투자운영 위험률의 추정

Abstract

Operational risk is defined as a risk causing loss to investors resulting from inadequate processes, human resources, system, or external events. However, most operational risks happen in many cases combined with market risks or credit risks, and for the cases which rarely happen but cause a big loss, it is hard to develop an evaluation model on the incidence of operational risks due to the lack of data. This paper presents Bayesian probability model as a method of solving problems caused from the lack of data. The hazard rate of the event of loss in investment operation is estimated using exponential distribution on the period of occurring operational risks as a likelihood function and using gamma distribution in natural conjugate with the exponential distribution as a prior distribution on the parameters of the exponential distribution. It suggests a method showing subjective perspectives of experts as the prior distribution using bootstrap method, and presents a method estimating the ratio of risk after deriving an estimation function from the prior distribution and the likelihood function and using estimated Bayesian values of posterior distribution.