Explore the vast range of titles available.

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.

Image recognition systems offer the promise to learn from images at scale without requiring expert knowledge. However, past research suggests that machine learning systems often produce biased output. In this article, we evaluate potential gender biases of commercial image recognition platforms using photographs of U.S. members of Congress and a large number of Twitter images posted by these politicians. Our crowdsourced validation shows that commercial image recognition systems can produce labels that are correct and biased at the same time as they selectively report a subset of many possible true labels. We find that images of women received three times more annotations related to physical appearance. Moreover, women in images are recognized at substantially lower rates in comparison with men. We discuss how encoded biases such as these affect the visibility of women, reinforce harmful gender stereotypes, and limit the validity of the insights that can be gathered from such data.

Since 2010, independent expenditures have grown as a source of spending in American elections. A large and growing portion comes from “dark money” groups—political nonprofits whose terms of incorporation allow them to partially obscure their sources of income. I develop a new dataset of about 2,350,000 tax documents released by the IRS and use it to test a new theory of political spending. I posit that pathways for anonymous giving allowed interest groups to form new networks and create new pathways for money into candidate races apart from established political parties. Akin to networked party organizations discovered by other scholars, these dark money networks channel money from central hubs to peripheral electioneering groups. I further show that accounting for these dark money networks makes previously peripheral nodes more important to the larger network and diminishes the primacy of party affiliated organizations in funneling money into candidate races.

Why is housing supply so severely restricted in US cities and suburbs? Urban economists offer two primary hypotheses: homeowner self-interest and political fragmentation. Homeowners, who outnumber and have organizational advantages over renters, are said to lobby against development to protect their property values. The fragmentation hypothesis emphasizes that development's negative externalities are borne locally while most of the benefits accrue regionally or nationally, leading localities to block housing. This paper offers another explanation: ordinary people simply do not believe that adding more housing to the regional stock would reduce housing prices. Across three original surveys of urban and suburban residents, only a minority of respondents say that a large, positive, regional housing supply shock would reduce prices or rents. These beliefs are weakly held and unstable (suggesting people have given the issue little thought), but respondents do have stable views about who is to blame for high housing prices: developers and landlords. Large, bipartisan supermajorities support price controls, demand subsidies, and restrictions on putative bad actors, policies which they believe would be more effective than supply liberalization for widespread affordability. We discuss the implications of these findings for efforts to expand the supply of housing.

Can politically polarizing events bear dividends for extremist lawmakers? Evidence from California legislative financial disclosures suggests they can. During the state’s numerous budget shutdowns of the last 30 years, extremist legislators outside their party median could expect greater fund-raising hauls than their more centrist counterparts. The results suggest that polarizing events such as California’s perennial budget impasses can make extremist positions more appealing to the polarized political elites who generally fund political campaigns. Regardless of the motivation, however, these results suggest a strong incentive to prolong political discord by extremists—a troubling outcome in cases where supermajority votes are required.

Communication is a fundamental step in the process of political representation, and an influential stream of research hypothesizes that male and female politicians talk to their constituents in very different ways. To build the broad dataset necessary for this analysis, we harness the massive trove of communication by American politicians through Twitter. We adopt a supervised learning approach that begins with the hand coding of over 10,000 tweets and then use these to train machine learning algorithms to categorize the full corpus of over three million tweets sent by the lower house state legislators who were serving in the summer of 2017. Our results provide insights into politicians’ behavior and the consequence of women’s underrepresentation on what voters learn about legislative activity.

This dissertation, broadly, focuses on how the ability to make political donations anonymously changed American politics. Culminating the 2010 Citizens United v. FEC decision, the rise of nonprofit corporations as a conduit for campaign money means that a large portion of spending in American elections cannot be connected to any individual donors creating a system akin to the Australian ballot for money in politics. I explore how this change affected three facets of American politics; how donors behave when they can give anonymously, how being able to shield ones donors affects the type of candidates an interest group supports and finally how the legal green-light for nonprofits to spend in elections changed which nonprofit organizations became financially involved with each other.First, I developed a complete accounting of grants made between nonprofit organizations built from over 2 million digitized IRS forms made public in the summer of 2016 as the result of a lawsuit. My dissertation is not only the first project that examines these filings at scale, but also the first time this full network has been mapped. Using a network science algorithm that partitions the full graph into meaningful communities, I develop a theory of what I term dark parties or groups of nonprofits linked financially that make independent expenditures in Congressional elections.Next, I show that while dark money organizations form networks similar to those of traditional political parties, the types of candidates they prefer are vastly different. In a chapter of my dissertation, I show that these organizations prefer candidates farther from the ideological center and are especially active during the primary elections that traditional parties tend to eschew.Using a mixed-methods approach, I show that being able to give anonymously has important consequences not just for interest group behavior, but for donor behavior as well. I examine a list of donors that I uncovered from court filings to a nationally active dark money organization that spenton two ballot initiatives in California during the 2012 election. This list is the only publicly available list of dark money donors in circulation today and the first time such a list is studied by an academic researcher. I show that the donors to this organization, which supported two conservative positions, were much more liberal in their non-anonymous political giving than donors who gave transparently. This finding shows that the ability to obscure ones identity lets a donor behave differently than they would when their donations are subject to public scrutiny.Finally, while ample literature on the effects of disclosure exists, examinations into the motivations of why donors choose anonymity in their political giving remains unstudied. I present two survey experiments that seek to answer this question. First, I present survey results from the 2018 Cooperative Congressional Election Study that show that past giving to candidates from the opposite party that one normally supports correlates with an increased willingness to pay a premium to keep one’s political giving secret. Next, turning to another survey experiment, I find that potential voters are more likely to react negatively to an actual argument by opponents of a ballot measure when they know the names of the actual donors to a dark money group that opposed it. Combined, these results indicate both a social pressure rationale for obscuring one’s political giving and a strategic goal of distancing an electoral campaign from controversial donors.Taken as a whole, this research answers a broader question related to the balance of power between political parties and interest groups. Political parties perform a myriad of functions crucial to the maintenance of government that our democracy as presently conceived would be unthinkable without them. Despite their ubiquity, however, parties are notoriously hard to define. Parties exist beyond the formal structure of party officers and official state chapters, encompassing a myriad of outside actors who while not bearing the official stamp of the organization are crucial to its mission. The balance of power between these interest groups broadly defined and the formal party organizations are dictated by a myriad of factors–such as legal limitations, resource constraints and differing electoral goals.

The incomplete nature of legislation bestows on the executive branch the residual rights of control over implementation of public policy. The Trump Tax Bill of December 2017, which gave U.S. state governors a one-time opportunity to distribute a geographically-targeted federal tax incentive, provides a useful case-study to untangle the determinants of accountability. All 50 Governors were given the opportunity to designate census tracts within their state for preferential tax treatment. Within 120 days of passage, governors selected up to 25% of their eligible tracts, a short window that allows confident measurement of the political situation when the favor was distributed. We model a governors’ designation of tracts to maximize competing goals of mobilizing their voters, persuading swing voters, rewarding co-partisan legislators, and pursuing the programmatic goal of alleviating poverty. We then estimate the likelihood that an eligible tract is selected as a function of both the economic characteristics of the tract and the political characteristics of the governor and the relevant state and federal legislators. Our results suggest that the executive accountability engendered by eligibility for reelection is weakened by the dual constituency hypothesis, especially in cases where programmatic intent conflicts with the governor’s political motives.

Since the passage of the 1989 Roberti-Roos Act, the Golden State has been home to the nation's toughest gun laws, according to the Brady Campaign to Prevent Gun Violence.