Abstract
<title>Abstract</title> <p> Racial disparities in marijuana enforcement have become a leading justification for recreational cannabis legalization, because a possession arrest is not only a criminal-justice event but an economic one: it can trigger pretrial detention, a lasting criminal record, and the well-documented penalties that records impose on employment, earnings, housing, and credit. We ask whether legalization narrows the Black/White gap in marijuana arrest rates, treating the disparity ratio itself as the estimand rather than the arrest level of any single group. Using the FBI Uniform Crime Reporting “Arrests by Age, Sex, and Race” master files linked to Census race-specific population denominators, we build a jurisdiction-by-year panel spanning 2015 to 2020 that covers the 50 states and the District of Columbia, eleven of which adopted recreational cannabis laws during the window. We estimate staggered-adoption difference-in-differences models using the Callaway and Sant’Anna (2021) and Sun and Abraham (2021) estimators, which are robust to treatment-effect heterogeneity, alongside two-way fixed-effects benchmarks. Legalization reduces the log Black/White ratio of marijuana-possession arrests by roughly 0.34 to 0.39 log points, a narrowing of about 29 to 32 percent that is statistically significant and stable across estimators, control-group definitions, leave-one-state-out re-estimation, and controls for reporting composition. Placebo outcomes, driving under the influence and larceny-theft, show no comparable change. Decomposing the ratio, arrest rates fall for both groups but proportionally more for Black residents, so the gap narrows through compression of a very large Black arrest rate rather than through equalization. Our conclusion, that legalization narrows but does not close the possession gap and leaves broader drug-enforcement disparities intact, independently corroborates the comprehensive national estimates of Meinhofer, Rubli, and Cunningham (2026) through a deliberately different and transparent research design. We draw out the implications for cannabis policy as racial-equity and economic-mobility policy. <bold>JEL:</bold> K42, I18, J15, J78 </p>