As several states ponder prison releases as a budget crisis measure, the National Center for State Courts - Center for Sentencing Initiatives embarks on a new effort to determine the impact of incarceration on subsequent offending. What is the marginal effect of incarceration - length of stay - on recidivism? Does incarceration shorten or lengthen residual criminal careers? Can policymakers be informed by research of the consequences of releasing some offenders before their sentences end, or of diverting them from prison in the first place? Are the answers to the last questions different in Texas, with 1 in 22 people under correctional control and 30% of that group in prison, versus a state like Oregon with 1 in 33 under control and 25% of that group in prison? (See One in 31, Pew Center for the States.) As Austin's own Dr. Bill Spelman put the issue in 2000 (27 Crime and Just. 419): "Over the past twenty years, the fifty American states have engaged in one of the great policy experiments of modern times . . . the states doubled their prison population, then doubled it again, increasing their costs by more than $20 billion per year. . . . [W]hether more prisons reduce crime matters less than how much." Enumerating the numerous methodological difficulties and varied results of studies, he concludes that "the results are sufficiently precise to conclude that most states should take a hard look before committing the money to build more prisons." (Spelman 422).
It turns out - of course - that there are robust academic debates in the substantial body of empirical research such as that displayed in volume 23 (2007) of the Journal of Quantitative Criminology, leading off with "Revisiting Incapacitation: Can We Generate New Estimates?" (Reuter & Bushway) The criminological approach relies on estimating lambda, the frequency of offending (so that, e.g., a year of incarceration/incapacitation prevents 10 crimes for an offender whose lambda = 10). The second, economics-based approach, encompasses incapacitation effects plus deterrent effects of prison (minus the effects of offenders being replaced by new offenders on the streets, as in drug dealing, and minus any criminogenic effects of incarceration) by generating overall estimates of the impact of prison on crime (and thus the ability to do cost-benefit analysis of the new prison beds). My favorite piece of the latter sort is "The Silence of the Lambdas: Deterring Incapacitation Research" (Miles & Ludwig 2007), which details the problems and limitations with estimating lambda and the lengths of criminal careers, and discourages criminologists from pursuing such lines of research. The National Center's current effort, which I have been privileged to observe as it takes shape, will systematically review the state of current knowledge and commission new research to try and provide policymakers with a more definitive understanding of the dilemma that some states already face, but virtually all should ponder.