Combining probability distributions in operational risk management

Abstract

The major difficulties and challenges which financial institutions face are closely related to the estimation of their level of exposure to operational risk. In their efforts to estimate this level, they construct a probability distribution using expert judgment. In financial institutions however, no common method exists to combine probability distributions which are derived from multiple experts. Current methods are often considered to be too complex, time consuming and lack to address problems such as effectiveness, efficiency and satisfaction. Moreover, handling large volumes of data requires a great deal of coordination, especially when derived from multiple expert judgments. In this paper we propose a method for combining probability distributions which are derived from multiple experts. First, we discuss relevant background literature on operational risk and expert judgment. Next, we propose a method to elicit and combine probability distributions which are derived from multiple expert judgments in operational risk management. We applied this method to two case studies in a large Dutch Financial Institution and present our research results. The results indicate that this method is more effective, efficient and leads to more satisfaction when implemented in practice. Finally, the main research issues as well as future developments are discussed.

Goossens, Louis H. J., and Grinsven, van. Jürgen H. M., Combining Probability Distributions in Operational Risk Management. 10th International Probabilistic Safety Assessment & Management Conference. June 2010, Seattle, Washington USA