Abstract
The recent legalization of marijuana in several states has led to increased public interest regarding the effect of legalization on crime. Yet, there is limited empirical evidence relating the legalization of marijuana use and distribution to criminal activity. This paper uses a difference-in-differences design to estimate the effect of marijuana dispensary openings on local crime rates in Denver, Colorado. We find that the opening of dispensaries actually decreases violent crime rates in above median income neighborhoods, an important finding in light of increased political debate surrounding legalization. We also find robust evidence that non-marijuana drug-related crimes decrease within a half-mile of new dispensaries but do not simultaneously increase within a half-mile to mile of new dispensaries, with one possible explanation being that legal marijuana sales and hard drug sales are local substitutes. Finally, in line with previous research, we find that vehicle break-ins increase up to a mile away from new dispensaries.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Google News had 20 separate news articles on legalizing marijuana on August 10, 2017.
Freisthler et al. (2017) use the same data, but they do not exploit the latitude and longitude coordinates. Instead, they aggregate to the census block level and only use data from 2013 to 2015.
Another possible explanation is that police presence increases immediately after the opening of a new dispensary. However, it is unlikely that police presence increased around the opening of all 188 dispensaries in our data, and in conversations with the Denver police department, officers indicated there were no efforts to increase police presence around dispensaries.
This second category is often called “common cause.”
Because Federal law still prohibits the sale and use of marijuana, the majority of transactions are still cash based.
Freisthler et al. (2017) do not have detailed information on when dispensaries opened.
Marijuana remains illegal at the Federal level.
To be sure, we estimated our primary model for each major crime type in the data. We do not report results for all crime types for lack of space, but results are available upon request.
However, as a robustness check we tested whether conversions had an effect on crime and found no statistically significant effect. Results available upon request.
,Results are roughly the same if we exclude the month the dispensary opened, e.g., March 2012.
One might wonder why we do include a variable indicating whether a dispensary is licensed for medical or recreational dispensing. All dispensary openings prior to 2014 are medical dispensaries by law. After 2014, when recreational dispensing became legal, nearly all new dispensaries were recreational. Therefore, the dummy for medical versus recreational is absorbed by the fixed effects. In practice, most dispensaries are now licensed for both types of dispensing. We also examined the effects of dispensaries converting from medical to recreational on crime in the same manner as Conklin et al. (2017). We did not find statistically significant results, likely because conversion dates are imprecise.
Search “denvergov content marijuana facility location guide,” for a list of the restrictions on dispensary locations.
We can get a sense of the distribution of application processing times by comparing dispensary opening dates to the initial application dates for dispensary openings that occurred within 12 months of the initial application date at the beginning of legalization. For instance, three dispensaries opened August 2010, three opened September 2010, six opened October 2010, four opened November 2010, two opened December 2010, and four opened in January of 2011 all with the initial application dates of February 2010. This suggests that the processing time for applications is between 6 and 11 months.
We find similar results using negative binomial regressions or zero-inflated Poisson where the alternative estimators converge. However, given the low count levels of certain crimes and the high-dimensional fixed effects, not all estimators converge for all categories of crime, e.g., marijuana offenses. Additionally, OLS yields similar marginal effects for all crime categories.
Marijuana crime results are not presented because the Poisson routine does not converge for this category of crime.
We also added \({n\mathrm {disp}}_\mathrm{{im}}^2\) to the model, but the coefficient was also statistically insignificant. If we relax the fixed effects, the coefficient on \({n\mathrm {disp}}_\mathrm{{im}}\) becomes statistically significant, which indicates that most of the variation in \({n\mathrm {disp}}_\mathrm{{im}}\) is absorbed by the fixed effects.
Dispensary fixed effects were sufficient in Chang and Jacobson (2017) because they evaluated the closures of multiple dispensaries on a single date. Thus, there were no neighborhood-specific changes in their sample period.
Crime counts are normalized by month of sample because the pre-treatment period is different for each treated location. This is done by regressing crime counts on month of sample fixed effects and subtracting the estimated fixed effects from the observed crime counts.
The section entitled, “Dispensary location selection: potential for external validity?” in the appendix presents basic summary statistics for neighborhoods with and without dispensaries. Notably, we find that neighborhoods with dispensaries have statistically significantly higher crime counts but do not exhibit statistical differences in crime rates or in key demographics.
We also tested a 1/4 mile radius, but for many of the crime categories the Poisson routine does not converge. For all that did converge, the effects were statistically insignificant.
