Certainty, Severity, and their Relative Deterrent Effects: Questioning the Implications of the Role of Risk in Criminal Deterrence

Abstract. In the late 60s, Gary Becker incorporated into his formal model of deterrence theory an explicit statement that the theory´s components—certainty and severity of punishment—are more or less influential than one another depending on an individual´s preference for risk. The certainty of punishment is more influential than the severity of punishment in the decision of whether or not to commit crime if an individual is risk acceptant; if a criminal is risk averse, then the severity of punishment is more important than the certainty of punishment. Many aggregate deterrence studies arrive at estimates that reveal varying effects of the certainty and severity components of deterrence theory, with the certainty of punishment carrying the greater, and many times the only, weight. Leaning on Becker´s extension of deterrence theory, empiricists assume that criminals have a preference for risk. Assertions that arrests and convictions are greater deterrent tools imply important worldly consequences because they indicate to governmental authorities where resources should be invested to insure the best deterrent payoff. In this paper, I question both the need to take risk into consideration in aggregate level deterrence studies and the empirical evidence that has been offered in support of attaching greater weight to the certainty of punishment. I show, first, that deterrence theory, from an applied policy standpoint, is encumbered through the explicit consideration of risk preferences. Next, I work through the algebra of the statistical formulations of deterrence models and demonstrate that the greater weight associated with certainty could well be an artifact of the mode