The minimum legal purchase age for alcohol and crime

On 1 December 1999, New Zealand lowered the minimum legal purchase age (MLPA) for alcohol from 20 years to 18 years. That set in place an interesting natural experiment for how legal access to alcohol affects young people's behaviour. I've blogged before about research showing the effects on hospitalisations (where there appears to have been an increase, but it's unclear how large or otherwise the increase was) and alcohol-involved motor vehicle crashes (where the evidence is weak). What about the effects on youth crime in New Zealand? After all, the research I referred to in this post a couple of months ago seemed to demonstrate about a 15.7 percent increase in crime at the age where alcohol becomes legally available (age 16 in Germany).

In a recent article published in the journal Oxford Bulletin of Economics and Statistics (open access), Kabir Dasgupta, Alexander Plum, and Christopher Erwin (all Auckland University of Technology) looked at how youth crime changed when the MLPA decreased (see also their non-technical summary in The Conversation). They use administrative data from Statistics New Zealand's Integrated Data Infrastructure, and looked at monthly periods for young people in 1994-1998 (when the MLPA was 20 years) and 2014-2018 (when the MLPA was 18 years). Specifically, they look at what happens to crime rates before and after age 20 for both groups, using a regression discontinuity design. They also employ a difference-in-differences approach, comparing youth aged 18-19 years with youth aged 20-22 years, before and after the change in MLPA.

Their first method (regression discontinuity) is probably best summarised by Figure 1 from the paper. First, the results for when the MLPA was 20 years (in 1994-1998):

Notice that there is a big decrease in court charges (their measure of crime) at age 20, but that drop is entirely driven by a decrease in age-dependent traffic offenses (which are mostly violations of drink-driving laws, where the breath alcohol limit differs for drivers younger than age 20). Indeed, their regression analysis shows that:

Overall, focusing on the more comparable age-independent crime indicators, we do not find any significant change in alcohol-related crimes when the MLPA was 20.

As for the more recent sample, when the MLPA was 18 years (in 2014-2018):

The pattern looks fairly similar, with the same drop in crime at age 20, driven by a large decrease in age-dependent traffic offenses. At first glance, there doesn't appear to be an up-tick in crime at age 18 (which is the MLPA in this sample). However, based on the regression analysis Dasgupta note that:

At the relevant MLPA threshold, we do not find any statistically discernible change either in the overall measure or in the age-independent measure of alcohol-related crime...

Looking at alcohol-induced traffic convictions, we do not observe any significant effect for the age-independent measure, which excludes BBAC limit violations. There is however a statistically significant (at the 1% level) jump in age-dependent traffic convictions at the 18-year age threshold. Specifically, an increase of approximately 7.5 convictions per 100,000 population ... indicates that gaining alcohol purchasing rights triggers a rise in infractions of mandated BBAC limits applicable to youth below 20. Compared with the sample mean of groups just under the relevant MLPA, the linear RD coefficient represents a 27 % (7.5×100/27.4) increase in alcohol-induced age-dependent traffic convictions.

So, there is a small increase in traffic violations at age 18 in this sample, which is likely to be mostly an increase in age-related drink driving offences. In further analysis, Dasgupta et al. show that:

...there is a significant increase in age-dependent traffic convictions among individuals residing in non-urban (rural) locations and those living in socio-economically more deprived neighbourhoods.

So, the results are concentrated among youth who live in poorer and more rural locations. However, before we get too carried away, we need to recognise that Dasgupta et al. are comparing a sample from 1994-1998 with a sample from 2014-2018. That wouldn't be my first choice of comparison group, not least because youth drinking norms have been changing over time (see here and here). It would be better to compare groups that are closer in time, and that's what Dasgupta et al. do in their difference-in-differences analysis (which they only present as a bit of a robustness check late in the paper). In that analysis, they report that:

...we do not find any significant evidence of an increase in criminal activities in the post-1999 period for young individuals who gained purchasing rights for the first time.

Dasgupta et al. don't actually show the results of that analysis in the paper, which is a bit disappointing, as I would have found it more compelling than the results they actually did report. Overall, they conclude that their results show:

...little evidence that late adolescents commit more alcohol-related crimes upon crossing over the legal purchasing age in New Zealand.

