Sex Offender Residential Movement Patterns: A Markov Chain Analysis

This article introduces a new approach to the analysis of changes in sex offender residences over time. Using a Markov chain framework, we analyze residential movement patterns of registered sex offenders in Hamilton County, Ohio, over a three-year period (2005-2007). Results indicate a 46 percent reduction in offenders violating spatial restriction zone policy as compared to a counterfactual case where offenders move as a function of housing distributions. Strong legacy effects are also found as offenders previously in violation of restriction policies move into other restricted zones at a higher rate than offenders who were previously in compliance with the policy. Parcels that previously were home to registered offenders also continue to attract offenders in future periods. Although we find differences in the probabilities of attracting offenders for parcels outside and inside restricted zones that are consistent with offender restrictive policies, these differences are actually significantly smaller than what holds under the counterfactual. Parcels in restricted zones continue to attract offenders at a higher rate than expected, despite the policy restrictions.