Empirical examinations of rational and behavioral explanations for asset pricing anomalies

Recently, researchers have documented patterns in equity returns that appear at odds with the efficient markets hypothesis. The economic magnitude of these anomalies and the demonstrated inability of the CAPM to eliminate them has led some to conclude that models based on behavioral biases are necessary to account for the presence of these patterns. This thesis is composed of two chapters that consider rational and behavioral explanations for capital market anomalies.; Chapter 1 shows that rational models based on macroeconomic risk and behavioral models based on investor mis-reaction to firm-specific risk have difficulty in explaining price and earnings momentum. I show that while price and earnings momentum profits are predictable based on aggregate financial market variables, this variation is not consistent with parsimonious specifications for the intertemporal CAPM and Fama French model. Furthermore, momentum profits' predictability is not consistent with a simple time varying underreaction story. More importantly, this essay documents that momentum profits are significantly negatively related to aggregate cash flow news as proxied for by news about macroeconomic aggregates such as industrial production growth and consumption growth. Evidence suggests that this is not the result of the so-called 'discount rate' effect. This is difficult to explain with both rational models based on macroeconomic risk and behavioral models based on investor mis-reaction to firm-specific risk.; Chapter 2 presents a test of a particular set of behavioral models, namely the static overconfidence and biased self-attribution hypotheses developed in Daniel, Hirshliefer and Subrahmanyam (1998) (DHS). I appeal to a large literature in market microstructure that suggests that order flow reflects the actions of informed traders and a noise component due to liquidity trade. Using this idea, I consider long-run price reactions after periods of large positive and negative firm-specific order flow. The relation between order flow and future returns can be thought of as a test of the DHS hypotheses. The results indicate that returns are negatively related to firm-specific order flow. This result is consistent with the DHS static overconfidence model, but not with their biased self-attribution model.