Scientific Proof that Drug Decriminalization in Portugal Saved Lives and Killed People

Here are the most recent data available from the European Monitoring Centre for Drugs and Drug Addiction, which address drug-related deaths in Portugal. Portugal decriminalized all drugs in 2001, and these factually accurate data can be used to prove that Portugal’s policy has been a complete success or a complete failure, assuming the analyst has no intellectual integrity. EMCDDA is one of those annoying organizations that provides full information without political spin, so clearly you can’t rely on the chart the way they print it up


If you have been commissioned to do a study for the Glibertarian Institute, then the chart below this paragraph is for you. Note that it is very important to start your analysis in 2001 to give the false impression that the new policy caused the drop in drug-related deaths, rather than coming into force several years after drug-related deaths were already in decline. Note also that you need to cut the data series off when it looks the best rather than go too far forward in time. Finally, do not mention that Portugal has expanded methadone maintenance and other addiction treatment, as that would present a highly plausible rival explanation: You can’t have that (what would your donors say?).

If on the other hand you are trying to land a job at the Draconian Foundation, you want to edit differently, as I have done in the chart below this paragraph. By adding back in the 1999-2000 data and deleting 2002-2006, you can use the remaining data to show that the effect of decriminalization was to reverse a pre-existing decline in drug-related deaths.

Now some advice for both parties. If you get confronted with the whole data series and your ideology or intellectual capacities forbid you to acknowledge that life is complicated, there are a number of rhetorical strategies I would suggest. If you work at the Glibertarian Institute, note that the population of Portugal has been rising and therefore the increase in number of deaths in recent years is due to increasing population size. Do not however acknowledge that this would also make the pre-decriminalization drop in number of deaths all the more impressive. If you got that job you wanted at the Draconian Foundation, just do the reverse.

Another handy strategy is to look around at other countries until you find somewhere else that had a similar change under different policies. But remember to confine your comparative work to that subsection of the data series that you wish to discount. So Glibertarians, point out that drug-related deaths have also been rising in Austria and Denmark in recent years, so what’s going on hardly reflects a failure of Portuguese policy. But don’t be so infra dig as to point out that drug-related deaths were dropping in Bulgaria and France from 2001-2003, which could make it look like Portuguese policy wasn’t the causal factor. Draconians can again just do the reverse.

Final words of wisdom to both sides: Describe the policy as legalization rather than decriminalization — no sense stopping at one misrepresentation when your ideological agenda is at stake!

Author: Keith Humphreys

Keith Humphreys is the Esther Ting Memorial Professor of Psychiatry at Stanford University and an Honorary Professor of Psychiatry at Kings College London. His research, teaching and writing have focused on addictive disorders, self-help organizations (e.g., breast cancer support groups, Alcoholics Anonymous), evaluation research methods, and public policy related to health care, mental illness, veterans, drugs, crime and correctional systems. Professor Humphreys' over 300 scholarly articles, monographs and books have been cited over thirteen thousand times by scientific colleagues. He is a regular contributor to Washington Post and has also written for the New York Times, Wall Street Journal, Washington Monthly, San Francisco Chronicle, The Guardian (UK), The Telegraph (UK), Times Higher Education (UK), Crossbow (UK) and other media outlets.

20 thoughts on “Scientific Proof that Drug Decriminalization in Portugal Saved Lives and Killed People”

  1. But I don't care if drug legalization "saves lives"; I mean, ok, on an all things being equal basis I'd prefer that it did, but that's not my primary concern.

    What I care about is,

    1. Liberty.


    2. The lives of people who AREN'T stupid enough to take drugs.

    Of course, legalizing drugs increases the former. And the graph doesn't relate who was dying, or under what circumstances. So none of your graphs are of any particular use to me.

    Got any graphs of drug war induced deaths?

  2. Brett, since the Portuguese policy leaves the "drug war" in place – drug dealing is still a crime – the expected impact on "drug war induced deaths" is zero.

  3. Sorry, why would it be honest for the Portuguese policy to be called legalization? Only laws against personal possession were nixed, drug users can still be hauled in front of a "dissuasion commission", and there are no legal commercial distribution channels.

    Kudos on a classic piece of Kleiman/Humphreys pox-on-both-houses third-way-ism. Note that you convincingly disparaged two views of one aspect of Portuguese drug policy (deaths due to overdose), but said nothing about anything else (HIV, treatment seeking, police relations, property crime). God forbid you (Humphreys in particular) should just lay down your preferred policy all at once. Instead we have to infer it from what he doesn't like.

  4. Brett appears to believe that the highest social good is the maximization of liberty of stupid people to kill themselves. This generally appears to be the standard for libertarians: liberty means the right to fuck up in ways that ruin your life and the lives of people around you.

  5. Bloix appears to believe that the highest social good is ruining the lives of people in far more serious ways than drugs generally can, such as imprisoning them, leaving them with criminal records, denying them student loans, breaking up their families, and causing them to become victims of crimes caused by the drug war.

  6. Comparisons with Portugal will only go so far, no matter what they indicate. As any Brazilian can tell you, the Portuguese are…different. (Brincadeira, viu!) Gosh, Humphreys and Kleiman have spent so much time thinking about drug policy and its consequences, you'd think they'd be eager to see what happens in California with legal recreational cannabis. You know, a little experiment.

