Public positions,private giving: Dark money and political donors in the Digital Age

Abstract

Dark money—campaign funds raised by 501(c)(4) designated non-profit corporations whose donors are exempt from disclosure—has become an increasingly large fraction of outside spending in American elections at both the state and the federal level. This paper makes use of the only publicly available donor list for a dark money group in existence today—that of “Americans for Job Security,” who contributed $11 million to two conservative-leaning ballot initiative campaigns in California during the 2012 elections. In comparing the ideological scores of donors of this dark money group to traditional donors to the two conservative propositions, I find a strong liberal tilt of donors to Americans for Job Security—indicating a social pressures motivation behind concealing one’s donation via a dark money group. These results also show disclosure laws have an effect on a donor’s calculus to contribute to a political cause.

Introduction

When the 2010 Citizens United v. FEC decision revolutionized the way American political campaigns were financed, the result was not just a sharp increase in the quantity of money in politics but a dramatic decrease in how much American voters could learn about it. In the election cycles that followed the decision, a new form of political spending began to emerge as a force in American elections. “Dark money,” political spending by IRS designated 501(c)(4) non-profit groups, exploded as a major funder of political campaigns at all levels of American government. These groups are allowed to spend money to influence an election, but do not have to disclose their donors with the same rigor traditional political groups such as PACs or parties are subjected to.

The new proliferation of anonymous money in American politics has important implications not only for theories of political participation but also for normative theories on campaign finance disclosure and its effect on the electoral process. Dark money brings new challenges to the established literature around both social influence in political participation and voters’ use of information shortcuts to inform their choices and tune their own preferences to those of a cause or candidate. Further, donating money to a campaign as a form of political engagement may be altered by the option of anonymous giving in order to avoid any of the social pressures once faced by voters in the days before the Australian Ballot.

This article takes advantage of a rare and unique disclosure in the opaque world of high-dollar campaign finance. Analyzing the only public list of dark money donors currently in existence, I utilize publicly available data to uncover the names of donors to a multi-million dollar campaign associated with the conservative position on two California ballot initiatives during the 2012 election. “Americans for Job Security”—a 501(c)(4) political non-profit affiliated with the Koch brothers’ network of independent expenditure groups (Novak and Maguire, 2013)—raised $11 million for the express purpose of defeating a tax-increase, (Proposition 30), and hindering union political fundraising abilities via a “paycheck protection” measure, (Proposition 32). Having uncovered the identities of dark money donors to the Americans for Job Security effort in California, I compare the ideological scores of these clandestine donors to traditional donors to both initiative campaigns and find donors giving to the dark money effort were significantly more liberal than those who gave to these conservative initiative campaigns through traditional, transparent channels. Like other studies of clandestine organizations (e.g. McMillan and Zoido, 2004; Skarbek, 2011), this happenstance disclosure of such a group’s inner-workings lets us assess the motivations of dark money financiers, whose combined efforts across all federal elections in 2012 at least $313 million.

The disconnect between the “public” pattern of giving when donating transparently and the “private” support for this seemingly incongruous ideological position speaks to many theories about both political behavior in social settings and the impact of disclosure on money in politics. Were a donor’s ideology the sole motivating factor that influenced one’s political giving, we would expect to see no divergence in the ideological profiles of transparent donors to these two initiative campaigns and the funders of Americans for Job Security. The disconnect between the public ideology of these dark money donors and their clandestine giving, however, indicates that other pressures beyond one’s own ideological convictions motivate political participation as expressed by political giving.

More broadly, the major overhaul of campaign finance law that took place following Supreme Court’s 2010 decision in Citizens United v. FEC have given rise to a new era of anonymous giving in American politics unseen since the Federal Election Campaign Act of 1974. Coupled with this are advances in technology that have made uncovering an individual’s political giving, and more importantly, publishing that information, rather trivial affairs. If dark money is a refuge for financiers of controversial political opinions (with the definition of “controversial” shifting radically between communities), then certain large swaths of political money are “going underground.”

Furthermore, this trend would have serious implications for latent-measures of ideology calculated from political giving. It is reasonable to assume that donors to dark money groups are more active and more sophisticated than the vast majority of those contributing in American elections. As such, the selection of some of these donors toward giving to non-disclosing entities may bias point estimates that rely on transparent giving to calculate spatial measures of a lawmaker’s ideological leanings—especially if those donors opting into anonymous giving are doing so in the face of cross-pressures from their social environments.

Campaign money, donors, and disclosure

The dilemma advanced by the advent of dark money lies at a fascinating intersection of the “social pressures” literature—the literature on disclosure and in the literature about donor motivations. Past research has examined the effect of social settings on political preference formation (Klar, 2014), political tolerance (Mutz, 2002), and self-reported voting behavior (Karp and Brockington, 2005). More recently, scholars have begun to explore how these social pressures might couple with disclosure laws and influence one’s decision to give money to a political cause or campaign.

