Commodity risks and the cross-section of equity returns

Abstract

The article examines whether commodity risk is priced in the cross-section of global equity returns. We employ a long-only equally-weighted portfolio of commodity futures and a term structure portfolio that captures phases of backwardation and contango as mimicking portfolios for commodity risk. We find that equity-sorted portfolios with greater sensitivities to the excess returns of the backwardation and contango portfolio command higher average excess returns, suggesting that when measured appropriately, commodity risk is pervasive in stocks. Our conclusions are robust to the addition to the pricing model of financial, macroeconomic and business cycle-based risk factors.

Is idiosyncratic volatility priced in commodity futures markets?

“…This finding is not surprising in the light of evidence that suggests that the backwardation/contango cycle acts both as a priced risk factor in equity markets and as a leading indicator of future economic activity (e.g. Baker & Routledge, 2012;Bakshi et al, 2015;Brooks, Fernandez-Perez, Miffre, & Nneji, 2016;Fernandez-Perez, Fuertes, & Miffre, forthcoming;Koijen, Moskowitz, Pedersen, & Vrugt, 2015).…”

Commodities as an Asset Class

“…We find that the value factor (HML) and investment factor (CMA) are able to explain part of the cross-section of returns of individual commodities. This is consistent with the studies that suggest that commodity risks are somehow priced in relation to stock markets (see e.g., Boons et al, 2012;Brooks et al, 2016), suggesting an increasing integration between commodity and stock markets. The predicted expected returns of the Fama and MacBeth (1973) two-step procedure are calculated with: E r e i = β i, f λ f ,t , where β i, f is the estimated beta with the time-series regressions (the first step) of Fama and MacBeth (1973).…”

“…Furthermore, we show that commodity and stock markets are somewhat integrated because asset pricing factors help explain the cross-section of commodity returns. This result provides further evidence for the literature that tries to link these markets (see e.g., Alves and Szymanowska, 2019;Boons et al, 2012;Brooks et al, 2016;Hou and Szymanowska, 2013;Lutzenberger, 2014;Salisu et al, 2019).…”

“…9 This provides us both with time-series asset pricing tests in the first stage, and gives an estimate of the risk premium of the risk factors in the second-stage cross-sectional regressions. This approach also is commonly used in the related literature (see e.g., Brooks et al, 2016;Daskalaki et al, 2014;De Roon and Szymanowska, 2010;Hollstein et al, 2021;L übbers and Posch, 2016).…”

“…have insignificant intercepts and significant risk premiums for risk factors, it is not possible to point at one single model (or factor) which performs best. In line with Bakshi et al (2019), Brooks et al (2016), and Jagannathan and Wang (1996), we visualize the performance of the studied models in explaining the cross-sectional variation in expected individual commodity returns by plotting the predicted expected excess return versus the realized expected excess return of individual commodities in Figure 3.1. The predicted expected excess return is calculated as the sum of the products of risk premiums and corresponding risk loadings.…”