Diagnosing Disclosure: A Social Scientific Perspective on Campaign Finance Disclosure

Introduction

Scholars across the social sciences have long been concerned with the outsized influence of the wealthy in the American political process. Recently, ambitious quantitative analyses have, in fact, shown that policy outcomes more closely reflect the preferences of the nation’s most affluent, especially when those preferences are distinct from those of the poor.[1] In this context, leading scholars have pointed to the system of campaign finance as a key mechanism of political inequality.[2] The tendency of the affluent to “vote with dollars” may mute the voices of average Americans and distort policy outcomes.

Despite the newly invigorated research agenda around political inequality and the role of money in it, our national elections continue to be financed almost exclusively by wealthy individuals. Recent Supreme Court cases like Citizens United v. FEC[3] and McCutcheon v. FEC[4] have only reinforced the perception that money buys political influence. Indeed, in a recent poll, a large majority of Americans—Democrats and Republicans alike—agreed that money has “too much influence” in the current system and that the wealthy have a greater say in elections.[5] Americans of all political stripes agree, too, that we need an overhaul of the campaign finance system, including greater transparency vis-à-vis the sources of donations made to outside spenders. In fact, the same percentage of Democrats and Republicans (76%) responded that outside groups should have to disclose the source of their funds.[6]

Although a large majority of Americans endorse stronger disclosure and believe that the campaign finance system must be reformed, we remain unable to answer very basic questions about the contours of who finances American elections. This is true, of course, for the so-called “dark money” sources that have flooded federal elections in recent years. But, it is also true of the enduringly largest share of money in the federal system—traditional “hard” money donations made to candidates and political committees from individual contributors. For instance, we have only rough estimates of how many unique individual donors actually make contributions (either itemized or unitemized) in each election cycle; we cannot readily identify those donors who donate the maximum amount to each candidate or committee; and we remain unable to trace donors across election cycles as they make multiple donations and, potentially, develop long-lasting relationships with candidates and members of Congress.[7] In other words, despite the now thirty-five year old federal disclosure system set in place by the Federal Election Campaign Act (FECA), we cannot readily identify who donates and how often or who the most important or influential players in this system might be.

Although the system of disclosure that currently governs the hard money system often wins praise even from critics of campaign finance regulations, I suggest in this essay that, at least from a social science perspective, even this system falls short of meeting some basic goals of transparency. Before describing how this system currently operates, I first suggest why we should even continue to care about hard money disclosure in an age of dark money, and offer a brief social scientific perspective on the study of money in politics. Having situated disclosure in this context, I then outline the contours of campaign finance disclosure in the federal system for individual contributors. I then suggest a few policy recommendations drawn from my work with Katherine Shaw.[8]

I. Why Disclosure?

In an age of “dark money” when candidates often fly across the country to meet with billionaires in an effort to secure multi-million dollar super-PAC donations, why care about the old hard money system at all? Although outside money affected nearly every House and Senate race in 2014, outside money has affected Congressional races unevenly.  Although in 28 races (about 6%) outside groups did spend more than all of the candidates themselves, in the remaining 443, outside money constituted only a fraction of the money raised and spent by candidates.[9]  This is, no doubt, an exceedingly significant development, although there is some discussion about how outside groups are related to the traditional channels of election finance.[10]  Nonetheless, for a significant portion of candidates and possibly more so for “safe” incumbents, the hard money system remains the most important source of campaign cash, even as outside sources have undoubtedly affected the electoral calculus for candidates of all types.  And within the hard money system, it is individual contributors that constitute the lion’s share of donations to candidates for Congress (see Figure 1).  In addition, improvements to this system, especially bite-sized ones like those I propose below, might simply be more politically feasible in an age of partisan polarization and ongoing stalemate at the FEC.[11]

Figure 1:  Individual and Political Action Committee Contributions to House and Senate Candidates as a Percentage of Total Candidate Receipts, 1984-2008[12]

Although an individual donation of $200 may pale in comparison to the multi-million dollar donations to super-PACs and non-profit groups that have garnered media attention in recent election cycles, itemized donations, in the aggregate, continue to be the lifeblood of the federal campaign finance system, especially for Congressional races.  This more social scientific understanding of political contributions—as aggregates with telling characteristics vis-à-vis both their sources and destinations—contrasts both with the Supreme Court’s conceptualization of political corruption in recent decisions and with many media accounts of the campaign finance system.

Consider, for instance, the idea of quid pro quo corruption or, in the Court’s words “dollars for political favors.”  This “bribery-like” form of corruption,[13] at least implicitly, invokes an individual or corporation as a singular actor making contributions to candidates and members of Congress for private gain.  Not only is this the sense in which the Court currently understands corruption, but it also often animates the public discourse on campaign finance spending.

