Model Analysis Of Asset Pricing Theory And The Applied Research In China

The evolution and development of asset pricing theory irreplaceably influenced the operation and decision-making of macroeconomic, and substantially affected the strategy of investors and the operation of the capital market. From the different background framework of the hypothesis, asset pricing theory can be divided into two categories:asset pricing theory of rational expectations and asset pricing theory of irrational expectations. The implication of the former is that there are only rational investors in the market, who make rational investment decisions according to the change of the market, and asset prices are determined by the market fundamental, and whenever asset prices deviate from the fundamental, they will be adjusted back to the fundamental automatically. The implication of the latter is that there exists irrational investors, who may have "herd mentality" or "excessive volatility", and asset prices are no longer determined by market fundamentals, which implies the existence of excessive volatility or bubble in asset prices caused by behavior errors, market frictions or even wrong pricing behavior. The widely used tests of asset pricing model of rational expectations are Fama-MacBeth cross-section regression, Fama-French three-factor model and GMM. Test of asset pricing model of irrational expectations that is both classical and popular is bubble test, especially supADF test. Tests of asset pricing models could help investors understand the internal mechanism of the change of asset prices, make effective portfolios, and guide investors to make correct investment decisions. Based on the interactive development process of asset pricing model and the testing methods, this paper expounds the research ideas and test procedures of various test methods to analyze the cross-sectional earnings of China’s stock market, to screen the rational and irrational periods, and to explore the applicability of asset pricing models in each period, combining with the new developments of asset pricing factor model and bubble tests. This paper interprets main testing methods for asset pricing, including Fama-MacBeth cross-section regression method, the Fama-French three-factor model test method, the generalized moment estimation and the newly-proposed supADF bubble test. Based on the reviews and summarization of the theoretical nature and related frontier developments of CAPM,(five) factor model, consumption-based CAPM and excessive volatility theory, this paper focuses on promoting theory and practical significance, especially estimation and testing methods. This paper analyzes how these financial econometrics testing methods promoted the development of asset pricing theory model, and how the asset pricing theory model in turn continuously helped to put forward new testing methods and promote the development of Econometrics by innovating in methods and refusing models, indicating that the two supplement and promote each other.The innovations of this paper can be summarized as:the theory analyses grouping design and rolling regression idea in Fama-MacBeth cross-section regression method and Fama-French three-factor model. Dividing asset portfolios can reduce systemic errors in the stock returns, and applying the idea of rolling regression updates the beta value of portfolio monthly, and improves the precision of the portfolio. The cross grouping among factors can effectively separate different factors that can affect the stock returns. The innovation points of the empirical research:this paper selects3000as the critical point of the Shanghai composite index, and divides the whole sample period into three stages according to the properties of the sample, and the volatility of Shanghai composite index, Applying Fama-MacBeth grouping design and rolling cross-section regression, results show that CAPM model is only applicable to SSE a-share data during2003.1to2006.12, the rest of the two periods both show distinct unsystematic risk or speculation, which makes the model inapplicable. Applying Fama-French three-factor model to test different stages, results show that the model is only applicable to SSE a-share data during2010.1to2014.9. There are not only risks but also characteristic factors behind the stock market value factors. But the book value ratio on the stock yields only relates to the stock characteristics. As there are more irrational factors during the financial crisis, the book value reveals obviously weaker explanation ability. We finally test bubbles in China’s stock market, specially the time-point of bubble existence, based on the latest proposed supADF testing method by using the Shanghai a-share index., and find that the origination and collapse time of bubble is2007.1and2008.1. It confirms that2007to2009is the stage of irrational expectations in Chinese stock market, and further explains why the CAPM and Fama-French three factors model cannot hold.To sum up, by testing the adaptability of three factor model in stationary and nonstationary stages in China’s stock market, we find that three factor model has a much higher explanatory power than conditional three factor model in stationary stage, though conditional three factor model based on the variation of beta or separation of systemic risk is proposed in domestic (He,2010). On the other hand, the conclusion of Tian (2014) that stock yield of China is only explained by market risk premium and market size effect is only applicable for China’s stock market in excessive volatility stage; three factor model is more powerful in the stationary stage after financial crisis. As a result, the testing of asset pricing model should not be compared as a whole. We should analyze the applicability of different testing approaches and choose the best one according to the characteristic of the real financial data, so as to reach the best testing effect.