The Research On The Long-term Memory Of Stock Market In China

Efficient market hypothesis asserts that, in an efficient market, stock returns arerandom series with no memory statistically, while the fluctuations of returns have norules. So the investors can’t predict the future trends based on historical prices.However, more and more studies have found that the stock return series showslong-term memory. The existence of long-term memory is contrary to the efficientmarket theory. And the traditional model of short-memory is challenged. What’s more,the return rate and volatility will act up to certain rules. The investors can achievereturns in excess of average market returns with the rules. China’s stock market hasbeen working just for two decades. What are the basic characteristics of currentdevelopment and efficiency? How Shanghai and Shenzhen stock market affect eachother? These questions are worthy of further studies.In this paper, firstly explained in detail the definition and connotation oflong-term memory on the basis of previous research,and made comparative analysisamong three methods which are used to distinguish the long-term memory, describeseveral common long-term memory model and their application to provide theoreticalbasis for the following empirical analysis. Then, took the closing prices of ShanghaiComposite index and Shenzhen Component index from2002.1.4to2011.6.30and didsome simple analysis of their return rates, such as stationarity test andheteroscedasticity test to observe some basic features. In addition, used theauto-correlation diagram, classic R/S method and modified R/S method to test thememory of return rate and volatility, applied the ARFIMA model, EGARCH model,FIGARCH model and ARFIMA-FIGARCH model to the return rate and volatility, andgot the appropriate conclusions. Finally, based on the long-term memory model,thecorrelation and spillover effects between Shanghai and Shenzhen stock market wereresearched.The research shows that China’s stock market return series don’t obey normaldistribution but peak top and thick tail, while the random errors of return hassignificant heteroscedasticity. From the tests, we found that both the return rate andvolatility of Shanghai and Shenzhen stock market show significant long-term memoryand it was more striking in Shanghai than Shenzhen. This indicates that China’s stockmarket is still very immature and much less than an efficient market. By the analysis between Shanghai and Shenzhen, we draw the conclusion that the return coefficientschange over time, and correlation is very strong, what means that the transmission ofinformation is efficient. Meanwhile, the two markets have bi-directional volatilityspillover effects and fluctuations of Shanghai impact Shenzhen strongly, with high rateof variance contribution. And conversely, Shenzhen market has low rate of variancecontribution to Shanghai.