Working Paper: CEPR ID: DP7689
Authors: Jan J. J. Groen; Paolo Pesenti
Abstract: In this paper we seek to produce forecasts of commodity price movements that can systematically improve on naive statistical benchmarks, and revisit the forecasting performance of changes in commodity currencies as efficient predictors of commodity prices, a view emphasized in the recent literature. In addition, we consider different types of factor-augmented models that use information from a large data set containing a variety of indicators of supply and demand conditions across major developed and developing countries. These factor-augmented models use either standard principal components or partial least squares (PLS) regression to extract dynamic factors from the data set. Our forecasting analysis considers ten alternative indices and sub-indices of spot prices for three different commodity classes across different periods. We find that the exchange rate-based model and especially the PLS factor-augmented model are more prone to outperform the naive statistical benchmarks. However, across our range of commodity price indices we are not able to generate out-of-sample forecasts that, on average, are systematically more accurate than predictions based on a random walk or autoregressive specifications.
Keywords: commodity prices; exchange rates; factor models; forecasting; PLS regression
JEL Codes: C23; C53; F47
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Changes in commodity currencies (F31) | Future commodity prices (G13) |
| Exchange rate-based model (F31) | Predictions of commodity prices (G13) |
| PLS factor-augmented model (C51) | Predictions of commodity prices (G13) |
| Exchange rate-based model (F31) | Outperforming naive benchmarks (C52) |
| PLS factor-augmented model (C51) | Outperforming naive benchmarks (C52) |