Hate Radio and the Genocide

Job market season is in full swing, and faculty is getting treated to what feels like week after week of five-course meals in a pretty good restaurant. I am organizing part of the circus this year as chair of the recruiting committee, and so I, if you want to stretch the analogy, I guess I am a bit like the Maitre d'... One of my (and everyone's) favourites this year was David Yanagizawa from the IIES in Stockholm. His job market paper looks at what explains the extent of the Rwandan genocide. It's a grim topic (maybe that's why he doesn't look very cheerful on his website pic). In particular, David is interested in the extent to which reception of a particular Hutu hate ratio station made people more likely to participate in attacks on the Tutsi minority. He uses the implications of geography for radio reception as a source of variation. The chart from his JMP illustrates the point. If you are in the "shadow" of a moutain range, you will not receive the radio signal, but chances are that you are very similar in other ways to people on the other side of the mountain. David finds that, if you were on the wrong side of the mountain and didn't get reception, you were also much less likely to go out and wield a machete against your neighbours. Radio in his model acts as a co-ordination device, and lowers the threshold for engaging in violence. While one can pick many nits with the paper and (in particular) the model, I thought it was one of the most imaginative exercises I had seen in a long time -- interesting question, interesting source of variation, potentially important results.