Trade policy analysis in Brazil: Assessing welfare impacts with revised Armington elasticities

Evaluating the impacts of trade policy through simulation can offer valuable insights, especially for developing countries still confronting significant trade barriers. However, the credibility of these insights might be compromised if the simulations depend on parameters derived from developed economies or are constrained by ad-hoc assumptions within the literature. To quantify these implications, we focus on Brazil as a case study, utilizing a Melitz-style model that facilitates sector-specific estimation of both an upper-tier "homeforeign"macroelasticity and a lower-tier "foreign-foreign"microelasticity. While 66%-77% of microelasticities surpass macroelasticities, the precision falls short in confirming the "rule of two"- the assumption that the microelasticity is twice the macroelasticity. In a simulated trade liberalization scenario, our findings show that welfare changes fall below those of models adhering to this rule, yet exceed those of approaches overlooking differentiation between the two elasticities. Furthermore, incorporating imperfect factor mobility results in more moderated predictions for welfare outcomes.