A dynamic approach to estimating and testing separability in U.S. demand for imported and domestic meats

The paper develops dynamic model of demand for domestic and imported meats and compares the performance of the dynamic model called the general AIDS error correction model (GAECM) with the static linear approximate almost ideal demand system (LA/AIDS) model. A test for separability is developed based on the long-run solution to the dynamic model. The results indicate rejection of the static model in favor of the dynamic model. Homogeneity and symmetry are shown to hold for the dynamic model but are strongly rejected for the static model. Finally, separability between domestic and imported meats is rejected for all models at the 5% level of significance. However, the results of the dynamic model based on the corrected likelihood test with a low p-value (p = 0.038) indicate that rejection of separability is borderline.