One of the key lessons of the Iraq surge is that putting more boots on violent ground tends to reduce the violence. The same lesson applies in American cities, but for the most part, the lesson hasn’t been learned. For reasons that mystify me, the same state and federal governments that shower money on urban school systems give nearly nothing to urban police forces. That gets it backward: the correlation between more money and better schools is weak at best; the correlation between more cops and less crime is very strong.
There's much more to Professor Stuntz's post, including:
Overstretched big-city police forces tend to make lots of drug arrests, because those arrests are easy to make—and too few arrests for violent crimes, which require more manpower to investigate. Over time, those police forces have come to see drug punishment as a substitute for punishing violent crime. As the crime statistics of the last generation show, that substitution doesn’t work.
Perhaps this is true, but disentangling violence from drugs is a tall order: