Asset Mispricing

Abstract: This dissertation studies the pricing of stocks in capital markets. It comprises five chapters, where the first serves as an introduction. The subsequent four chapters are each written as self-contained research papers. While the theory of efficient markets serves as the theoretical foundation, I approach the research from a conceptual starting point that recognizes market mispricing.The first paper investigates a testing methodology of market efficiency based on fundamental valuation. The methodology is based on an investment strategy where stocks with high (low) V/P-ratios are assigned into long (short) portfolios. We conjecture that under the assumption of independence between the portfolio assignment and systematic risk, a positive return from such investing strategy is inconsistent with market efficiency. We estimate fundamental values based on a flexible residual income valuation model via the state-space framework and implement the investment strategy on a sample of U.S. stocks spanning 1980–2017. The implementation shows a significant positive monthly return. Moreover, the results are substantiated in a standard five-factor model. In sum, these results appear anomalous with respect to market efficiency, at least as given by the five-factor model.The second paper examines whether improvements in earnings forecasting translate into improvements in implied cost of capital estimates of expected returns. I attain high-performing earnings forecasting via a machine learning approach. In particular, I implement and evaluate six popular machine learning methods to forecast earnings. The evaluation demonstrates that the machine learning algorithms can generate earnings forecasts that consistently outperform state-of-the-art benchmarks. Moreover, I estimate the implied cost of capital on a sample of U.S. stocks spanning 2000–2017. The general result indicates that improvements in earnings forecasting do not translate into improvements in return predictability. While issues with the implied cost of capital methodology could explain the results, another possible explanation is market mispricing.The third paper compares the performance between the implied cost of capital and factor model approaches in estimating the cost of capital in an inefficient market. I conduct the comparison in a Monte Carlo simulation experiment. The simulation results indicate that the implied cost of capital approach is more robust to market inefficiency.The fourth paper analyzes investor learning of cash flow expectations in the context of market efficiency. I argue that the bias-variance tradeoff translates into inefficiencies in market pricing. Moreover, in a simple model, I prove that these inefficiencies can be exploited by an investor aware of whether market prices exhibit a bias or suboptimal variance.