The forecasting approach to efficiency in the wool market

This paper develops and presents estimates of a simultaneous equations model of the Australian wool market, the world's largest producer and exporter of fine wool. The model contains functional relationships for unhedged inventories, consumption of raw wool, and the activities of both hedgers and speculators in wool futures. Expectations are represented by the adaptive hypothesis. This model extends the work of Leuthold and Hartmann (1979) and Leuthold and Garcia (1988) by including expectations in the spot-futures model, and that of Goss and Giles (1986) by including composite equations for hedger-speculators, extending the expectations hypothesis to the consumption equation, and by using the model to test the efficient markets hypothesis. Wald tests and likelihood ratio tests for unit roots in wool cash prices are conducted and in no case can the hypothesis of a single unit root be rejected. Estimation is by three stage least squares, with correction for first order serial correlation. The model provides good intra- and post-sample forecasts of most variables, especially of unhedged inventories and consumption of wool, both important spot market relationships. The model-derived forecast of the spot price is inferior to the forecast implicit in the futures price, although a compositive predictor clearly outperforms the futures price as an anticipation of subsequent cash prices. Nevertheless, it is suggested that the efficient markets hypothesis should not be rejected, because there is evidence that futures market agents are learning to use the information contained in the model.