Date of Award

Winter 2009

Project Type


Program or Major


Degree Name

Doctor of Philosophy

First Advisor

Michael D Goldberg


This dissertation examines returns in equity markets and the ability of extant models to account for their behavior. I first consider the so-called "expectations hypothesis" (EH) and find that it fails empirically even when structural change is incorporated into the analysis. I then examine the consumption Capital Asset Pricing Model (CAPM), which has been the workhorse of financial economics for more than two decades. In its canonical version, researchers find that the model fails empirically along several dimensions. This failure has led them to explore several modifications, which they find perform better than the canonical model, although mostly in calibration exercises that ignore the time variation in the data. I show that one of the key features of this time variation is that the ex ante return on stocks in excess of the risk free rate undergoes extended periods of time in which it is largely positive and other periods of time in which it is largely negative. I also show that the canonical CAPM and all of its proposed modifications have essentially the same implication for such sign reversals. This result enables me to propose a simple test that confronts the entire class of consumption CAPM models with the time series data. I find that the canonical model and its modifications are unable to explain the pattern of sign reversals we observe in the data. This leads me to consider an alternative model that replaces expected utility theory with endogenous prospect theory and the Rational Expectations Hypothesis with an Imperfect Knowledge Economics representation of forecasting behavior. I find that my alternative model performs better than the consumption CAPM in accounting for the time variation of stock returns.