Long Memory Models Research Based On Chinese Stock Market

As financial market is the core of modern economy, its complexity and volatilitywidely exist in the development phase of economy and society. Since2012, financialtransition and structural adjustment have become one of the key words in Chinaeconomic and financial development. How to take the advantageous position during theproceeding of finance internationalization and take the positive measures to reply to thevolatility of the capital market have become the urgent affairs in our development offinancial market.Many studies show that financial time series has long memory property. There stillexists remarkable autocorrelation sequence even the time interval is long. History willaffect on the characteristics of time series model in a long time, classic models cannotdescribe the financial time series accurately any more. On the other hand, the traditionalnormal distribution shows some limitations. This paper aims at the two problemsmentioned above, uses China stock market assets return rate series as an example, bycomparing the difference between different distribution fitting effect of long memorymodels, find tools that can describe the memory Chinese stock market well, in order tomanage Chinese stock market risk and help investors make decisions in theinvestment.This paper discusses how to use the long memory models to study the time series,making the Chinese stock market return series as an example and using several typicallong memory models with different distributions to analyze the volatility of time series.By comparing the differences of fitting effect of each model, the paper finds out whichmodel can be the best tool to describe whether the Chinese stock market has longmemory property, to measure the Chinese stock market risk and help investors makeinvestment decisions. The empirical results show that Shanghai stock market returnsseries present long memory. Besides, time and event have an influence on long memory.Finally, there is a summary of the paper and some insufficient and further researchdirections.