"Taste heterogeneity, labor mobility and economic geography" - A critical reconsideration and correction

This article reconsiders the linear new economic geography model under heterogeneous agents developed by Tabuchi and Thisse (2002) by applying an analytical technique introduced by Ludema and Wooton (1999). Two problematic aspects are identified first, the bifurcation pattern for countries which differ in amenities is incorrect. I show that the degree of agglomeration is highest when trade costs are high. Besides this minor problem, the second critical issue concerns the welfare analysis. It is shown in this note that this model exhibits a latent tendency for overagglomeration when trade costs are high and underagglomeration when trade costs are low, bringing it in line with other welfare analyses of new economic geography models.