Comparative Study On Volatility Forecasting Capability Of Shanghai And Shenzhen 300 Index

In recent years,with the acceleration of financial globalization,the rapid development of China's financial market has also been affected by various policy changes and information,and the variety of financial instruments and transaction rates have also grown rapidly,which has greatly increased risk and volatility of China's financial markets.So how to accurately portray and predict the volatility of financial markets is particularly important.Many scholars have constructed complex financial models which are expected to better estimate volatility.The autoregressive conditional heteroskedasticity(ARCH)model proposed by Engle Fitting the volatility into a new phase is a good way to portray the phenomenon of volatility clusters in financial markets.Subsequently,his student Bollerslev improved it and proposed a generalized autoregressive conditional heteroskedasticity(GARCH)model,which can fit the long-term autocorrelation process of heteroscedasticity with only few lag orders.It's found that the GARCH model is more accurate in measuring the general risk.However,with the progress of financial globalization,financial crises have also occurred frequently,such as the crisis of the European exchange rate system and the US subprime mortgage crisis that swept the world.Frequent financial crises are mainly due to inaccurate characterization of financial market volatility under extreme conditions and lack of corresponding risk measures.Therefore,how to effectively measure and controll financial risks under extreme conditions has become an important issue in financial research.Chou proposed a conditional autoregressive range(CARR)model by combining the range difference with the GARCH model.It is found that the model can better describe the volatility of financial markets under extreme conditions,and its risk measure in extreme cases is more accurate.In the financial market,there are many events that show the rate of return sometimes fluctuates violently,and the jump of asset profitability is very harmful.Because of its randomness and violent nature,it often leads investment institutions or individuals to huge losses,serious bankruptcies,even more will lead to the collapse of financial markets and thus affect the stability of society,so this paper uses the GARCH-Jump model to improve the traditional GARCH model to characterize the rate of return jump.The value at risk(VaR)is then used to measure and compare the risks under these models.Since the CSI 300 Index can reveal changes in market stock prices,it can also reflect the overall trend of the market.Therefore,this paper uses the 14-year daily closing,minimum and maximum price of Shanghai and Shenzhen 300 Index to calculate the rate of return and the range.The GARCH model,the GARCH-Jump model and the CARR model were built using the data of the previous 13 years,and then the corresponding VaR was calculated for the established model at different confidence levels using the data of the last year.This paper aims to explore the estimation and prediction of the volatility of the Shanghai and Shenzhen 300 Index by different models,and compare the model's superiority to the volatility estimation.VaR is used to compare the risks under different models,and to predict and portray the Shanghai and Shenzhen stock markets.The volatility of the stock market in Shanghai and Shenzhen provides a certain reference for reflecting,controlling and preventing the risk level of the stock market.This paper mainly studies from three aspects:(1)Estimation and prediction of the volatility of the Shanghai and Shenzhen 300 Index by the traditional GARCH model,GARCH-Jump model,and CARR model.(2)A comparative analysis of the volatility prediction of the Shanghai and Shenzhen 300 Index based on the GARCH-Jump model and the traditional GARCH model and CARR model.(3)Based on the VaR calculation results under these models,the accuracy and efficiency of the risk estimation of the Shanghai and Shenzhen stock markets are compared.This paper draws the following conclusions through the above three aspects of research:(1)Under the condition of 95% confidence level,the VaR failure rate of the Shanghai-Shenzhen 300 index yield based on the GARCH model is generally less than 5%.Under the confidence level of 99%,the VaR failure rate under the GARCH model is generally greater than 1%,it can be concluded that the GARCH model only depicts the risks in the general situation in the Shanghai and Shenzhen stock markets,among GARCH model,the more accurate description of the risk of the Shanghai and Shenzhen stock markets is MGARCH-GED.(2)Under the condition of 95% confidence level,the VaR failure rate of the Shanghai and Shenzhen 300 index yields based on the adjusted GARCH-Jump model is close to 5%,and the risk of the Shanghai and Shenzhen stock markets is very accurate.Under the confidence level of 99%,the VaR failure rate under the GARCH-Jump model is greater than 1%,possibly because the model jump strength parameter still does not accurately extract the price changes.(3)Under the condition of 95% confidence level,the VaR failure rate of the Shanghai-Shenzhen 300 index based on the CARR-W model is very close to 5%,so the risk of the Shanghai and Shenzhen stock markets is very accurate.However,under the 99% confidence level,the VaR under the CARR-W model has a large deviation from the risk of the Shanghai and Shenzhen stock markets.The failure rate of CARR-E at 95% and 99% confidence level is only 1.64% and 0.41%,which seriously overestimates the risk because the exponential distribution has no memory,and the tail of the distribution is exponentially decayed.It is also a typical non-thick tail distribution,so it's impossible to accurately estimate the thick tail characteristics of the financial time series.After calculating the CARR model,the likelihood function value is larger than GARCH and GARCH-Jump models,MAE and RMSE are smaller than GARCH and GARCH-Jump models,we can infer that the CARR model is more accurate than other models.(4)It is reasonable to use GARCHM-GED or adjusted GARCH-Jump to measure the general risks in the Shanghai and Shenzhen stock markets,and the CARR model is used to measure the risks in extreme situations in the Shanghai and Shenzhen stock markets.In general,these three models can effectively control and manage the risks in the Shanghai and Shenzhen stock markets.