Research

Abstract:  Non-fatal strangulation (NFS) is a dangerous form of intimate partner violence (IPV) and a strong predictor of homicide. We collect information on state NFS statutes and link them to FBI Supplementary Homicide Reports, 1990–2019, to estimate their causal effects on intimate partner homicide (IPH) rates. Using the two-stage difference-in-differences estimator (Gardner et al., 2025), which accommodates staggered adoption and heterogeneity, we find that NFS laws reduce female-victim IPH by 14% and male-victim IPH by 27% among those aged 18–49, with no detectable effects for those aged 50–70 or for homicides committed by strangers. Event-study profiles show flat pre-trends and sustained declines following enactment of NFS laws. Using the National Incident-Based Reporting System data, we estimate that NFS laws increase the share of IPV incidents classified as aggravated assaults—especially when the victim is a woman—and increase arrests conditional on IPV aggravated assaults, providing a two-step mechanism by which NFS laws disrupt the escalation of violence and protect lives.

Abstract:  We study the housing market response to a national disclosure policy that mandated the prominent display of energy efficiency information at listing. Using the near-universe of housing transactions in England and Wales with a highly flexible, generalized difference-in-differences specification, we estimate willingness-to-pay premiums for more energy-efficient properties of 1–3% nationally and up to 6.5% in London. We develop a conceptual framework to interpret the empirical findings, decomposing the hedonic weight on energy efficiency into valuation and attention components. Information salience is a key channel: buyers respond to the efficiency rating made prominent at listing. In most settings, observed premiums substantially exceed the present value of energy savings, pointing to an investment component alongside consumption value. These large valuation effects from a simple disclosure mandate point to information-based reform as a powerful and underutilized climate policy tool.

Earlier version circulated under the title: "Do Homebuyers Value Energy Efficiency? Evidence From an Information Shock"

Abstract: This paper studies a 2013 reform in which the city of Camden, New Jersey dissolved the city police department and replaced it. While crime did fall in Camden following the reform, evidence from a synthetic control approach suggests that there was no additional effect attributable to the reform beyond the expected improvement given state and national trends of a general fall in crime. However, we do find evidence of a 50% improvement in clearance rates, particularly for violent crimes.

Abstract: This paper studies how Medical Marijuana Laws (MML) affect the composition of crime. Using a generalized synthetic control estimator and data from the pre-recreational legalization period (1996–2012), we document three findings. First, MML lead to substantial increases in trafficking and possession of drugs other than marijuana, suggesting reallocation within criminal markets rather than simple crime reduction. Second, criminal responses differ markedly by age: younger criminals substitute toward marijuana trafficking while older criminals exit drug markets, with some shifting to property crime. Third, MML are associated with a 50% reduction in opioid use among older arrestees, consistent with marijuana-opioid substitution. We interpret these patterns through the lens of criminal human capital: legalization disrupts existing networks, and the rational response depends on age-specific incentives to reinvest. Data on police numbers rule out changes in enforcement intensity as an explanation.

Abstract: Due to felon disenfranchisement laws (FD), more than 7.4% of Blacks are disenfranchised compared to only 1.8% of other Americans. We study the political consequences of this racial disparity. Our difference-in-difference setup exploits the unpredictable timing of changes in FD, while allowing for time-varying state heterogeneity. We find FD legislation causes a 1 - 2 point reduction in turnout overall, but double that amongst Blacks. This is too large to be a mechanical effect, implying FD substantially reduces turnout among enfranchised Blacks. Further results show that FD reduces the number of Black U.S. Representatives and also lead to more conservative state policy.