Commodity price movements feature prominently in attempts to model the channels through which `the South' affects Northern economies. It is also important to determine how commodity prices are themselves affected by changing rates of economic activity and demand in the North. So far this has been analysed in terms of specific commodities; attempts to investigate this issue using aggregate commodity groups have been rudimentary.
In Discussion Paper No. 319, Prathap Ramanujam and Research Fellow David Vines develop and estimate a simple structural model consisting of supply, demand and price equations for four aggregate commodity groups food, beverages, agricultural raw materials, and metals and minerals. The authors' model incorporates stock data for the first time: forward-looking agents adjust their holdings of commodity stocks and of financial assets in response to changes in world interest rates, which can thus affect commodity prices.
Ramanujam and Vines use their model to study the effect of industrial production in the developed countries on primary commodity prices. Their estimates reveal a large short-run effect on commodity prices, as a consequence of commodity stocks' buffering behaviour. The results also suggest smaller, but still significant permanent effects, which vary across the commodity groups. The metals and mineral equation gives the most significant coefficients in the price equation, so the authors focus on that commodity group in their simulations of stockholding and buffering behaviour.
A 1% rise in OECD GDP causes a 1.5% increase in the demand for metals and a 1.5% permanent rise in real metal prices. This result contradicts earlier studies, which found no significant impact, Ramanujam and Vines note. In the short term neither supply nor demand are very price-responsive, so the short-run price `overshoots', rising by around 4%. Commodity stocks fall in the short run, then return to their previous level in the long term. Without the `buffering' effect of reductions in stockholdings, prices would overshoot even more.
A 1% increase in metals supply has similar, but slightly smaller effects in the opposite direction. Initially prices fall by about 3% and in the long run by 1%. Increases in stocks `buffer' some of the increase in supply, dampening the short-run downward overshooting of prices.
In response to an increase in world interest rates, real metal prices initially fall quite sharply. The negative effect on commodity stockholdings as agents buy into financial assets completely dominates reduced commodity supply caused by increased costs of production. Over time prices rise again, and in the long run the effect is a small increase in price. On the other hand the reductions in commodity stockholdings occur only slowly in response to the interest rate increase. Lower prices feed through gradually into higher demand and lower supply, until stocks fall to the level at which people are prepared to hold them willingly, even though the rate of interest on financial assets is higher than before. This simulation illustrates the importance of financial factors in commodity price behaviour, Ramanujam and Vines conclude, and confirms the need to model the interactions between commodity and financial markets


Commodity Prices, Financial Markets and World Income: A Structural Rational Expectations Model
Prathap Ramanujam and David Vines

Discussion Paper No. 319, June 1989 (IM/IT)