Introduction
The strongest U.S. finding is a reduction in cannabis-possession arrests; evidence on trafficking, violent crime, and property crime is mixed. Cannabis legalization changes crime in a direct way by removing some cannabis conduct from the category of criminal offenses. Whether it changes other forms of crime is a more difficult question. A fall in possession arrests is not the same as a fall in drug trafficking, and a change in property-crime rates is not necessarily evidence of a change in violence associated with illicit markets.
This distinction also connects to a broader question about enforcement institutions: reducing a category of arrests does not necessarily change the organizational incentives that keep drug-enforcement practices in place. See the discussion of those incentives in Why Drug Policy Enforcement Fails.
The most useful criminological approach is outcome-specific. It asks what happens to possession arrests, illicit trafficking, violent crime, and property crime—and how those outcomes might respond to different reforms. Medical legalization permits cannabis access for qualifying patients. Adult-use, or recreational, legalization permits possession and use by adults under state law, usually subject to age and quantity limits. Decriminalization reduces or removes criminal penalties for possession without necessarily legalizing production or sale. Retail implementation is the later opening of licensed stores and the development of regulated commercial supply. These are related but distinct interventions: decriminalization changes enforcement, while retail implementation changes legal supply.
The U.S. evidence does not support a single verdict that legalization either reduces or increases “crime.” It supports a narrower conclusion: possession arrests decline clearly, while effects on illicit markets and conventional crime depend on policy design, timing, geography, and how crime is measured. In short:
- Possession arrests: clear reductions, especially among adults.
- Trafficking-related violence: possible reductions in specific border-market contexts, but no direct proof of reduced trafficking volume.
- Violent and property crime: mixed, context-dependent evidence; nonsignificant results do not prove no effect.
Possession arrests: the clearest effect
Possession arrests are the outcome most directly affected by legalization. When state law permits adult possession, police have fewer legal grounds to arrest adults for that conduct. A reduction in arrests is consequential because it reduces criminal-justice contact, but it should not automatically be interpreted as evidence that cannabis use or all drug-related offending has declined.
Gunadi and Shi (2022) analyzed U.S. cannabis-possession arrest data from 2010 through 2019 using a difference-in-differences design. In practical terms, this design compares changes over time in legalization states with changes in comparison states. Their adult category was people aged 18 and older; it was not necessarily the same as the legal purchase age of 21. They reported a 76.3 percent reduction in adult possession-arrest rates in recreational-legalization states without prior decriminalization and a 40.0 percent reduction in states that had already decriminalized cannabis. The smaller additional reduction in previously decriminalized states is consistent with the possibility that decriminalization had already removed much of the criminal-justice response to possession.
The study did not find a statistically clear change in Black–White possession-arrest-rate disparities. A large aggregate decline in arrests can therefore coexist with persistent disparities in who is stopped, searched, cited, or arrested. Legal change may reduce arrests without eliminating unequal enforcement of remaining restrictions, such as public use, underage possession, or sales.
That distinction is part of the broader racial-politics problem in drug enforcement, where formal legal reform and unequal criminal-justice exposure can coexist. See Sentencing Disparities and the Racial Politics of Drug Enforcement for related context.
The observed finding is a reduction in arrest rates. One interpretation is that legal change removed a category of low-level criminalization. Whether use, supply, or other crime changed requires different outcomes. Arrest data reflect police decisions and legal classifications as well as behavior.
Illicit trafficking: when legal supply displaces illegal markets
Legal supply could reduce illicit trafficking by taking customers and revenue away from illegal producers and sellers. It might also reduce violence associated with protecting illicit markets. But displacement depends on the market that develops. High taxes, limited store access, product shortages, quantity restrictions, interstate demand, and continuing federal prohibition can leave space for unlicensed suppliers. A state can legalize possession while illegal production and sales remain important.
Gavrilova, Kamada, and Zoutman (2019; published online in 2017) examined medical marijuana laws and violent crime near the Mexican border. Their publisher abstract reports reductions in violent crime in states bordering Mexico, with stronger reductions in counties close to the border and in crimes related to drug trafficking. It also reports reductions in crime in the nearest border state when an inland state adopted a medical marijuana law.
These are observed crime findings in a specific geographic and policy setting. The authors interpret them as consistent with medical-law reforms weakening markets traditionally controlled by Mexican drug-trafficking organizations. That interpretation is theoretically plausible, but the study does not provide direct proof of a reduction in trafficking volume. Violent-crime measures are indirect indicators of market activity, and changes in enforcement, reporting, migration, or local conditions could also matter.
Retail implementation is central. Decriminalization can reduce possession arrests without legalizing supply. Adult-use legalization can establish a legal framework before stores open. Retail implementation may then alter prices, availability, consumer purchasing, and competition with illegal sellers. Treating enactment and retail opening as the same event can obscure the mechanism being tested.
Violent crime: reductions and increases in different settings
Violent crime may respond to cannabis reform through competing pathways. Legal markets could reduce violence by displacing illicit-market disputes or reallocating police time away from possession enforcement. Consumption patterns might also change, although any substitution between cannabis, alcohol, and other drugs is an empirical matter. Conversely, new commercial activity, illegal diversion, or local market disruption could produce risks in particular places. These are hypotheses, not established consequences. The border-market evidence concerns one pathway to violence; it does not establish what happens to aggregate violent crime across the United States.
Dragone, Prarolo, Vanin, and Zanella (2019) used county-level difference-in-differences and spatial regression-discontinuity designs around the Washington–Oregon border. They reported reductions in rape and property crime on the Washington side relative to Oregon and to pre-legalization years. This is evidence from a border setting, not evidence of a nationwide reduction. Border areas may experience spillovers in consumer movement, dealer activity, police practices, and reporting.
