Abstract

The compound growth rate of China’s assets under management（referred to as AUM）from 2012 to 2015 is 51%，the total AUM reached 60 trillion yuan by the end of June 2016. The operation and development of asset management industry is more and more important to China’s economy，society and people’s life. Meanwhile，the asset management institutions are also facing various difficulties，problems and challenges，including assets pricing and valuation，the risk analysis and forecast etc.，in where asset pricing model and risk factor theory play a key role.

After half a century of development，asset pricing model and risk factor theory has experienced the evolution from single factor to multiple factors，rational to behavioral，linear to nonlinear. The Capital Asset Pricing Model（CAPM）is a single factor linear model based on rational expectations. Along with a large number of empirical studies found that the simple use of the market factor cannot fully explain the excess return of assets，and the explanation ability of market factor is declining，the correctness of CAPM has been doubted. At the same time，CAPM has also been doubted by theorists. The Arbitrage Pricing Theory（APT），which was published by S.Ross，considers that the return of assets is linearly related to a set of factors，and lays a theoretical foundation for the multi-factor asset pricing model. In 1980s，the study of “anomaly” reached a peak，including the size effect and leverage effect etc. After that，the Fama-French three-factor model considered the size and value factors as systematic factors. During this period，scholars also found phenomenon which in violation of the rational expectations，such as momentum effect etc. R.Shiller verified the existence of excessive fluctuation in stock price，which leads to the emerging and development of behavioral finance. The momentum factor was introduced to the Fama and French three-factor model by M.Carhart，and recognized as a systematic factor. Although E.Fama and K.French released a new five factor model in 2014，new factors named as RWM and CMA are still remains to be verified.

During the process of studying，systemic factors are gradually classified into two categories：Static factors and dynamic factors[1]. The static factors mainly refer to the various macroeconomic variables，which are coming from economic activities and social operation of human beings，and they can produce positive or negative impacts on the price of risky assets by the market factor. Investors cannot decide or change static factors，only accept them passively. The dynamic factors mainly refer to the various basic properties of assets，such as size，value，momentum，they can explain the excess returns of risky assets，all of them are long-term effective，tradable，manageable and interpretable，and can be studied by mimicking portfolio.

The practice of asset pricing model and risk factor theory can be divided into two types：Active management and passive investing strategy. The most important applications of passive investing strategy are construct indices and index funds. Certain index fund represents a certain style of investment portfolios，the purchase of the index fund is equivalent to investing a package of same style stocks. By doing this，investors can obtain the corresponding incomes from certain factors，and avoid the liquidity risk and operating risk by invest in a single stock. In order to obtain active management profits，active managers need to analyze the performance of the factors and volatility of stocks etc.，adjust the positions by certain rules.

The researches on the factor theory has profoundly changed the methods of assets allocation. Traditional assets allocation only adjusts positions between stocks and bonds，but after the Internet bubble burst in 2001，equities of most funds fell dramatically，investors began to purchase funds of different strategies instead of the allocation on the types of risk assets. Alternative investment strategies like private equity fund，hedge fund and commodities were recognized by more and more investors. But in financial crisis period of 2008-2009，the average loss of hedge funds，which claimed to be immune to the volatility of market，was 20%-30%. Investors have to rethink and look for the strategies that are low correlation with the market factor，and investing in different risk premia has gradually become the protagonist of assets management. Since then，the diversification of factor investing became popular，asset allocation method changes from the investment strategies to factor investing.[2]

Although the applications of asset pricing model are very wide，in addition to the investment can also be used to fund performance evaluation，performance attribution，risk analysis，portfolio analysis and so on，but there are still many problems in the research and practice of factor model. First of all，most of the empirical researches on factor models using linear regression. This requires all the return data to obey the normal distribution，but whether it is a single stock or portfolio historical return contains a lot of sample error and trading noise，fundamentally unable to satisfy the classical assumptions of linear regression，so the linear regression results may be biased，thereby affecting the overall results. Secondly，the linear regression is very sensitive to the selection of data sets. Using different time scales data，even if the same method may get different results，the risk factor coefficients of same portfolio may be significant in one period，and not significant in another period. This phenomenon reflects the heterogeneity of investors，and there is a big difference between the behavior of large asset managers and retail investors，whether in the trading frequency or information processing. China’s stock market is dominated by retail investors，so it is full of uncertainties and has obvious heterogeneity. Finally，so-called dynamic analysis of linear regression in previous researches only adds a fixed rolling window to estimate the coefficients，cannot reflect the internal structure changes of market caused by emergency information.

First of all，this paper reviews the theoretical development of factor theory，factor investing and asset pricing models. Secondly，this paper analyzes the formation mechanism of systematic factors，and discusses several problems in present research. Then，in order to overcome these problems，this paper employs wavelet analysis and dynamic model averaging method to study the multi-factor assets pricing models. At last，the paper analyzes the momentum and contrarian effect，and factor investing strategies in China’s stock market，then puts forward the investment strategies which would performance well in China’s stock market.

Key Words：Asset Pricing Model；Factor Theory；Factor Investing；Wavelet Analysis；Dynamic Model Averaging

[1] This Classification Method Was Introduced by Ang A. Asset Management：A Systematic Approach to Factor Investing. Oxford University Press，2014.

[2] Bender，J.，P.Hammond，and W.Mok. Can Alpha Be Captured by Risk Premia？. The Journal of Portfolio Management，2014，40（2）：18-29.