That's not quite true for when the drinking age is 18, but if most of the increase in crime at age 18 is drink driving violations that relate to a lower breath alcohol limit for those aged under 18, it isn't clear that there's a strong case for an increase in harm. However, although statistically insignificant, there is an apparent increase in crime overall at age 18 in the regression discontinuity results. It is relatively small, at 4 convictions per 100,000 population, or a 7.6 percent increase. That's not insubstantial. I'd be much more cautious about claiming that there is 'little evidence' of an increase in crime. With the prevalence of youth drinking way down, it's not at all clear how much statistical power an analysis that looks at all young people has. It would be interesting to see what the results would be, if limited only to youth who drink.

Read more:

• Changes in the drinking age and hospitalisations - big effect or small effect?
• Drinking behaviour, drink driving, and more on the drinking age
• The drinking age, prohibition, and alcohol-related harm in India
• Legal access to alcohol, teenage drinking, and crime

The drinking age, prohibition, and alcohol-related harm in India

India is an interesting research setting for investigating the effects of policies, because states can have very different policies in place. Consider alcohol: According to Wikipedia, alcohol is banned in the states of Bihar, Gujarat, Mizoram, and Nagaland, as well as most of the union territory of Lakshadweep. In states where alcohol is legal, the minimum legal drinking age (MLDA) varies from 18 years to 25 years. And the laws change relatively frequently. Mizoram banned alcohol most recently in 2019.

Indian states provide a lot of variation to use for testing the effects of alcohol regulation. And that is what this 2019 article by Dara Lee Luca (Mathematica Policy Research), Emily Owens (University of California, Irvine), and Gunjan Sharma (Sacred Heart University), published in the IZA Journal of Development and Migration (open access), takes advantage of. They first collated exhaustive data on alcohol regulations changes at the state level, focusing on prohibition and changes in the MLDA. They note that:

Between 1980 and 2008, the time frame for our analysis, the MLDA ranged from 18 to 25 years across the country, and some states had blanket prohibition policies. In addition, we identified six states that changed their MLDA at least once; Bihar increased its MLDA from 18 to 21 in 1985, and Tamil Nadu repealed prohibition and enacted an 18-year-old MLDA in 1990, then subsequently increased it to 21 in 2005. Andhra Pradesh and Haryana both enacted prohibitionary policies in 1995 (the MLDA in Andhra Pradesh had been 21, and 25 in Haryana) only to later repeal them in 1998 and 1999.

In all, Luca et al. have data on law changes in 18 states over the period from 1980 to 2009, and for 19 states in a more limited number of years. They then look at a number of different outcome variables, drawn from the 1998-1999 and 2005-2006 waves of the National Family Health Survey, as well as crimes and mortality data. They first show that:

...men who are legally allowed to drink are more likely to report drinking, and the relationship is statistically significant. Given that the mean of alcohol consumption for men in the data is approximately 24%, this 5 percentage point change in likelihood of drinking is substantial, representing a 22% increase in the likelihood of drinking.

So, alcohol regulation does affect drinking behaviour (which seems obvious, but is much less obvious for a developing country like India than it would be for most developed countries). Having established that alcohol consumption is related to regulation, Luca et al. then go on to find that:

...husbands who are legally allowed to drink are both substantially more likely to consume alcohol and commit domestic violence against their partners...

...policies restricting alcohol access may have a secondary social benefit of reducing some forms of violence against women, including molestation, sexual harassment, and cruelty by husband and relatives. At the same time, changes in the MLDA do not appear to be associated with reductions in criminal behavior more broadly. We find suggestive evidence that stricter regulation is associated with lower fatalities rates from motor vehicle accidents and alcohol consumption, but also deaths due to consuming spurious liquor (alcohol that is produced illicitly).

In other words, there is evidence that stricter alcohol regulations are associated with lower levels of alcohol related harm, particular domestic violence and violence against women. Now, these results aren't causal although they are consistent with a causal story. Interestingly, Luca et al. choose not to use instrumental variables analysis (which could provide causal evidence), because the regulations proved to only be weak instruments (and they were also worried about violations of the exclusion restriction, because changes in alcohol regulation might have direct impacts on criminal behaviour). Luca et al. still assert that their results 'suggest a causal channel', and to the extent that we accept that, it highlights the importance of alcohol regulation in minimising alcohol-related harm in a developing country context.

