The short-run effects of marijuana dispensary openings on local crime


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 to check access.

Fig. 1


  1. 1.

    Google News had 20 separate news articles on legalizing marijuana on August 10, 2017.

  2. 2.

    We also find that dispensaries are located in neighborhoods with higher crime counts. However, we do not find differences in neighborhood demographics. See Table 11 in the Appendix for details.

  3. 3.

    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.

  4. 4.

    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.

  5. 5.

    This second category is often called “common cause.”

  6. 6.

    Because Federal law still prohibits the sale and use of marijuana, the majority of transactions are still cash based.

  7. 7.

    Freisthler et al. (2017) do not have detailed information on when dispensaries opened.

  8. 8.

    Marijuana remains illegal at the Federal level.

  9. 9.

    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.

  10. 10.

    However, as a robustness check we tested whether conversions had an effect on crime and found no statistically significant effect. Results available upon request.

  11. 11.

    ,Results are roughly the same if we exclude the month the dispensary opened, e.g., March 2012.

  12. 12.

    A 6-month pre-/post-period robustness check is presented in the Appendix Table 10.

  13. 13.

    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.

  14. 14.

    Search “denvergov content marijuana facility location guide,” for a list of the restrictions on dispensary locations.

  15. 15.

    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.

  16. 16.

    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.

  17. 17.

    Marijuana crime results are not presented because the Poisson routine does not converge for this category of crime.

  18. 18.

    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.

  19. 19.

    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.

  20. 20.

    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.

  21. 21.

    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.

  22. 22.

    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.

  23. 23.

    The average cost of a theft is $3532 (McCollister et al. 2010; Miller et al. 2004). We could not find information on the costs of drug and alcohol possession and sales crimes.


  1. Anderson MD, Hansen B, Rees DI (2013) Medical marijuana laws, traffic fatalities, and alcohol consumption. J Law Econ 56(2):333–369

    Article  Google Scholar 

  2. Becker GS (1968) Crime and punishment: an economic approach. J Polit Econ 76(2):169–217

    Article  Google Scholar 

  3. 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

    Article  Google Scholar 

  4. Blake D, Finlaw J (2014) Marijuana legalization in Colorado: learned lessons. Harv L Pol’y Rev 8:359

    Google Scholar 

  5. BLS (2016) Quarterly census of employment and wages [cd-rom]

  6. Brinkman J, Mok-Lamme D (2017) Not in my backyard? Not so fast. The effect of marijuana legalization on neighborhood crime

  7. Castro T (2010) Lapd chief: Pot clinics not plagued by crime. Los Angeles Daily News

  8. Chang TY, Jacobson M (2017) Going to pot? The impact of dispensary closures on crime. J Urban Econ 100:120–136

    Article  Google Scholar 

  9. C.K. (2017) Jeff sessions orders tougher drug crime charges. The Economist

  10. Conklin J, Diop M, Li H (2017) Contact high: The external effects of retail marijuana establishments on house prices. Real Estate Economics

  11. 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

    Article  Google Scholar 

  12. 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

    Article  Google Scholar 

  13. Friedman M (1972) Prohibition and drugs. Newsweek

  14. Friedman M (2017) Booker introduces bill to legalize marijuana nationwide. Politico

  15. Groff E, McCord ES (2012) The role of neighborhood parks as crime generators. Secur J 25(1):1–24

    Article  Google Scholar 

  16. Grogger J, Willis M (2000) The emergence of crack cocaine and the rise in urban crime rates. Rev Econ Stat 82(4):519–529

    Article  Google Scholar 

  17. Gurman S (2017) Task force on marijuana law offers little on new policies. PBS NEWSHOUR

  18. Hakim S, Shachmurove Y (1996) Spatial and temporal patterns of commercial burglaries: the evidence examined. Am J Econ Sociol 55(4):443–456

    Article  Google Scholar 

  19. Hunt J (2017) Marijuana devastated Colorado, don’t legalize it nationally. USA Today

  20. Ingold J (2010) Analysis: Denver pot shops’ robbery rate lower than banks’. The Denver Post

  21. Kau JB, Rubin PH (1975) New estimates of the determinants of urban crime. Ann Reg Sci 9(1):68–76

    Article  Google Scholar 

  22. Kepple NJ, Freisthler B (2012) Exploring the ecological association between crime and medical marijuana dispensaries. J Stud Alcohol Drugs 73(4):523–530

    Article  Google Scholar 

  23. Komarek TM (2018) Crime and natural resource booms: evidence from unconventional natural gas production. Ann Reg Sci 61(1):113–137

    Article  Google Scholar 

  24. 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

    Article  Google Scholar 

  25. 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

  26. 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

    Article  Google Scholar 

  27. NYTimes (2014). Repeal prohibition, again. New York Times

  28. Rephann TJ (2009) Rental housing and crime: the role of property ownership and management. Ann Reg Sci 43(2):435–451

    Article  Google Scholar 

  29. Santos M (2017) What actually happened to violent crime after Washington legalized marijuana. The News Tribune

  30. Thornton M (1991) The economics of prohibition. University of Utah Press Salt Lake City

  31. Thornton M (2007) Prohibition versus legalization: do economists reach a conclusion on drug policy? Indep Rev 11(3):417–433

    Google Scholar 

  32. Waldman P (2017) Will Jeff sessions launch a war on weed? If so, it could accelerate marijuana legalization. The Washington Post

Download references

Author information



Corresponding author

Correspondence to Jesse Burkhardt.



Appendix Tables and Figures

See Figs. 23 and Tables 78910.

Fig. 2

Number of marijuana dispensary openings per year

Fig. 3

Trends in marijuana-related crimes

Table 7 Dispensary-level crime rate summary statistics
Table 8 Test of parallel trends
Table 9 Local effects of marijuana dispensary openings: 1/3 mile radius OLS estimates
Table 10 Local effects of marijuana dispensary openings: 1/2 mile radius for 6 months before and after opening date

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.

Fig. 4

Assessor neighborhoods and dispensaries in Denver

Table 11 Crime and demographic summary statistics by neighborhood

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

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).

Download citation

JEL Classification

  • K10
  • K14
  • K42
  • R50