Understanding Dynamic Conditional Correlations between Commodities Futures Markets

We estimate dynamic conditional correlations between 10 commodities futures returns in energy, metals and agriculture markets over the period 1998-2014 with a DCC-GARCH model. We look at the factors influencing those correlations, adopting a pooled mean group (PMG) estimator. Macroeconomic variables are significantly correlated with agriculture-energy and metals-energy dynamic conditional correlations; while financial variables are relevant in the agriculture-energy correlations and poorly significant in the metals-energy ones. Speculative activity is generally not statistically significant. Correlations started increasing in the years before the financial crisis and decreased at the end of our period of analysis