Empirical Research On Macro Multifactor Pricing Models In China's A-share Stock Market

fter nearly 30 years of development,the stock market has become a highly active market subject in China's capital market.It is the main market for domestic companies to finance and investors to invest.As a direct financing channel,the stock market should play an important role in optimizing the allocation of resources and promoting the development of national economy.China's A-share market has undergone many turns of bulls and bears,with different fluctuations.More and more rational investors begin to pay attention to the macroeconomic operation and the policy information of the central bank.Under this background,only by deeply and accurately understanding the relationship between macro-economy and stock market and the pricing mechanism of macro-economic variables in stock market,can we better formulate investment strategies to achieve the investment purpose of minimizing risk and maximizing returns.The purpose of this study is to screen out the macro-factors that can have a significant impact on asset price changes from many domestic macro-economic variables,and to establish an empirical model based on the classical multi-factor asset pricing model and test its explanatory effect.The main questions are as follows:Can macro-economic variables be used to reflect the A-share market?The pricing factors of systemic risk,which macro factors can be used as explanatory variables to explain the cross-sectional changes of stock market returns.Therefore,this paper first introduces the theory of multi-factor asset pricing model and macroeconomic Multi-Factor Pricing Model in detail,and then analyses the three theoretical foundations of macroeconomic multi-factor model:dividend discount model,arbitrage pricing model(APT)and inter-period capital asset pricing model(ICAPM).Modern finance holds that the essence of investment is how to make decisions between risk and return,while dividend discount model,arbitrage pricing model(APT)and inter-period capital asset pricing model(ICAPM)provide a rigorous theoretical basis for risk measurement.The main difference between APT and ICAPM lies in the different motivation behind the selection of specific factors:APT often obtains risk factors through statistical analysis of historical return rate,while ICAPM focuses on the state variables that can describe the conditional distribution of future return rate.At the same time,this paper elaborates the role and status of macroeconomic factors as explanatory variables in the empirical asset pricing model.The macroeconomic multi-factor model has undergone more than 30 years of evolution.Scholars in the field of asset pricing fully realize the advantages of macroeconomic factors over other factors:it has intuitive economic implications.Righteousness and rigorous logic.According to the analysis in the literature review,this paper establishes two main empirical contexts:one is to select macroeconomic indicators or construct factors within the framework of arbitrage pricing model APT;the other is to select macroeconomic indicators that can affect future investment opportunity set under the framework of inter-period capital asset pricing model ICAPM and extract their expectations.Information is used as a pricing factor.1.Empirical Analysis Based on APT FrameworkBecause APT model can not determine the specific composition of factors,this paper uses two methods of factor selection to construct factors under the framework of APT model,trying to explore a macro Multi-Factor Pricing Model more in line with China's A-share market.(1)Using macro-economic indicators to construct factors,a macro-multi-index pricing model is proposed and tested empirically,which achieves the innovation of empirical test of classical methods in China's A-share market.This chapter studies whether the main domestic macroeconomic indicators can be used as pricing factors to determine the expected return of stock assets.Referring to the macro five-factor model proposed by Chen,Roll and Ross(1986),the author constructs 10 macro-economic variables as risk factors,such as industrial production growth rate,unexpected inflation rate,expected inflation rate,credit spread and term spread,total consumption growth rate and commodity index return rate.Factor pricing model.Fama-MacBeth(1973)two-stage regression results show that six macroeconomic factors except commodity index yield factor have passed the significant test in cross-sectional regression,but the factor risk premium has positive and negative.Similar to the conclusions of Chen,Roll and Ross(1986),macroeconomic factors other than the return on commodity indices can systematically determine the expected return on stock assets.Among them,the unexpected inflation rate has the most significant explanatory effect,and has a positive risk price.The multi-factor model with industrial production growth rate,unexpected inflation rate,credit spreads,term spreads and total consumption growth rate as explanatory variables has lower pricing error and stronger explanatory effect on cross-sectional differences of stock return than Fama-French three-factor model.The way in which these factors determine the expected return of stock assets is consistent with the model predicted by the dividend discount model.Classical market factors are not statistically significant in all multi-factor models,nor can they improve the model's ability to explain cross-sectional differences,so they do not provide more pricing information than macroeconomic factor models.