The Rap on Crime

You may have heard that Congress held hearings last week focusing on profane and violent lyrics in popular rap music. The Daily Show had a great segment on straight-laced Congressmen bashfully quoting and walking through some pretty hardcore lyrics. Despite the hilarity surrounding the Representatives struggling through the profanity, the implicit message behind their words is more serious and crystal clear: violent and profane imagery in rap spurs on violence and other illegal behaviors in society at large.

While some artists (most notably Dancing with the Stars ringer Master P) agree with this, several people at the hearing (read: industry executives) pointed out that Congress may be getting the direction of causality wrong. So the question is: Does rap induce violence or simply reflect it?

I wonder if we can chip away at this question with some data. It shouldn't be hard to get yearly statistics on different kinds of crime committed, sales of rap albums, and a bunch of socioeconomic indicators (per capita income, average age, population breakdown by race, etc). With multiple years of data, one could control for state-specific effects that don't change over time.

However, this alone likely won't be enough: we still can't say whether its changes in rap album sales or changes in violence (or both) that drive the story. What we really need is a variable that predicts rap sales but does not influence violent crimes through another independent channel. I wonder if rap, like grunge, started in one part of the country and slowly diffused out. If so, it might be possible to interact geographic distance from "epicenter" and share of young and minority population to form an instrumental variable (though the epicenter argument probably applies differentially to the spread of old school rap, and thus the local average treatment effect we would derive would not be policy relevant). One might also want to look at lagged rap sales as well.

Any thoughts?