The perceived value of integrating small partial- equilibrium structural models of individual livestock industries into a comprehensive single-sector model is to take advantage of the interrelationships that are usually expressed by cross elasticities on both the supply and demand sides of these industries. Model integration should provide a more realistic representation of the livestock industries and an improved mechanism for industry analyses. However, model integration could also lead to increased error in model simulation that could reduce the value of the larger model for those purposes. Using forecasting as an example application, this paper investigates how the increased endogenisation of cross-commodity relationships in alternative structural econometric models of the Australian livestock industries affects the simulation performance of the larger model. Forecast accuracy and encompassing tests were used to investigate the value of model integration by comparing the accuracy of the models' forecasts and by testing for differences in the information contained in those forecasts. The general result was that combining the models did not adversely affect the forecasts from the integrated model and the encompassing tests indicated that the forecasts of the integrated and single models contained different information. Because the forecasts of the integrated model were not impaired relative to the single model forecasts, model integration was considered to be useful for forecasting and other types of economic analysis in the livestock industries.