Unexpected Correlations: Lead and Crime, Coffee and AIDS

While I'm no great fan of Freakonomics in particular, I am a fan in general of economists uncovering hitherto unsuspected causations. There was a great one in the Washington Post on Sunday, about the link between lead exposure and crime. It seems that by far the best thing you can do if you want to bring your crime rate down is to switch to unleaded gasoline and then wait for 20 years.

Here's Emily Oster, with another interesting correlation: Uganda's rate of HIV infection, it turns out, is very highly correlated to the amount of coffee that it exports. The famous ABC campaign in the country did precede a fall in HIV infection, but it also preceded a fall in coffee exports, so maybe public education doesn't reduce AIDS in sub-Saharan Africa as much as we thought it did.

There's lots more where that came from in this talk, including the fact that if a sub-Saharan African nation doubles its exports, it will quadruple its rate of AIDS infections. Poverty reduction and AIDS reduction are not always, it would seem, the same thing.