As the midterm elections approach, money is flowing into the coffers of candidates, parties, and outside groups at unprecedented rates. Meanwhile the Federal Election Commission (FEC) has, by many accounts, softened its enforcement of campaign finance laws. Clever political operatives and their lawyers stand ready to exploit the regulatory vacuum. This post focuses on one particularly inventive and consequential practice: sharing and selling political data.
The Cambridge Analytica scandal opened our eyes to the fact that personal information is valuable to campaigns. Cambridge Analytica used the information of over 80 million Facebook users to create advertisements, build models to predict behavior, and deliver the right messages to the right people, apparently without Facebook’s knowledge. Many organizations do not want the information of their customers and users exploited this way, but others do. Companies, campaigns, and parties have been collecting and exchanging personal information for years.
Super PACs create especially perverse incentives in this marketplace. To see why requires some background. Super PACs are political organizations that can raise and spend unlimited amounts of money to influence federal elections, so long as they do not coordinate expenditures with candidates. They cannot make contributions to candidates because contributions can lead to corruption (or the appearance of corruption). This means a super PAC cannot write a check for, say, $2,000 to the Trump 2020 campaign. What about data? Can a super PAC collect personal information about voters and then give it to a candidate? The answer is no. This would constitute an “in-kind” contribution. According to the FEC, mailing lists, polling data, and lists of activists constitute items of value subject to contribution rules. If these constitute contributions, other kinds of voter information must too.
This creates a dilemma for super PACs. Increasingly they possess data that would help their preferred candidates. For some super PACs, generating and sharing data with candidates is part of their purpose. Right to Rise gathered data on behalf of the Jeb Bush campaign, and Ready for Hillary PAC did the same work for Hillary Clinton. If a super PAC can’t give its valuable information to a campaign, what can it do?
Sell it. The law allows super PACs to sell their information to campaigns, as long as campaigns pay the “usual and normal charge.” The “usual and normal charge” is the “price of those goods in the market from which they ordinarily would have been purchased.” If a campaign receives a good “at a charge that is less than the usual and normal charge,” that discount is a contribution. In short, if a super PAC sells the information to a campaign at market price, it acts legally, but if it sells information below the market price, it makes an illegal contribution.
What is the market price of information? A nonpartisan company (like political data firm Aristotle) that collects information and sells to any interested customer, regardless of political affiliation, might generate a market price: what is the most anyone is willing to pay? But super PACs don’t work like this. They are political organizations looking to support particular candidates. They are not willing to sell to just anyone; they will entertain offers from only a small group of buyers. To illustrate this dynamic, the Trump re-election campaign is offering its database for rent to any “political group, org, nonprofit or business” that is “not hostile to the President” (emphasis added). For super PACs, then, the market isn’t exactly competitive. Many new super PACs exist to support only one candidate. If a super PAC is only willing to sell to one candidate, then the market isn’t competitive at all.
The FEC isn’t blind to this problem, but its solution is inadequate. The FEC requires that the data being sold have an “ascertainable fair market value” and that the parties involved engage in a “bona fide, arm’s-length transaction.” As with many areas of campaign finance, enforcement is a challenge. “Fair market value” is notoriously difficult to define, and one wonders how often transactions between candidates and super PACs supporting them are really “arm’s-length.” The FEC has declined to provide instructions on how one should value data, acknowledging in 2004 that “[r]easonable persons can disagree about how . . . to determine the value of [a] mailing list.”
The question of market value runs deeper yet. Suppose a super PAC agrees to sell data to a candidate for $100,000. If a for-profit, politically neutral firm would sell the same data to the same candidate for the same amount, then we might reason that $100,000 is the market price, so the super PAC’s sale should be permissible. But not so fast. A for-profit firm would pocket the $100,000, but the super PAC will not. It will take that $100,000 and spend it to support the candidate. Supporting the candidate, after all, is the super PAC’s purpose. Suppose the super PAC has good information about the campaign’s needs and strategies. Consequently, the super PAC can spend the $100,000 just as effectively as the campaign. Suddenly the transaction doesn’t look like a market value exchange. The super PAC didn’t give data to get $100,000. It gave data, got $100,000—and then essentially gave the $100,000 back. The data sale that the law permits is indistinguishable from a contribution that the law forbids.
A defender of the regulation might respond as follows: Super PACs cannot coordinate with campaigns. They cannot ask a candidate, “How would you like us to spend our money?” Instead, they have to act independently. Sometimes they have good information about the campaign’s needs and strategies, but other times they don’t. This is all true, but it doesn’t solve the problem. Suppose the super PAC has only a rough sense of how best to support the candidate. The super PAC sells the data for $100,000, spends the $100,000 to support the candidate, but because of its imperfect information conveys only $50,000 in value to the candidate. The same problem arises. The candidate appears to have paid $100,000 for the data, but in fact she paid $50,000—below market price. The super PAC effectively made a contribution of $50,000. The typical “independent expenditure” by a super PAC is already dubious, but data selling is even worse: the campaign gets the valuable data, and then the super PAC spends the campaign’s money in support of the campaign.
The rabbit hole continues to descend. Instead of selling data for money, super PACs and campaigns can trade data for data. The FEC allows candidates, parties, and super PACs to swap voter information so long as the databases are of “equal value.” The FEC has not provided clear guidance on how to ensure that this requirement is satisfied. Some organizations, like the non-profit Citizens United, use a commercial firm to rent their data to candidates. Presumably the firm ensures that buyers pay a reasonable price, but again, if the seller is only willing to sell to certain buyers, then prices might be distorted. In any case, the use of a broker to facilitate a swap might be the exception rather than the rule (see Marc Elias’s comments on p. 162). In the absence of an independent valuation, the parties to the exchange are the ones determining if the databases are of equal value.
Some lawyers, we are told, discourage their super PAC clients from swapping data for fear of violating coordination rules, but this practice is not universal. In the past, FEC commissioners have argued that the self-interest of long-standing political organizations like the NRA or the Sierra Club ensures that databases sell or trade at market value. For organizations like these, a database of donors, for example, is a valuable asset because it is a primary source of income, and the organization has an interest in ensuring its own sustainability. But single-cycle super PACs upend this conventional wisdom. They exist to benefit candidates this cycle, and their purpose is to maximize their benefit to candidates before disbanding after the election.
Reforms could close some of these loopholes. Currently, parties engaged in data swaps do not have to disclose any details to the FEC. Since data is valuable and swaps can involve illegal coordination, mandating disclosure seems like an obvious first step. If data is sold rather than swapped, FEC filings say only general things like “email list purchase.” Requiring parties to file a form describing the information being exchanged or sold could help keep candidates and super PACs honest.
Another reform would prohibit super PACs from selling data to or swapping data with a campaign unless a broker acts as an intermediary. A broker could appraise the data, which could at least mitigate concerns about discount prices. A more aggressive reform would entirely forbid sales or swaps between super PACs and campaigns. Without better regulation, data sales and exchanges erode the reasoning at the heart of Citizen’s United, that “independent expenditures . . . do not give rise to corruption or the appearance of corruption.”