Research Statement

Transparency and accountability are at the core of our democracy. "The people" cannot have political equality if they cannot see what the government is doing. However, more transparency is not always better. Burdensome transparency requirements can slow implementation of important policies.

Agency transparency is particularly important. The federal government is comprised of hundreds of agencies and bureaus with millions of employees implementing thousands of discrete policies, yet voters have few avenues with which to hold them accountable. Agency transparency, which enables accountability, is fascinating, substantively important, and under-studied.

By contrast to agency transparency, legislative transparency is minimal at best. Many of my current projects engage the data-rich environment of campaign finance, exploring transparency among aspiring (and incumbent) legislators as they run for office. How does campaign finance transparency affect voters and donors? The rest of this statement contains brief discussions of my current and future work on campaign finance transparency and agency transparency and accountability.


My work in campaign finance transparency so far uses large-N quantitative methods to measure the “chilling” effect of campaign finance disclosure requirements, and mixed methods to estimate the informational benefits to the public of learning about campaign finance information and candidate compliance with campaign finance regulations. Under it all is a core separation of powers issue that exists for every campaign finance enforcement body, and is interesting to me as an administrative law professor. Campaign finance regulators oversees individual elected officials running as incumbents, and those same members hold the regulator’s purse strings.

Campaign Finance Compliance

The inter-branch dynamic comes into play in my paper with Christian Grose, Campaign Finance Transparency Affects Legislators' Outcomes and Behaviors. We examine one particularly interesting moment in the history of the Federal Election Commission. Shortly after the FEC was created, agency accountants decided to conduct random audits on campaign finance compliance for 10% of house seats. We interviewed the retired chief auditor, who participated in the audits, and gathered the audit data from the archives. Our results suggest that audited incumbents performed worse at the polls in the next election, particularly if the audits uncovered violations, compared to incumbents not subject to audit. The effect is large: five percentage points among audited violators. The audits also improved legislative attention to the districts. Audited candidates who knew that the auditors would be reporting compliance issues back to the FEC made more trips home before the next election. Awkward inter-branch issue aside, random audits seem, on net, to have improved legislative behavior. This paper makes a large contribution to several literatures. First, we inform the literature on scandal, which has long downplayed the importance of campaign finance scandals to voters. Prior studies on scandals have not had the benefit of randomization, as we do here, and so we are able to show that actually, voters do care about campaign finance scandals. Second, we contribute to the literature on congressional elections and congressional behavior by showing that audits from a regulator can improve attention to the legislator’s home district. Finally, we believe there is an interesting compliance story here, aside from the campaign finance story. If we broaden the interpretation of the scandal studies to be about compliance – some incumbents failed to comply with regulations, like campaign finance regulations, others failed to comply with common norms, like not cheating on one’s spouse – we observe that these random audits alerted voters to compliance problems with their incumbent legislators, and the voters reacted accordingly.

In my work in campaign finance compliance, it has become clear to me that the compliance literature and the congressional and campaign finance literatures are not in conversation and should be. I am currently collecting data on another compliance-related project, called You Saw it Here First. It is a descriptive project tracking the subsequent careers of candidates for city council or mayor in New York. I ask, "If a candidate had campaign finance violations as a city council or mayoral candidate, are they more likely to have subsequent problems with legal compliance in their future careers?" My theory is that campaign finance violations are a type of compliance problem, and they should signal to voters that candidates may have trouble with the clerk-like, compliance tasks inherent to public work, or that they may be corrupt. New York politics has had dozens of corruption scandals in the past decade, so this is a fun data collection so far. If the association exists, then one might argue that campaign finance disclosures, when accompanied by audits, provide information about candidate compliance in addition to the policy-related information benefit the Court emphasizes in Buckley v. Valeo.

