Research

Manuscript and Supplementary Material

Abstract: To what extent does variation in access to local governance explain partisan sorting in the United States? Prior research shows that voters do not sort into neighborhoods based on partisan concerns alone. However, political attitudes are related to many residential preferences such as tastes for local public goods, making partisanship alone a weak treatment. To overcome this limitation, I exploit discontinuous changes in the provision of municipal public services at the institutional boundaries between incorporated and unincorporated areas. Examining the near-universe of U.S. cities, I document a decline in electoral support for Democrats in areas just outside city limits. The gap in partisanship is particularly prominent outside major metropolitan areas but is less detectable in states where the contrast in local governance between cities and their outskirts is weaker. Using geolocated voter registration data, I show that partisan sorting is related not only to a city's demographic and socioeconomic characteristics but also to its taxation and expenditures. These findings shed new light on how the contours of local political institutions inform geographic partisan sorting.

Replication Data

Abstract: In theory, a country will impose tariff barriers to protect the domestic industries and firms that are less competitive relative to foreign imports. This study investigates whether Revealed Comparative Advantage (RCA) is related to tariff protection in the three large North American economies. I find little evidence for the hypothesis that higher RCA values always correspond to lower tariff levels. The effects of RCA on tariffs are heterogeneous across sectors; consumer goods are likelier to see higher tariffs as RCA increases than agricultural or other goods. These results challenge the theory that export-competitive goods will necessarily receive less tariff protection.

Abstract: Many accounts of distributive and pork-barrel politics implicitly assume that incumbents have full discretion over which constituencies receive funding. However, many government resources are allocated via a competitive grant process. If would-be recipients such as local governments are deterred from applying for funding based on partisan concerns, then the distribution of grant winners can be naturally skewed toward the incumbent's allies even if the downstream grant award process is politically neutral. I test this theory using three long-running infrastructure grant programs in the United States that publish complete lists of applicants as well as winners. I find evidence in favor of an endogenously formed applicant pool. In two of the three cases, partisan alignment is positively associated with a locality's propensity to apply for a grant. This association is large for places with lower baseline administrative capacity but attenuates as capacity increases. Using a factorial difference-in-differences design, I show that applications to a prominent federal transportation grant can be sensitive to a change in the president's party. Joe Biden's 2021 inauguration was followed by a relative increase in applications from Democratic-leaning localities compared to Republican-leaning ones. This paper contributes to the distributive politics and fiscal federalism literatures by emphasizing the places that select into distributive politics rather than the ones that simply receive funding.

Abstract: We investigate the increase over time in out-of-district congressional campaign contributions by individual donors relative to contributions in their home districts. After confirming secular trends in the relative incidence of these kinds of contributions, we demonstrate that the pattern holds for both small and large donors, but is considerably muted for the latter. Next, we quantify the extent to which these secular trends are driven by changes in the behavior of recurrent individual donors versus replacement of more locally-focused donors by more nationally-focused ones. Our analysis indicates that both mechanisms play roughly equivalent roles. Finally, we examine the extent to which shifts in the electoral security of a contributor’s member of Congress induce changes in national focus, and whether any such relationship has itself evolved over the past four decades; we recover only modest effects. The conjunction of our findings findings hints at adjustment by individual donors in response to changing macropolitical realities, and only secondarily to changes in donors’ local political environments.

Abstract: Can cities improve public safety in mass-transit systems through durable, non-personnel infrastructure investments? While policing and temporary enforcement surges can reduce crime, they can also generate distributional costs and undermine state legitimacy. It remains unclear whether permanent changes to the built environment can deliver comparable safety gains in enclosed, transient settings where anonymity is high. We argue that visible, permanent improvements to station conditions deter opportunistic offenses by increasing natural surveillance, reducing concealment, and raising perceived detection risk, without relying on additional discretionary enforcement. We test this argument using the staggered rollout of New York City's Metropolitan Transportation Authority Re-NEW-vation Program across the subway network (2022--2026). With a station-by-quarter panel and administrative crime complaints on subway premises, we estimate causal effects using complementary difference-in-differences designs, including two-way fixed effects and specifications that allow for cohort heterogeneity and treatment dynamics. Across estimators, we find consistent evidence that upgrades are followed by substantial declines in station-level crime complaints. Our project speaks to how cities can co-produce public safety through non-personnel environmental interventions.