Do Recreational Cannabis Laws Narrow the Black/White Gap in Marijuana Arrests? Evidence from Staggered Difference-in-Differences

Technical Summary: "Do Recreational Cannabis Laws Narrow the Black/White Gap in Marijuana Arrests?"

Problem and Motivation
Racial disparities in marijuana enforcement in the United States are well-documented: despite similar usage rates, Black Americans are arrested for possession at significantly higher rates than White Americans. This paper posits that these arrests are not merely criminal-justice events but economic ones, as they trigger pretrial detention, criminal records, and subsequent penalties in employment, earnings, housing, and credit. While recreational cannabis legalization is often justified as a mechanism to reduce these disparities, the empirical question remains whether such laws actually narrow the Black/White arrest gap or merely lower arrest levels for all groups while leaving relative disparities intact. The authors argue that because the policy debate is framed in relative terms, the disparity ratio itself must be the estimand, rather than the arrest levels of a single group.

Data and Empirical Strategy
The study constructs a jurisdiction-by-year panel spanning 2015 to 2020, covering all 50 states and the District of Columbia.

• Data Sources: Arrest counts are drawn from the FBI Uniform Crime Reporting (UCR) "Arrests by Age, Sex, and Race" master files, linked to race-specific population denominators from the U.S. Census Bureau.
• Treatment: Eleven states adopted recreational cannabis laws within or just before the sample window (staggered adoption). The remaining 40 jurisdictions serve as the control group.
• Outcome Variable: The primary outcome is the log of the Black/White ratio of age-aggregated marijuana possession arrest rates per 100,000 residents. The authors use a log transformation to interpret effects as proportional changes in the disparity ratio.
• Identification Strategy: Recognizing that conventional two-way fixed-effects (TWFE) estimators can be biased in staggered adoption settings with heterogeneous treatment effects, the authors employ heterogeneity-robust estimators:
  • Callaway and Sant'Anna (2021): A group-time average treatment effect (ATT) estimator using never-treated states as controls.
  • Sun and Abraham (2021): An interaction-weighted event-study estimator.
  • TWFE: Reported only as a benchmark.
• Falsification and Placebos: To isolate the effect of cannabis-specific policy changes from general shifts in enforcement or reporting composition, the study analyzes:
  • Offense-specific outcomes: Marijuana sale/manufacture (which remains illegal) and the aggregate of all drug-abuse violations.
  • Non-drug placebos: Driving under the influence (DUI) and larceny-theft.
• Controls: The authors deliberately avoid controlling for time-varying covariates that may be outcomes of legalization (e.g., police budgets, incarceration rates) to avoid the "bad control" problem. Instead, they control for the number of reporting agencies and coverage population to address potential biases from the voluntary nature of UCR reporting.

Key Results

1. Reduction in Disparity: Recreational legalization reduces the log Black/White ratio of marijuana possession arrests by approximately 0.34 to 0.39 log points. This corresponds to a statistically significant narrowing of the disparity ratio by roughly 29% to 32%.
2. Specificity of Effect: The reduction is concentrated in marijuana possession. There are no statistically significant reductions in disparities for marijuana sale/manufacture, the aggregate of all drug-abuse violations, or the placebo outcomes (DUI and larceny-theft). This suggests the effect is specific to the decriminalization of possession rather than a general shift in policing intensity.
3. Mechanism (Decomposition): The narrowing of the gap is driven by a proportional reduction in arrest rates for both Black and White residents, but the decline is significantly larger for Black residents.
   • Black possession arrest rates fall by approximately 70%.
   • White possession arrest rates fall by approximately 60%.
   • The gap narrows through the compression of a very large initial Black arrest rate, not through the equalization of rates or an increase in White arrests.
4. Robustness: The findings are stable across different estimators, control group definitions (never-treated vs. not-yet-treated), leave-one-state-out re-estimations, and controls for reporting composition. The results hold even when weighting by population, though the effect is smaller in the most populous states (e.g., California), suggesting the unweighted average may overstate the national equity gain.

Significance and Claims
The paper claims to provide an independent, transparent replication of the qualitative conclusion reached by Meinhofer, Rubli, and Cunningham (2026), using a different sample window, outcome construction, and falsification design. The authors argue that their results strengthen confidence in the literature, countering jurisdiction-specific studies that found persistent or widening disparities.

The authors characterize the significance of legalization as a "real but partial" equity intervention:

• Success: It mechanically removes the most common cannabis charge, thereby compressing the racial disparity associated with that specific offense by roughly 30%. This reduction plausibly lowers the differential exposure of Black residents to the economic harms associated with criminal records.
• Limitations: The gap narrows but does not close; a substantial absolute disparity remains. The improvement does not extend to offenses that remain illegal (sale/manufacture) or to drug enforcement broadly. Furthermore, legalization does not address the stock of existing criminal records.

Consequently, the paper concludes that while legalization is a necessary component of a racial-equity and economic-mobility agenda, it is not a complete remedy. The authors suggest that to fully realize the economic benefits, legalization must be paired with complementary policies such as automatic expungement (to address the stock of records) and continued scrutiny of the enforcement of remaining offenses.