The Analysis And Application Of Idiosyncratic Volatility Among Stocks

On the correlation between the stock risk and revenue has always been the hot spot of the investors and academia, according to the classical theory of capital asset pricing model(CAPM) think that the risk of return on assets is completely determined by the systemic risk, investors can be offset by building a diversified investment portfolio risk, and idiosyncratic risk(not systemic risk) will not be reflected in stock prices. That company has nothing to do idiosyncratic risk and expected return, don’t need to get extra risk compensation. But in recent years some scholars study found that the company idiosyncratic risk and expected return there are significant negative correlation relationship, namely the heterogeneous volatility puzzle. Based on sample interval 30 January 2002 to June 30, 2015 stocks on the yield data in China’s stock market has carried on the empirical research. Based on individual stocks on yield fluctuation, the beta coefficient of CAPM model and the heterogeneity of individual stocks volatility portfolio analysis, verify the a-share market is low heterogeneity volatile combination heterogeneity of high gain volatility puzzle. the company scale, the book value ratio, turnover rate, the momentum effect cannot explain the negative correlation with the expected earnings of heterogeneity fluctuations. Finally the Fama-Macbeth cross section two step regression method, for the two-dimensional grouping construct multiple portfolio in the cross-section regression, found in either group combination sets, heterogeneity fluctuation factor is negative, namely high heterogeneity volatility risk and didn’t get high yield positive compensation, also verify the heterogeneity of individual stocks volatility there is a negative relationship with the expected return.At the same time we also applied the vector autoregressive model(VAR) to study based on the heterogeneity of wave packet portfolio returns and 300 index. 30 January 2002 to June 30, 2015 the logarithm yield rate data as the research object, through unit root series stationarity, establish combination yields and 300 index vector autoregressive model; Using Granger causality test to verify the heterogeneity of wave packet excess returns and the 300 index between the existence of the real dynamic causality; Then using impulse response function and variance decomposition to analyze the portfolio income and the strength of the causal interaction between the market index.