However plausible their argument on the lack of trade deflection in China is, the product-level feature of their dataset does not allow them to explore the linkage between firms’ previous exporting status in different markets and trade deflection. It is highly likely that Japanese firms which were serving more markets compared to their Chinese counterparts could deflect their trade to alternative destinations. In addition, some Chinese firms might have deflected their shipments to some of their trading partners, which would be hidden when the exports are aggregated to product level. In the presence of sunk start-up costs of exporting, it is difficult for the exporters to sell their products in alternative markets if they have not setup ongoing trading relationships in multiple markets. This argument is impossible to analyze without breaking down the firm exports by exported products and export destinations. This study employs a unique four dimensional (firm-product-destinationyear) Brazilian firm-level export data to investigate the effect of sunk start-up costs of exporting on the trade flows of the firms to alternative markets when they suffer AD duties in a particular destination. Analyzing the firm-level responses of AD duties on trade deflection will provide a better understanding of which destinations are potential export markets and whether the past trading relationships matter to deflect trade for the firms whose products are targeted by AD measures. In this regard, our rich dataset provides a unique opportunity to explore the variation in exports within firms across different destinations when there is a change in trade barrier for a particular product in a particular export destination. To date, there is too few studies in the AD literature using firm-level data estimate the effect of antidumping protection on the productivity of domestic import-competing firms in the illustrates the relationship between EU and US antidumping measures and foreign direct investment through a microeconometrical analysis of Japanese firms’ plant establishments in the electronics industry. In a noticeably detailed analysis investigates the plantlevel responses to AD measures for the protected plants in the examines the pricing effect of AD duties and the exporters’ response to a threat of retaliation stemming from domestic AD actions remains the only firmlevel study of AD policy to analyze the value of export sales and the extensive margin of exports. Although related, our paper conceptually distinguishes from  by the fact that their study focuses on the effect of France’s AD duties on the exports of the domestic protected firms, whereas this paper analyzes the effect of AD duties which targets the exporters in the international market. In addition, they exploit a three dimensional panel which does not differentiate the product categories for the firms which exports multiple products. Whereas, with a four dimensional panel data for firm, product and destinations, the empirical analysis carried out in this paper is a significant improvement over the previous studies which investigate the trading effects of AD measures. It should be highlighted that AD duties provide a useful way of examining trade deflection. Antidumping duties yield substantial changes in trade flows given the fact that they are on average 10 to 20 times higher than the most favorednation (MFN) tariffs.6 Besides, AD duty is a product and a market specific trading cost for a firm. For example, if Mexican AD agencies impose an AD duty on Brazilian cotton shirt exporters, neither the other textile shirt exporters of Brazil nor the cotton shirt exporters of Argentina will be affected by this discriminatory policy adjustment. Hence, if a firm sells multiple products to a destination, it is obliged to pay AD duties only for the particular product which is targeted by the importer country. Since our analysis is based on the attractive source of variation in the value of exports within firm-product combinations across export destinations, these product specific shocks for the firms in different export markets perfectly fits with our research question. Alongside this, Brazil is a well-suited country for such an analysis for number of reasons. First, highly disaggregated firm-level data of Brazilian exports makes Brazil an outstanding case for this research. Second, Brazilian exported products were frequently targeted by AD duty over the period of our sample. There are 51 AD cases filed against Brazil in this period, 40 of which resulted ‘n rulings against Brazil. Moreover, these affirmative cases correspond to 120 unique six-digit HS products. Finally, countries which imposed AD duty on Brazilian exported products accounts for almost 50% of the Brazil’s total exports over the sample. This allows us to expect a dramatic impact of AD duties imposed by these countries on the trade flows of Brazilian firms to third countries. Table 1 documents the products subject to AD duties and the duty imposing countries between 1994 and 2000. Our main findings can be summarized as follows: firms only deflect exports to countries where they have an already established trading relationship. In particular, we find that, on the intensive margin, firms increase their export to alternative countries in which they were already exporting the targeted product when they suffer an AD duty in a particular export destination.