Skewness Based Investment Strategies on Commodity Futures

Abstract

Mitton & Vorkink (2007) and Barberis & Huang (2008) have shown that investors exhibit lottery (or skewness) preferences where they prefer assets that have small probabilities of delivering large payoffs. Skewness preferences have also been shown to manifest in commodity futures through selective hedging (Stulz 1996). Selective hedging stands for a practice, where subjective views are incorporated into hedging decisions to maintain exposure to upside events. As such hedgers in commodity markets sell more futures (less) futures when the futures returns are negatively (positively) skewed (Gilbert, Jones & Morris 2006). Subsequently, positively skewed futures are expected to become overpriced and negatively skewed futures underpriced (Fernandez-Perez, Frijns, Fuertes & Miffre 2018). The performance of alternative skewness based strategies on commodity futures are analyzed in this thesis.

The commodity futures risk premium has recently been decomposed into a spot premium and a term premium by Szymanowska, Roon, Nijman & Goodberg (2014). The spot premium represents the near-term price risk of the underlying asset and is captured by either long or short positions in the front contract. The term premium is interpreted as deviations from expectations hypothesis, where markets might for example expect changes in convenience yields. The term premium can be captured by calendar spreads, where a short position is taken on the front contract and a long position is taken on a farther maturity contract.

Following Fernandez-Perez et al (2018) a skewness strategy that captures spot premia is implemented. The spot strategy ranks commodities into quintiles by realized daily return skewness over the previous year and buys front contracts on the lowest quintile and shorts front contracts on the highest quintile. In addition, novel term premia strategies are also implemented. The term premia strategies use the same ranking instrument as the spot strategy and the first term based strategy buys calendar spreads on commodities with below-median skewness and the other strategy buys spreads on commodities with above-median skewness. The term premia strategy is implemented in alternative maturity series, where the farther contract matures in at least 4, 6, or 8 months. The data contains 25 commodities that are traded in the US and the period ranges from 03/1996 – 10/2019. Data is obtained from Datastream.

The spot skewness strategy delivers a significant average excess return (13,2% pa) and significant alpha (9,8-12% pa) when regressed on commodity pricing models. The results of the spot skewness strategy are driven by the poor performance of the short leg, positively skewed commodities, and support the theories of lottery preferences and selective hedging. The term premia skewness strategies deliver smaller but significant average excess returns. However, the risk-adjusted returns are found to be insignificant when term-structure based factors are controlled for. As such, the skewness signal does not contain novel information about term premia.