The Process of Policy Innovation: Prison Sitings in Rural North Carolina

We gauge the relative impact of economics, demographics, and politics on the decisions of 79 rural North Carolina counties whether to site a prison in the period 1970-2000. The results of this model demonstrate that, contrary to the expectation that counties site prisons in response to economic distress, the demographic characteristics of each county affect the relative likelihood of a prison siting more than its economics does. The influential demographic predictors are those Inextricably bound up with development options-the education levels of its citizens-and those that limit its ability to pursue controversial projects-its not-in-my-backyard (NIMBY) constituencies-rather than those that measure its racial diversity. Therefore, prison siting is neither a simple story of economic determinism nor one of environmental racism. We use a proportional hazards regression to model this innovation adoption in response to the challenge to select methods that take the potential time dependence of adoptions into account. A duration model is also particularly suitable for cases such as this one for which the innovation adoption is better understood as a process than as an event.