The Verdict: Alcohol vs Cannabis by Every Metric
The final comprehensive comparison. After weighing every harm criterion — physical, psychological, societal — one unambiguous conclusion emerges.
In 2010, Professor David Nutt and his colleagues published what remains the most comprehensive scientific ranking of drug harm ever produced. Using Multi-Criteria Decision Analysis across 16 distinct harm dimensions, they assigned each substance a composite score. The result was unambiguous: alcohol scored 79 out of 100 — the highest of any drug assessed, surpassing heroin (55), crack cocaine (54), and methamphetamine (33). Cannabis scored 15 out of 100. [Nutt et al. (2010, The Lancet)]
In 2026, the same framework was updated with new data in a CIHR-funded study published in the Journal of Psychopharmacology. The conclusion did not budge: alcohol remains the most harmful substance to the individual user, to others, and to society as a whole. The gap between alcohol and cannabis is not marginal — it is a chasm, bigger than the gap between alcohol and any other drug. [Journal of Psychopharmacology (2026, CIHR-funded)]
Overdose Risk
Alcohol can and does kill in a single session. The median lethal blood alcohol concentration is around 0.40% — easily reachable through binge drinking, particularly in combination with other depressants. Alcohol poisoning kills thousands globally each year. Cannabis cannot cause fatal overdose by any known pharmacological mechanism. There are no cannabinoid receptors in the brainstem regions that control respiration; no matter how much is consumed, the body will not stop breathing. [Nutt et al. (2010)]
Organ Damage & Carcinogenicity
The World Health Organisation classifies alcohol as a Group 1 carcinogen — the highest category, shared with asbestos, formaldehyde, and tobacco. It causes seven types of cancer. The Global Burden of Disease study published in The Lancet concluded that no amount of alcohol consumption is safe for overall health. Cannabis is not classifiable as a human carcinogen. The primary health risk — lung irritation from smoking — is entirely eliminated by vaping, edibles, or tinctures, which are standard in legal markets. [GBD/Lancet] [WHO]
Dependence & Withdrawal
Anthony et al.'s epidemiological analysis found that 15.4% of alcohol users develop dependence at some point in their lives, compared to 9.1% of cannabis users. But the distinction in withdrawal severity is more significant than the prevalence. Alcohol withdrawal — delerium tremens — carries a 5% mortality rate: fatal seizures and cardiac arrest are a real risk of stopping drinking. Cannabis withdrawal causes insomnia, irritability, and reduced appetite. It is uncomfortable. It is not lethal. [Anthony et al.] [DSM-5]
Mental Health
Alcohol is a potent depressogen. It disrupts serotonin and dopamine systems, lowers seizure threshold, and is consistently associated with elevated rates of depression, anxiety, and suicide. The relationship is causal, not merely correlational. Cannabis can trigger psychotic episodes in individuals with latent predisposition — a real but conditional risk that affects a minority of users. The cognitive impairments associated with cannabis are reversible; those caused by alcohol — particularly frontal lobe damage and alcohol-related dementia — are permanent. [Nutt et al. (2010)] [Hirvonen et al.]
Violence & Societal Harm
Alcohol is the only recreational drug with a proven pharmacological link to violence. It reduces serotonin, impairs frontal lobe function, narrows cognitive focus, and increases aggression — a biochemical recipe for violence. Hoaken and Stewart's meta-analysis in Addictive Behaviorsconfirmed alcohol as a direct causal agent in aggressive behaviour. Cannabis has the opposite pharmacological profile: it is a sedative, anxiolytic, and muscle relaxant. Multiple studies across US states and Canada have found that cannabis legalisation is associated with reductions in violent crime and domestic violence. [Hoaken & Stewart (2003)]
The Safety Margin
The Margin of Exposure framework, designed by toxicologists to compare the safety of chemicals in consumer products, provides the most telling single number in the entire comparison. Alcohol's MOE is below 10 — indicating high risk. Cannabis's MOE is greater than 10,000 — indicating negligible risk for a typical user. Lachenmeier and Rehm's 2015 paper in Scientific Reports showed that alcohol's MOE is the lowest of any psychoactive substance — meaning it has the narrowest gap between a typical dose and a toxic dose. Cannabis's MOE is effectively infinite by comparison. [Lachenmeier & Rehm (2015, Scientific Reports)]
The Cumulative Calculus
When we sum every dimension — overdose, organ damage, cancer, dependence, mental health, violence, economic cost, family breakdown, brain damage, and societal burden — the picture that emerges is not one of comparable risk. Alcohol is in a class of its own: the most harmful substance known to humanity by every objective scientific measure. Cannabis is among the least harmful, clustered with substances that carry minimal risk to the user and negligible risk to others. [Nutt et al. (2010)]
The full Harm Matrix compares all six harm domains side by side. The Substance Matrix extends the comparison to 18 substances. Our Physical Effects, Social Impact, History, and Policy pages each examine a different dimension of the comparison in depth. The conclusion is consistent across every page.
There is a thought experiment that crystallises the absurdity of our current drug laws. Imagine, for a moment, that alcohol and cannabis were both discovered today. A new substance arrives on the market: it is a Group 1 carcinogen, causes seven types of cancer, is directly lethal in overdose, triggers homicidal violence, shrinks the brain, destroys families through FASD, creates physical dependence with fatal withdrawal, and is responsible for nearly 10,000 deaths a year in the UK alone. Would regulators license it for sale at every corner shop, supermarket, and petrol station, and allow it to be advertised on television during football matches?
And if a second substance arrived — one that cannot cause fatal overdose, is not carcinogenic, does not trigger violence, causes no permanent brain damage, produces only mild and non-lethal withdrawal, and has therapeutic applications recognised by national health services — would regulators respond by imprisoning anyone who used it? The question answers itself. Our drug laws are not a reflection of science. They are a relic of history: of racism, corporate protectionism, and political expediency. The evidence has been clear for decades. The only thing missing is the courage to act on it. [Nutt et al. (2010)] [Lachenmeier & Rehm (2015)]
Sources: [Nutt et al. (2010, The Lancet)] | [Journal of Psychopharmacology (2026)] | [Lachenmeier & Rehm (2015, Scientific Reports)] | [GBD/Lancet] | [WHO] | [Hoaken & Stewart (2003)] | [Hirvonen et al.] | [Anthony et al.] | [DSM-5]