Capital punishment and deterrence: a portfolio approach

Abstract

Econometric studies of the deterrence effect of capital punishment have developed models of the supply of criminal homicides which include a variable to measure the probability of capital punishment. The present study, in contrast, employs a portfolio approach in which a model is developed realting changes in the rate of homicides across states to changes in the rate of a portfolio of all reported criminal offences. The effects of executions on changes in homicide rates are measured by changes in the nature of the relationship between homicide and the portfolio and abnormal variations in the residual of a model developed in the absence of executions. Using data from 1983–1988 across states and the District of Columbia, both the nature of the relationship and the residual variation are found to be uniquely altered in a manner consistent with the deterrence hypothesis. The results of the present study provide insight into why some studies of capital punishment may have failed to find appropriate disturbances in the rate of homicide. Disturbances may be observed only after accounting for changes in the crime portfolio. That is, those conditions that cause crime rates to rise in general may induce increases in the homicide rate that overwhelm the negative effect produced by executions.