Can Futures Prices Predict the Real Price of Primary Commodities?

Can futures markets provide useful real-time forecasts of period average commodity prices? We consider seventeen primary commodities across energy, metals, and agricultural markets and find that futures-based forecasts of period averages outperform the (end-of-period) random walk forecast for the majority of commodities. We document that the prior mixed evidence on the usefulness of futures-based forecasts was driven by the time-sampling of the futures and no-change benchmark data, as well as the forecast evaluation period examined. We show that non-parametric approaches based on the most recent trading data (in lieu of averaging) are the most accurate. Results suggest that academics, policymakers, and industry can consider utilizing futures prices as forecasts of commodity prices.