Optimization Of Calculating VaR And CVaR With Conditional Autoregressive Range Model

With the rapid development of science and technology and the rising economic level,various types of Financial derivatives in the financial market are emerging one after another,and financial markets are becoming more and more complex.Therefore,risk measurement becomes an indispensable tool for studying financial markets.VaR and CVaR,which are the mainstream method of risk measurement in the financial market,have been welcomed by the majority of risk managers.The key part of forecasting VaR and CVaR is modeling the distribution of return on assets and capturing the volatility process.The most popular model for forecasting volatility is the generalized autoregressive conditional heteroskedasticity model,which is simply refered to as GARCH model.However,the GARCH model is based on the rete of return,it can't describe the price fluctuations during the period,and the estimation of the volatility is not accurate enough.Hence,the conditional autoregressive range(CARR)model which combines the GARCH model with price range is purposed.Based on the CARR model and the traditional calculation methods of VaR and CVaR,this paper proposes a new method for calculating VaR and CVaR which is suitable for low-frequency data.Then,with the daily data of the CSI 300 stock index from January 6,2009 to March 22,2019,on the one hand,we verify the superiority of the CARR-VaR and CARR-CVaR models over the traditional VaR and CVaR models,and the GARCH model.On the other hand,assuming the distribution of the disturbance term to follows an exponential distribution,a Weibull distribution,a Gamma distribution with an unit mean,and making the distribution of the residual in the return model follows a normal distribution,a student T distribution and a GED distribution with a mean of 0 and a variance of 1,explores the performances of different types of CARR models in calculating VaR and CVaR.It provides a certain reference for financial risk regulators to measure risk.The CARR model is a very meaningful volatility estimation model,which makes up for the shortcomings of the traditional GARCH model in extracting price information during the period,and improves the accuracy of risk management.The calculation method of CARR-VaR and CARR-CVaR proposed in this paper are proposed makes full use of the advantages of CARR model on characterizing volatility and combines with the improved calculation method of VaR and CVaR,effectively improves the accuracy of financial risk measurement,provides a reference for the application of CARR model in the risk measurement field of China's stock market.At the same time,the dual monitoring method of VaR and CVaR also helps to manage risks more efficiently and accurately,and reduce unnecessary economic losses.