Wu, Wen, and Wilson (2021) examined Oregon counties using Uniform Crime Reporting data from 2007 through 2017. They reported increases in several serious-crime rates in Oregon relative to comparison states, including overall violent and property crime and offenses such as aggravated assault, burglary, larceny, and motor-vehicle theft. The finding conflicts with the reductions reported by Dragone and colleagues. It should be presented as a study-specific result, not as proof that recreational legalization generally increases serious crime.
Clearance rates should also be kept separate from violent-offense rates. A policy may affect police investigations or the evidentiary significance of cannabis without changing the number of offenses reported to police. Arrests, reported offenses, victimization, and clearance rates are not interchangeable measures of violence.
Property crime: mixed evidence and heterogeneous mechanisms
Property crime is not a single behavior. Burglary, larceny, and motor-vehicle theft involve different opportunities, victims, and markets. Cannabis reform could reduce some property crime if illicit-market activity declines or if legalization changes economic opportunities. It could increase some offenses if retail locations become targets or if local conditions change. Those mechanisms need not operate uniformly across neighborhoods or states. Property crime requires separate treatment because its opportunities and mechanisms differ from those of drug-market violence.
Dragone and colleagues reported property-crime reductions near the Washington–Oregon border. Wu and colleagues reported increases in several property-crime measures in Oregon. These findings do not cancel one another out, but neither can be generalized without qualification.
Sabia, Dave, Alotaibi, and Rees (2024) used national data covering 2000 through 2019 and a generalized difference-in-differences approach. They reported reduced marijuana-related arrests and little evidence that recreational marijuana laws increased violent or property crime. Their crime results include arrest measures, so they should not be paraphrased as proof that reported violent or property offenses did not change. “Little evidence of an increase” means that the study did not obtain a statistically clear positive estimate under its design; it does not prove that the true effect is exactly zero.
The national analysis is useful because it examines many states and several datasets. At the same time, an average effect can conceal localized changes. A policy might have different consequences in a border county, a tourist district, a rural cultivation area, and a neighborhood near a retail outlet. The appropriate conclusion is that property-crime estimates remain sensitive to context and measurement.
Why studies find different effects
The first source of disagreement is policy definition. Medical legalization, adult-use legalization, decriminalization, and retail implementation change different legal and economic incentives. A study of medical access near the Mexican border cannot be treated as a direct estimate of a mature adult-use retail market.
The second is timing. Enactment, effective dates, licensing, retail openings, and market stabilization may occur years apart. Early estimates may capture transition, enforcement adjustment, or temporary supply conditions rather than a long-run equilibrium.
The third is comparison. States that legalize are not randomly selected, and neighboring areas may experience spillovers. States also differ in policing, demographics, economic conditions, pre-policy crime trends, and the completeness of crime reporting. These apparently conflicting results are easier to interpret once policy timing, comparison groups, and outcome measurement are placed side by side.
The fourth is measurement. Arrests are partly measures of enforcement. Reported offenses depend on victims’ reporting and agencies’ recording. Uniform Crime Reporting data can vary in coverage and quality by jurisdiction and year.
Finally, statistical methods matter. Difference-in-differences estimates depend on assumptions about pre-policy trends and treatment timing. When effects differ across states or evolve over time, model choice can affect the estimate. Each result should therefore be interpreted as evidence under a particular design, not as a timeless property of legalization.
Conclusion
Cannabis legalization clearly reduces adult possession arrests: Gunadi and Shi report reductions of 76.3 percent without prior decriminalization and 40.0 percent with prior decriminalization for their 18-and-older adult category. That finding concerns criminalization and enforcement, not every form of drug-related crime.
Evidence on illicit trafficking is more indirect. Medical-law research near the Mexican border finds violent-crime reductions consistent with market displacement, but it does not directly measure or prove reduced trafficking volume. Evidence on violent and property crime is mixed: a border study reports reductions, an Oregon study reports increases in several serious-crime rates, and a national study reports little evidence of increased violent or property crime arrest measures.
The strongest criminological conclusion is conditional. Legalization changes the legal opportunity structure—the set of behaviors the law permits or penalizes—around cannabis possession. Its broader consequences depend on whether legal supply displaces illegal supply, whether retail access is available and competitive, how police reallocate enforcement, and which crime outcome is measured. The contribution of this evidence is not a universal crime estimate, but a map of which outcomes change directly when criminal penalties, legal supply, and retail access change. Different reforms change different forms of criminal-justice contact and may produce heterogeneous effects across markets and places.
References
Dragone, D., Prarolo, G., Vanin, P., & Zanella, G. (2019). Crime and the legalization of recreational marijuana. Journal of Economic Behavior & Organization, 159, 488–501. https://doi.org/10.1016/j.jebo.2018.02.005
Gavrilova, E., Kamada, T., & Zoutman, F. (2019). Is legal pot crippling Mexican drug trafficking organisations? The effect of medical marijuana laws on US crime. The Economic Journal, 129(617), 375–407. First published online 2017. https://doi.org/10.1111/ecoj.12521
Gunadi, C., & Shi, Y. (2022). Association of recreational cannabis legalization with cannabis possession arrest rates in the US. JAMA Network Open, 5(12), e2244922. https://doi.org/10.1001/jamanetworkopen.2022.44922
Sabia, J. J., Dave, D. M., Alotaibi, F., & Rees, D. I. (2024). The effects of recreational marijuana laws on drug use and crime. Journal of Public Economics, 234, 105075. https://doi.org/10.1016/j.jpubeco.2024.105075
Wu, G., Wen, M., & Wilson, F. A. (2021). Impact of recreational marijuana legalization on crime: Evidence from Oregon. Journal of Criminal Justice, 72, 101742. https://doi.org/10.1016/j.jcrimjus.2020.101742