Empirical Research On Asset Pricing Model In China's A-share Stock Market

Asset pricing model is an economic theory which describes equilibrium-price formation mechanism of capital asset.Research on assets pricing model is of great importance to market participants and the development of the national economy.Asset pricing model can be widely used in many applications in financial area and is the benchmark by specifying the normal returns when asset prices have fully reflected all available informations in testing of market efficieny.An efficient capital market is an important component of a smoothly functioning economy system in a country.Due to the incomplete knowledge about the mechanism of asset pricing,however,to establish an asset pricing model which can capture all systematic patterns of expected return would be quite difficult.The capital markets in the United States have developed for more than a century and the American scholars have developed many influential theories in asset pricing.The classical C APM provides a simple and intuitive descrption of expected return and risk under market equilibrium by using a single pricing factor.But a large number of empirical studies have shown that this model misses much variation of expected return related to other variables.As results,ICAPM and APT extend the risk of expected return from a single factor to a number of factors.In the framework of multi-factor pricing philosophy,Fama-French three-factor model,q-factor model and Fama-French five-factor model come out as better models in explaining systematic variation patterns of expected return in the United States stock market.The A-share stock market is the main body of China's stock market where China's domestic companies finance via issuing equity securities and investors do their equity investment.Therefore,the A-share market should play an important role in allocating resources and contributes much to the development of China's national economy.Because the shorter period of China's own asset pricing research development,an effective theoratical system in this area has not been well established yet.Existing asset pricing models developed in the United Statesare are used in China's financial applications directly.However,China's stock market is quite different in many aspects with United States' stock market.The models which are developed in the United States are not necessarily applicable in China's financial applications.Therefore,examing the performance of existing US-models in A-share market,or developing applicable asset pricing model for the A-share market is a quite meaningful topic and a necessary task for China's financial industry.This task also motives this doctoral dissertation.For this task this dissertation sets up two research objects:First,find out some systematic patterns of expected returns in the A-share stock market.The classical CAPM says that the only risk measure is the market ? which is the single determinant of the expected return across different assets.However,this assertion has not been well supported by empirical data.Actually,stock return in US-market have shown significant variation related to market captalization,book-to-market ratio,profitability and asset growth.Some of these patterns of return have been seen in the A-share market,some not and some need to be examine.This dissertation examines those patterns in A-share market.Second,find out or develop an applicable asset pricing model which can explain those return patterns in the A-share market.Specifically,this dissertation examine 4 candidate asset pricing model with A-share stock market data,identifies the best one among those by comparing their explanatory power for the stylied return patterns.Toward these two research objects,this dissertation finishes an empirical study on A-share market with asset pricing theories developed in the USA as theoretical guide.The main results and findings are:(1)Performing a thorough review of important theoretical and empirical papers in asset pricing area which aims to a better review of related literature.This dissertation starts with a discussion of the relationship between market efficiency and asset pricing model(the joint-testing hypothesis),points out the importance of asset pricing model to asset pricing theory development.Based on this,I then review the development history of asset pricing theories started with Markowitz's portfolio theory,show the evolution of single-factor model to multi-factor model under the mean-variance analysis framework,and rely on the multi-factor efficiency with which Fama(1996)rebuild the Merton(1973)ICAPM as the theoretical foundamen of this paper.On other side,I also review many important classical empirical studies and recent empirical findings.This bears fruit that I summarize essential research ideas and methods which provide powerful weapon for the empirical parts of this paper and further research for other scholars in China.(2)Testing the CAPM by using A-share market data from 1998 to 2014 which provide better result due to better application of classical empirical methods.This dissertation performs testing on the five implications of CAPM by using both time-series regression and cross-sectional regression and gets a consistent result,i.e.the market ? doesn't suffice to explain the expected return in the A-share market.Methods used in testing differentiate this dissertation with other existing paper in China,specifically,the choise of dependent variable and explanatory variables,and the use of GRS test.In time-series regression,the majority of existing papers sort stock on their market ?.Because the market ? doesn't actually have influence on return,this sorting cannot produce effective testing.This dissertation sorts stock with two independent variables,besides ? also on market cap.This sorting produce relative large variation of stock returns and many regression intercepts which are significant different from zero.Meanwhile,the GRS test reject the zero-intercept hypothesis on 1%level which means CAPM doesn't fully explain stock return in A-share stock market.In the cross-sectional regression,this dissertation are different with existing literature in China in the way that I divide the total risk of stock return into ? risk and non-? risk.By implementing the Fama-MacBeth regression it is shown that? doesn't have any significant explanatory power for stock return,but the non-? risk.This result is important for the further empirical research of this dissertation,because the further empirical parts are identifying particular presenting variable of the non-? risk.The existing literature which use both time-series regression and cross-sectional regression as tool of testing CAPM cannot get consistent results,but this is not a problem in this dissertation.This consistent result also supports the multi-factor asset pricing logic of ICAPM.(3)Discovered and Identified 5 systematic return patterns(anormaly variable)in A-share stock market through which new knowledge of return patterns in A-share stock market are achieved.I exmine the explanantory power of 10 fundamental variables for cross-section of stock return in A-share market under two circumstances:consider the updated information in half-year reports or not consider.Variable which have significant effect in both cases are market cap and book-to-market ratio.Variable which only have significant effect when updated information in half-year reports are considered are profitability growth,asset growth per share and stock number growth.Earning-to-price ratio,profitability,asset growth,equity growth and liability growth are not explanatory by itself in both cases.From perspective of explaining period,the effect of updated information in half-year reports are untouched among existing literature,and thus a new way to test which I provide.From perspective of scope of candidate variable,the profitability related variables and asset growth related variables are main new variable which existing literature have not researched systematically.The significant explanatory power of profitability growth,asset growth per share and stock number growth I have found are new knowledge in the A-share stock market return patterns.Based on those findings,I suggest some stock selection strategies which have been shown to be extrem effective in the retrospective test for 124 months.(4)finding out a relative applicable asset pricing model for A-share stock market which is a main contribution of this dissertationBased on research on Fama-French three factor model,q-factor model and Fama-French five factor model,I suggest a six factor model which includes an additional pricing factor besides five existing factors in Fama-French five factor model designed to help capture return variations related with asset growth related variables.A hourse race of 4 candicate model is performed by using 5 LHS portfolio excess return as dependent variables,the six factor model shows the best performance as the GRS test cannot reject it on 1%level for all the 5 LHS portfolio and the adjusted R2 averages about 93.4%.This result means the six factor model is the model which is relative suitable for the A-share stock market among four candidate models.Besides that,the horse race also provides information regarding the explanatory power of other three models in A-share market,the average adjusted R2 ranges from 95.3%to 96.3%.It is important to be aware of that the effectiveness of the explanatory power of the six factor model needs widely robustness test and those work are tasks to be finished by future researchs.With advanced asset pricing theories and empirical methods from the US as foundament,this dissertation work on asset pricing model for China's A-share stock market.The idea and methodologies obtained during the research process enrich the content and scope of existing literature.The empirical findings have substantial implication for China's practice and application in finance area.