Theoretical Research And Empirical Analysis On Long Memory Of Chinese Stock Market

After influenced by COVID-19 at the beginning of 2020,the Chinese stock market kept reaching new peaks as the epidemic was brought under control,and its effective return on the investment pushed the enthusiasm for the stock market to the new heights.Reviewing on the develop history of the Chinese stock market,you can clearly see a continues and prosperous booming history.However,under this grand occasion of “China Stock Mania”,the system of the Chinese stock market still remains imperfection with the low effectiveness in the financial markets,which is far from developed countries.By studying the long memory of the Chinese stock market,we can understand the development of it further,dig into the potential features of the stock market data,and provide some reference for stock price forecasting,risk preventing and investor decision.By using the closing price of Shanghai and Shenzhen composite index,this paper carried out the long memory analyzes on the low frequency,high frequency and extreme data.The main contents are as follows:1.Discussed the evolution and difference between the efficient market hypothesis and the fractal market hypothesis,and gave the analysis combined with the current situation of Chinese stock market.As we all know,it can be found that the Chinese stock market is not completely consistent with the efficient market under the efficient market hypothesis,combined with the current situation such as the imperfect relevant system of Chinese stock market,the uneven quality of shareholders,and the inefficiency of the market.With the proposing and developing of the fractal market theory,the fractal market hypothesis subsequently arised can better explain various phenomena of the Chinese stock market,which is close to the real market.This market hypothesis made it possible to predict the stock price.The research on the long memory of the stock market could also be used to judge the future trend of the stock market.2.Carried out the long memory analyzes on the processed sequences of the daily,weekly and monthly closing price of Shanghai and Shenzhen composite index from 2000 to 2020.By using the R/S method,MR/S method and DFA method to test and compare,the Hurst index H was all greater than 0.5;by using the R/S method,the Hurst index H range was[0.60,0.72];by using the MR/S method,the Hurst index H range was[0.57,0.67],and by using the DFA method,the Hurst index H range was[0.52,0.59].The logarithmic return series of the two indices have long memory,and with the lengthened value period,the short-term disturbance weakened,and its long memory also became stronger,which indicated that the whole Chinese stock market has long memory.3.In order to explore the characteristics of extreme volatility of the Chinese stock market,the time window was positioned for 4 days,and the maximum and minimum sequence of returns of Shanghai and Shenzhen composite index were obtained.By using the R/S method,MR/S method and DFA method to calculate the Hurst index H,the range of the result was[0.75,0.82],much larger than the value of Hurst index H calculated from traditional data.It shows that the extreme value sequence has stronger long memory,and the in-depth study of it can play a certain guiding significance for the prevention and control of the extreme volatility of the stock market and financial abnormities.4.High frequency financial data has been proved leaping.The fourth chapter of this paper used a kind of stochastic process with the characters of hopping and long memory at the same time,constructed the test statistics by using the asymptotic behavior of power variation,and used the one minute closing price of Shanghai and Shenzhen composite index for empirical test.The results show that two sequences have long memory,providing an alternative model for the research of high frequency financial data.