Research On The Nonlinear Feature Of The Stock Market In China

The subject of this thesis is to study the Chinese stock market's non-linear fea-tures. We investigate the volatility and risk of Shanghai Composite Index of Chinastock market. From the perspective of nonlinear models, we describe the China stockmarket's non-linear characterizations. Then, base our findings we analyse the impli-canons of VaR for finacial risk management in China stock market. The full thesis isdivided into five chapters:In the first chapter, we briefly introduce the research background and the presentsituation. In the second chapter, we introduce two familiar models (ARCH modeland GARCH model) and the relevant knowledge of VaR. But in financial markets, theheavy-tailed will underestimate VaR for finacial risk management. So we introduce1VARFCH model to describe the stock market and the heavy tail index estimation inchapter 3. Chapter 4, we do some simulations about the introduced model and someempirical analysis of Chinese stock markets. We obtain that in the 95% confidencelevel, even if the innovation is fit to student-t distribution of AR-GARCH model, theVaR estimated is underestimate. And then with the AR-NARFCH Model to estimatethe value of risk, we found that when -y=1/2, and the innovation follows student-tdistribution, the model can cover the actual loss. But it's too conservative when thedegree of freedom of student-t is from 3 to 5. While the degree of freedom of student-tis 6, the estimation of the value of risk is very near. Further more, we estimate the tailindex of stock return series through Sum-plot method. The stock return series exposeright skewness and the tail index of about 3. The last chapter is for the potentialproblems and where need to be improved.