An Empirical Research On The Relationship Between Heterogeneous Beliefs And Stock Returns In The US Market

A key assumption of standard finance model is that all investors have identical expectations regarding future returns of all securities. However, in the real financial world, investors differ in valuations as a result of the uncertainty of future returns, private information and heterogeneous priors etc. In the presence of short-sale constraints, stock prices reflect a more optimistic valuation as pessimistic investors are kept out of the market. Thus stocks are overpriced in the short term and generate lower future returns. The higher the dispersion in expectations, the higher extent stock prices are overpriced and the lower future returns are generated.This paper examines the relationship between heterogeneous beliefs and stock returns in the US market. Among previous empirical work, there are a lot of theoretical debates on the proxy for heterogeneous beliefs. I postulate that the dispersion in brokers'recommendations could be a more appealing proxy. I examine the impact of dispersion in recommendations on stock returns by constructing portfolios using the stocks listed on NYSE, AMEX and NASDAQ stock markets from 1994 to 2008. The empirical results are consistent with the hypothesis that stocks with higher disagreements generate lower future returns. First, I perform time-series tests on portfolio returns using Fama-French three factor model. Over the period from 1994 to 2008, stocks with low dispersion in brokers'recommendations earn a significantly higher return than stocks with high dispersion. In order to exclude the ambiguity posed by size effect, I reexamine the overvaluation effect by controlling for size. The dispersion effect remains robust and becomes stronger in small-sized stocks. Since small stocks suffer from higher short-sale costs, the variation of return differential among different size quintiles supports the overvaluation hypothesis. In my sub-period analysis, the role of dispersion in recommendations in explaining stock returns remains significant in each of the three sub-periods and gets weakened after 2002, which might be caused by the sharp decline in over-optimism after 2002, improvements in companies'transparency and reduction in short-selling costs during recent years. Finally, cross-sectional regressions are performed using Fama-Macbeth two-pass methodology, which provides further evidence being supportive for the negative relationship between heterogeneous beliefs and stock returns. The return differential estimated in this paper is higher than previous results, which suggests that dispersion in brokers'recommendations could be a reasonable measurement of difference of opinions among investors.Compared to the US market, Chinese market is still showing low maturity. In addition, individual investors take a dominant part in public investments. However, individual investors are dispersed in educational background and income status, which may contribute to the fact that heterogeneous beliefs significantly exists in Chinese market. Furthermore, short sales have been completely abandoned in Chinese market. Thus the role of heterogeneous beliefs in explaining stock returns becomes quite important. Besides, with the final implementations of short-selling mechanism and stock index, it becomes vital for government to make wise policies for securing investments in an environment with weaker investment restrictions.