Information Theoretic Method for Estimating the Number of Crimes Averted by Incapacitation

Results indicated that although a fair amount of heterogeneity was found among offenders, there were few discernible differences across vendor subgroups defined using demographic characteristics. With the exception of gender, where incarcerating males was found to avert slightly more crime than incarcerating females, differences were negligible and inconsistent across race and ethnicity groups. There was a fair amount of variation among the estimated annual crimes averted by incapacitation across various States and across the two offense categories--crimes against persons and property related crimes. The limited set of analyses simulating the elasticities of the incapacitation effect to an increased prison term suggests that the gains to be made by further increases in prison terms are disproportional. The elasticities of the crimes against persons averted are larger than 1.0 for a few individuals, but are largely clustered around 0.9. For property related crimes, the point of diminishing marginal returns may have been crossed. Simulations suggest that for most individuals, a percent increase in prison term will yield a less than 1 percent increase in the number of crimes averted. Based on these simulations, it is reasonable to conjecture that, at least for property related crimes, reducing the incarceration rates of the large number of inmates may result in little or no reductions in the number of crimes averted by a station. Although the amount of production in prison term that may yield little or no reductions in public safety will vary tremendously among individuals and will require more detailed analysis, prospects of reducing incarceration expenses without reducing public safety is very appealing. Data were collected from 38,624 prisoners released from 15 State prisons in 1994 for a period of 3 years. Figures, tables, references, appendix