Aranceles preferenciales; Comercio internacional; Integración; Modelos con datos panel; Modelos de conteo.; Modelos de selección binaria; Política comercial; Tratados de Libre Comercio
Exports concentration in a reduced number of economic sectors has been always a concern for trade policy makers, mostly in developing countries. According to the economic theory and some recent empirical evidence, market access in international trade is improved through free trade agreements and preferential tariff rate programs which could affect positively export diversification. Following a similar approach used by Debaere and Mostashari (2005) and Volpe and Gómez (2007), this paper estimates the effects on Colombia’s export diversification induced by the lower tariffs offered by Mexico and the broader preference margin vis-à-vis third countries in the Mexican market derived from the G3 Free Trade Agreement (FTA). For this purpose, dynamic Probit and dynamic Poisson models are estimated using an unbalanced panel data available from 1995 –year when the G3 FTA entry into force– to 2010, the latest year available. Moreover, this paper also attempts to estimate the effects of Venezuela’s withdrawal from G3 in 2006 and the effects of the bilateral crisis between Colombia and Venezuela in the second half of 2000s. In general, results show that lower tariffs faced by Colombian products in Mexico are associated with higher export diversification but this impact is sensitive to the economic cycle and a negative impact on diversification can arise depending on the specification of the model. Also, broader preferential margins in Mexico induce higher export diversification for Colombian products. However, there exists a strong dependence of export diversification in Colombia on the initial conditions and the past of the export portfolio offered to Mexico. On the other hand, Venezuela’s withdrawal from G3 FTA in 2006 had no effect on export diversification from Colombia to Mexico meanwhile the bilateral crisis showed –in some specifications– positive effects on the export diversification to Mexico, the latter revealing a trade deviation pattern. Macroeconomic control variables –including the real exchange rate– showed to be irrelevant to the model.