Research On The Structure Break And Long-term Memory Of China's Stock Market

China's stock market is a complex system.Research on its behavioral characteristics has always been a hot issue and a focus area.Most of the early studies were based on the linear paradigm of efficient markets,but they were constantly questioned by the actual market conditions.Subsequently,fractal market theory came into being and created a new situation for the study of financial markets.This article is based on the fractal market theory and combines structural break theory to explore the structural break and long-term memory characteristics of the Shanghai and shenzhen stock market,and thus reveals the complexity of China's stock market and provide new suggestions for its healthy development.The main content of this paper is as follows:Firstly,Monte Carlo experiments are used to clarify that the existence of structural breaks can easily cause overestimation of memory parameters.Subsequently,it is detected that there are 6 variance structure breaks in both the returns of Shanghai Composite Index and Shenzhen Stock Index,and the relationship between structural breaks and major economic events is discussed.Secondly,starting from the non-normality and auto-correlation,the irrationality of the weak effective theory in China's stock market is demonstrated.On this basis,the single-fractal and multi-fractal measurement methods are introduced to estimate the long-term memory of the China's stock market,which fully shows that the Chinese stock market is a fractal market.In addition,although the memory parameters in each sub-sample interval based on the detected structural breaks are phased but all smaller than the whole sample interval,which proves that structural breaks may cause misjudgments on long-term memory parameters.Finally,the article verifies the predictability of the Shanghai and Shenzhen stock market price direction by using the symbol sequence method,and verifies the predictability of their trends by using the time-varying Hurst index.All of these proves that the most realistic value of the non-effective stock market is the ability topartially predict future stock prices.