After cannabis legalization in Nova Scotia, alcohol sales fell 2.4% at alcohol-only stores but rose 3.1% at stores also selling cannabis, suggesting legalization triggered both substitution and co-purchasing.
Drug policy researchers, alcohol policy analysts, Canadian provincial regulators
Opposite effects: -2.4% at alcohol-only vs +3.1% at dual stores
What the researchers found
Comparing 17 months before and after cannabis legalization in Nova Scotia, alcohol-only stores saw an initial 2.91% sales decrease with minimal ongoing growth (0.06%/month), while stores selling both cannabis and alcohol saw a 0.55% initial increase with stronger growth (0.29%/month). Post-legalization alcohol sales averaged 2.4% below pre-legalization at alcohol-only stores but 3.1% above at cannabis-selling stores. Combined sales were 1.2% below pre-legalization. Beer was more affected than spirits or wines.
Why it matters
Whether cannabis substitutes for or complements alcohol use has major public health implications. This natural experiment in Nova Scotia, where some government stores sold both substances, shows that the answer may be both: overall alcohol sales declined slightly, but co-location drove increased alcohol purchasing.
The numbers in context
Alcohol-only stores: -2.4% average; cannabis-selling stores: +3.1% average; combined: -1.2%; initial response at alcohol-only: -2.91%; cannabis sellers: +0.55%; beer most affected
How the study worked
Comparative interrupted time series models analyzing monthly alcoholic beverage sales in Canadian dollars at Nova Scotia liquor stores during 17 months before and after cannabis legalization (May 2017-February 2020). Contrasted alcohol-only stores versus stores that also began selling cannabis.
What this study cannot tell us
Observational interrupted time series cannot establish causation. Only one Canadian province studied. 17-month post-legalization window may not capture longer-term trends. COVID-19 pandemic began near the end of the study period. Sales data may not perfectly reflect consumption.
How to read the evidence
Well-designed interrupted time series with natural experiment controls provides moderate evidence, limited by single-province scope and relatively short follow-up.
When this study was published
2025 publication analyzing sales data May 2017-February 2020
The bigger picture
The co-location of cannabis and alcohol sales in government stores is a unique Canadian policy feature. The finding that proximity to cannabis sales increased alcohol purchasing suggests that retail design choices within legalization frameworks have unintended consequences for alcohol consumption patterns.
Questions still open
- Does the co-purchasing effect persist over longer time periods? Would separating cannabis and alcohol retail locations reduce the complementary purchasing effect?
Common questions
Did cannabis legalization reduce alcohol use?
Why did co-located stores see higher alcohol sales?
Read the original research
Alcohol sales changes in a Canadian province after recreational cannabis legalization.
The International journal on drug policy, 142, 104840
Citation
Armstrong, Michael J. (2025). Alcohol sales changes in a Canadian province after recreational cannabis legalization.. The International journal on drug policy, 142, 104840. https://doi.org/10.1016/j.drugpo.2025.104840