Persistence of Financial Risk

This paper presents a new method for discussing persistence of financial risk based on volatility impulse response function. The new definition of volatility persistence and common persistence is established. Further more, the corresponding topics are introduced into the area of nonlinear common persistence through wavelet neural network. The proposed methods here are easy to be conducted, and can be extended to discuss the persistence of higher moments risk. The empirical results show that there exists volatility persistence in Chinese stock markets, which can not be removed by linear combination. However, based on wavelet neural network, a nonlinear combination is found to reduce the volatility persistence. This phenomenon is defined as nonlinear common persistence, which is very helpful in understanding of portfolios investigation