This study obtains parameter estimates relating to the United States and foreign demand for high protein meals during the period 1955-1972.

The world economy of high protein meals was formulated in a theoretical framework which provided a basis for the statistical estimation. The core of the analysis is a simultaneous equation model composed of eight linear stochastic behavioral relations and four linear identities. The three-stage least squares procedure was employed to estimate the unknown structural parameters of the statistical model.

The statistical analyses of this study contribute to the quantification of demand interrelationships between soybean meal and other high protein meals, and to the demand relationships between the United States and foreign markets. Elasticity measures were computed from the estimated structural coefficients. The results of the present study suggest that domestic, as well as foreign demand, for soybean meal during the 1955-72 period were price inelastic. In addition, this study suggests that demand for soybean meal has become more elastic in the recent years. A technical complementary relationship was found between corn and other high protein meals in the United States. Soybean and other high protein meals appeared to be very close substitutes as livestock feeds both in the United States and foreign markets. Furthermore, the study indicates that both export and import demands for protein meals were price inelastic. Particularly, the United States demand for imports of other high protein meals which was estimated to be highly price elastic.

Derived reduced forms were computed from three-stage least squares structural parameters. The unrestricted least squares reduced forms were estimated using ordinary least squares procedure. Effects of a specific change in the predetermined variables on all the jointly determined variables were examined using the three-stage least squares reduced form. The price of corn and animal units was found to have a great impact on the consumption of high protein meals.

Economic forecasting was evaluated and performed on the basis of the reduced form equations, using both derived reduced form equations and ordinary least squares estimated reduced form equations. When compared with actual values, within the sample period, ordinary least squares estimated reduced form equations tend to be superior than reduced form equations computed from three-stage least squares structural coefficients. Nevertheless, comparison of the two methods in forecasting test using observations for 1973 and parameters estimated from observations for the period 1955-72 was inconclusive. Finally, this study suggests that a structural change might have occurred in recent years.