Student performance and legal access to alcohol

I recently read an interesting 2013 paper (ungated earlier version here) by Jason Lindo, Isaac Swensen, and Glen Waddell (all from University of Oregon), published in the Journal of Health Economics. In the paper the authors investigated the effect of attaining the minimum legal drinking age (which is 21 in the U.S.) on college students' academic performance. This is a fairly important question, since we'd like to know if drinking makes young people worse off (and if so, by how much), so knowing if it interferes with their studies (and if so, by how much) is one way of getting an answer to the broader question.

Lindo et al. used student-level academic transcript data from the University of Oregon, and compared "a student's grades after turning 21 to what would be expected based on his average prior performance and accumulated experience". Here's what they found:

The results from our preferred approach indicate that students' grades fall below their expected levels by approximately 0.03 standard deviations upon being able to drink legally, a modest amount compared to the 0.06 to 0.13 standard-deviation effect estimated in earlier research. The effect is statistically significant, manifests in the term a student turns 21, and persists into later academic terms. In addition, we find that the effects on academic performance are especially large for females, low-ability males, and males who are most likely from financially disadvantaged backgrounds.

It is worth repeating their main result with some emphasis added: Being legally able to drink is associated with lower academic performance by 0.03 standard deviations. In other words, this is one of those cases where the effect is statistically significant, but the size of the effect means that it is economically meaningless.

Having said that we don't know what the mean GPA or standard deviation of GPA are as they aren't reported in the paper, so we can't really evaluate how 'large' the effect is. However, the authors note that this is "the equivalent of causing a student to perform as if his or her SAT score were 20 points lower". Given the SAT has a standard deviation of about 100, this is equivalent to lowering their SAT scores by 0.2 standard deviations, i.e. the difference between being in the 61st percentile and the 64th percentile of the SAT distribution. That is, a pretty small effect.

It would be interesting to replicate this sort of analysis for New Zealand though, but in the context of the drinking age changing from 20 to 18 in the 1990s (if the academic data are available). It may be that there are greater effects on attaining the minimum legal drinking age for younger people, but I wouldn't bet too much on it.

Overall, file this paper under 'nothing much to see here'.

Drinking behaviour, drink driving, and more on the drinking age

I found this article on texting yourself to moderate your drinking behaviour interesting. From the article:

A University of Auckland researcher is about to begin a full research trial where the participants will get a text message like this, written by themselves, to remind them not to drink too much...

"My premise is most of us are reasonably intelligent people. So why don't we tap into that and allow people to create their own messages."

The premise is that people compose text messages to themselves that they will receive later in the night, reminding them not to over-indulge, and that messages from themselves are more likely to be successful than messages from others. Karen Renner (PhD candidate at the University of Auckland claimed that her initial study showed a 23 percent reduction in alcohol-related harms. I can't find any published study thus far, but if you're interested her research protocol is recorded here. It appears she is using YAAPST (Young Adult Alcohol Problem Severity Test), which "is a sensitive measure for mild alcohol-related consequences, such as hangover, feeling sick, being late for work/school, etc." Reducing hangovers by 23 percent is a useful outcome. I look forward to the results of the wider study, which you can join as a participant here.

However, while thinking about this study it is worth noting that even as "relatively intelligent" people we are notoriously bad at self-control. If we were rational decision-makers, we would realise that drinking too much increases the risk of getting ourselves (or others) into trouble. That's why behavioural economists advocate for pre-commitments - prior (and irreversible) actions that commit us to a certain course of action - Dan Ariely discusses a few pre-commitments here (wearing the granniest pair of granny underwear to ensure you won't bed a guy on the first date, priceless).

I wonder whether a text message to yourself constitutes a large enough pre-commitment to modify your behaviour. There is little cost to you of ignoring a text message, not matter how coarse the language you use. In other words, it's not binding (see here for other examples). An interesting alternative might be, if you're not home by some self-imposed curfew your home computer gives money to charity (unless you're home to stop it!), or to your ex or other person you'd rather not give money to. There's got to be an opportunity for a new app in there (and now I've posted the suggestion, if you develop one you ought to cut me in!).

Speaking of drinking behaviour, there is a recent paper in the Journal of Health Economics by Frank Sloan, Lindsey Eldred and Yanzhi Xu (all of Duke University), which looks at the behavioural economics of drink driving (gated, and I can't see an ungated version anywhere online). Using survey data from the U.S., they investigated a number of questions about the behaviour of drink drivers, including:

Question: "Does the cognitive ability of persons who report they drank and drove in the past year differ from those who did not? Perhaps drinking and driving is a byproduct of cognitive deficits."