(2)Although the model interpretation effect of the first method is very good,it still has limitations.Firstly,the dividend discount model can be used as the theoretical basis to determine which macroeconomic variables can be the alternative risk factors,but the selection of variables is still the same.The main innovation of this paper is to use panel data of grouped macroeconomic indicators to extract risk factors.Because investors'asset selection behavior is affected by all publicly released macroeconomic information,such factor selection strategies also enable us to evaluate comprehensively which macroeconomic variables affect the horizontal of stocks.Cross-sectional yield differentials have explanatory power,rather than a priori choice of a small part of them.This paper divides 113 macroeconomic indicators into 8 groups,extracts the principal components from each group of macroeconomic data.The purpose of this is to give this principal component more economic meaning.This is because according to the definition of principal component analysis(PCA),the principal components obtained by this method are necessarily orthogonal to each other.The natural orthogonality between the principal components extracted by full sample PCA method makes it almost impossible to give economic meaning to the high-dimensional factors obtained by using this method.The macro principal component factors obtained after grouping can be directly used as pricing factors and as explanatory variables.Fame-Macbeth(1973)two-stage regression method is used to test the pricing ability of these principal component factors and the explanatory ability of cross-sectional differences in stock market returns.In addition,this paper uses factor simulation combination method to construct explanatory variables in order to improve the explanatory effect of explanatory variables.II.Model Establishment and Empirical Test of Intertemporal Capital Asset Pricing Model(ICAPM)In this paper,the two empirical studies on the factor structure based on macroeconomic indicators are carried out under the framework of APT.Next,this paper extends the single-period situation of APT model to the cross-period situation of ICAPM,regards the unexpected information of macroeconomic indicators and factors as explanatory variables that can influence the change of investment opportunity set,and tests whether APT model can be an influencing factor for investors to make cross-period investment decisions.The proposal of ICAPM fully takes into account the investors'cross-term asset selection behavior,and the change of investment opportunity set directly affects the cross-period hedging demand of investors in ICAPM.Therefore,the systemic risk that investors take comes not only from market fluctuation,but also from the unexpected information of macroeconomic variables that affect the investment opportunity set.In this sense,ICAPM provides another method and perspective for selecting risk factors.Based on the theory of Intertemporal Asset Capital Pricing Model and Petkova's(2006)method,this paper focuses on the pricing ability of unexpected information of macroeconomic indicators.Empirical results show that the unexpected information of short-term interest rates and default spreads have passed the significant test,indicating that the unexpected fluctuations of both will affect investors'investment decisions in the next period,and they are macroeconomic indicators that investors should pay attention to when making investment decisions.When using the principal component factor,the fiscal factor,monetary factor,output factor and price factor all pass the significant test,which shows that the unexpected information contains the systemic risk affecting asset returns.This indicates that when the unexpected information of the above principal component changes,investors will decide whether or not according to their demand for cross-period hedging.It will invest in this period,which will affect the change of stock returns.Comprehensive empirical results show that the model using unexpected information of principal component factors as explanatory variables has better explanatory ability,and more comprehensively describes that unexpected changes in macro-economy will affect stock price changes through influencing investors'investment decisions.Based on the summary of the research theory and methods of macro-multifactor pricing model,and according to the analysis context of macro-multifactor pricing model,this paper mainly summarizes how to extract factors from macro-economic indicators as explanatory variables to construct macro-economic multifactor pricing model.Based on different methods and perspectives,we can get the factor conclusion.Different macro-multi-factor models are constructed.The analytical methods and frameworks summarized in the research process enrich the content and scope of the existing literature.The empirical results have some enlightenment for market participants and policy makers.