A few days ago, Levitt blogged about an interesting study finding that violent movies reduce crime (at least in the short run). The reason, according to the study’s authors, Gordon Dahl and Stefano DellaVigna, is simply that more violent movies means fewer drunken louts on the streets. It is simply an incapacitation effect.
One way of testing this hypothesis would be to look for the opposite type of experiment: What puts more drunken louts on the streets?
Two words: college football. Or, at least, that is the focus of a new study by Daniel Rees and Kevin Schnepel linking crime and sporting events. They analyze daily crime data, but instead of analyzing the changes around the release of new movies, they look to see what happens on game day. Their findings are quite striking, and they report large rises in assaults, vandalism, and disorderly conduct on game days. As might be expected, this effect is large in the city of the home team, but basically non-existent in the city of the visitors.
You might be worried that this rise in arrests reflects more police on the street on game day (and hence more arrests per crime), rather than simply more crime. But the authors provide a clever response, noting that upset losses by the home team have a particularly large effect on violent assaults, while expected losses have little effect. Unless police chiefs are also successfully forecasting football outcomes, it seems that this alternative explanation doesn’t hold water.
The effects here are pretty large, and the study is quite convincing. It is worth noting that these results occur despite the fact that the football programs they analyze ban the sale of alcohol in the stadium.
There’s a nice back story here, too. Schnepel is an undergraduate student, who got to thinking about this question while sitting in Rees’ class on the economics of crime. Soon enough, an idea became some explanatory regressions, and then a co-authorship, and now, an interesting paper. And all this happened in time for Schnepel to boast about it on his grad school application (which I look forward to reading!). [Correction: Schnepel is in fact a current graduate student in economics at the University of Colorado, Denver. Though we'd still love to get a look at his application.]
Putting the Rees-Schnepel and the Dahl-DellaVigna studies together, the policy implications seem pretty obvious: we want to engage aggressive young men in activities that amuse them (like movies), but don’t provide an outlet for violence (like a football game). This seems pretty obvious, but then it got me wondering: why do violent movies lead us to sit quietly and stare at the screen, while football leads us to get out of our seats and begin the biffo? A general theory here may well suggest broader policy implications from these sorts of studies.