Costs and Benefits of Campaign Finance Transparency

My projects focus on the costs and benefits of disclosure. For over forty years, the courts have assumed that both costs and benefits exist, but they have not cited anything more than "anecdata" to support these claims. Doug Spencer and I have attempted to quantify the costs of campaign finance disclosure using large observational datasets from all fifty states (In the Shadows of Sunlight) and with Spencer and Chris Elmendorf, in a project examining the first election in Seattle featuring public finance in the form of vouchers (Mind the (Participation) Gap). We have not found evidence of anything beyond negligible "chilling" of contributions that would otherwise have been made.

However, observational studies require airtight identification to be convincing in terms of causal inference. Importantly, the costs of disclosure -- risk of harassment and loss of privacy -- are costs in name only, because in most jurisdictions, donors can avoid the costs entirely by giving to a so-called "dark money" organizations. Dark money presents selection challenges to scholars studying the effects of campaign finance observationally, as I explain in my recent Annual Review piece, Campaign Finance Disclosure.

Because identifying a causal relationship to estimate the costs of disclosure is difficult given the current legal framework, in another strand of my research, I measure the costs of disclosure in an experimental setting. Elmendorf and I will field a survey experiment in 2018 to determine whether a chilling effect exists, and we have designed it in such a way to also measure a thawing effect. What is a thawing effect? People who are thawed by disclosure are giving because it is a publicly visible act. Without disclosure, they would not give. We suspect that the 2016 campaigns detected a thawing effect among politically active voters. The net effect of chilling and thawing is something researchers have not yet analyzed, but it should further skew the donor pool in ways that have real-world effects on representation.

The benefits of campaign finance disclosure are probably broader than the Court conceives. In addition to You Saw it Here First, I have three early-stage projects exploring the nature of these benefits. In To See or Not to See? I present results from a survey experiment showing that, once we control for political engagement, people who choose to view campaign finance information vote more "competently" than people who are equally engaged but did not have access to the campaign finance information. Another early-stage project, Show Me the Money, builds on existing draft and experiment that shows that voters punish candidates supported by dark money and reward those who “overdisclose”. In the new data I'll gather this fall, using a conjoint experiment, I give the respondents more information and explore how much voters are willing to trade off on their preferred policies in order to elect more transparent candidates. I suspect voters will not compromise on major policy issues, but that they might on minor policy issues. I also expect inter- and intra-party differences in terms of willingness to reward and punish candidate choices around transparency. Finally, I will analyze whether voters treat a candidate's chosen level of campaign finance disclosure (mere compliance with the law, over- or under-compliance) as a valence characteristic or as a partisan issue. Elites are divided on whether we need more or less campaign finance disclosure (a partisan issue), but voters consistently say they want more of it (a valence issue), so I will test competing hypotheses.

Campaign Finance: Regulatory Gaps and Microtargeting

This is a heady time for campaign finance disclosure research. As the public has understood the mechanisms by which dark money emerges, and its effects in everything from advertising on television to disinformation on social media, my work has gained public attention. Turnover at the Supreme Court has replaced (or threatens to replace) disclosure optimists (Scalia, Kennedy) with disclosure skeptics (Gorsuch, Kavanaugh), and as a result, I suspect this line of research will continue to draw interest for years to come. Doug Spencer and I saw the “dark money” problem coming when we started working on the data analysis for Citizens United, States Divided in the wake of the decision. In that project, we compared states that were “treated” by the holding in Citizens United that allowed unlimited corporate and union independent expenditures to states that already allowed unlimited corporate and union independent expenditures. We find that while independent expenditures increased in both treated and control states between 2006 and 2010, the increase was more than twice as large in the treated states and nearly all of the new money was funneled through nonprofit organizations and political committees where weak disclosure laws and practices protected the anonymity of the spenders.

