Subjective Beliefs and Asset Prices

Asset prices are forward looking. Therefore, expectations play a central role in shaping asset prices. In this dissertation, I challenge the rational expectation assumption that has been influential in the field of asset pricing over the past few decades. Different from previous approaches, which typically build on behavioral theories originated from psychology literature, my approach takes data on subjective beliefs seriously and proposes empirically grounded models of subjective beliefs to evaluate the merits of the rational expectation assumption. Specifically, this dissertation research: 1). collects and analyzes data on investors' actual subjective return expectations; 2). builds models of subjective expectation formation; 3). derives and tests the models' implications for asset prices. I document the results of the research in two chapters. In summary, the dissertation shows that investors do not hold full-information rational expectations. On the other hand, their subjective expectations are not necessarily irrational. Rather, they are bounded by the information environment investors face and reflect investors' personal experiences and preferences. The deviation from fully-rational expectations can explain asset pricing anomalies such as cross-sectional anomalies in the U.S. stock market. In the first chapter, I provide a framework to rationalize the evidence of extrapolative return expectations, which is often interpreted as investors being irrational. I first document that subjective return expectations of Wall Street (sell-side, buy-side) analysts are contrarian and counter-cyclical. I then highlight the identification problem investors face when theyform return expectations using imperfect predictors through Kalman Filters. Investors differ in how they impose subjective priors, the same way rational agents differ in different macro-finance models. Estimating the priors using surveys, I find Wall Street and Main Street (CFOs, pension funds) both believe persistent cash flows drive asset prices but disagree on how fundamental news relates to future returns. These results support models featuring heterogeneous agents with persistent subjective growth expectations. In the second chapter, I propose and test a unifying hypothesis to explain both cross-sectional return anomalies and subjective return expectation errors: some investors falsely ignore the dynamics of discount rates when forming return expectations. Consistent with the hypothesis: 1) stocks' expected cash flow growth and idiosyncratic volatility explain significant cross-sectional variation of analysts' return forecast errors; 2). a measure of mispricing at the firm level strongly predicts stock returns, even among stocks in the S&P500 and at long horizon; 3). a tradable mispricing factor explains the CAPM alphas of 12 leading anomalies including investment, profitability, beta, idiosyncratic volatility and cash flow duration.