Research On The Integrated Risk Measurement Of Chinese Commercial Banks And Its Sensitivity

In this paper, based on the financial data of Chinese commercial banks and the open data of the financial market, we model the market, credit and operational risk distributions respectively by Monte Carlo simulation with the risk factors model and the loss distribution approach. The joint risk distribution is directly constructed with the method of copulas. And VaR is used to measure and assess the total risk of Chinese commercial banks. Specifically, we examine the sensitivity of risk estimates to business mix and inter-risk correlation. The empirical results demonstrate that the more complicated copula-based approach is well for the integrated risk measurement of Chinese commercial banks. Add-VaR systematically overestimates total risk while N-VaR underestimates total risk. We find that the total risk is sensitive to the chosen level of market exposure. At the same time, because the operational risk distribution has much heavier tails than a normal distribution, we also find that risk is especially sensitive to the chosen level of operational exposure and correlation with market and credit risk. This could have important implications for risk managers and regulators since not only is operational risk difficult to measure, but it also requires special care in aggregation.