Três ensaios sobre o uso de medidas antidumping

Abstract

This thesis consists of three studies that analyzed the use of antidumping (AD), focusing on the Brazilian economy. The first article verified how the main economies have used the AD after the Uruguay round, between 1995 and 2012. Therefore, some techniques were employed in order to reveal the main users of this instrument and the trend of using associated to each one of them. We also tested the hypothesis that countries could converge in terms of the number of AD cases and if spatial effects could be involved in this process. Finally, a criterion was established in order to identify countries that usually favor the most competitive sectors through AD practice. The results revealed that Brazil was the only heavy user that was counter to the general downward trend verified on the AD cases. Meanwhile, China was not only the main target of this mechanism but the cases against this country showed a growth trend also. Furthermore, we found convergence in the use of this instrument only when the economic weight of these countries was considered and our index indicated that Turkey and the European Union seem to be favoring the most competitive industries through the AD practice. The second article analyzed the main determinants of opening of AD cases using a panel data containing information of 46 AD users during 1995 to 2013. For this purpose, an extensive review of the empirical work of this area was performed in order to develop a comprehensive model, containing several variables considered relevant. The estimates carried out by Poisson or Negative Binomial, revealed that a decrease in imports, an increase in foreign income, a reduction in domestic income, currency devaluation or an improvement in the current account could reduce AD cases. Meanwhile, countries that belong to the higher income groups, which adopt high import tariffs, which are frequent targets of AD cases and are intensive exporters of metal products, chemicals and plastics, tend to open more AD cases. Also, countries that do not belong to the OECD and those from Europe or Central Asia tend to use less AD measures. The third article analyzed the influence of macroeconomic factors on the opening of AD cases in Brazil and in Argentina. Notably, such research had been done only to one of these countries (Brazil) and the authors rejected any macroeconomic influences. Since this type of analysis usually does not have large samples, the relevant variables were selected using the Sala-i-Martin (1997) test. This procedure allowed us to surpass the other articles regarding the inclusion of variables. As recommended, the models were estimated by Poisson regression. It was evidenced by the results that both Brazil and Argentina are affected by macroeconomic factors. However, the traditional (aggregated) models would tend to deny or minimize such influence on Brazilian economy and overestimate this effect in Argentina. Thus, the use of disaggregated data seems to be useful in this type of analysis.