Multicountry-models and international trade: a critical survey

This work aims at analyzing selected empirical trade models to compare their main characteristics and their potential in explaining and forecasting international phenomena. This critical survey provides a general framework of the main issues and challenges which are faced by researchers engaging in modelling international trade flows. We firstly introduce some criteria to select the group of models analyzed in this work. They are empirically estimated models with a significant sectoral and country detail, generally used for scenario simulation analysis and forecasting. Then, we analyze how multi-country models have developed from a pooling approach where global imports are aggregated and shared-out into single-country exports with fixed shares, to a more sophisticated approach taking advantage of bilateral trade flows information. Significant differences across existing models persist and affect their results. We argue that some characteristics – such as the endogenization of trade shares, the high disaggregation of commodity flows, the path-dependence of bilateral trade influenced by relative competitiveness, and the inclusion of the trade of services -- are crucial challenges to modelers for making the international trade channel more effective and realistic.