An Empirical Research On Volatility Of Chinese Stock Market Based On Markov Swtiching Vector Autoregressive Model And Investor Sentiment

As the largest investor in the world,China plays an important role in sustaining global economic growth.As an emerging securities market,Chinese securities market has made remarkable achievements in just over 30 years.Not only the trading technology has reached the world’s advanced level,but also a perfect regulatory and legal system has been established.However,compared with the mature securities market in the west,Chinese stock market still has some defects that can not be ignored.As a result,the whole market is prone to irrationality.Individual investors lack the understanding of the information of listed companies and thus rely on market sentiment to evaluate companies and amplify the impact of investor sentiment.Therefore,it is of great significance to study the relationship between the fluctuation of Chinese stock market and investor sentiment.Investor sentiment is a very important link since the birth and development of behavioral finance.In recent years,many foreign scholars have done in-depth research on it and found that investor sentiment is indeed an important reason affecting stock market fluctuations.However,considering the national conditions and the differences between Chinese stock market and the mature stock market of western developed countries,the research methods can not be completely copied.On the basis of studying foreign research results,we constructs composite investor sentiment index considering domestic conditions,market characteristics and data availability in this thesis.In researching the volatility of Chinese stock market,many domestic scholars began to turn to the research of Markov regime switching model.which replaced ARCH model.In this thesis,the volatility of the stock market,is studied by using the Markov regime switching vector autoregression model(MSVAR).On this basis,investor sentiment is considered as an endogenous variable to characterize the volatility of the stock market in China.First of all,we select 7 single sentiment indicators and adopt principal component analysis to construct composite investor sentiment index CISI.Composite investor sentiment index is composed of four principal components,which accumulatively contain 88.127%of the information of the original 7 variables.Therefore,the four principal components can well summarize the original 7 groups of data.At the same time,the correlation between CISI index and CSI 300 stock index reached 0.657,and the correlation was significant,indicating the effectiveness of the index.Secondly,the lag order of MSVAR model is determined as 3 by establishing VAR model in this paper.and the established VAR model passes the stability test of the model,laying a foundation for the subsequent MSVAR model.Finally,according to SC information criterion.AIC information criterion,HQ criterion and logarithmic likelihood value.MSIH(2)-VAR(3)model is selected as the optimal Markov regime switching vector autoregressive model for modeling,and the regime attribute,regime transition probability matrix and regime probability graph of model parameters are obtained.After analyzing the attributes and probability charts of each zone system,we come to the conclusion that Chinese stock market is mainly in the period when the CSI 300 stock index is on the high side and investor sentiment is relatively high.The period of high sentiment is longer than the period of low sentiment,and each period is relatively stable without sudden and drastic changes.At present,Chinese stock market is in a period of relatively high investor sentiment,which has lasted for a long time before,and the overall trend of Chinese stock market is to be depressed in the future.