Operational Risk Measured by Bayesian Networks with a Poisson-Gamma Joint Distribution in a Financial Firm.

Abstract

Main objective is to quantifying capital requirements of Operational Risk based on Bayesian inference by using an operational risk advanced measurement model, particularly when historical information is not available for a typical Mexican financial institution. The model employs a conjugated Poisson-Gamma distribution and feeds from experts interviews information so parameters can be measured. Monte Carlo simulations based on an interval for experts expected value of a loss event were generated from which following results were collected: 1) operational risk value can be gotten with insufficient information at a 95% of confidence, 2) expected losses tend to increase when experts expected events increase as well, 3) a positive correlation between operational risk and experts expected events exist, 4) frequency and severity of losses are smaller at the beginning and higher as operational risk value is been approached, then both decrease again. Described results depend highly on assumptions model and experts opinion and information available. Methodology proposed stands for an operational risk advanced measurement, so a specific strategy can be formulated for the firm to avoid losses and therefore operational risk.