This paper offers some preliminary steps in the marriage of some of the theoretical foundations of the new economic geography with spatial computable general equilibrium models. Modeling the spatial economy of Brazil using the usual assumptions of CGE models makes little sense when one state, São Paulo, accounts for 40% of GDP and where transportation costs are high and accessibility low compared to European or North American standards. Hence, handling market imperfections becomes imperative as does the need to address internal spatial issues from the perspective of Brazil's increasing involvement with external markets such as MERCOSUL, EU, NAFTA. The paper builds on the B-MARIA-27, a multiregional CGE model of the Brazilian economy; non-constant returns and non-iceberg transportation costs are introduced and some simulation exercises carried out. The results, limited in this paper to short-run considerations, confirm the asymmetric impacts that transportation investment has on a spatial economy in which one state (São Paulo) is able to more fully exploit scale economies vis a vis the rest of Brazil. The analysis also reveals the importance of parameter estimation in handling imperfectly competitive markets.
Read full abstract