Contrast these conceptions with the implicit understanding of political influence used by social scientists.  Social scientists model the impact of aggregate interest group contributions on roll call votes.[14] Or we estimate if, and how, aggregate candidate spending (across hundreds or thousands of candidates) predicts candidate success.[15] In all cases, we take for granted that the appropriate explanatory variable is the impact of an aggregate of thousands, or millions, of contributions or expenditures on political outcomes, whether these aggregates are internally coordinated or not.  But more than the implicit understanding embedded in this methodology, social scientists have explicitly theorized the relationship between the aggregate “active public” and the policymaking process.  In the words of political scientists Verba, Schlozman, and Brady:

When a group is active—especially one with identifiable, politically relevant characteristics—it becomes visible to an elected representative and is incorporated into his or her salient constituency.  Politicians attend to their constituencies and know who is watching what they do . . . .  Even in the absence of explicit directives—and constituents often do not send detailed messages—elected officials anticipate the needs and make inferences about the preferences of potentially active constituents.  Thus, it matters not only how participants differ from the non-participants in their opinions . . . but who there are.[16]

The authors methodically demonstrate how these active publics differ systematically from the inactive; and, again, later work has linked this participatory distortion with policy outcomes.[17]  “Salient constituencies” are embedded in the logic of the policymaking process, regardless of the intentions or coordination of the individual participants.  Indeed, even theories of oligarchy assert the primacy of a collective on outcomes, even if this collective is relatively small.[18]  (In another vocabulary, Lawrence Lessig conveys the logic of social science analysis with his concept of dependence corruption, albeit with a normative claim embedded.)[19]

The more simplistic, atomistic understanding of political corruption, however, often drives the current debate about campaign finance in general and disclosure reform in particular.  Calls for reform that focus on revealing big money donors to non-profits are no doubt important, especially given the Court’s tacit approval to bring these sources of donations into the orbit of campaign finance disclosure.  But, if we think in terms of aggregates, we might also do well to critically re-evaluate those aspects of the current regime that are of enduring importance.  Reforms in this arena may be more politically feasible, and they may even help to guide future efforts as “dark money” is brought into the sphere of disclosure.  In the next section, I use my own research on the campaign finance system to identify several systemic flaws with our current, entrenched disclosure system—flaws that, in fact, mirror an atomistic understanding of political corruption.  I then outline, in the last section, a few modest but high-impact proposals for reform that are, perhaps, more tractable in the current political climate.

II. Repurposing Disclosure

With the enduring importance of the hard money system in mind, I turn now to describing the current system of disclosure vis-à-vis individual donors—again, the largest source of hard money donations for federal candidates.  I suggest that this system may both reflect and reproduce atomistic conceptions of political influence.  To address these deficiencies of the disclosure data in my own work on campaign finance, I have repurposed federal disclosure records to make inferences about how individual contributors—in the aggregate—allocate their money in federal elections and, perhaps more importantly, how these patterns have changed over time.

Although federal law mandates disclosure (discussed below), the form of disclosure, again, is not readily amenable to social scientific analysis.  In a perfect world, we would need the sort of data that can tell us at least two things: who contributes at the level of the individual and how these contributors are situated in their social context.  The raw FEC records, however, are recorded and released per contribution—that is, it is exceedingly difficult, without significant time and effort, to identify unique individuals and their characteristics within these records.  Social scientists are increasingly interested in how money operates in politics, but we lack the data to answer basic questions about who finances American elections despite the massive amount of information currently offered on the FEC’s website.

In the federal hard money system, candidates and political committees must collect and then report the sources and sums of all contributions over $200.  For individual contributors, the donor must report his or her name, address, and occupation/‌employer. These disclosure records are then sent to the FEC, which checks and releases the records to the public.  The scale of this system is enormous—millions of transactions have been processed by the FEC in recent election cycles.[20]  These disclosure records are then made available to the public through the FEC’s website where users may download summary[21] and “detailed”[22] files, which contain the full population of contributions made in a particular election cycle (or as close as we can get to that population). Although for some players the burdens of compliance are high, the disclosure records are, in theory, a valuable, continuous account of one significant form of political participation.  There are few if any, other continuous records of American political participation (save, perhaps, voter registration records) that would allow such a complete portrait of how relatively elite Americans have “voted” with their dollars for nearly forty years.

Despite this, the limitations of the FEC records are numerous—so numerous, in fact, that I will focus here on just a few of what I see as the most pressing issues in the system of hard money disclosure.  I illustrate some of these issues in Table 1 with a representative sampling of the contributions made by the now very well-known super-PAC contributor, Sheldon Adelson. Mr. Adelson contributes from two separate addresses—one in Massachusetts and one in Nevada.  And his occupation varies over time.  Sometimes he identifies as a “multi-business owner”, other times as “self-employed”, and still others as an executive of the Venetian resort.  Such variation within contributors is, indeed, common within the records.

        Table 1:  Examples of Raw Disclosure Records from the FEC Individual Contribution File[23]

Perhaps as no deliberate method of evasion (although sometimes the patterns suggest otherwise), many of the affluent Americans who make itemized donations have multiple “occupations” and employers—business holdings, corporate directorships, as well as substantial occupational and residential mobility.  There are also a high number of missing values and effective “nonresponses” in this field.[24]  From a research perspective, these deficiencies make it exceedingly difficult to assign a likely occupation to unique individual contributors, to standardize occupations into broad industry classifications, and, ultimately, to track the contours of political giving by industries in the aggregate.

