Date of Award

Spring 2023

Degree Type

Restricted to Claremont Colleges Dissertation

Degree Name

Economics, PhD


School of Social Science, Politics, and Evaluation

Advisor/Supervisor/Committee Chair

Gregory DeAngelo

Dissertation or Thesis Committee Member

Thomas J. Kniesner

Dissertation or Thesis Committee Member

Pierangelo De Pace

Terms of Use & License Information

Terms of Use for work posted in Scholarship@Claremont.

Rights Information

© 2023 Abdullah Alswelh


Marijuana, Recreational marijuana laws, Medical Marijuana Laws, Traffic Fatalities, Alcohol Consumption, Current Population Survey data, Two-way fixed effects, National Survey of Drug Use and Health

Subject Categories



Medical and recreational marijuana laws have proven to be contentious issues as they are being considered in numerous states in recent years. This dissertation investigates three different aspects of the effect of marijuana legalizations. In the first chapter of this dissertation, my co-author 1 and I replicate Anderson et al. 2013 paper "Medical Marijuana Laws, Traffic Fatalities, and Alcohol Consumption" and extend the analysis using a more appropriate estimator given the differential rollout of the treatment (Callaway and Sant' Anna, 2021). From 1990 to 2010, medical marijuana laws (MML) were adopted by 14 states as well as the District of Columbia. Using traffic fatalities data from Fatality Analysis Reporting System (FARS) at the state level from 1990 to 2010, we replicated Anderson et al. (2013). The original work found that after the states implemented the MML, there was a reduction in traffic fatalities. However, we find a heterogeneous treatment effect by employing the differential-timing difference-in-differences (DTDD) estimator. Most of the negative results are derived from western States (early adopters), while the reverse effect is observed in eastern states. In the second chapter of this dissertation, my co-author2 and I explore the effect of Recreational marijuana laws (RMLs) on Harder Drug Use and Crime. Unlike Medical Marijuana Laws, RMLs legalize the possession of small quantities of marijuana for recreational use. From 2000 to 2021, RMLs have been adopted by 18 states and the District of Columbia. Opponents argue that RML-induced increases in marijuana consumption will serve as a “gateway” to harder drug use and crime. Using data covering the period 2000-2019 from a variety of national sources (the National Survey of Drug Use and Health, the Uniform Crime Reports, the National Vital Statistics System, and the Treatment Episode Data Set) this analysis is the first to comprehensively examine the effects of legalizing recreational marijuana on hard drug use, arrests, drug overdose deaths, suicides, and treatment admissions. Our analyses show that RMLs increase adult marijuana use and reduce drug-related arrests over an average post-legalization window of three to four years. There is little evidence to suggest that RML-induced increases in marijuana consumption encourage the use of harder substances or violent criminal activity. In the last chapter of this dissertation, I examine the effect of recreational marijuana laws (RMLs) on labor market outcomes using 50 states and the Districts of Colombia using data from the Current Population Survey data (CPS) from 2001 to 2019. This analysis employs difference-indifferences (DD) with two-way fixed effects (TWFE) and Callaway and Sant’Anna 2021 Differential-timing difference-in-differences (DTDD) Estimator to explore the changes in average weekly wage and average weekly hours of work after states adopted RMLs. I include in this analysis the marijuana use among adults to unpack the change in labor market outcomes using the National Survey of Drug Use and Health (NSDUH) from 2002-2018. I find that RMLs increase marijuana use among adults. In addition, I find no evidence that RMLs affect labor market outcomes (average weekly wage or average weekly work hours) using difference-in-differences (DD) with two-way fixed effects (TWFE) and Callaway and Sant’Anna 2021 Differential-timing estimator.