Sunday, October 17, 2021
The title of this post is the title of this notable new working paper authored by Scott Callahan, David M. Bruner and Chris Giguere. Here is its abstract:
U.S. drug policy presumes prohibition reduces crime. Recently states have enacted medical marijuana laws creating a natural experiment to test this hypothesis but is impeded by severe measurement error with available data. We develop a novel imputation procedure to reduce measurement error bias and estimate significant reductions in violent and property crime rates, with heterogeneous effects across and within states and types of crime, contradicting drug prohibition policy. We demonstrate uncorrected measurement error or assuming homogeneous policy effects leads to underestimation of crime reduction from ending marijuana prohibition.
And here is a key paragraph from the paper's introduction:
Our results indicate that MMLs result in significant reductions in both violent and property crime rates, with larger effects in Mexican border states. While these results for violent crime rates are consistent with previously reported evidence (Gavrilova et al., 2017), we are the first paper to report such an effect on property crime as well. Moreover, the estimated effects of MMLs on property crime rates are substantially larger, which is not surprising given property crimes are more prevalent. We also find novel evidence consistent with our hypothesis that MMLs reduce violent crime rates more in urban counties compared to rural counties, contrary to previous estimates (Chu and Townsend, 2019). We attribute this result to greater conflict between producers in urban counties under prohibition. Overall, our results are consistent with the need for market participants to create de facto property rights under prohibition, often through the use of violence. Our results are also consistent with prohibition causing a diversion of scarce policing resources, which when reallocated have the greatest impact on more pervasive types of crime and in locations where crime rates are higher. These findings demonstrate both the importance of accounting for heterogeneous policy effects on crime and the necessity to correct for measurement error in crime data when conducting policy analysis.