Essays on Empirical Asset Pricing

Abstract

This thesis is comprised of three Chapters. Chapter 1 investigates the pricing implication of inequality risk in the cross-section of Chinese stock returns. We find a significant negative inequality risk premium regardless of whether tests are performed on individual stocks or portfolios. Moreover, we find evidence that inequality risk exposure is mildly positively associated with idiosyncratic volatility and turnover rate, while it is negatively associated with the E/P ratio and ownership concentration. This supports the conjecture that poorer investors (proxied by retail investors) tend to hold assets with higher inequality betas.

Chapter 2 investigates the pricing implication of macroeconomic factors in the cross-section of Chinese stock returns. Our results suggest that the Chinese stock market now prices macroeconomic risks similarly to more developed markets, challenging its previous characterization as a "casino." Most of the macroeconomic factors tested have significant loadings on the stochastic discount factor and economically significant risk premiums. Furthermore, we find that the Fama-French five factors become superfluous in the presence of the macroeconomic factors. However, the performance of the conventional macroeconomic factors is unstable over different test assets, highlighting the need for further research to identify macroeconomic factors that could better describe the cross-section of Chinese stock returns.

Chapter 3 aims to improve the explanatory power of macroeconomic factors in the cross-section of Chinese stock returns. To do so, we construct a set of machine-learning-enhanced macroeconomic factors using a novel procedure that integrates the theory of the ICAPM with Elastic Net regression. Our results show that these machine-learning-enhanced factors capture more useful pricing information than traditional ad hoc sparse macroeconomic factors, outperforming the PCA and Chen-Roll-Ross (1986) factors in most scenarios.