Long-memory Research Of S&P/CITIC Stock Index

The long-memory research has always been the focus in the capital market area, especially in the fund market. Designing and operating the funds of different styles need to consider the characteristic of the stock market of different styles, whose structure and price movement will have influence on it. Within the researches on asset price volatility, long-memory researches had broken the assumption of Efficient Markets Hypothesis, and provided a new perspective for asset pricing and risk management. In order to reflect the characteristic of S&P/CITIC Pure Style Index, the paper will carry on the empirical analysis from the perspective of the long memory.The paper will start with the concept of Fractal Market Hypothesis and the development of the long-memory research. Secondly, two main methods from two different angles, R/S method and MF-DFA method, will be used to examine long memory in the part of empirical analysis. Finally, the path of the index will be predicted by Moving Hurst Index.So far, the research has drawn three main conclusion. First, based on the R/S method, Chinese fund style asset return series has a genuine long-memory and non-periodicity features. Second, based on the MF-DFA method, Minor fluctuation of these series have a persistent feature while large fluctuations have inverse. And this two points above can be used by fund companies and managers to form moderated style-drifted strategies to gain short-term abnormal return. Third, as a indicator of judging the market operation, Moving Hurst Index can predict the future trend of the market roughly.