Asymmetric Of Stock Returns Based On Bayesian Quantile Threshold Autoregressive Model

Moderate or reasonable fluctuations of stock price benefit the market to maintain moderate liquidity, which is a necessary condition for the healthy operation of the stock market. Conversely, excessive or unreasonable fluctuations of stock price will be distorted stock price mechanism, resulting that price can't truly reflect the listed company's intrinsic value and hinder the stock market to optimize the allocation of resources the core function effectively. In a long time, the domestic and international stock market share price volatility frequently, many studies showed that when "good news" and "bad news" and extreme events appear, the stock market generally exist different degree of volatility asymmetry phenomenon, and the general performance of the "leverage effect", namely bad news to the volatility of the stock market impact than the same degree of good news.The quantile regression theory framework based on considering points a regression method not only can describe the response of the central tendency of the variables, but also can describe the tail variable behavior. The integrated use of ADF test, BIC test, stability test method, from the stock market return spillover effects, extreme events, good news and bad news angle constructs the quantile threshold auto regression model to study the stock market returns asymmetric dynamic dependent phenomenon. At the same time, with the framework of Bayesian inference to analysis quantile threshold autoregressive model, set the parameters with random, according to the statistical structure analysis of the model to choose appropriate priori distribution of parameters, to infer the posterior distribution of the parameters, and designed a hybrid algorithm for Gibbs-MH sampling simulation, so as to obtain the distribution characteristics and parameter estimation.The results of the study show that:At the lower quantiles the S&P500 stock index returns presents positive spillover effect to SZC. At the lower and middle quantiles S&E600 stock index returns presents positive spillover effect to SZC; As to S&E600 in the whole site S&P500 stock index returns presents relatively stable positive spillover effect, significantly negative spillover effect of SZC in the high quantiles to S&E600. The spillover effect of SZC stock index returns and S&E600 stock index returns to the S&P500 not has been found. As to the asymmetric dynamic dependent:the SZC stock index returns, the lag of extreme positive returns in middle and upper quantiles present a significant positive effect, extreme negative returns in high quantiles presents a significantly negative effect, the middle part quantiles not presents significantly impact the effect of lagged returns, the good news in addition to low quantiles presents negative impact while other quantiles are presents significant positive impact which present the opposite results of negative news. For S&E600 stock index returns, lagged extreme positive returns in the low quantiles performance significantly negative effect, in the high position performance significantly positive effect, which is opposite to the lagged extreme negative conclusion. The same results has been found in middle quantiles presents insignificant impact. Good news at the lower quantiles presents significant negative effect while significant positive effect in high quantiles, the results coincided contrary to the bad news. The S&P500 stock index returns, we gain the same conclusion with S&E600 stock index returns. The same conclusion we also verified by Bayesian statistical inference.