Modelling operational losses: a bayesian approach.

The exposure of banks to operational risk is increased in the recent years. The Basel Committee on Banking Supervision (known as Basel II) has established a capital charge to cover operational risk other than credit and market risk. According to the advanced methods defined in “The New Basel Capital Accord” to quantify the capital charge, in this paper we shall present an Advanced Measurement Approach based on a Bayesian network model that estimates an internal measure of risk of the bank. One of the main problems to face to measure the operational risk is the scarcity of loss data. The methodology proposed solves this critical point because it allows a coherent integration, via Bayes’ theorem, of different sources of information, such as internal and external data, and opinion of ‘experts’ (process owners) about the frequency and the severity of each loss event. Furthermore, the model corrects the losses distribution considering the eventual relations between different nodes of the network, that represent the losses of each combination of business line/event type/bank/process and the effectiveness of the correspondent internal and external controls. The operational risk capital charge is quantified by multiplying the VaR per event, a percentile of the losses distribution determined, and an estimate of the number of losses that may occur in a given period. Furthermore, it becomes possible to monitor the effectiveness of the internal and external system controls, in place at the bank. The methodology we shall present in this document has been experimented, as a pilot project, in one of the most important Italian banking group, Monte dei Paschi di Siena (MPS).