The Mistreated Model: Some Technical Comments on Porojan's Paper on 'Trade Flows and Spatial Effects'

The paper by Porojan on 'Trade flows and spatial effects' in a recent issue of this journal is criticized on three methodological grounds: it includes two of the gravity model independent variables as relative rather than absolute numbers; the dummy variables included are not binary but interval; and the 'sample' of countries creates a non-normal distribution of inter-country distances, undoubtedly creating heteroscedasticity problems. Because of these three problems, the model is mis-specified, and the results thereby obtained are unreliable. Copyright Kluwer Academic Publishers 2003