Operational risk modelling and organizational learning in structured finance operations: a Bayesian network approach

The business unit level operational risk manager is responsible for measuring, recording, predicting, communicating and controlling operational risks within their organizational units. In support of the risk manager’s role, Bayesian networks have been recommended as a tool for operational risk management. In this research, we describe the development of a Bayesian network model for predicting operational risks for a structured finance operations unit, operating within one of Australia’s major banks. In constructing the Bayesian network model, we adopt a previously developed structured causal network model (Sanford and Moosa, 2009). Given the limited availability of operational data, we rely on the elicitation of subjective a priori probabilities, sourced from domain experts. Sensitivity analysis is performed to validate the model against the beliefs of management and operational staff. To ensure ongoing improvement of the network model and to support organizational learning within the domain, we also investigate and discuss the learning functionality of the Bayesian Network tool.