Two Essays on Asset Return Predictability

兩篇關於資產收益可預測性的論文

Student thesis: Doctoral Thesis

View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Awarding Institution
Supervisors/Advisors
Award date16 Aug 2021

Abstract

This dissertation is composed of two essays on asset pricing. The first essay investigates the predictive power of reciprocal return risk premium on the option returns in the cross-section. The second essay examines the predictability of three covariance risk premium components on the equity market return.

The first essay presents a robust new finding that reciprocal return risk premium, defined as the difference between the expected reciprocal of return under physical measure and under risk-neutral measure, significantly predicts the option returns in the cross-section. After controlling for numerous standard firm characteristics, the positive relation between reciprocal return risk premium and option returns remains. The high minus low spread of the delta-hedged option portfolios sorted on the reciprocal return risk premium can not be explained by standard factors, and the double sorting trading strategies controlling for a collection of firm characteristics are still profitable. Both the good and bad uncertainty contribute to the predictability. The positive relation between reciprocal return risk premium and option returns is mainly driven by the RVIX2 component, which is by definition a natural measure for return variation under risk-neutral measure and a good tracker for tail risk embedded in the out-of-money put options.

In the second essay, I decompose the market variance risk premium(VRP) into three covariance risk premium components (good, bad, and mixed covariance risk premiums) based on the signs of the index constituent individual stock returns. I find that the three covariance risk premium components capture different time-varying risk compensations for the realization of the equity returns. Hence the three components convey distinct information about future market excess returns. More specifically, I find that the documented strong predictive power of VRP on the market excess returns in the short run mainly resides in the good covariance risk premium component. Using the three covariance risk premiums jointly can improve the predictability. The empirical results are remarkably robust across different sample periods or inclusion of traditional alternative predictors as control variables.

    Research areas

  • reciprocal return risk premium, return predictability, option-implied information, covariance risk premium