Student performance and legal access to marijuana

This post might be viewed as a sequel to this 2016 post on access to alcohol. In a recent article by Olivier Marie (Erasmus University Rotterdam) and Ulf Zölitz (Maastricht University), published in the journal Review of Economic Studies (ungated earlier version here), the authors report on an interested natural experiment in the Netherlands:

We exploit a temporary policy change in the city of Maastricht in the Netherlands, which locally restricted legal access to cannabis based on individuals’ nationality.

Specifically, access was restricted so that only citizens of the Netherlands, Germany or Belgium could purchase marijuana, and citizens of other countries could not. The authors then look at the impact of that policy on student performance (for over 4400 students) in the business school at Maastricht University. The find that:

...the temporary restriction of legal cannabis access had a strong positive effect on course grades of the affected individuals. On average, students performed 10.9% of a standard deviation better and were 5.4% more likely to pass courses when they were banned from entering cannabis shops... Sub-group analysis reveals that these effects are somewhat stronger for women than men and that they are driven by younger and lower performing students.

This is a useful study because it is difficult to find natural experiments where the change in legal access affects some groups but not others - most studies compare differences between areas with and without legal access, where spillover effects are likely to be meaningful. I note that the results suggest an effect that is quite a bit larger than the Lindo et al. paper on the impacts of alcohol on student performance I discussed back in 2016, but not too much larger than this 2011 study by Carrell et al. (ungated earlier version here), also on alcohol.

Some of Marie and Zölitz's other results are interesting as well, such as:

...previous research has documented that cannabis consumption most negatively influences quantitative thinking and math-based tasks... Therefore, we split all courses depending on whether they are described as requiring numerical skills or not and test whether such course grades are differentially affected. We find that the policy effect is 3.5 times larger for courses requiring numerical/mathematical skills: a result in line with the existing evidence on the association between cannabis use and cognitive functioning.

In their other results, they show the good news that there are no teacher effects - that is, there was no difference in student outcomes between those whose teachers were unable to purchase marijuana after the policy change and those who were still able to. So I guess if you want to pass an economics paper, it pays to stay off drugs, but it doesn't matter if your lecturers do.

[HT: Marginal Revolution, last August]

Read more:

• Student performance and legal access to alcohol