At issue is that with the popularization of the internet, information about who gave money to a campaign has never been easier to access and disseminate. These advances in transparency may dissuade some donors from supporting a candidate or initiative because of the backlash they may find themselves subject to. For instance, donors to California’s Proposition 8, a same-sex marriage ban on the 2008 general election ballot that was ultimately successful, are an often-cited example of donors facing real economic consequences for their political giving. According to Johnson et al. (2010):

An enterprising and anonymous programmer mashed up the names and geographic locations of the donors with Google maps, producing www.eightmaps.com, a site where any visitor could see who in what neighborhoods contributed to the campaign. As a result many individual supporters were targeted with insults, threats and boycotts.

Studying the phenomenon with a survey experiment, La Raja (2014) found that individuals reported less willingness to contribute if their contributions were to be made public. Additionally, he found the effect especially acute for those who faced strong cross-pressures from those in their social group. These results mirror the conclusions of Carpenter et al. (2014), whose “failed” field experiment found a strong reluctance from Congressional candidates in 2012 to allow those researchers to inform voters as part of an experimental condition that their donations would be made public on the internet (though by law, all donations above the federal threshold must be disclosed online). However, using actual individual donation data at the state-level and exploiting variations in disclosure law, Wood and Spencer (2016) found a negligible effect on more rigorous disclosure laws and political giving. Importantly, the authors found scant evidence that those facing strong cross-pressures—as defined by those with a pattern of giving significantly distinct from that of their home ZIP code—were any less likely to donate in the face of enhanced disclosure.

Dark money introduces a major dilemma for those researching the effect of disclosure on political giving. As Wood (2018: 16) describes, “[i]n any race in which dark money is present, so, too, is the opportunity to select into a disclosure condition.” As such, making any causal estimates on the effect of disclosure on political giving using typical campaign finance datasets is fraught with peril as the types of persons giving publically and those either giving below the disclosure threshold (less than $200 for federal elections and varying amounts for state-level elections) or to a dark money group are two different types of people (Wood, 2018). Some contemporary accounts by journalists offer anecdotal evidence in support for this idea. For instance, in describing the foundation of “Crossroads GPS,” a 501(c)(4) created by GOP consultant Karl Rove to compliment his existing super PAC “American Crossroads,” Vogel (2014: 53) relates an interview with a ranking member of the organization who said “some donors didn’t want to be disclosed, and therefore a (c)4 was created.”

What then would motivate a donor—especially one of the more elite and munificent donors involved with an organization like Crossroads GPS—to opt for donating clandestinely? Using traditional campaign finance datasets does not allow for researchers to assess the motivations of those who chose both to donate and to do so anonymously because, by definition, those individuals remove themselves from observation. Indeed, the advent of dark money introduced a best-of-both-worlds option for donors wishing to give to a campaign but reticent to do so publicly. These elite donors are theoretically subject to the same cross-pressures as are the more modest contributors to American political campaigns. As such, rather than curtail their activities as the subjects in La Raja (2014) opted to do, they may opt for campaign finance vehicles that can allow them to give expressively (i.e. Ansolabehere et al., 2003; Francia et al., 2003), but avoid backlash.

Dark money

Coined in 2010 by Bill Allison of the Sunlight Foundation ¹ the term “dark money” refers to certain IRS designated non-profit groups that exist in a regulatory limbo opened by the 2007 Supreme Court case Wisconsin Right to Life v. FEC (see Eskridge et al., 2007 for a discussion of the case; and Beckel, 2016 for its implications on the spread of dark money). While this decision opened the tap for these 501(c)(4) groups (their designation in the tax code) to donate money to political causes, the flood was unleashed by the twin cases of 2010’s Citizens United v. FEC and Speech Now v. FEC. According to an analysis by the Center for Responsive Politics, a non-partisan think tank monitoring campaign finance, the amount of dark money spent in federal elections has been increasing exponentially since 2010 (see Figure 1), with spending in the 2014 midterms surpassing the dark money totals of both the 2012 congressional races and topping the 2010 midterms by about $30 million (Maguire, 2014). The 2016 election saw at least $178 million in dark money spending with most of it aimed at Congressional races and 76% of that amount coming from conservative groups (Maguire, 2016). These numbers, however, are merely an estimate as they represent only spending that falls within certain timeframe prior to an election that must be disclosed to the FEC. In actuality, the number is likely much higher.

Figure 1.

Estimated dark money spending. Source: CRP.

