Forecasting Commodity Prices: Futures Versus Judgment
Electronic Access:
Free Download. Use the free Adobe Acrobat Reader to view this PDF file
Summary:
This paper assesses the performance of three types of commodity price forecasts—those based on judgment, those relying exclusively on historical price data, and those incorporating prices implied by commodity futures. For most of the 15 commodities in the sample, spot and futures prices appear to be nonstationary and to form a cointegrating relation. Spot prices tend to move toward futures prices over the long run, and error-correction models exploiting this feature produce more accurate forecasts. The analysis indicates that on the basis of statistical- and directional-accuracy measures, futures-based models yield better forecasts than historical-data-based models or judgment, especially at longer horizons.
Series:
Working Paper No. 2004/041
Subject:
Agricultural commodities Commodities Commodity prices Financial institutions Futures Oil Prices
English
Publication Date:
March 1, 2004
ISBN/ISSN:
9781451846133/1018-5941
Stock No:
WPIEA0412004
Pages:
28
Please address any questions about this title to publications@imf.org