There are three basic assumptions to forecasting:
1. The future will be similar to the past.
2. Forecasts are rarely accurate.
3. Accuracy decreases with time.
Given those somewhat dismal assumptions, a lot of progress has been made in the forecasting of commodities over the past 30 years. Forecasting commodity prices has become an interdisciplinary mix of math, statistics, physics and computer science. Both academic and technical advances have made it much easier to produce complex mathematical models. Still many challenges remain.
Unforeseen structural changes to markets render models obsolete. This has been particularly true in recent years as corn, sugar and vegetable oil have become fuel as well as food sources. Huge inflows of speculative money have distorted old supply/demand/price relationships. New global demand (China, India) and changes in climate will continue to restructure market dynamics in the future. As a result, mathematical models while very useful, remain subject to the winds of change. Forecasting commodities can still be as much of an art as it is a science. However, with knowledge of their limitations, forecasts can still be used profitably to manage risk.
For a better understanding of commodities, analysis and forecasting, we suggest the following reading list:
- Forecasting Commodity Markets, Julian Roche, 1995 Probus
- Schwager on Futures – Fundamental Analysis, Jack Schwager, 1995 John Wiley
- Agricultural Price Analysis & Forecasting, John Goodwin, 1994 John Wiley
- Modeling & Forecasting Primary Commodity Prices, Walter Labys, 2006 Ashgate
- Commodities & Commodity Derivatives, Helyette Geman, 2005 John Wiley
- Agricultural Prices & Commodity Market Analysis, John Ferris, 1998 McGraw-Hill