• trade costs;
  • heterogeneity;
  • uneven development;
  • gravity equation


In our study, we tested the hypothesis of the bidirectional causality between trade costs and economic development using data on Italian provinces. Using different methods to control for multilateral resistance, we applied a gravity equation to estimate sectoral exports to 188 countries over the period 1995–2004. Provincial trade costs were constructed as the sum of five province-specific elasticities, including distance, adjacency, and common money. We found that Italian provinces are heterogeneous with respect to trade costs. These costs are influenced by lagged provincial per capita income and industrial structure. In turn, trade costs influence future provincial per capita income. This bidirectional relationship between trade costs and income is broadly consistent with the two-way causation process emphasized by the New Geographical Economics.