Since nobody else has noted it on here yet I figured I might as well: Griffe's poked his head back up for the first time since the Summers flap, with an essay on "Politics, Prison and Race". Nothing particularly new here -- using similar methods as he's used in past essays, he posits a model to explain the ironic fact that black/white incarceration ratios in the US correlate positively with the "progressiveness" of states and their governments.

The short answer is the same as with sex differences in the mathematical sciences -- due to the difference at the tails of the distributions, the higher you set the threshold the greater the disparity will be. More "progressive" states tend to set higher thresholds for incarceration, ergo the higher B/W imprisonment ratio. It's straightforward, and given the crudeness of the model it fits the data passably well.

I know I'm not the only poster here who finds this sort of "ideal gas model" approach remarkably frustrating in its limitations. Suggestive though it may be, one could sit and poke holes in it all day, and ultimately the macro model persuades nobody until you've given it microfoundations...