When a local prosecutor sends a convicted felon to prison, the cost of keeping him locked up--an average of $31,286 per year--is paid for entirely by the state, not the county where the prosecutor holds office. The problem with this setup, some argue, is that prosecutors end up enjoying a "correctional free lunch," meaning they can be extremely aggressive in their charging decisions without having to worry about how much it will cost the local taxpayers who elected them. If prosecutors were forced to take the cost of incarceration into account, the theory goes, there might not be 1.36 million people in America's state prisons.
Of course, if the locals pay all the costs of an incarceration, you might get "under zealous" prosecutors. Why? Because the benefits of prosecution don't all flow to the locals. Some of the benefits arguably go to people in other parts of the state and, maybe, other parts of the country.
I don't worry about this incentive in the other direction, though, when I look at police behavior and D.A. behavior. I think the downside from too few prosecutions is much less than the downside from too many. It seems unlikely that prosecutors would substitute by prosecuting more marijuana crimes and fewer murders.
I do have one problem, though, with the proposal from W. David Ball that the Slate author mentions. Neyfakh writes: "Ball argues that states should take the money they're currently spending on their prison systems, distribute it among counties based on their violent crime rate, and allow local decision-makers to spend it as they see fit." Professor Bell is proposing that the more violent crime a county has, the more money it would get. Do you see a problem?