Behavioral preferences and beliefs in asset pricing

This thesis studies topics in behavioral asset pricing. The proposed contributions are both theoretical and empirical in nature. First, Chapter 1 provides a brief introduction for the rest of the thesis. Next, Chapter 2 develops a new model of the effect of beta ambiguity in markets, both with and without non-ambiguous alternatives. The model predicts a drop in the premium of assets with an ambiguous beta once assets without ambiguity become available. We conclude the chapter by showing that inflation risk premium in the equity market drops once TIPS become available. Next, Chapter 3 presents a new model of information incorporation by an agent with a cost of updating beliefs. The model predicts both over- and under-reaction of prices to information. In addition, the model relates the strength of over- and under-reaction to signal extremeness and to beliefs variance. The chapter is concluded with empirical evidence of momentum and reversal that are manifested in the cross-section of equity. Finally, Chapter 4 proposes a new way of testing models of ambiguity aversion. We show theoretically that different models of ambiguity aversion generate different predictions in term of the preference for the timing and graduality of ambiguity resolution. We then develop an experimental framework and document preference for late and gradual resolution of ambiguity.