After reviewing extant literatures, this dissertation considers the equity risk premium, a conception related with expectation; thus, it could also be termed as the expected equity risk premium. There is historical evidence that the equity investment return exceeds government bond investment, which is defined as historical equity risk premium- a totally different concept from the expected one. In summary of the literatures, the estimation results of the equity risk premium based on earnings forecast by the ex ante approach are more stable and reliable compared with results by the ex post approach.
After analyzing the history and current situation of the Chinese securities market, this dissertation determines the research objects of Chinese A-shares and H-shares market, considers HS300 index and HSCEI as a proxy of A-shares and H-shares market, respectively; chooses a period starting from year 2005 as a sample period; takes market consensus of earnings forecast as inputs; and constructs a three-stage DDM to measure the equity risk premiums of the A-shares and H-shares markets. For comparison, this research also calculates the historical equity risk premiums by the ex post approach.
By analyzing empirical results, the historical results are greatly influenced by sample period selection and are volatile with high standard deviation and poor stability. Thus, this dissertation holds that the ex post approach is not suitable for the current Chinese stock market and the historical results can’t be used as inputs to stock pricing models directly. The expected equity risk premiums, however, are almost not influenced by sample period selection and are stable with a low standard deviation.
Thus, the expected equity risk premiums by the ex ante approach are a reasonable and reliable estimation of the true value of the equity risk premium of the Chinese stock market and could be used for stock pricing and measuring the risk preference level of the whole market.
This research adopts time series analysis tools, mainly GARCH models, to analyze the expected equity risk premium series of A-shares. The main findings are the series is stationary and there is ARCH effect in the mean equation that can be captured by GARCH models. By comparing different GARCH models, EGARCH(1,1)-M is selected for its better fitting and forecasting abilities of the expected equity risk premium series of A-shares. These are helpful in predicting the market trend.
Based on the logic of DDM, the research of A/H shares price premium puzzle decomposed the influencing factor into three factors (the difference of the equity risk premiums between these two markets, the difference of earnings forecasts, and the difference of risk-free rates) and conducted econometrical analysis through impulse response function and variance decomposition based on the VAR model, to analyze influence of above mentioned factors. The main finding is, the explanatory power of difference of the equity risk premiu