cannabis legalization does not automatically increase public harm; the structure of markets and rules determines outcomes, according to a University of Bath review published in The Lancet Psychiatry. The authors analysed reforms across five continents using population surveys, health records and longitudinal studies from January 1, 2000 to March 18, 2025, and concluded that commercialised supply models raise adult use and disorder while tightly regulated or non-commercial systems do not.
Scope and methods The review is the largest international assessment of cannabis policy changes to date. Researchers compared decriminalisation, medical access, state-controlled sales, non-profit distribution and commercial retail. They tracked three main outcomes: prevalence of cannabis use, cannabis use disorder, and psychiatric outcomes including psychosis, anxiety and depression. The team modelled data from multiple countries and jurisdictions to isolate policy effects from broader trends.
European decriminalisation: small or no rises in use Across multiple European examples, decriminalising possession did not produce large increases in use. Portugal’s 2001 decriminalisation showed a 0.3 percentage-point rise in past-year use from 2001–2007, compared with 0.4 points in Spain and 8.4 points in Italy over the same period. Czechia’s 2010 decriminalisation did not change the average age of initiation. When the UK downgraded cannabis from Class B to Class C (2004–2009), surveys found no material change in use among 15–17 or 18–20 year-olds and a decline in hospital admissions for cannabis-induced psychosis.
Global examples: neutral short-term effects After South Africa’s 2018 Constitutional Court ruling that private adult use and cultivation were unconstitutional to criminalise, adolescent cannabis-related hospital admissions between 2015 and 2020 did not increase. Georgia’s 2018 policy change showed no change in annual prevalence among 18–29 year-olds. A cross-country study covering 38 countries from 2001–2010 found no consistent link between liberalisation and adolescent use.
Uruguay: regulated, non-profit channels cut harm Uruguay implemented national legal access without a commercial retail market. Access routes included home cultivation (up to six plants), not-for-profit social clubs and state-licensed pharmacies. The government banned advertising, limited branded products and capped THC in some products. Post-legalisation comparisons show reductions in youth use across measures: frequent use (10+ days/month) fell by 1.8 percentage points and high-risk use declined by 5.0 points. Among 18–21 year-olds, high-risk use dropped by 3.5 points. Uruguay’s model reduced both prevalence and high-risk use in the groups the debate most often centres on.
North America: commercial markets increase adult use In Canada, national legalisation in 2018 mainly led to private retail models at the provincial level. Adult use rose from 22% in 2018 to 27% in 2020 and stabilised at 26% by 2023–24. US states with commercial supply also report higher adult prevalence and increased cannabis use disorder in some analyses. The review links this trend to commercial incentives: companies seek higher sales through lower prices, stronger products and marketing. Some jurisdictions documented increases in emergency department visits related to psychosis as dispensary density rose; Colorado showed a notable association between dispensary-per-population ratios and psychosis-related visits, although a larger multi-state study found no clear association.
Mechanism: commercial incentives change product and demand The review identifies the commercial motive as the key driver of observed harms. Private firms increase sales by lowering prices, developing higher-THC products, and advertising. Those actions raise exposure and the risk of heavy, problematic use. Steve Rolles of Transform Drug Policy Foundation summarised the mechanism: profit incentives push markets toward higher consumption and product formulations that raise health risks.
Harms of continued prohibition The review also compares reform harms to harms from continued criminalisation. In England and Wales, a 2025 London Drugs Commission report found Black people are five times more likely than white people to be stopped and searched for suspected drug offences; cannabis possession drives many of those stops. A 2018 UK study estimated over one million police hours per year spent enforcing cannabis bans and an annual criminalisation cost of more than £890 million in England and Wales. Convictions create long-term barriers to employment, housing and education and sustain illegal markets that generate violence.
Policy implications and recommendations The review’s data point to a clear policy prescription: legal access can reduce criminal justice harms and can be implemented without raising population-level health burdens, but design choices matter. – Non-commercial and tightly regulated models show the most evidence of neutral or beneficial outcomes. Examples include regulated home growing, membership-based social clubs, and state-controlled distribution, as in Uruguay, Malta and Germany’s not-for-profit frameworks. – Where retail is allowed, governments must control product potency, ban or tightly limit advertising, restrict product forms that appeal to young people, and monitor retail density and price incentives. – Policy should measure and report outcomes (use prevalence, disorder rates, psychiatric admissions, criminal justice contacts) at regular intervals to allow course correction.
Conclusions The University of Bath review compiles 25 years of data and finds that cannabis policy effects depend on market structure. Commercialised supply correlates with higher adult use and more cannabis use disorder in several jurisdictions, while decriminalisation and non-commercial legal models show little to no rise in harm and, in some cases, reduced high-risk use. Policymakers choosing reform can reduce criminal-justice harms without increasing population-level health burdens if they prioritise strict controls on supply, potency and marketing rather than unrestricted commercialisation.