While I do not often write about “breaking news” type events, a fruitful conversation with former FEC chair Ann Ravel resulted in a law review piece about disinformation on social media that explores the fine-grained institutional details at the FEC that allow for and were unable to stop the flood of disinformation in the 2016 election, which continues today. All of them are tied to campaign finance disclosure and disclaimer requirements. Our most urgent and important recommendation concerns counter-speech. In the Alvarez case, the court ruled that the remedy for false political speech is not censorship, but rather counter-speech. Online advertising is attractive to candidates for several reasons, one of which is microtargeting. Microtargeting is almost impossible on television and radio, which reach broad swaths of the public. Moreover, when a candidate uses television or radio to reach the public, the station discloses information about the ad purchase to the government (and public). That allows for counter-speech by her opponents. The same is not true online. On social media platforms, advertisers can reach very small segments of the voting population, and there is no disclosure of the targeted audience required, frustrating opponents’ attempts at counter-speech. The platforms’ current efforts at transparency do not solve this fundamental problem.

Microtargeting is possible because of individually-identifiable voter information. What if candidates were deprived of it? Chris Elmendorf and I think that several interesting and democratically-beneficial outcomes would result. In Elite Political Ignorance, we explore solutions to the well-documented gap between what elected officials think their constituents want and what their constituents actually want. The solution may be less, rather than more, individually-identifiable information about voters. We propose several interventions to deprive candidates of the ability to target voters for turnout or suppression. One particularly interesting proposal is a publicly funded voucher program with a carefully-designed disclosure framework to help candidates and representatives better understand, in real time, what policies their districts prefer. The Seattle voucher program that we study in Mind the Gap (mentioned above), is the first voucher program of its kind in the United States. It requires full disclosure of donors and recipients. We argue that tying vouchers to a district but not an individual would be a better approach.


I am riveted by bureaucratic transparency and accountability. I could spend the rest of my career studying FOIA, open meetings requirements, and notice and comment. The bureaucracy is big and voter attention is limited. Moreover, voters' ability to hold bureaucrats accountable is limited and usually happens one of two ways: through their Member of Congress, or via presidential election. Because it is so hard to hold agencies accountable, our ability to extract information from the agencies is paramount. Yet for decades, presidents have layered in responsive appointees at the top of agencies, a politicization that may affect agency responsiveness to voters.

One potential consequence of the increased layers of political appointees may be decreased responsiveness to FOIA requests. FOIA can be politicized in other ways, as well, like the agency decision to locate the FOIA office in the secretariat, rather than, say, the management or information office. David Lewis and I studied this phenomenon in a quasi-experimental audit study in our paper Agency Performance Challenges and Agency Politicization. We sent two FOIA requests to 126 agencies and timed agency responsiveness to our requests. One was typical (we requested the FOIA log) and one was more sensitive (we asked about political appointee involvement in FOIA matters). We found that politicized agencies are slower to respond to requests, even when we controlled for agency size and workload. We did not detect quality differences among those who responded. Rather, it seemed the offices were not performing well, either because political actors focus more on other agency activities or because of poorer management agency-wide.

If agencies underperform, they are likely to face a congressional oversight hearing. However, not all agency actions trigger oversight hearings. Congress is busy and will therefore only hold oversight hearings under certain conditions. Janna Rezaee, Sean Gailmard, and I are studying the conditions under which agencies are more likely to face hearings. Analyzing decades of oversight hearings, preliminary results from our project Bureaucratic Agency Problems and Legislative Oversight, suggest that oversight hearings are more likely to occur when the committees and agency have policy conflict, when the agency is particularly active in policymaking, and when the house and senate committees experience policy conflict (as under divided government).

I plan to continue my work studying bureaucratic transparency and accountability in the coming years. Our elected officials, who oversee the bureaucracy, have limited attention and hundreds of agencies and bureaus to oversee. Formal procedures bring transparency and paper trails for congressional investigations, so I believe that agencies may successfully obscure malfeasance through using informal procedures. Informal procedures, especially informal adjudication, have few publication and reason-giving requirements, and many agencies provide limited avenues to contest informal decisions. While we have limited data on cross agency use of informal adjudications, it is commonly asserted that agencies use informal adjudication much more than any other procedure. Data availability is so limited due to weak disclosure requirements on informal adjudicatory procedures, and probably as a result, informal adjudication is under-studied. It is ripe for future study, and I’m just the person to do it – I love tracking down “impossible” data and interviewing bureaucrats, skills that will serve me well in researching this phenomenon.