Empirical Research. Three-factor Capm Model In The Shanghai A-share Market

In 1950's, Markowitz published'Portfolio Selection'based on efficient market hypothesis (EMH), creating the modern investment theory. After that, Sharp(1664), Linter(1665) and Moisson(1966) respectively put forward the capital asset pricing model (CAPM) under the general market equilibrium, associating return with risk. The strict assumptions coming from EMH put the CAPM in face of the challenge from empirical tests. Anomalies including "Size Effect", "Value Premium" and "Calendar Effect" make CAPM less persuasive in explaining the performance of the market. To improve the model, Fama and French first attributed the risks of asset investment to market factor, size factor, and value factor, of which the first represented the systematic risk of the market and the other two referred to the characteristic risks contained in the certain portfolio. The corresponding FF three-factor model then successfully explaining the difference in the returns on various assets, and thereafter, acquired the support from empirical efforts.In this article we test the Shanghai Stock Exchange (SSE) under the context of FF three-factor model with the sample of all the stocks from April 2000 to October 2010. As in the advanced markets of U.S. and EU, size effect and value premium are also observed here. We examine the continuous profitability of the sample corporations to see whether it is able to explain the value premium as predicted by Fama and French. The result is positive:corporations with higher book-to-market ratio show poorer profitability, then indicating higher risk. For size effect, we originally put forward the potential answers from the strategy of diversification that possibly adopted by large firms and the "share numbers effect" in the secondary market. The strategy of diversification may lower the risk of the firm's business and the cost of operation, then improving its economic performance. And a large number of shares outstanding would decrease the possibility for the stock to be manipulated by some powerful speculators.The three-factor model well explains the difference in returns on size-BE/ME portfolios. Although the zero-intercept assumption is rejected, they don't make contributions to the difference in return spread. With characteristic risk factors standing independent of the market risk,βs converge into a narrow interval around 1. Small stock portfolios perform more sensitively to size factor and high-BE/ME portfolios perform more sensitively to value factor.