Proposition 30 and Proposition 32

As it had done for about the last 100 years, California sent a raft of initiatives to voters during the 2012 General Election. Among them were Proposition 30 and 32, a plan to raise taxes and a plan to curb the political spending of unions respectively. Both initiatives were hotly contested with about $190 million spent on both sides of each initiative campaign. ² Both propositions neatly cleaved California’s powerful special interest groups along ideological lines, with Democratic allied unions supporting Proposition 30 and opposing 32 and Republican business groups taking the opposite tack. Ultimately, voters passed the tax increase and defeated the union fundraising restrictions by wide margins in each.

Like many elections post Citizens United, California’s 2012 general election also saw an explosion of dark money spending. In particular, one $11 million donation to fund the opposition to Proposition 30 and facilitate passage of Proposition 32 caught the eye of California regulators. Just days before the November election date, the Fair Political Practices Commission, California’s equivalent of the Federal Elections Commission, sued “Americans for Responsible Leadership,” a dark money group based in Virginia, to force the group to turn over a list of donors (Barker and Meyer, 2014). While this donation was legal under federal elections laws revised following Citizens United, a California law passed in May of that year made donor disclosure mandatory if donors knew they would be contributing to a political cause (Megerian and Dolan, 2012). The case wound its way up the judicial totem pole until the California Supreme Court forced the group to hand over its records on November 4, 2012, just two days before voters were to go to the ballot box.

While the results of the lawsuit were unremarkable for average voters—the funders of Americans for Responsible Leadership, “turned out to be other non-profits, whose individual contributors remained secret,” according to the Los Angeles Times (Megerian and Dolan, 2012)—buried in the case file was a partially redacted list of 136 donations to “Americans for Job Security,” the dark money group where the funds spent in the California races originated (Figure 2). A dark money powerhouse nationwide, Americans for Job Security, a Virginia-based 501(c)(4) non-profit, made about $16 million worth of independent expenditures in federal elections during the 2012 cycle, according to the Center for Responsive Politics, making it the fourth most “active” dark money organization in that election. A representative from Americans for Job Security worked with two California political consultants to raise money through his organization, specifically noting in solicitations that donor identities would be “non-reportable,” according to court documents.

Figure 2.

A partial list of the Americans for Job Security donor list obtained via a Public Records Act request by the author to the California Fair Political Practices Commission.

I was able to obtain this list from the California Fair Political Practices Commission via a request under California’s Public Records Act in November of 2013. Court documents and reporting by news outlets such as the Los Angeles Times, the Huffington Post and the Center for Responsive Politics (see Blumenthal, 2013; Novak, 2013; York, 2013) indicate this is a full accounting of the donors giving money to Americans for Job Security for these two California campaigns. Further, as illustrated by Figure 3, most of these donations came from other California residents, lending further veracity to this claim.

Figure 3.

Dots indicate the zipcode listed for an AJS donor.

Decoding the names proved an arduous but ultimately achievable feat. First, I matched the provided zip code and partial street addresses (see Figure 2) with donor databases made available by the California Secretary of State’s Office and the Federal Elections Commission. Matching the first names provided in many of the cells from the Americans for Job Security list, I was able to find corresponding public records that included the donors full name (e.g. from home purchases, Securities and Exchange Commission filings or business websites). The result was 83 unique donations (the other 53 donations were too fully redacted to decipher the donor’s identity with any certainty) made by 66 unique donors (i.e. several donors gave money more than once).

In July 2016, Hedge Clippers (Hedgeclippers.org 2016)—a non-profit news site allied with several California unions and progressive organizations—published a more complete list that largely corresponded with my own, but included an additional contributor (66 in my analysis vs. 67 in the Hedge Clippers report). The results of the Hedge Clippers analysis were reported on by the Los Angeles Times and San Francisco Chronicle (see Blume et al., 2016; Gutierrez, 2016). As of this writing, these results have gone unchallenged in either the media or in court—lending further validity that the donors named were indeed true donors to Americans for Job Security.

Given that 38.97% of donations to Americans for Job Security were too redacted in my analysis to discern the identity of the giver, questions over selective censorship naturally arise. ³ However, as Figure 4 shows, the vast majority, some $23.4 million or about 83% of the $28.4 million raised by Americans for Job Security was able to be linked to a donor. Nonetheless, concerns about censorship and selection bias are addressed in the appendix.

Figure 4.

Size of AJS donations: redacted vs. non redacted.

After combining the identified donors from both my research and the Hedge Clippers report, their political leanings could then be discerned by matching their names to the Database of Ideology, Money in Politics and Elections (see Bonica, 2014). Bonica’s spatial ideological scores use campaign contributions to derive the ideological positioning of a donor based on which candidates that donor contributed to, making it ideal for this analysis. Matching based on names and zip codes produced ideological scores for all 67 unique individual donors.

