ads that make “false statements about your identity or qualifications”, “falsely imply affiliation with, or endorsement by, another individual, organization, product, or service”, or “conceal or misstate information about the business, product, or service.” Facebook, likewise, prohibits “deceptive, false, or misleading content, including deceptive claims, offers, or business practices.”
Several states also have laws preventing misleading campaign statements. These laws have been successfully challenged, on occasion, but the U.S. Supreme Court has not weighed in definitively on their constitutionality. Given the First Amendment’s protection of political speech, and especially campaign speech, these laws (at a minimum) would be read narrowly, if they are constitutional at all. But the internet platforms are not similarly constrained. Of course, evaluating “political truth” is a fraught business—and perhaps one that is simply not worth the effort if the costs of enforcement, the damage to a portal’s reputation, or even the chilling effect on robust communication prove to be too high. But the platforms can, if they so desire, develop more sophisticated and indeed, restrictive, rules for misleading statements than ones that would be permitted by the First Amendment.
This would require the development of bipartisan or nonpartisan decision-making boards that can quickly evaluate claims once an objection has been lodged. Once such an institution is in place, we should, of course, expect the normal charges of bias and unfairness to be lodged against those institutions, just as occurs with other fact-checking organizations, such as PolitiFact and Factcheck.org. But just as these firms police truth in the commercial sphere, so too they could police it in the political sphere, perhaps only at the margins when the statements are demonstrably false. Even that might have a chilling effect on some true statements that push the envelope but do not break it. However, the platforms can raise the bar on the truthfulness of claims made in paid ads, deciding to leave the borderline claims to other spaces on the internet or to unpaid communication.
As hard as it might be to evaluate the truthfulness of online communications, regulating their tone might prove even more difficult. But these portals already maintain some restrictions on the tone and content of ads, even if not described as such. Twitter, for example, prohibits ads with “inflammatory content [that] is likely to evoke a strong negative reaction or cause harm.” Facebook prohibits “shocking, sensational, disrespectful” content from its ads. In addition to similar restrictions, Google prohibits ads that have “content that may be deemed as capitalizing on or lacking reasonable sensitivity towards a natural disaster, conflict, death, or other tragic event.” All such standards would be voided for vagueness by courts under the First Amendment if passed by the state.
Given the capacious language and the wide discretion of these portals’ guidelines for censoring ads, it is unclear whether additional authority would be necessary to regulate negative, personal attack ads, for example. Some refinement as to what might constitute “disrespect” or “evoke a strong negative reaction” in the context of a political campaign is in order. YouTube provides some unhelpful, but telling, specificity in its Community Guidelines: “We’re not asking for the kind of respect reserved for nuns, the elderly, and brain surgeons. Just don’t abuse the site.”
Regulating tone and preventing misrepresentation is somewhat different than promoting fairness in on-line campaigning. Regulating communication in order to prevent lopsided political debate has a long pedigree in American law, even if it has gone out of fashion in recent years. Through policies such as the “Fairness Doctrine” and “Equal Time Rule” the Federal Communication Commission has endeavored, at times, to ensure that debate on matters of public concern on television or radio is not slanted toward one point of view. Because the airwaves are a public utility of sorts, such government action was thought to be warranted to ensure, in particular, that networks did not feed their viewers only one point of view or allow one candidate to have a platform unavailable to her opponent.
One can envision several analogs to these rules that internet portals could apply to themselves. First, to avoid the “digital gerrymandering” described above, the portals should adopt policies that provide for equal access regardless of partisan affiliation. In other words, Republican candidates and groups ought to be able to buy ads on the same terms as Democrats. This has not proven to be a problem, and perhaps might never be, but a commitment to nonpartisanship would be an important signal to send to the actors in the political ad marketplace.
However, if the portals wish to get serious about the polarizing impact of political advertising and the ideological echo chambers of the internet, then they might think about how alternative points of view might be presented alongside paid advertisements for one party or candidate. (Cass Sunstein made a similar proposal fifteen years ago to encourage political news websites, in general, to force their readers to engage with opposing points of view.) Of course, any such strategy might defeat the business model for such ads: Who would pay for an ad if by doing so one also pays for an opponent’s ad or links to his website, for example? Therefore, any such standard or practice would still need to be sufficiently beneficial for the purchaser to buy the ad, despite the fact that an alternative point of view would somehow be available to the same viewer.
This proposal may seem unrealistic and undesirable, and perhaps it is. But a modified version might be more palatable to both the platforms and the advertisers. If many of the dangers of online campaign communication derive, in particular, from the degree of targeting these new modes enable, then a deal might be struck between the degree of targeting allowed and the obligation to abide by certain norms of fairness and tone. In other words, as the risks posed by the unique communication capabilities of the internet increase, the greater the obligation to avoid echo chambers becomes.
Until this past summer, neither Google nor Facebook allowed advertisers to target individuals with a specific communication. Candidates and parties could not give a voter file to the platforms and request, for example, that a certain advertisement appear on the user’s laptop or phone while they were using the relevant site. Unless the candidates had a preexisting relationship with the voter, such as knowing their email address or having taken some action (such as giving money) on a candidate website, the campaign was not able to use the platforms to send unsolicited, individually targeted communications. For the most part, candidates and parties had to work through third party companies, such as Acxiom and Targetsmart, to achieve a second-best solution. The campaigns gave voter files to those companies. Those firms, in turn, reverse engineered an algorithm, based on certain demographic and geographic characteristics, which would capture as many people on the list as possible, as well as others with a similar profile. Such targeting was not at the individual level, but was likely to capture a group that was similar to the list of individuals provided
Everything has changed in the past few months. Both Facebook and Google now allow advertisers to target consumers individually. A campaign’s capacity to send individually tailored political messages is now only a function of the detail and richness of its voter dataset and its willingness to segment the political market to mobilize different voters with different messages.
At this critical turning point, it is important to consider policies that might mitigate the dangers of individual-level targeting, while also seizing the opportunities for reform that the internet enables. Reforms such as those described above, as to disclosure, misrepresentation, fairness, and tone, could be required in exchange for greater targeting. As messaging gets increasingly particularized, the firms could require that greater restrictions in the public interest would apply. Of course, any unique constraints on political advertising must be calibrated to the business interests of the portals seeking such advertising and the likelihood that such advertisers would not simply move to other avenues of communication. However, trading individual-level targeting for mild constraints on advertising would be a deal that, for most campaigns, is one with benefits that exceed the costs.
Fifteen years ago Larry Lessig put forth a theory that “code is law”, arguing that the architecture of the internet can regulate freedom in ways analogous to formal statutes. Quite soon in our political life, we will recognize that code, developed by private but powerful actors, will regulate electoral communication in ways similar to official government bodies. Despite the seemingly libertarian and even anarchic tendencies of the internet, this new world will still be a regulated one. Campaign speech will flourish on the internet, but the avenues for the most effective communication will still be subject to standards set by the private firms that police them, rather than the government.
The pressing question is whether the same standards these portals use for commercial advertising should be used for political advertising. In the fading age of broadcast television, the two were treated differently, and the same will likely be true in the internet age. The challenge is to match these standards to the new forms of microtargeted communication enabled by new technologies. The unique challenges posed by these new forms of communication are real, but private regulation, in many ways, may be better positioned to preserve democratic values than the government ever could. In this moment of transition, it is critical that we get this balance right.