Essays in Expectation Formation and Asset Pricing
Xu, Zhengyang
2020
Abstract
This dissertation examines the role of investors' belief formation in asset valuation. In the first chapter, I document that subjective bond risk premia implied by survey forecasts of future Treasury yields are acyclical at the one-year horizon. This is in stark contrast to large countercyclical variation in objective risk premia fitted from in-sample predictive regressions of future bond excess returns. This difference in risk premia implies a wedge between subjective and objective expectations of future short rates, which I show is predictable by trend and cycle components of macroeconomic forecasts. I show that these empirical findings can be explained with a learning model in which the agent filters latent trend and cycle components of fundamentals in real time, while an econometrician analyzing the data ex-post has full knowledge of the data-generating processes. The model also yields predictions, consistent with the data, on the joint behavior of the unconditional yield curve slope, the cyclicality of short-rate and macroeconomic expectation wedges, and the cyclicality of objective risk premia. My results suggest that equilibrium models of bond risk premia should target acyclical subjective risk premia and expectation formation, rather than ex-post in-sample fitted risk premia from predictive regressions. The second chapter (co-authored with Stefan Nagel) builds on recent evidence that lifetime experience shapes individuals’ macroeconomic expectations and it explores the asset-pricing implications of this evidence for the aggregate US stock market. We study an economy in which a representative agent learns---but with fading memory---about the constant underlying endowment growth rate. The agent downweighs observations in the distant past but is otherwise Bayesian in evaluating uncertainty. The model explains both standard asset pricing facts and investor expectations within a simple and tractable framework, in which subjective belief dynamics are constrained by survey data. In the model, fading memory implies perpetual learning and permanently high subjective uncertainty about long-run growth, but the subjective equity premium is virtually constant. In contrast, an econometrician who knows the true long-run growth will find a high and strongly countercyclical objective equity premium which is predictable. Consistent with this theory, we show empirically that experienced payout growth (an exponentially weighted average of past growth rates) is negatively related to future stock market excess returns, predicts survey expectation errors, and is positively related to aggregate analyst forecasts of long-run earnings growth. In the third chapter, I argue that econometricians find high returns from trading on the profitability anomaly because investors in real time failed to spot profitable firms that are difficult to analyze. I document that the Fama-French three-factor alphas of the profitability anomaly only exist among firms with high information frictions, proxied by young age, high forecast dispersion, high past return volatility, and/or high option-implied volatility. The results are robust to excluding micro-firms and using different measures of profitability. Short-sale constraints, liquidity, and financial distress do not fully account for the alphas. I show that the empirical pattern is consistent with a noisy rational expectations equilibrium model in which investors use profitability as a noisy signal to learn about future firm payoffs.Subjects
Belief formation Equity Fixed income General equilibrium
Types
Thesis
Metadata
Show full item recordCollections
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.