Answer: Possibly not. Drink drivers differed in cognitive ability from non-drink-drivers in only one of their three measures of cognitive ability (self-reported memory).

Question: "Is such behavior attributable to lack of knowledge of DWI laws? One reason for lack of knowledge is that the cost of acquiring the requisite information may be higher for some individuals in part because of lower cognitive ability."

Answer: Drink drivers actually have higher knowledge of the DWI laws than non-drink-drivers, so lack of knowledge doesn't explain drink driving.

Question: "Do drinker drivers lack self-control, as indicated by a lower propensity to plan for the future and by greater overall impulsivity?"

Answer: Yes. Drink drivers are more impulsive, and less prone to plan events involving drinking (such as selecting a designated driver in advance). They also find that drink drivers have higher rates of time preference (they place higher values on the present relative to the future than others), and some evidence of time inconsistency and hyperbolic discounting.

One last bit of interest I noted from the article was this:

According to our survey findings, the probability of arrest for DWI, conditional on driving after having had too much to drink is 0.008. Considering the probability of prosecution and conviction for DWI, the probability of a DWI conviction given a drinking and driving episode is about 0.006.

Given that low probability of receiving a penalty, even a rational decision-maker might find that the costs of drink-driving (penalty for drink-driving multiplied by the low risk of being caught and penalised) was lower than the benefits. So it need not be the case that we assume that drink-drivers are irrational to explain their behaviour.

In terms of reducing drink driving, while pre-commitment might seem attractive as a solution (since drink drivers are more likely to lack self-control), it might not work well since drink drivers are less likely to plan ahead and create a pre-commitment not to drink-drive. Forcing the pre-commitment onto recidivist drink drivers (like ignition interlocks, and related driver licensing changes) would seem like an appropriate intervention based on these results.

On a somewhat related note, the week before last I posted a comparison of two studies examining the effects of the change in the drinking age on hospitalisations in New Zealand. Just days after that post, a new paper by Taisia Huckle and Karl Parker (both of Massey University) looking at the effect of the change in the drinking age on alcohol-involved crashes was released by the American Journal of Public Health (ungated PDF version here). Here's the abstract:

Objectives. We assessed the long-term effect of lowering the minimum purchase age for alcohol from age 20 to age 18 years on alcohol-involved crashes in New Zealand.

Methods. We modeled ratios of drivers in alcohol-involved crashes to drivers in non-alcohol-involved crashes by age group in 3 time periods using logistic regression, controlling for gender and adjusting for multiple comparisons.

Results. Before the law change, drivers aged 18 to 19 and 20 to 24 years had similar odds of an alcohol-involved crash (P = .1). Directly following the law change, drivers aged 18 to 19 years had a 15% higher odds of being in an alcohol-involved crash than did drivers aged 20 to 24 years (P = .038). In the long term, drivers aged 18 to 19 years had 21% higher odds of an alcohol-involved crash than did the age control group (P ≤ .001). We found no effects for fatal alcohol-involved crashes alone and no trickle-down effects for the youngest group.

Conclusions. Lowering the purchase age for alcohol was associated with a long-term impact on alcohol-involved crashes among drivers aged 18 to 19 years. Raising the minimum purchase age for alcohol would be appropriate.

The paper was covered by the NZ Herald here. Some of the problems with the paper have been discussed by Thomas Lumley at StatsChat and by Eric Crampton at Offsetting Behaviour, so I won't bother going over the same ground again.

Changes in the drinking age and hospitalisations - big effect or small effect?

I've been meaning to blog on this for a while. Last year there were two papers published both investigating the effect of the decrease in the legal drinking age [*] from 20 to 18 on alcohol-related hospitalisations in New Zealand. The decrease in drinking age kicked in on 1 December 1999, so there are plenty of data available now to evaluate its effects.

The first paper is this one in the Journal of Health Economics by Emily Conover of Hamilton College and Dean Scrimgeour of Colgate University (earlier ungated version here). The second paper is this IZA Discussion Paper by Stefan Boes of the University of Lucerne and Steve Stillman of the University of Otago (hereafter B&S).

Remarkably, both papers use almost the same dataset and almost the same methods. While it's not unusual for several research teams to be working on the same research question at the same time, it's surprising to see too such similar papers appear in quick succession. And with similar results, though interpreted in meaningfully different ways.

