Article ID: | iaor20116873 |
Volume: | 186 |
Issue: | 1 |
Start Page Number: | 39 |
End Page Number: | 59 |
Publication Date: | Jun 2011 |
Journal: | Annals of Operations Research |
Authors: | Hausken Kjell, Bier M |
Keywords: | learning |
We present a novel model capable of distinguishing between the effects of negative incentives (‘sticks’) and positive incentives (‘carrots’) for influencing the behavior of intelligent and adaptable adversaries. Utilities are developed for the defender and the terrorist. The defender is assumed to have a unit cost of defense, and unit costs of providing negative and positive incentives. The terrorist likewise has a unit cost of attack, which may either increase or decrease if the defender provides negative incentives, and enjoys a unit benefit of positive incentives. We show that the potential for perverse effects of counterterrorism (e.g., the emergence of hatred) can cause defenders to rely on positive incentives and decrease their reliance on negative incentives at equilibrium, with use of negative incentives completely eliminated in situations where these would be moderately effective when applied. With low potential for perverse effects of counterterrorism, the defender should rely on effective negative incentives.