Good volatility, bad volatility and the cross section of commodity returns

This article studies whether asymmetries in volatility help explain the cross section of commodity returns. We decompose realized variance into upside and downside components and construct a normalized difference measure, the relative signed jump (RSJ), following Bollerslev et al. (2020). A trading strategy that goes long the top tercile of commodities with the highest RSJ and shorts the bottom tercile delivers a statistically and economically significant annualized excess return of -6.29%. We also find that our tradable RSJ factor explains the cross section of commodity returns beyond well-established factors in a multivariate price setting context. Our results also show that the pricing ability of volatility asymmetries is distinct from other higher order moments such as realized skewness.