While the anti-Proposition 30 campaign and the pro-Proposition 32 campaigns were distinctly conservative affairs, donors to Americans for Job Security spanned the ideological spectrum. A density plot of the ideology scores (see Figure 5) shows two distinct clusters within the donors to Americans for Job Security, indicating that donors to this dark money group were not ideological moderates, but distinct clusters of liberal and conservative donors. In Bonica’s coding of ideology, liberal donors are assigned ideological scores less than zero and conservatives the opposite.

Figure 5.

AJS donor ideologies vs. transparent donors. Vertical line indicates median value.

Using the same procedure to attach ideological scores to donors giving to Americans for Job Security, I then matched donor information to the No on Proposition 30 campaign and Yes on Proposition 32 campaign to Bonica’s ideology scores. Of these donors, 321 of 568 unique donors to the Yes on Proposition 32 campaign were matched with a Bonica score as were 2367 of 2493 donors to the No on Proposition 30 campaign. Viewing the density plot of all donor ideologies (Figure 5), the disconnect between Americans for Job Security and those donating in the open becomes even more evident.

Taking the analysis further, a two-tailed t-test shows a statistically significant difference in the ideological mean of the Americans for Job Security donors (M=0.101, SD=0.88) and disclosed donors to the No on Proposition 30 campaign (M=1.3, SD=0.433); $t(51.5)=9.81,p<.0001$ . Similarly, Americans for Job Security and disclosed donors to the Yes on Proposition 32 (M=0.943, SD=0.43) had statistically significant mean ideologies; $t(55)=6.76,p<.0001$ .

Further, the results from a two-sample Kolmogorov-Smirnov test allow us to reject the null hypothesis that those who gave to Americans for Job Security and those who gave openly come from the same population of donors in the case of the No on Proposition 30; $D=0.811,p<.0001$ ; and the Yes on Proposition 32 campaigns; $D=0.566,p<.0001$ . Concerns about whether this difference is motivated by a bias in the Bonica scores introduced by large donors who donate widely to ideologically disparate causes is addressed further in the supplementary materials.

Conclusion

Are dark money non-profits utilized to avoid backlash against donors for giving to causes antithetical to their publicly stated beliefs? In this case study of Americans for Job Security, the answer appears to be yes. Donors who in previous elections had given mostly to liberal candidates were able to funnel money into a resoundingly conservative cause without fear of discovery. While there were many conservatives giving to Americans for Job Security whose financial support for a campaign opposing a tax increase and hobbling unions’ political fundraising efforts would raise few eyebrows, the group also contained enough liberals whose participation may have exposed them personally to social backlash. While a perfect counter-factual is not available due to the intensely clandestine nature of dark money, only 637 donations were made to liberal candidates by transparent donors to the No on 30 and Yes on 32 campaigns between 2008 and 2012 out of a total 15,744 donations. Compare this to the 2,470 of a total 6,128 by the Americans for Job Security donors and a stark distinction emerges. While not every donor giving to Americans for Job Security generally funded liberal causes or candidates, a large enough proportion did in order to shift the mean to the left.

In this case study, liberal donors giving to these ardently conservative causes may have simultaneously feared backlash against their businesses or their reputations, while at the same time preferring the policy outcomes these two initiatives would bring. For these donors, dark money offered a convenient out. By donating behind a veil of secrecy, they were able to avoid the repercussions their giving might have brought if done through traditional and transparent channels. The result is a rolling back of the clock to the more opaque days of campaign finance disclosure—when donations were made more or less free of public scrutiny.

The rise of dark money also calls into question our spatial notions of ideology when it comes to political donors. Because we assume that donors donate honestly and that donations are sincere statements of ideological preference, one is able to use donors as bridges to compare the ideologies of politicians in various governing bodies (Bonica, 2014, 2018). However, what happens to this assumption when the public giving of donors varies drastically from the type of giving that goes on behind the veil of secrecy dark money groups afford? Furthermore, if the decision to give anonymously via a dark money group is biased systematically (e.g. only liberal donors give anonymously to conservative ballot initiatives), what does this do to the ideal point estimates that rely on political giving as their bridge points? If it is precisely the type of large-scale donors that would both donate to state and federal races that are choosing also to donate to dark money organizations, perhaps then the ideal point estimates that lean so heavily upon them are unreliable.

Dark money is by definition obscure and by design difficult to trace. The data for this study was only available due to a happy accident in poor redaction, not some sort of transparency push by either the California state or federal governments. While illuminating, we have no real way of knowing how the ideological composition of Americans for Job Security compares to the myriad other dark money groups that have operated since the 2012 election. This fact is troubling not only for the generalizability of these results, but also in the sense that a large percentage of spending in American elections can pass from donor to campaign and face absolutely zero scrutiny along the way.

Supplemental Material

Online_Appendix – Supplemental material for Public positions, private giving: Dark money and political donors in the Digital Age

Supplemental material, Online_Appendix for Public positions, private giving: Dark money and political donors in the Digital Age by Stan Oklobdzija in Research & Politics