Does increasing the legal age for buying alcohol reduce traffic accidents?

Does increasing the legal drinking age reduce traffic accidents caused by young drivers? The idea is that if you increase the legal age at which people can drink, young people are going to quietly abide by the law, not do anything stupid, read the bible, contemplate their sinful natures, and stay out of trouble.

Hang on though, one thinks: drink-driving is already illegal at any age, so what exactly does one expect to change when one restricts the sale of alcohol to 21 years and over, instead of having the current age limit of 18? If you were worried about them breaking the law before, why would you think changing the drinking laws would help? Breaking 2 laws is harder than 1?

In a recent letter to the Medical Journal of Australia, Jason Lindo and Peter Siminski, two economists from Texas and Wollongong respectively, point out that the more recent and more authoritative economics studies find that raising the age limit on buying alcohol does not help reduce serious traffic accidents at all. They do this in reaction to a completely one-sided account by medics who call for the drinking age increase, citing mainly cross-sectional studies (find attached the letter by the two economists and the reply of the authors of the offending article, which basically admits the cherry-picking that they originally engaged in:Lindo and Siminski 2014 with Toumbourou et al reply).

Lindo and Siminski point out that in New South Wales, changes to drinking laws did not change the accident rate of young people. Neither did a recent reduction in the drinking age in New Zealand, where the drinking age reduced from 20 to 18, increase accident rates amongst the 18-19 year olds (their behaviour was changing already, but not after the law change). Moreover, they point to studies that show that people indeed do substitute alcohol for other drugs that also affect their driving, which helps explain why there is on balance neither a positive nor a negative effect on traffic accidents from changing the age drinking laws. The studies they quote, which include the only studies on Australia on this topic, used the latest techniques based on analysing changes in behaviour of young people just before and after the introduction of the laws, which is what one wants to do. Prior studies are less convincing because they compare behaviour between regions within a country or over long time periods, which comes with the problem that regions and periods differ for many other reasons than merely the drinking age.

More generally, one can doubt the wisdom of a puritanical attitude to alcohol simply by looking at differences across countries. Central Europe, and in particular France, Italy, Spain, and the other Southern European countries, have much more relaxed attitudes to alcohol, with kids learning much younger to be responsible with alcohol. The more repressive attitude in the UK and here in Australia, on the other hand, is associated with binge drinking, very high rates of teenage pregnancies, and extreme risk behaviour. Once the kids do get access to alcohol, often by illicit means as the forbidden fruit is made so enticing, they dont hold back, which should make one wonder about the wisdom of declaring the fruit so forbidden.

Lando and Siminski thus try to inject a bit of common sense and self-reflection into our debates on alcohol laws, apparently having to fight a rather puritanical bunch of medics that insists we cannot trust young people and should ban them from buying alcohol till they are 21. Yet, we allow those between 18 and 21 to drive, to vote, and to die for us as soldiers in foreign battles, but we are supposed to declare them incompetent when it comes to drinking?

Lando and Siminski are hence right, both on the latest science that says there is no real relation between the drinking age and traffic accidents, and on the larger issue of consumer choice: if we abandon the idea that all voters are equal and that we should proscribe the behaviour of some of them, where do we stop? Should we lock up all young people from the age of 15 to 25 to prevent them from doing anything we did ourselves but do not want them to do? I have heard medics argue this at conferences….

So it is a very paternalistic and holier-than-though brigade that wishes to control the lives of others, without any regard to the joy they are destroying, using selective studies to argue their case. Why did the MJA publish the original one-sided piece by medics, one wonders? Economists are right to resist such reckless and blinkered destruction of consumer surplus.

Author: paulfrijters

Professor of Wellbeing and Economics at the London School of Economics, Centre for Economic Performance

7 thoughts on “Does increasing the legal age for buying alcohol reduce traffic accidents?”

  1. The blood alcohol limit for P plate drivers is 0.0, so changing the legal drinking age would not be expected to have any effect anyway.


  2. I could not see where the authors admitted to cherry picking. They admitted to emphasising peer reviewed studies. Presumably the more recent unpublished research that Lindo and Siminski cite will soon be published and so the authors will no longer be able to ignore it.

    There is no doubt in my mind that road safety researchers (including health researchers in this area) risk becoming a puritanical cult. You may remember a n article I wrote here last year about how distance travelled is an important, and ignored, determinant of road risk. I circulated some of my ideas to leading traffic researchers. One very well-known researcher, while agreeing with me in principle, said

    “I have resisted saying all this in the media because I am very concerned by the public reaction…Putting these messages out which create more uncertainty will make the public more resistant to regulation.”!

    As you rightly say: their attitude is that the public cannot be trusted to make judgments when it comes to road safety.

    Anyway, on the issue of when to allowing drinking, I have often wondered about the wisdom of granting a drinking license and driving license (not to be confused with a drinking and driving license) on the same day. The Americans give you a drinking license 5 years after your driving license. I rather think it might make more sense to delay the driving licence, but parents would no doubt hate this!


    1. Hi Chris,


      Of course the authors dont say “yeah, we cherry-picked”. But they do respond by saying that part of the literature is inconclusive and they acknowledge knowing about the various studies that oppose their earlier argument. Hence, by implication, they knowingly cherry-picked first time round by only having a one-sided narrative ignoring the contrary evidence.

      Yes, there is something to be said for making it harder to get a driving license, particularly for motorbikes. The main group you then make life difficult for would be the rural population: hard to work as a farmer or in a small village and not be allowed to drive.


      1. It gets worse. The Boes & Stillman piece that they dismiss very specifically finds that the method they use fails in this context. Stillman uses NZ data to show that despite an RDD effect at the drinking age, diff-in-diff on an actual change in the drinking age shows basically no effect. It’s not the kind of working paper you should just ignore if your whole method hangs on the RDD.

        See here for discussion:


    2. Chris,
      “I have resisted saying all this in the media because I am very concerned by the public reaction…Putting these messages out which create more uncertainty will make the public more resistant to regulation.”

      I fear it is not just road safety advisers who fall into sin using this type of faux-game theory logic. Unfortunately, the Climategate emails show that many in the Catastrophic AGW crowd have ruinously harmed the integrity of the climate science profession by committing the same misdeed.

      In everyday life, the world really does works better if everyone plays by the Eighth Commandment.



  3. Agree that the naive view that banning things leads to good outcomes is dominating the ‘alcohol is bad’ industry. In fact, no one gives a hoot about alcohol – they don’t like the binge drinking-violence phenomena and the drink driving.

    But making alcohol more elusive and expensive makes it more attractive to use it as an excuse to rebel. If beers are $5 each, instead of $2 each, then I won’t drink a little regularly, I’ll make sure if I drink I get drunk.

    My views are here


  4. Hi Eric,

    Agreed. I find it a bad look for the MJA.
    Interestingly, Hicki was on the original medics article but he didn’t join in with the reply.


Comments are closed.