Short-term Effects of Executions on Homicides in Five States: A Dose Response Function Analysis

• In Event: Considering Context in Capital Punishment

Abstract

Studies using time series panel data have reached mixed results regarding whether the death penalty and/or executions deter homicides. They are also subject to critiques that the findings from these studies are not robust to model specifications such that even small changes in specifications yield dramatically different results. Land et al. (2009) claim a principal reason for the sensitivity of the findings to model specifications is that there are very few state-years (about one percent of all state-years) in which there have been six or more executions. Focusing on Texas, a state that has used the death penalty with sufficient frequency, they find evidence of a modest short-term deterrent effect of reductions. In this study, we conceptualize a type of dose-response function in the sense that the response (decrease in numbers of monthly homicides) goes up with the intensity of the dose (numbers of executions per month). If the dose is not sufficiently high, the response will be null or at least not statistically significant. We replicate Land et al. (2009) with executions - homicides monthly data from 1994 to 2014 in Florida, Oklahoma, Missouri, and Virginia using dynamic regression (DR) - linear transfer function (LTF) models.

Authors

• Kenneth C Land, Duke University

• Raymond H. C. Teske, Sam Houston State University

• Hui Zheng, Ohio State University