Introduction

over the past twenty years, the fifty American states have engaged in one of the great policy experiments of modern times. In an attempt to reduce intolerably high levels of reported crime, the states doubled their prison populations, then doubled them again, increasing their costs by more than $20 billion per year. The states and the Federal government have given up a lot to get to this point: That $20 billion could provide child care for every family that cannot afford it, or a college education to every high school graduate, or a living-wage job to every unemployed youth. But crime levels appear to have (at last) responded, dropping to their lowest level in years. Thus recent history provides a prima facie case for the effectiveness of prisons.

Not everyone has found this evidence persuasive. Some argue, quite convincingly, that the recent crime reductions had nothing to do with the prison buildup. Crime dropped because the job prospects of povertystricken youths have improved, or because police have become more effective at getting weapons off the street, or because neighbors are beginning to watch out for one another again. As usual, correlation does not guarantee causation. If we are to determine the role of the prison buildup in the recent crime reductions, we will need to take a more systematic approach.