Geographic Profiling of Lone Wolf Terrorists : The Application of Economics, Game Theory and Prospect Theory

This paper presents an economic analysis of the choices made by lone wolf terrorists. Using RAND-MIPT data about the fatalities that are inflicted by different attack methods, the paper develops an analysis on a foundation of orthodox utility theory and Markowitz-Tobin approximations. This approach permits a computable opportunity set within a risk-reward or mean-variance framework. Optimal choices can be determined using the Markowitz quadratic programming technique. The framework may provide a useful foundation for an economic perspective on ‘offender profiling’ applied within a terrorism context. Mapping attack methods into mean-variance space provides a more definitive categorisation of the riskiness of attack methods from the terrorist’s perspective and suggests the possibility of identifying the terrorist’s revealed risk preference. Inferences about the unknown offender may be drawn that complement other aspects of the investigative process. One of the key challenges of law enforcement is drawing inferences about the offender’s location and the location of potential targets. Superimposing a game theoretical payoff matrix over a geographic location where payoffs are partially informed by the terrorist’s choices and risk preference may contribute another, economic, perspective to this part of the law enforcement process. Prospect theory may also contribute useful insights into the geographical profiling problem