Optimal management of operational risk using a Bayesian Decision Network model.

The research represents the extension of a Bayesian Network (BN) model for operational risk quantification into a Bayesian Decision Network (BDN) incorporating decision and utility nodes for the purposes of managing operational risk within a derivatives trading environment. The resulting Bayesian Decision Network can be employed to simulate different operational events under alternative management actions to identify those operational risk management policies under which the expected utility for the organization is maximized. Industry practitioners in financial institutions are now turning to the problems of quantifying operational risks for the purposes of capital adequacy under the Basel II accord. Much of the proposed approaches to modelling and managing operational risk are still contentious. This research will provide a detailed case study of the application of a new technology, Bayesian Decision Networks, to support the management of operational risk. This will be the first research carried out on modelling operational risk using Bayesian Decision Networks carried out in Australia, in partnership with an Australian Financial Institution. The current literature in this area has a very short history with few current publications. This research will add to the limited extant literature