F. Albert Wang

Strategic trading, asymmetric information and heterogeneous prior beliefs

Abstract We develop a multi-period trading model in which traders face both asymmetric information and heterogeneous prior beliefs. Heterogeneity arises because traders agree to disagree on the precision of an informed traders private signal. In equilibrium, the informed trader smooths out her trading on asymmetric information gradually over time, but concentrates her entire trading on heterogeneous beliefs toward the last few periods. As a result, the models volume dynamics are consistent with the U-shaped intraday pattern at the close. Furthermore, the model predicts a positive autocorrelation in trading volume, and a positive correlation between trading volume and contemporaneous price volatility.

Stock Market Mispricing: Money Illusion or Resale Option?

We examine two hypotheses to explain stock mispricing: i) the money illusion hypothesis (Modigliani and Cohn (1979)) and ii) the resale option hypothesis (Scheinkman and Xiong (2003)). We find that the money illusion hypothesis may explain the level, but not the volatility, of mispricing in the U.S. market. In contrast, the stock resale option hypothesis, which stems from heterogeneous beliefs about future dividend growth rates and short-sale constraints, can explain both the level and the volatility of mispricing. The evidence suggests that while the two hypotheses complement each other in explaining the level of mispricing, the resale option hypothesis provides a more coherent explanation for asset price bubbles, in which extraordinarily high price levels are often accompanied by excessive volatility and frenzied trading.

Information Asymmetry, Liquidity, and Stock Returns

We develop a new measure for the probability of informed trading, called PCP. Using double-sorted portfolios, we find that excess returns increase from low to high PCP portfolios. In regression analysis, the effect of PCP on returns is significantly positive after controlling for illiquidity effect. The point estimate of the coefficient on one PCP is 3.788, suggesting that a difference of 10 percentage points in the PCP between two stocks leads to a difference in expected returns of 4.5 percent annually. Thus, the effect is also economically significant. Our results support the information asymmetry hypothesis that information risk is priced.