Uma análise para o efeito-fronteira no Brasil

Abstract

This paper studies the border effect in the Brazilian international trade. We consider a structural gravity model developed by Anderson e Wincoop (2003) and the empirical specifications proposed by Baldwin e Taglioni (2006, 2011). The econometric method utilized in this paper is robust against several identification problems such as a large number of trade observations equal to zero, heteroskedasticity and unknown specification of the error distribution. Our results suggest that the higher border effect in the Brazilian international trade, largely corroborated by the existing literature, is due to misspecification of the gravity equation, specially because of the exclusion of the Multilateral Resistance Term (MRT) suggested by Anderson eWincoop (2003).