An econometric analysis of trade diversion under NAFTA

We provide an econometric analysis of whether or not the tariff preferences extended to Canada and Mexico under the North American Free Trade Agreement (NAFTA) may have resulted in trade diversion. A review of previous studies, both descriptive and econometric, suggests that trade diversion has occurred especially as evidenced by Mexico's increased shares of US imports apparently at the expense of several Asian countries. We use a conceptual framework based on a partial-equilibrium model of differentiated product industries under monopolistic competition for many countries. The model is implemented empirically using a fixed-effect panel analysis of US imports at the Harmonized System (HS) 2-digit level for the period, 1992-1998. Of the 70 sets of regressions that were run, the coefficients of the tariff rates were statistically significant in 15 cases. The strongest evidence of trade diversion was found mainly for US imports of textile and apparel products. We also estimated regressions for selected commodities at the HS 4-digit level. The results suggest trade diversion for textiles, apparel, and some footwear products but not for trade in motor cars and vehicles and television receivers, which may have been more influenced by changes in foreign direct investment and outsourcing rather than tariff preferences.