| Investors’heterogeneous belief (i.e., opinion divergence), focused in this paper, is specific form of investors’heterogeneity. The traditional asset pricing theory assumes that investors are homogeneous. It implies that all market participants have a similar trading horizon, receive the same information, and have homogenous expectation. Such strong assumption ignores the difference among the investors. For example, different investors have different knowledge (prior belief), information obtaining ability and analytical skills, risk tolerances. These differences will almost certainly result in heterogeneous expectations and judgments on the asset price. Thus, different investors have heterogeneous beliefs, which influence the investment decision of market participants and asset prices.The theoretical literatures about heterogeneous belief asset pricing are all discussed under the short-sale constraints. In Chinese A-share stock market, short-sale was strictly restrained before March 31,2010, but since that date, short-sale constraints on partial stocks have gradually loosened. The Chinese stock market is an ideal experimental platform because of the absolute short-selling constraints existing in long period. Meanwhile, as the short-sale constraints are eliminated gradually, opportunities have been provided for us to analyze further whether the heterogeneous belief asset pricing validity is weakened.As early as 1977, Miller proposed the conceptual model for the heterogeneous belief asset pricing theory. Since then, the study had remained in the theoretical model derivation stage in long times. Because heterogeneous belief is related to cognition and behavior of different investors, it cannot be measured directly. Empirical researches on investors’ heterogeneous beliefs are insufficient until the beginning of the 21st century. In the existing literature, indirect variables are used as proxies for heterogeneous beliefs. For example, analysts’forecast dispersion, stock return volatility, bid-ask spread, unexpected trading volume, and commissioned order spread on an intra-day basis proxies commonly used in the literature. In this paper, author has reviewed the constitution methods of different measures and made comparisons of their theoretical basis and applicability.Author establish a turnover decomposition model by using the idea of unexpected trading volume (Garfinkel,2009) to obtain the degree of heterogeneous beliefs. The heterogeneous beliefs are directly extracted from trading activity, ensuring that 1) the heterogeneous beliefs obtained have been translated into actual trading activity and exerted an impact on asset prices and 2) such problems as incomplete information or overlap with other exogenous measurement indices of heterogeneous beliefs are avoided. We are thus able to investigate the effect of heterogeneous beliefs on stock returns and draw the following conclusions.First, portfolio strategies developed according to the degree of heterogeneous beliefs display intuitively that the heterogeneous beliefs surely have the stable influences on asset price. To be specific, in a market with short-sale constraints, the opinion of pessimists would not be incorporated into the current price of a stock, whereas an optimistic opinion held by investors is easy to express through trading. Ultimately, this would lead to a higher current price than the intrinsic value of the firm. Higher heterogeneous beliefs lead to higher current prices and -lower future returns. So, the expected return in the subsequent month of a portfolio with a low degree of heterogeneous beliefs is significantly higher than that of other portfolios, and this difference is most obvious with stocks of smaller companies. When factors of size and value are controlled, the expected average return in the subsequent month of a portfolio based on small-company value stocks and with a low degree of heterogeneous beliefs is significantly higher than that of portfolios based on large-company growth stocks and with a high degree of heterogeneous beliefs. But, momentum effect has not been discovered in portfolio analysis. The research conclusion has the practical guiding significance for fund portfolio structuring in the investment practice. (See Chapter 4 for details)Second, from the perspective of asset pricing models, the pricing validity of heterogeneous beliefs has been tested with the econometric method. When heterogeneous beliefs index is introduced into the CAPM, the Fama-French three-factor model and the Carhart four-factor model, all results of regressions show that the magnitude of heterogeneous beliefs has a significantly positive correlation with the current month stock returns and a significantly negative correlation with the subsequent month stock returns. The time series regression controlled liquidity shows that heterogeneous beliefs have stable effect on stock returns. In the robustness test, analysts’ forecast dispersion cannot explain subsequent stock returns in Chinese market. When further considering idiosyncratic volatility, momentum reversal and other factors, the conclusion is still valid. (See Chapter 5 for details)Third, the degree of investor’s heterogeneous beliefs varies in bull market and bear market, and has the asymmetrical influences on the asset price. In a bull market, investors are active in trading and provided with the abundant information. They are more motivated to unearth information in order to acquire higher returns. Meanwhile, the abundant information and the active trading cause investors to make decisions easily based on different information sets. So investors have higher degree of heterogeneous beliefs in bull market. In the bear market, trading is inactive, the quantity of information is small and information distribution is slow. This leads to information convergence and lowers investors" heterogeneous beliefs. Years from 2000 to 2009 are divided into 5 bull or bear market periods, and difference magnitude of heterogeneous belief in these two markets can be seen through statistical result. The regression analysis has indicated that heterogeneous beliefs have the asymmetric influences in bull and bear markets. The positive effect on current stock prices in bear market is stronger than that in bull market. And, the reducing effect on future stock returns in bear market is stronger than that in bull market. (See Section 6.1 for details)Finally, author discussed the pricing implications of heterogeneous beliefs after short-sale was allowed. The theoretical literatures about heterogeneous beliefs asset pricing are all discussed under the short-sale constraints. If the pricing implications of heterogeneous beliefs will disappear after short-sale was allowed? Market friction and trading costs can prevent the expression of bearish opinions, so the negative relation between heterogeneous beliefs and future returns is till significant in American stock market. Whether heterogeneous beliefs have explanatory power on future returns in Chinese A-share market is an empirical problem. China has started the pilot securities margin trading, and short-selling stocks and non-short-selling stocks have existed simultaneously in the market since March,2010. Based on the data from 2011 to 2013, this paper has empirically verified through the group comparison and the regression analysis that heterogeneous beliefs have the weaker explanatory ability for the short-selling stocks’ current returns and has no significant influences on their future returns after the short-sale constraints were lifted. (See Section 6.2 for details)... |
PDF Full Text Request