Economic importance of correlations for energy and other commodities

Abstract

This paper uses a non-parametric model to examine time-varying correlations for energy and other commodities and tests their economic importance by combining dynamic correlations with the momentum strategy. Our empirical analysis offers three new findings. First, the profits from a trading strategy that accounts for the patterns in correlations are higher than the profits that do not account for the time-varying correlations. Second, profits are asymmetric to higher and lower levels of correlation. Profits are maximized at higher levels of correlation. These profits are robust to the financialization period and the backwardation and contango phases of the commodity market. Finally, the economic source of profits reveals that profits are largely explained by the basis and momentum factors, the exception being grains and softs portfolios at higher correlation levels.