Forward Moments and Risk Premia Predictability
40 Pages Posted: 6 Mar 2017 Last revised: 30 Jul 2018
Date Written: January 30, 2018
We estimate a term structure of risk-neutral forward skewness and examine its information content for equity market risk premia in conjunction with that of the already established forward variance. We use Partial Least Squares to extract from each term structure a single predictive factor that is motivated by the theoretical implications of affine no-arbitrage models. The resulting forward variance factor, FVF (forward skewness factor, FSF), reflects changes in the slope (curvature) of the term structure. The empirical analysis reveals that both factors exhibit significant in-sample and out-of sample predictability for the equity and the variance premium. More importantly, FSF exhibits equity premium predictability that is stronger than and incremental to that provided by FVF. However, FSF does not outperform FVF in the case of variance premium predictability.
Keywords: Forward Moments; Implied Volatility Surface; Partial Least Squares; Predictability of Stock Returns; Equity Premium; Variance Premium
JEL Classification: G10; G11; G12
Suggested Citation: Suggested Citation