References
Anderson MD, Hansen B, Rees DI (2013) Medical marijuana laws, traffic fatalities, and alcohol consumption. J Law Econ 56(2):333–369
Becker GS (1968) Crime and punishment: an economic approach. J Polit Econ 76(2):169–217
Bennett T, Holloway K, Farrington D (2008) The statistical association between drug misuse and crime: a meta-analysis. Aggress Violent Behav 13(2):107–118
Blake D, Finlaw J (2014) Marijuana legalization in Colorado: learned lessons. Harv L Pol’y Rev 8:359
BLS (2016) Quarterly census of employment and wages [cd-rom]
Brinkman J, Mok-Lamme D (2017) Not in my backyard? Not so fast. The effect of marijuana legalization on neighborhood crime
Castro T (2010) Lapd chief: Pot clinics not plagued by crime. Los Angeles Daily News
Chang TY, Jacobson M (2017) Going to pot? The impact of dispensary closures on crime. J Urban Econ 100:120–136
C.K. (2017) Jeff sessions orders tougher drug crime charges. The Economist
Conklin J, Diop M, Li H (2017) Contact high: The external effects of retail marijuana establishments on house prices. Real Estate Economics
Freisthler B, Gaidus A, Tam C, Ponicki WR, Gruenewald PJ (2017) From medical to recreational marijuana sales: marijuana outlets and crime in an era of changing marijuana legislation. J Prim Prev 38(3):249–263
Freisthler B, Ponicki WR, Gaidus A, Gruenewald PJ (2016) A micro-temporal geospatial analysis of medical marijuana dispensaries and crime in long beach, California. Addiction 111(6):1027–1035
Friedman M (1972) Prohibition and drugs. Newsweek
Friedman M (2017) Booker introduces bill to legalize marijuana nationwide. Politico
Groff E, McCord ES (2012) The role of neighborhood parks as crime generators. Secur J 25(1):1–24
Grogger J, Willis M (2000) The emergence of crack cocaine and the rise in urban crime rates. Rev Econ Stat 82(4):519–529
Gurman S (2017) Task force on marijuana law offers little on new policies. PBS NEWSHOUR
Hakim S, Shachmurove Y (1996) Spatial and temporal patterns of commercial burglaries: the evidence examined. Am J Econ Sociol 55(4):443–456
Hunt J (2017) Marijuana devastated Colorado, don’t legalize it nationally. USA Today
Ingold J (2010) Analysis: Denver pot shops’ robbery rate lower than banks’. The Denver Post
Kau JB, Rubin PH (1975) New estimates of the determinants of urban crime. Ann Reg Sci 9(1):68–76
Kepple NJ, Freisthler B (2012) Exploring the ecological association between crime and medical marijuana dispensaries. J Stud Alcohol Drugs 73(4):523–530
Komarek TM (2018) Crime and natural resource booms: evidence from unconventional natural gas production. Ann Reg Sci 61(1):113–137
McCollister KE, French MT, Fang H (2010) The cost of crime to society: new crime-specific estimates for policy and program evaluation. Drug Alcohol Depend 108(1):98–109
Miller TR, Zaloshnja E, Lawrence BA, Crandall J, Ivarsson J, Finkelstein AE (2004) Pedestrian and pedalcyclist injury costs in the united states by age and injury severity. In: Annual proceedings/association for the advancement of automotive medicine, vol 48, p 265. Association for the Advancement of Automotive Medicine
Morris RG, TenEyck M, Barnes JC, Kovandzic TV (2014) The effect of medical marijuana laws on crime: evidence from state panel data, 1990–2006. PLoS ONE 9(3):e92816
NYTimes (2014). Repeal prohibition, again. New York Times
Rephann TJ (2009) Rental housing and crime: the role of property ownership and management. Ann Reg Sci 43(2):435–451
Santos M (2017) What actually happened to violent crime after Washington legalized marijuana. The News Tribune
Thornton M (1991) The economics of prohibition. University of Utah Press Salt Lake City
Thornton M (2007) Prohibition versus legalization: do economists reach a conclusion on drug policy? Indep Rev 11(3):417–433
Waldman P (2017) Will Jeff sessions launch a war on weed? If so, it could accelerate marijuana legalization. The Washington Post
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 Appendix Tables and Figures
See Figs. 2, 3 and Tables 7, 8, 9, 10.
Number of marijuana dispensary openings per year
Trends in marijuana-related crimes
1.2 Dispensary location selection: potential for external validity?
In the Appendix, we explore whether dispensaries are located in unique areas of Denver. There are 283 neighborhoods that are defined by the Denver Assessor, and these neighborhoods are meant to group areas of homogeneous single-family residences based on household characteristics and demographics. The neighborhoods are about the size of census block groups. A map of the neighborhoods along with the location of each dispensary is provided in Fig. 4.
Table 11 presents the balance of observables across neighborhoods with and without dispensaries. This table shows that neighborhoods with dispensaries have statistically significantly higher crime counts across all crime categories. This is unsurprising given the strict zoning restrictions placed on dispensary locations. On the other hand, we find no statistical differences in month-to-month crime trends between neighborhoods with and without dispensaries, indicating the crime rate over time is relatively homogeneous across Denver. Perhaps most surprisingly, we find there are no statistical differences between several key demographics including race, education, and income between neighborhoods with and without dispensaries.
Overall, Table 11 suggests that zoning regulations place dispensaries in locations with higher crime counts on average, but these locations are not statistically different from non-dispensary locations based on other key observables. This suggests that we may see similar effects if dispensaries were placed in other parts of Denver.
Assessor neighborhoods and dispensaries in Denver
Rights and permissions
About this article
Cite this article
Burkhardt, J., Goemans, C. The short-run effects of marijuana dispensary openings on local crime. Ann Reg Sci 63, 163–189 (2019). https://doi.org/10.1007/s00168-019-00931-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00168-019-00931-0