Young people do dumb things. Drunk people do dumb things. When people are both young and drunk, they do even dumber things. [*] In other words, when young people gain access to alcohol, their behaviour changes. I've posted before about various studies that compare young people before and after attaining the minimum legal drinking age. Studies have found that gaining legal access to alcohol lowers grades (but only by a little), increases hospitalisations, and increases alcohol-involved motor vehicle accidents.
Add to that list of effects of legal access to alcohol this recent article by Fabian Dehos (RWI-Leibniz Institute for Economic Research), published in the Journal of Health Economics (sorry, I don't see an ungated version online). Dehos looks at the impact of German teenagers attaining the minimum legal drinking age (of 16) on alcohol consumption and crime.
Dehos combines nationally representative survey data from the German Federal Centre for Health Education (covering 2005, 2007, 2008, 2011, and 2015) and from the European School Survey Project on Alcohol and Other Drugs (covering 2007 and 2011). The total sample size is over 20,000 people aged between 14.5 and 17.5 (1.5 years either side of the minimum legal drinking age).
Dehos then employs a regression discontinuity design, testing whether there is a discrete jump in alcohol consumption in crime that happens exactly at age 16. In terms of consumption, one of the key insights is summarised in Figure 1 Panel (a) from the paper:
Notice that alcohol consumption jumps up at age 16, both in terms of consumption within the last 30 days, and consumption within the last 7 days. That suggests strongly that young people start drinking more frequently when they get to age 16 (or, at least, they are more likely to report drinking in a survey when they get to age 16). It is especially noticeable that there is no jump in lifetime consumption of alcohol at age 16, which also suggests that drinking becomes more frequent at age 16. Dehos then shows quantitatively that there is:
...a 35% increase... of the overall drinking intensity at age 16. Similarly, there is a 20% increase in the number of drinking days within the last 7 days right at the cutoff...
Turning to crime, the key insights are summarised in Figure 4 Panel (a) from the paper:
Notice again the discrete jump in alcohol-influenced crimes at age 16. There is no such increase in non-alcohol-influenced crimes, which should give us some confidence that there is something special about age 16 that leads to an increase in alcohol-influenced crimes. It should be no surprise what the leading contender is for explaining that change! Quantitatively though, Dehos finds that:
The overall criminal engagement of alcohol-induced crimes increases by 11.7 offenses per 10,000 person-years at age 16... The jump of alcohol-induced offenses at the MLDA, thus interprets as a 15.7% increase... in criminal engagement under the influence.
Those are quite substantial changes in alcohol consumption and alcohol-related crime, and much larger than the effects found in the research I cited in the posts linked earlier. A possible explanation for that is that alcohol access at age 16 has a much greater impact on young people than alcohol access attained at age 21 (as in the US research that found a tiny effect on student academic performance) or at age 18 (as in the New Zealand research on hospitalisations and motor vehicle accidents).
On the other hand, perhaps police change the way that they handle crimes committed by people aged 16 (compared with younger offenders). Dehos tries to rule this out by showing that there is a statistically significant discontinuity at age 16 for a survey question on whether young people "had problems with the police within the last year because they had drunk". That is unconvincing to me (in fact, it makes the case for police treatment of offenders changing at age 16 even stronger). He also looks at a question on whether young people "were involved in a fight because they had drunk", and also finds a statistically significant increase at age 16. In this case, the effect cannot be related to police treatment of offenders, so that provides a little more confidence that Dehos is finding a real effect of alcohol consumption on crime (even though the true effect might be somewhat smaller than what he finds).
Overall, the takeaway messages from this paper are that alcohol consumption increases when alcohol becomes legally available to young people (no surprises there), and that young people are more involved in crime when alcohol becomes legally available to them. That by itself doesn't mean that the drinking age should be increased, only that there are benefits (in terms of lower crime) in doing so, which will need to be carefully weighed up against any costs. On the positive side, with decreasing alcohol use among young people (see here and here), perhaps youth crime will start to decrease?
*****
[*] Before anyone gets offended, remember that I was young once too. One of my favourite song lyrics is from Adema's The Way You Like It: "Nowadays, no one remembers when they were young and stupid". I do remember being young and stupid, so when I say that young and drunk people do dumb things, I'm speaking from personal experience.
No comments:
Post a Comment