Few studies have investigated the effects of medical marijuana laws on crime and even fewer have investigated the effects of retail marijuana laws on crime. These studies have mostly employed state-level panel data (Alford, 2014; Morris, TenFyck, Barnes & Kovandzic, 2014; Gavrilova, Kamada & Zoutman, 2014). I aim to study the effects of both medical and retail marijuana laws using city-level panel data. In order to model crime, along with medical and retail marijuana indicator variables, I include socioeconomic and demographic time-varying factors that are known contributors to crime. By using a two-way fixed effects two stage least squares approach, I control for unobserved constant heterogeneity and endogeneity. The results indicate that medical marijuana laws have a decreasing effect on property crime and retail marijuana laws have an increasing effect on property crime. Additionally, they show that the physical dispensaries are the driving factors causing this relationship. However, the results should be interpreted cautiously because there may be unobserved time-varying characteristics that my model does not capture.
