Last week, a New York Times editorial argued that privately-owned prisons “can cost more to operate than state-run prisons,” even though these prisons allegedly cherry-pick inmates by excluding sick inmates who may cost more. In questioning the wisdom of recent privatization efforts, the editorial cites a report published by the Arizona Department of Corrections, which found that “inmates in private prisons can cost as much as $1,600 more per year, while many cost about the same as they do in state-run prisons,”

This analysis, however, does not take into consideration the effect that privately owned prisons have on public corrections costs. A 2007 Vanderbilt study, for instance, found that “state correctional systems which use private prisons in addition to public prisons experience lower rates of growth in the cost of housing their public prisoners” in comparison to states without private prisons. In particular, between 1999 and 2004, the average cost of housing prisoners in a public facility grew by about 5 percent in states without a private prison. But, average costs increased by less than 2 percent in states with some prisoners housed in privately run prisons.

As the Vanderbilt study shows, the merit of privately-owned prisons should not only be measured by a cost-per-inmate comparison. Rather, as an April 2011 Reason report argues, Private-Public Partnerships (PPPs) can “provide an effective, cost-saving alternative for governments seeking to address significant capacity needs while taking pressure off their corrections budget.”

Nevertheless, there is some evidence suggesting that private prisons do cost less to operate than government-run prisons, producing savings between 5 and 20 percent. Of 28 studies reviewed by the Reason Foundation, 22 found that savings averaged about 15 percent, while the other six studies found the costs to be about the same. In Texas, for example, in 2008, it cost Texas taxpayers $47.50 a day on average to keep an offender in a state prison, whereas privately operated prisons cost the state only $36.10 per day.

Equally important, though, is determining the best way to measure, evaluate and reward the performance of the prison system. In addition, there is the question Megan McArdle addressed yesterday of when the government should contract out. An interesting and innovative British proposal profiled last year would create a system in which each warden or contractor would have a portfolio of inmates—as if they were investments. As the report says, “the goal is to achieve a positive return by reducing recidivism through the efficient allocation of resources and implementation of effective practices both during the incarceration and parole phases of the offender’s progression through the system.” In effect, the plan would fund prisons partly based on their results.

As McArdle wrote “I’d be much more sanguine about the prospects for an agency in either the prison or welfare system that got paid only if their “clients” stayed employed, out of jail, and drug free for a year after leaving the system.”