The Economics of Commodity Markets

5

Cross-Commodity Linkages

This chapter deals with a cointegration analysis commodity by commodity in their respective sub-category, i.e. agricultural products, industrial metals, precious metals, and energy markets. The main focus lies in the determination of the long-term relationship – if any – between these specific categories of commodities, and the identification of the common factors that could explain the convergence of the price series towards a fundamental value. First, we recall the main econometric steps to proceed with a cointegration analysis. Then, we carry out our investigation for each sub-category.

5.1 A PRIMER ON GRANGER CAUSALITY TESTING AND COINTEGRATION

5.1.1 Granger Causality Testing

In conjunction with the analysis of the matrix of cross-correlations, the econometrician may resort to Granger causality tests as well. These tests allow us to infer causality ‘in the Granger sense’ between a set of dependent and independent variables selected by the user, and may be useful in econometric modeling prior to the regression analysis. When applied to commodity markets, Granger causality tests will tell us the nature of the inter-relationships between the various markets and categories of commodities.

Recall that a process P1t Granger causes P2t at the order p if, in the linear regression of P2t on lagged prices , at least one of the regression coefficients of P1