NBER Working Paper
We study the effect of drug decriminalization on overdose mortality using Oregon’s Measure 110 and Washington’s Blake decision. Synthetic-control estimates show that both decriminalization regimes sharply reduced drug arrests and were followed by sustained increases in overdose mortality relative to matched counterfactuals. The estimates imply approximately 1,186 excess deaths in Oregon and 1,895 in Washington from 2021 through 2023. We also revisit the role of fentanyl supply shocks. A fentanyl-share control may partly capture enforcement-driven declines in non-fentanyl NFLIS reports, while alternative fentanyl measures leave positive overdose effects.

SSRN Working Paper
This paper studies the intergenerational effects of a higher unemployment rate at the time of school-leaving. Using data from the National Longitudinal Survey of Youth and an instrumental variables approach, we find no evidence that early labor market conditions for mothers affect their children’s outcomes. These null effects are precisely estimated; for a one standard deviation increase in the state unemployment rate at a mother’s time of school-leaving, we can rule out detrimental effects on their child’s low birthweight status, educational attainment, and age 29-30 employment larger than 0.15 standard deviations.

NBER Working Paper
While investments in schooling generate large private and external returns, negative peer interactions in school may generate substantial social costs. Using data from four national sources (Uniform Crime Reports, National Incident-Based Reporting System, National Crime Victimization Survey, National Electronic Injury Surveillance System) and a variety of identification strategies, this study comprehensively explores the effect of in-person schooling on contemporaneous juvenile violence. Using a proxy for in-person schooling generated from anonymized smartphone data and leveraging county-level variation in school calendars — including unique, large, localized changes to in-person instruction during the COVID-19 pandemic — we find that in-person schooling is associated with a 28 percent increase in juvenile violent crime. A null finding for young adults is consistent with a causal interpretation of this result. The effects are largest in larger schools and in jurisdictions with weaker anti-bullying policies, consistent with both concentration effects and a peer quality channel. Back-of-the-envelope calculations suggest that relative to closed K-12 schools, in-person schooling generates $233 million in monthly violent crime costs.