  7. Funny post, but what *is* going on in Portugal? Did the potency of heroin being imported go down and then up or something like that?

  8. Most commenters here seem to be missing the point. The post is not actually about drugs abuse in Portugal. It's illustrating the ways that statistics can be used to say whatever a pundit wants them to say. To quote Mark Twain, "There are three kinds of lies: white lies, damned lies and statistics."

  9. Another thing: I'll grant the main point, which is that statistical info is bent to partisan purposes, but then we already knew that, thanks Mr. Twain. But if there is also a more specific point about what Portugal's experiment shows, I question whether "deaths from drugs" is really the statistic that matters. Portugal has a population of about 11 million, and the annual "drug deaths" varies from 152 to 368 in the period covered. In light of the country's population, the variation in the number of deaths could easily be due to other factors, like Cardinal Fang mentions. Let's talk about prison budgets, crime rates, even–if we dare–the "cost to society" of drug use in lost productivity. Drug deaths? Hmm.

  10. Bloix doesn't appear to believe anything other than that Brett Bellmore's libertarian argument is not persuasive.

    Artor appears to believe that the hijacking of comments threads is a bad thing. He might be right about that.

  11. Wait, so did Portugal just decriminalize the possession of all drugs, or production as well? If it's just possession, then that's a rather half-assed approach that does nothing to deal with the criminality on the "supply" side.

  12. Brett, in the drug policy world, "decriminalization" means that possession for personal use is not a crime, but dealing is a crime.

    (Confusingly, that's what was called "Prohibition" when we applied it to alcohol.) So the trumpeting of Portuguese policy as proof that drug legalization would be a good idea is grossly dishonest. Bring it up with your libertarian friends.

    Warren, if you want to know what drug policy I support, I wrote a book called Against Excess. It's available on line.

    Matthew, if someone proposes the theory that driving your car off a cliff will improve your mood, scientific ethics does not require that you actually try it to find out. An experiment is designed to yield knowledge, and can easily be reversed if it doesn't work. Neither of those is true of Prop. 19.

  13. Mark, I'm not saying anything about Portuguese drug policy except that the graphs provided are worthless for assessing it. And that, even if they told us something coherent that we could agree on, it's not something particularly relevant to why *I* want drugs legalized.

    If a policy change resulted in a 1 for 1 substitution of increased deaths among drug users for decreased deaths among non-drug users gunned down in turf wars and misdirected non-knock raids, that would, from my perspective, be a vast step forward. I am not troubled by people dying from risks they voluntarily undertake, nearly as much as I am people making sensible choices, and dying anyway as collateral damage of the war on drugs.

  14. Er, where do your data come from? Dipping into the Euro website, a lot of the tables don´t have an entry for Portugal at all, including the summary of total drug-related deaths by country. A footnote cites ongoing problems in reconciling data from diferent agancies. Can we infer that the Portuguese data are actually of pretty poor quality, on top of the year-to year noise of a long-tail event?

    The long footnote to the two very different UK series suggests that the category ¨drug-related death¨ is inherently problematic. It´s not like homicides in crime statistics, which are more objective than other crime indicators. How much effort do emergency physicians put into the correct labelling of causes of death when the subjects arrived dead at the hospital?

  15. "An experiment is designed to yield knowledge, and can easily be reversed if it doesn’t work. Neither of those is true of Prop. 19."

    Cmon. We'll learn plenty from it and if there's any serious signs of trouble voters will kill it via the next ballot. Concerned CA voters can't let threats like pot sales and gay marriage stand for long.

  16. With 300 deaths representing 0.0028% of the population (and having no idea what contribution alcohol abuse/policy/prices had on these deaths), I don't see how anyone can infer anything at all from this stat (so I agree w/ the spirit of the post). If only we could measure the difference NOT gaining a criminal record has over the lifetimes of many thousands of non-problem illicit users who happen to get caught. My guess in is, in this job market, it's pretty big.

  17. I'm still wondering why the schizoid nature of the Portuguese phenomenon? Obviously there are other factors at work here. The law of unintended consequences is probably the strongest case against legalization/decriminalization/etc. I suppose that law could work backwards as well: criminalization has had many unintended consequences.

  18. Greetings RBCers:

    As Artor notes, I am having fun with statistics, and making fun of some people who have done exactly what I did above in their analyses of Portugal. If you trawl the European press you will see many an op-ed in respected outlets using drug-related death data to pronounce on the value of the Portuguese drug policy, as well as people grabbing subsets of data on other variables to support their ideological views.

    A broader debate on Portugal is worth having, even though my post was not intended to be contribute to it. To analyze that country's situation would require a good deal more data, which I may pull together someday but that wasn't my purpose here (but nonetheless, good debate among the commenters, informative to read).

  19. These data need to be normalized for population (converted to deaths per 100,000 of Portuguese population in each time period) or you can't say anything, for starts.

    Only if the population was stable would these particular numbers be worth considering further; as incomplete data, nothing intelligent can be said from them.

  20. It has to be stated that 2001 was the same year the US invaded the number 1 opium producing country in the world. I want to see the price/availability of heroin in Portugal at this same time. I'm noticing that the numbers seem to return to pre-2001 levels after we deemphasized Afghanistan. I'm betting drug-related deaths have little to do with legal status and a whole lot more to do with availability, which legal status can affect, but is far from the most important factor.

Comments are closed.