Research

Do drug-money seizures influence police behavior? In this paper, I exploit variation from the 2015 suspension of payments under the Equitable Sharing program, a federal asset forfeiture initiative that sent monetary kickbacks to local law enforcement for policing drug crimes. Using a difference-in-differences design, I find that participating local agencies cut drug-related arrests in response to the suspension of adoptions payments by close to 5%. The decline scales with pre-suspension participation and is strongest for possession offenses and municipal police departments. Despite the decline in drug arrests, I do not find evidence that violent or property crime increased in participating jurisdictions after the suspension. Taken together, my results suggest that civil forfeiture revenue meaningfully expands drug enforcement effort, but with little identifiable impact on overall crime rates.

I study how employers interact with the workers with a criminal record and how those interactions change when employers are hit by adverse shocks. Using confidential person-level linkages between the Criminal Justice Administrative Records System (CJARS) and the Longitudinal Employer-Household Dynamics (LEHD) infrastructure, I construct a longitudinal employer-employee panel that follows workers with and without criminal records across establishments, industries, and regions from 2000 onward. I first test whether workers with records are disproportionately concentrated in low-wage, cyclically sensitive industries. Next, exploiting national variation in industry employment growth (e.g., during the Great Recession), I test whether firms hit by industry shocks reduce hiring and increase separations of workers with records relative to nonrecord holders, controlling for various worker and firm-level confounders. Finally, I identify mechanisms and whether these responses widen the conviction gap in employment, earnings, and reliance on safety-net programs.

Hundreds of local law enforcement agencies across the United States now use unmanned aerial vehicles equipped with audio, video, and thermal imaging technology to respond to emergency calls for service. These rapidly-proliferating “Drones as First Responder” (DFR) programs are predicted to substantially improve the quantity and quality of information available to police, but no comprehensive evaluation of DFR programs currently exists. I have collected detailed geolocational flight logs for the near-universe of publicly-documented drone flights (~300k) conducted by 200+ local law enforcement agencies to examine how drone flight characteristics vary across departments. For a subset of these departments, I pair this with granular data on calls-for-service, crime, and use-of-force to provide the first comprehensive study of how DFR programs impact modern police work.