Search

Based on historical data of operational risk in Chinese commercial banks between 1994 and 2012, this paper introduces Lévy measure to describe discontinuous jumping behavior of operational losses, uses thinning method to simulate the dynamical process of operational risk losses and adopts Lévy Copula model considering frequency dependence and severity dependence simultaneously. In this paper, a dynamical operational risk model with time-varying parameters and time-varying correlation structure is given as well as the corresponding numerical experimental technology to calculate VaR and CVaR in different confidence levels. The empirical result shows that Lévy Copula model of which setting risk is decreased shows priority in describing the dependence structures between risk cells. In comparison with traditional Copula model, Lévy Copula model depicts dependence structure more explicitly and intensively lowering total risk capital and passes model robustness test as well. Furthermore, the dynamical Copula model displays trends of risk and reduces the estimation bias of risk capital deriving from time-varying parameters. [ABSTRACT FROM AUTHOR]