This paper investigates the role of commodity price information when evaluating inflation risk. Using a model averaging approach, we provide strong evidence of in-sample and out-of-sample predictive ability from commodity prices and convenience yields to inflation, establishing clear point and density forecast performance gains when incorporating disaggregated commodities price information. The resulting forecast densities are used to calculate the (ex-ante) risk of inflation breaching defined thresholds that broadly characterize periods of high and low inflation. We find that information in commodity prices significantly enhances our ability to pick out tail inflation events and to characterize the level of risks associated with periods of high volatility in commodity prices.
