Empirical Analysis Of Long Memory Of Volatility In Stock Markets In China

In this thesis volatility theories and some methods along with fitted models of long memory process discovered in the developed countries are used for empirical analysis of volatility of stock markets in China, and, then, some comparisons with fitted models are also made for the behavior of volatility in China stock markets to be well understood.Firstly, based on the relevant theories and the internal and external research statements, emphasis is put on the introduction to a series of testing methods, typical fitted models and parametric estimation means of the long memory process in this thesis.Secondly, four substitutes for volatility are discussed and the corresponding sample data series from stock markets in China are used to test long memory in volatility, and the result reveals the presence of long memory in volatility in stock markets, but the degrees of long memory corresponding to substitutes is different from each other.Thirdly, comparisons are made among four kinds of ARCH fitted models, which all show a high persistence in volatility and FIARCH gives a better description of volatility in stock markets in China.At last, an analysis of GARCH, SWARCH and SWGARCH is made respectively from the point of switching regimes models, and the empirical evidence shows that switching regimes models are also able to give a better fitness of volatility, but a contradiction in the degrees of persistence of volatility is also found between switching regimes model and ARCH.