Estimating Oil Risk Factors Using Information from Equity and Futures Markets

Abstract:     
This study makes four contributions to the commodity pricing literature. First, a new four-factor model for commodity prices is developed, where each factor has a simple economic interpretation. The model exhibits unspanned stochastic volatility and is demonstrated to fit oil futures prices with comparable or lower root mean square error than reported in previous studies. Second, it is shown that stock returns of oil-related firms may be included in the estimation to give significant additional information about the oil risk factors. Our factors explain a large portion of return variation in oil-related portfolios, with an average R2 of 70%; for comparison, the average R2 with the Fama-French-Carhart (FFC) model is 22%. In contrast with previous studies, the oil factors extracted using our methodology are associated with a significant risk premium and are shown to be systematic. Finally, we construct factor mimicking portfolios that permit the trading and hedging of our factors. The latter are shown to be significantly related to the FFC factors, can explain the risk premium in all but the momentum factor, and exhibit half the explanatory power of the non-market FFC factors among industry portfolios (excluding oil).

Number of Pages in PDF File: 66

Keywords: Oil, futures, asset pricing, APT, real options, CAPM, Fama-French factors, Markov chain Monte Carlo

JEL Classification: G12, G13

working papers series