This within contributor variation might not be problematic except that the FEC does not assign something like a unique donor identification number.  The records also lack an identifier with more discriminating power (like date of birth) that would allow for a relatively easy way to identify unique individuals over time.  For these reasons, the process of identifying unique individuals within the records is exceedingly difficult.  And without a donor identification number, we are left to merely estimate the number of Americans who have made donations; how large those donations are; and how often or for how long donors “participate” in this way.

Instead, in order to track individual donors, the records must first undergo extensive cleaning and standardization.  Even with these transformations, it is often infeasible to link individual records using only exact matches.  Imagine a contributor who has given under his full name “Robert” and a nickname “Bob”.  An exact matching algorithm would fail to join Robert and Bob’s contributions into a group because, of course, “Robert” is not the same as “Bob”.  Recent advances in data science, however, have made possible another method—probabilistic record linkage.  Probabilistic record linkage algorithms identify matching records by comparing multiple matching variables and assigning an overall match score that quantifies the likelihood that the pair represents a true match.[25]

Although it is now possible to identify most unique individuals using such techniques (after, of course, the extensive pre-processing mentioned above), we might consider reforming disclosure to produce better disclosure data that can, at least, more easily answer basic, but pressing, questions about who finances American elections.  In the next section, I offer a few recommendations for small, but potentially far-reaching, improvements to the current system of disclosure based on social scientific understandings of survey design.

III. Reforming Disclosure

Absent a significant shift in the political climate, sweeping reforms of the disclosure system are unlikely. But there are smaller, perhaps less politically loaded, reforms that could be implemented to improve the disclosure system we already have and that the Supreme Court has approved.  At the same time, there is even some evidence of support for better disclosure at the otherwise highly partisan FEC.  The FEC has recently moved to modernize the campaign finance disclosure system with the introduction of a new campaign finance data portal.  Among its many laudable improvements, the site offers users a more visually appealing, interactive platform to search and filter disclosure data.  One of the most significant improvements to the previous data platform is the ability to quickly locate the sources of candidates’ funds, vis-à-vis the locations, employers, occupations, and sizes of contributions from individual donors.  In addition, the FEC has released an OpenFEC API to better facilitate the use and re-packaging of FEC disclosure data.

As promising as these recent efforts may be, they merely mask underlying and far-reaching problems in the disclosure system and in the way that disclosure data is collected.  Improving the quality of disclosure data—and the public’s consumption of it—ultimately requires improving the quality of the data collection mechanism.  As such, the improvements I suggest below could easily complement improvements to the data dissemination efforts by the FEC.  In the case of the third recommendation (unique donor identification numbers), Congressional action may be necessary before implementation. Taken together, these simple reforms might allow us to more easily make inferences about political influence in American elections.

A. Standardized Disclosure Forms

One obvious reform is to institute the use of standardized disclosure forms by all political candidate and committees.  At present, no standardized disclosure form exists to collect basic donor information; the FEC handbook simply instructs political candidates to collect the names, addresses, and occupation/‌employer of donors.[26]  But a standardized disclosure form—one that solicits the required pieces of disclosure information in a consistent and thoughtful manner—could make a big difference in ultimately generating cleaner, more usable disclosure data.

B. Improved Question Format

The way a disclosure form solicits the required information matters, too.  It can, in fact, matter a whole lot. Research has shown that just small changes to the way survey questions are asked can elicit very different responses.[27]  With this in mind, the adoption of a standardized disclosure form with survey questions that are both valid and reliable—for instance, like those pre-tested and fielded by the Census Bureau[28]—could also produce better, cleaner, more usable disclosure data.  More importantly, cleaner, standardized data would allow us to more easily make inferences about donors in the aggregate by minimizing the within-contributor noise illustrated above, as well as the high number of missing values and nonresponses.

C. Unique Identification Numbers

A more ambitious proposal would be to assign unique identification numbers to unique individuals who make itemized contributions.  This would allow academic researchers, journalists, informational intermediaries (like the Center for Responsive Politics or the Sunlight Foundation), and even the public to conduct basic analyses with the individual contributor as the unit of analysis.  For instance, we could then estimate, with greater certainty, the size of the mean and median donation per contributor; the most common patterns of participation in the donor pool; and how donor partisanship has changed over time.  It should be noted that this proposal would require donors to supply additional identifying information with their contributions (like, for instance, date of birth).  Without such information, it is very difficult (if not impossible) to disambiguate donors with the same name who have given to a variety of federal candidates and committees over time.  In other words, there is no ex post solution to this problem without modifying the data collection mechanism.

Conclusion

Although recent calls for reform have centered on bringing dark money into the orbit of disclosure, I suggest in this piece that there are good reasons to also critically re-evaluate the entrenched system of disclosure than governs the hard money system.  In the aggregate, at least, hard money donations remain a crucial source of campaign cash for candidates, especially in Congressional elections.  If we think that aggregate donations matter as a mechanism of political influence, then we would do well to ask how our current disclosure system could best shed light on these funding streams.