Research

Job Market Paper

"The 'Gorilla in the Closet': Regulatory Enforcement Under Federalism" (with Romaine Campbell)

(Job Market Paper) In Progress

How does federal regulatory capacity affect state enforcement outcomes? We provide a model in which a stronger federal regulatory agency can either strengthen or weaken states' negotiating position with their regulated entities. The optimal federal enforcement for the states is one that maximizes state-level negotiated penalties. We apply the model's insights in the context of environmental regulation to test whether the US Environmental Protection Agency (EPA) is too strong or too lenient in two environmental programs: the Clean Air Act and Superfund. First, using an EPA database of state-issued penalties, we show that when EPA's budget was cut in 2011, state-issued penalties for Clean Air Act violations shrank by 15%. Second, using a dataset with information about environmental remediation projects under state jurisdiction, we show that firms are more likely to begin cleanup projects during Democratic federal administrations. Our remediation analysis identifies the mechanism: while firm cleanup behavior is affected by EPA strength, cleanups conducted by the state are not, providing evidence that the effects operate through firm-state bargaining. We conclude that over one third of EPA's effect on environmental penalties is through its spillovers to state enforcement outcomes, and that states would benefit from a stronger EPA.


Other Research

"The Welfare Effects of Eligibility Expansions: Theory and Evidence from SNAP" (with Charlie Rafkin)

Conditionally accepted at American Economic Journal: Economic Policy

Click here for survey instrument, and click here for literature review details.

We study the U.S. rollout of eligibility expansions in the Supplemental Nutrition Assistance Program. Using administrative data from the U.S. Department of Agriculture, we show that expanding eligibility raises enrollment among the inframarginal (always-eligible) population. Using an online experiment and an administrative survey, we find evidence that information frictions, rather than stigma, drive the new take-up. To interpret our findings, we develop a general model of the optimal eligibility threshold for welfare programs with incomplete take-up. Given our empirical results and certain modeling assumptions, the SNAP eligibility threshold is lower than optimal.

"Inefficient Social Disconnection" (with Amanda Pallais)

In Progress

Social disconnection has increased over the last 25 years, while traditional contexts for forming connections, like churches and in-person workplaces, have weakened. Through a field experiment, we show the effectiveness of an alternative approach to connection formation: mutual friend intermediaries. We hosted a four-week trivia competition where participants could either join friend-organized teams or register to be matched onto teams with strangers. We then randomized which teams could participate. When teams were organized by existing friends, teammates who weren’t initially friends were 20 pp more likely to be friends and 15 pp to be in text communication four months later due to participating, while participating reduced contemporaneous loneliness by 11 pp (24%). In contrast, the intervention did not generate lasting relationships for pairs matched onto teams by the competition. Yet, we propose mutual friends make inefficiently few introductions due to a market failure: it is difficult to compensate friends for organizing costs. Consistent with inefficient underprovision, survey evidence finds substantial untapped supply of friend introductions – 72% of people say they could connect friends, while only 5% recently have – alongside strong demand for such introductions. Finally, we discuss examples of organizations that encourage people to bring their networks together at scale, showing how the organizations best-positioned to scale this approach may not be those typically associated with community building

"Welfare Analyses of Firm-Based Government Policies" (with Valerie Chuang, Nathaniel Hendren, and Eric Zwick)

In Progress

"Superfund Sites and Cancer Incidence"

In Progress

Communities around state and federal Superfund sites are often concerned about increased risk of cancer from exposure to environmental contamination, but there is little causal evidence on this risk. Using data on public environmental remediation projects, I study how exposures to environmental contaminants relate to cancer incidence. Long disease latency periods mean that comparing outcomes before and after site cleanup is insufficient, so for identification, I borrow insights from the toxicology and epidemiology literatures on the specific diseases associated with each contaminant. I compare outcomes across sites, using only years before remediation begins, where the identifying variation is the specific contaminant involved in each site.