Conover and Scrimgeour (hereafter C&S) look specifically at the health impacts of the law change. They used NZ Health Information Service (NZHIS) data covering all hospitalisations in New Zealand over the period 1993 to 2006. They use both a difference-in-difference approach (DiD), and a regression discontinuity design approach (RDD), to evaluate the difference in alcohol-related hospitalisations for those aged 18-19 years before/after 1 December 1999 (they also look at impacts on those aged 15-17 years). They estimate their models separately for each sex, and in both cases their control group is those aged 20-23 years (though they also present results that use a control group of those aged 20-21 years).

Boes and Stillman (hereafter B&S) look at a broader set of impacts than simply hospitalisations, including self-reported alcohol consumption, and alcohol-related motor vehicle accidents. They also use NZHIS data, over the period 1996 to 2007, and use similar DiD and RDD approaches to C&S. The key age group is again those aged 18-19 years (and they also look at those aged 15-17 years), but they don't present separate results by sex, and their control group is those aged 20-21 years.

The results are interesting. Using their preferred DiD approach, C&S find:

...a significant increase in hospitalizations as a consequence of passage of the Sale of Liquor Amendment Act (1999). Among eighteen and nineteen year old males difference-in-difference estimates indicate a 24.6% (s.e. = 5.5%) increase in alcohol-related hospitalizations. For females in the same age group, the estimated effect is 22% (s.e. = 8.1%).

They find qualitatively similar results using RDD, and conclude (in the abstract) that this shows "a substantial increase in alcohol-related hospitalisations among those newly eligible to purchase liquor". And you'd have to agree, the relative effects seem quite large - more than a 20% increase in hospitalisations for both young men and young women is pretty substantial.

B&S find even larger relative effects:

Overall, our results imply that the reduction of the [drinking age] from 20 to 18 led to a 75-91% increase in alcohol related hospital admission rates for 18-19 year olds.

These are BIG effects, right? In relative terms, yes. But in absolute terms, B&S point out just how many additional hospitalisations their effects translate into:

Translating the relative impacts using population figures from 1999 implies that the reduction in the [drinking age] from 20 to 18 led to approximately an additional... 2.1-2.6 [admissions] per month for 18-19 year olds... in the immediate aftermath of the law change.

So, across the country as a whole that would be an additional 25-31 alcohol-related hospitalisations per year. And remember, C&S found smaller effects than B&S. So, the effect is only very SMALL in absolute terms.

But that's not the end of this story. Aside from the period of the data and the control group, part of the difference in the two papers lies in their choice of what constitutes an 'alcohol-related hospitalisation'. C&S use all ICD-9 or ICD-10 codes that mention alcohol. B&S use a smaller subset of the ICD-9 codes, limited to alcohol use disorder (code 3050). In comparing their results with C&S, B&S argue that "...we use a much narrower definition of alcohol-related admissions that intentionally excludes more chronic conditions, which we show below are unresponsive to the reduction in the [drinking age]."

But there is a problem with both studies. In both cases, the hospitalisation data used was based on a limited set of ICD-9 or ICD-10 codes that are directly related to alcohol. However, there are a large number of ICD codes that could be related to alcohol, but are not alcohol use disorder and don't explicitly mention alcohol. For instance, the most obvious example are injuries or assaults that would be coded to S or T or X codes, none of which mention alcohol. Having spoken to a couple of clinicians, they pointed out that the ICD codes are very specific - unless they are sure the symptoms arise directly from alcohol, then the alcohol codes are not used.

Given that, it is likely that the small effects (in absolute terms) described by B&S may in fact be big effects if all of the hospitalisations that result from alcohol (including injuries, assaults, etc.) were included. Some estimates suggest that an average of 18 percent of emergency room admissions at peak party time are alcohol related. So, the jury has to remain out on the question of whether the decrease in the drinking age had a big effect or small effect on hospitalisations. More research required. [**]

*****

[*] Technically, there is no legal drinking age in New Zealand. Instead there is a minimum age for purchasing alcohol. I'm using the term 'drinking age' in this post purely for simplicity.

[**] Wellington Hospital has been collecting data on alcohol-related emergency room admissions for a number of years, even when the primary diagnosis is not alcohol-related (see here for example). As I understand it, most hospitals do this now, although completeness might be an issue. These data could be leveraged (along with associated time/day of admission data) to get an estimate of the proportion of ICD codes that are not included in the C&S study that might be alcohol-related. Imperfect of course, but possible.