The Stock Market Yields Long Memory Of Volatility Empirical Research

In the recent 20 years, long memory of economic, financial timeseries array become economics, finance study research focus of field.Have taken the lead in finding the memory for a long time (long memory)in the series of hydrology time from Hurst from the tide data. The study of long memory about the capital market is undoubtedlyimportant for analyzing and understanding market structure, judging thetendency of the market and have influence of market risk and suchchanging in the future ,etc.. Because of the existence of long memory isnot merely violate to market efficient theory, and that will be struck to thetraditional empirical approach. At home, more and more scholars and experts start the research ofthe long memory on the stock market of our country in recent years. Butbecause we start relatively late, present research is still at the stage ofdigesting and assimilate, there is a lot of place need the improved.Existing research is mainly to utilize ARFIMA and general GARCHmodel to carry on research to the volatility series modeling of the return.But no matter ARFIMA model or some GARCH models have a verygreat defect. And because of our country stock market special backgroundand market characteristic, make above-mentioned model can not fineportrayal domestic long memory nature characteristic of the returnvolatility of stock market. So, base on the concrete national conditions ofChina and strengthen the theoretical research to the domestic longmemory of capital market, combine domestic characteristic of stockmarket is it is it revise or select to model for use suitable model seemsvery urgent to be and necessary to go on to come. Because of this, this text reviews and comments relevant researchdocuments both at home and abroad at first, describe method of testingand the thought of modeling of long memory from the theory, lay thefoundation for the following research. Afterwards, we use one new testingand modeling thought to take the empirical study our country stockmarket return volatility series. Namely decompose series first, then totesting and modeling series after decomposing, to get more effective andmore accurate conclusion about examining and modeling. Finally , combine the data of U.S.A. and stock market of HongKong , compare with research the domestic stock market and U.S.A. ,Hong Kong stock market, see if look for the difference and weak pointbetween our country and other stock markets of developed country, offersome suggestions for the fact that the market perfects and develops. Our text divide into four chapters, the structure of every chapter andbasic content are as follows: Chapter one --Introduction This chapter probes into the research background of the thesis andmeaning of the selected title, and main research contents and researchapproach at first. Then the author has introduced the existing method oflong memory testing and modeling at first, and made relevantcommentaries and found all there is a weak point to a certain extent inthese methods, namely they can't totally dispel the influence of theshort-term memory impact on long memory. Then the author has donedetailed exposition and explanation to the definition and the reason oflong memory. Finally, I explain the key concept originally in the article--The regime-switching, and research the difference between theregime-switching that this text used and that of traditional Markov. Chapter two – the empirical examination of long memory of thereturn volatility of stock market in our country In this chapter, I have used ICSS algorithm to look for thebreakpoint of Shanghai and Shenzhen stock market volatility series atfirst; Segment according to breakpoint of the series, divide into differentregime, is based on this, dividing the series into one only including longmemory and one not including long memory; then we use the rescale R/Sto examine two series which has been decomposed, the result ofinspection shows the success in decomposition process. We made certainexplanation to the breakpoint and the get some conclusions finally. Cha...