Tests And Applications Of Asset Bubbles And Asset Pricing Models

Since Fama and French proposed the three-factor asset pricing model,factor pricing models have been one of the hot and key issues in the field of empirical finance,and have become an important academic frontier in finance.By constructing a simple and easy-tounderstand model with characteristic factors,factor asset pricing model provides a useful perspective for understanding the law of asset price formation and is widely used in the investment community,which is one of the most closely integrated research fields between economics theory and practice.This thesis is dedicated to extending the applicability test and application of the asset pricing model,thus forming a methodological innovation and application innovation.The theory of asset pricing models is based on the efficient market hypothesis.Existing tests of asset pricing models in the literature usually assume the market is weakly efficient,i.e.,stock prices follow random walks,but ignore the existence of asset price bubbles.In this thesis,we apply the GSADF bubble test proposed by P.C.B.Phillips,a well-known econometrician,to identify bubble events in our capital market and test the asset pricing model by classifying the weak efficient period and the bubble period accordingly.This approach separates the joint tests of market efficiency and asset pricing models,and provides a deciphering idea to solve the problem of joint tests of asset pricing models.This thesis applies this approach to the "Chinese version" three-factor model by Liu et al.(2019)and the Fama-French five-factor model,and examine the applicability of the two models during the weak efficient period and the bubble period in China’s capital market,respectively.The results of the bubble test for the CSI 300 index show that there were two large bubble events in China’s A-share market during the period 2005-2017.The first bubble occurred on the eve of the global financial crisis in 2008,with a bubble duration of April2006-December 2007,followed by a market recovery period of about one year until December 2008 when it returned to weakly efficient again.The second bubble was the 2015 stock market crash.This bubble lasted for about 7 months,which originates in November2014 and collapses in June 2015,and the market again experienced a recovery period of about a year and slowly returned to weakly efficient in May 2016.Based on the bubble test results,this thesis tests the applicability of the "Chinese version" three-factor model by taking the duration of the second bubble event as the bubble period and selecting the corresponding weak efficient period.Liu et al.(2019)propose a "Chinese version" three-factor model by adapting the classical Fama-French three-factor model to the unique phenomenon of "shell-value contamination" and the data characteristics of asset returns in China’s capital market.The results of this thesis confirm that the overall explanatory power of the "Chinese version" three-factor model is better than that of the Fama-French three-factor model in China’s capital market.However,the explanatory power of the "Chinese version" three-factor model decreases significantly in the bubble period compared with the weak efficient period of the market,evidenced by the decrease in the marginal pricing ability of the three factors,among which the marginal pricing ability of the value factor decreases most significantly.Furthermore,this thesis makes improvements to the GRS test,a popular test for asset pricing model,and implements simulation by wild bootstrap.The GRS test is proposed by Gibbons et al.(1989,GRS),which is commonly used in the mainstream literature to assess the applicability and explanatory power of asset pricing models.However,the GRS test is based on the assumption that the disturbance terms are independently and identically distributed,and the real return data usually do not satisfy this assumption premise.When there is an ARCH effect in the disturbance term,the GRS test power of finite samples decreases.Therefore,this thesis proposes a GRS test under the wild bootstrap algorithm,which preserves the original variance structure of the return data by constructing bootstrap samples.We apply this method to conduct simulation experiments with the "Chinese version" three-factor model.The performance of the model under the traditional GRS test and the modified GRS test are examined and compared.The results show that the modified GRS test has significantly higher test power of finite samples compared to the traditional GRS test when there is an ARCH effect.Moreover,the modified GRS test is less supportive of the null hypothesis and the pricing model is less likely to pass the test.The modified GRS test can more accurately reflect the actual explanatory power of the model in the weak efficient and the bubble periods.Finally,this thesis applies the modified GRS test to the applicability of the FamaFrench five-factor model in China’s capital market and examines the dynamics of the marginal pricing ability of the five factors during the weak efficient and the bubble periods.Based on that,this thesis digs deeper into the marginal pricing ability of the factors and examines the theoretical content of the marginal pricing ability of the factors and its implied economic implications.The thesis focuses on the marginal pricing ability of the value and investment factors in the Fama-French five-factor model.By extracting the portfolios which the value and investment factors have pricing ability,we find that the main characteristics of the companies included in these portfolios are,their investment growth rate,profitability,and their ability to generate profits and market value from their net assets are higher than the market average.This thesis argues that the mechanism by which the value and investment factors act on their pricing ability is that this kind of companies is able to rapidly convert investment growth rates into endogenous growth.Specifically,these companies are able to quickly convert investor capital into company size expansion and product innovation.As the technological advantages and scale of their products continue to expand,the market value created by their original net assets also continues to grow with them,and their investment behavior can be effectively translated into economic benefits,thus realizing the value growth of the company.Accordingly,the companies for which the value factor and investment factor have pricing ability are not only the specific targets for stabilizing investment during the current epidemic prevention and control period in China,but also the orientation of long-term strategic investment in China.This kind of companies constitutes the new driving force of China’s economic growth.