A Bayesian network structure for operational risk modelling in structured finance operations


Working Abstract: It is only recently that banks have begun to take seriously the measurement and management of operational risks. The inclusion of operational risk within Basel II is evidence of its increased importance. Although much previous research into operational risk has been directed at methods for determining levels of economic capital, our research is concerned more locally with the of an operational risk manager at a business unit level. The business unit level operational risk manager is required to measure, record, predict, communicate, analyse and control the operational risks within their unit. It has been proposed by a number of finance researchers that Bayesian Network models may be practical tools in supporting the risk manager’s role. In the following research, we take the first steps in developing a Bayesian network structure to support the modelling of the operational risk. The problem domain is a functioning structured finance unit within a major Australian bank. The network model incorporates a number of existing human factor frameworks to support the modelling of human error and operational risk events within the domain. The network supports a modular structure, allowing for the inclusion of many different operational risk event types, making it adaptable to many different operational risk environments.

DOI: 10.1057/jors.2011.7

Extracted Key Phrases

4 Figures and Tables

Cite this paper

@article{Sanford2012ABN, title={A Bayesian network structure for operational risk modelling in structured finance operations}, author={Andrew D. Sanford and Imad A. Moosa}, journal={JORS}, year={2012}, volume={63}, pages={431-444} }